Chemosphere Lajeunesse 2011 Antidepressants in Trouts.Pdf

Chemosphere xxx (2011) xxx–xxx

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier.com/locate/chemosphere

Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects ⇑ André Lajeunesse a,c, Christian Gagnon a, , François Gagné a, Séverine Louis a, Patrick Cˇejka b, Sébastien Sauvé c a Environment Canada, Fluvial Ecosystem Research, Science and Technology Branch, 105 McGill Street, Montreal, Quebec, Canada H2Y 2E7 b Montreal Sewage Treatment Plant, 12001 Maurice-Duplessis Boulevard, Montreal, Quebec, Canada H1C 1V3 c Department of Chemistry, Université de Montréal, P.O. 6128, Downtown Postal Station, Montreal, Quebec, Canada H3C 3J7 article info abstract

Article history: This study examined the tissues distribution of selected serotonin reuptake inhibitors (SSRIs) in brook Received 10 September 2010 trout exposed for 3 months to continuous flow-through primary-treated effluent before and after ozone Received in revised form 23 November 2010 treatment. A reliable analytical method was developed for the quantification of trace amounts of antide- Accepted 4 December 2010 pressants in small tissue homogenate extracts. Levels of six antidepressants and four of their N-desmeth- Available online xxxx yl metabolites were determined using liquid chromatography–tandem mass spectrometry. Significant amounts of the SSRIs were found in fish tissue—in decreasing order: liver > brain > muscle. Sertraline Keywords: and its metabolite desmethylsertraline were the predominant substances observed in most tissues Antidepressants (0.04–10.3 ng gÀ1). However, less SSRIs (0.08–1.17 ng gÀ1) were bioaccumulated in the ozonated effluent. Wastewater LC–MS/MS The early molecular effects of these SSRIs on the Na/K-dependent ATPase pump activity in brain synap- Bioaccumulation tosomes where also investigated in vitro and in fish exposed to the municipal effluents. With respect to Metabolites their potential biological effects, in vitro exposure to selected SSRIs induced a reduction of the brain Na/K- Na/K-ATPase activity ATPase activity in synaptosomes in a dose-dependent manner. Results showed that Na/K-ATPase activity was readily inhibited by exposure to municipal effluent before and, to a lesser extent, after ozone treatment. Moreover, the Na/K-ATPase activity was significantly and negatively correlated with brain tissue concentrations of fluoxetine (r = À0.57; p < 0.03), desmethylsertraline (r = À0.84; p < 0.001), and sertraline (r = À0.82; p < 0.001). The present study reveals that SSRIs are readily available in fish, biologically active and corroborates previous findings on the serotonergic properties of municipal effluents to aquatic organisms. Crown Copyright Ó 2010 Published by Elsevier Ltd. All rights reserved.

1. Introduction (Jones et al., 2006; Lishman et al., 2006). Antidepressant drugs are among the PhACs likely to be found in Canada (IMS Health Over the last decade, pharmaceutical and personal-care Canada, 2006), United States (Schultz et al., 2010), and other coun- products (PPCPs) have been recognized as a major source of pollu- tries around the world (Calisto and Esteves, 2009). The commonly tion for the aquatic environment (Halling-Sørensen et al., 1998; prescribed antidepressants from the class of selective serotonin Daughton and Ternes, 1999). Their occurrence in most municipal reuptake inhibitors (SSRIs) have become the most successful class efﬂuent seems to indicate that conventional wastewater treat- of marketed antidepressants, with the number of new prescrip- ments have only a limited ability to remove such substances from tions written for these drugs now exceeding that of tricyclic anti- sewage. In the 1990s, it was discovered that some active drugs depressants (DeVane, 1999). have the capability to interfere with ecosystems in concentrations To date, most analytical methods for determining the presence as low as a few nanograms per liter (Halling-Sørensen et al., 1998). of antidepressants and their respective metabolites in environ- Until now, different pharmaceutically active compounds (PhACs) mental matrixes have been developed for surface waters and have been detected in the environment: analgesics, antibiotics, wastewater efﬂuents (Lishman et al., 2006; Lajeunesse et al., antiepileptics, antidepressants, synthetic and natural hormones 2008; Schultz and Furlong, 2008; Metcalfe et al., 2010), sediments (Schultz et al., 2010), and sewage sludge (Radjenovic´ et al., 2009). ⇑ Corresponding author. Tel.: +1 514 496 7096; fax: +1 514 496 7398. Due to the complexity of samples, these analyses typically employ E-mail address: [email protected] (C. Gagnon). liquid chromatography–tandem mass spectrometry (LC–MS/MS).

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 2 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx

Since the occurrence of antidepressants in municipal wastewater erin/luciferase enzymatic system were purchased from Sigma–Al- could lead to physiological and behavioural effects on aquatic drich Co. (St. Louis, Missouri, USA), while cis-tramadol13C-d3 was organisms (Fong, 1998, 2001; Calisto and Esteves, 2009), it is crit- purchased from Cerilliant Corp. (Round Rock, Texas, USA). The ical that these compounds be measured in organisms using reliable high-performance liquid chromatography–grade solvents (metha- analytical methods. Even so, only a few analytical methods have nol and acetonitrile) and ammonium hydroxide were provided been reported on antidepressants in aquatic organisms (Brooks by Caledon Laboratories Ltd. (Georgetown, Ontario, Canada). et al., 2005; Chu and Metcalfe, 2007; Ramirez et al., 2007; Metcalfe Reagent-grade hydrochloric acid, acetic acid, monobasic potassium et al., 2010; Schultz et al., 2010). Unfortunately, the required mass phosphate, and ammonium bicarbonate were provided by Ameri- of tissues often reached up to 2 g, making difficult the detection of can Chemicals Ltd. (Montreal, Quebec, Canada). Solid-phase extrac- SSRIs and their metabolites with small organisms or limited quan- tion (SPE) cartridges of 6 mL, 500 mg (Strata-SCX™) were tity of target tissues. purchased from Phenomenex (Torrance, California, USA). Stock As antidepressants are present at relatively low levels in the solutions of 100 mg LÀ1 of each substance were prepared in meth- environment (e.g. ng LÀ1), risk for acute toxic effects is unlikely, anol and stored at 4 °C in amber glass bottles. All corrosive and but chronic environmental toxic effects cannot be excluded (Péry pure standard chemicals were handled carefully under a ventilated et al., 2008; Carlsson et al., 2006). The potential ecotoxicological fume hood wearing appropriate protection. impacts of neuroactive drugs on aquatic organisms are not well understood at the present time. Moreover, the lack of information 2.2. Experimental procedures about the fate and long-term subtle effects of these compounds and/or their metabolites in aquatic organisms makes accurate risk 2.2.1. Instrumental parameters assessment difficult (Gagné and Blaise, 2004; Stanley et al., 2007; A Thermo Finnigan Surveyor liquid chromatograph (LC) quater- Paterson and Metcalfe, 2008). Hence, the measurement of the nary pump equipped with an autosampler was used in this study. Na/K-dependent ATPase pump activity in synapsomes represents As reported in a previous paper (Lajeunesse et al., 2008), analytes a more specific biomarker of the effects of SSRIs (Kanoh et al., were separated on a C18 column using a binary gradient made of 1998). The serotonin reuptake pump (or transporter) exchanges (a) ammonium bicarbonate (10 mM) pH 7.8, and (b) acetonitrile one K ion per molecule of serotonin. The loss of K ions outside at a flow rate of 200 lL minÀ1. Volume of injection was 20 lL for the synapse button is counterbalanced by the Na/K-ATPase pump biota extracts. which maintains Na and K ions respectively outside and inside of Eluted analytes were monitored by a Thermo Finnigan TSQ the membrane. This process requires energy (e.g. ATP ? ADP + Quantum Ultra triple quadrupole tandem mass spectrometer inorganic phosphate) that can be easily measured (Kakko et al., equipped with an electrospray ionization source. The capillary 2003). Thus, the presence of SSRIs decreases the serotonin- was held at 5.0 kV and the cone was operated at an optimized volt- dependent hydrolysis of ATP in neuron synaptosomes by the com- age for each analyte in positive-ion mode (+ESI). Additional instru- petitive inhibition of the transporters. At present, the modulation ment parameters and analytical details are available (Lajeunesse of Na/K-dependent ATPase activity in nerve synaptosomes by the et al., 2008). The specificity of all target analytes was insured by SSRIs in fish is largely unknown. Despite its possible usefulness, the use of two MS/MS transitions in multiple reactions monitoring this promising methodology alone cannot bring sufficient informa- (MRM) mode. All optimized mass spectrometer parameters are tion about the exact concentration of drug needed to elicit biolog- available as Supporting information in Table S1, Appendix A. ical effects in fish. Therefore, a combination of a biomarker of effects and a tissue SSRIs analysis would strengthen the link be- 2.2.2. Fish collection and preparation tween the presence of this class of contaminants in target tissues Brook trout (Salvelinus fontinalis), 20 cm long, were obtained and the resulting early biological effects. from a commercial fish hatchery (Station piscicole Trois-Lacs of In this study, we report the development of a reliable LC–MS/ Wotton, Quebec, Canada). A total of 50 fish per mesocosm (volume MS method for the trace detection of antidepressants in tissues capacity of 150 L) were exposed for 3 months at the Montreal of brook trout exposed to treated (ozone 15 mg LÀ1) and non-disin- WWTP to 10% and 20% v/v of final effluent (mean effluent flow rate fected effluent at the Montreal wastewater treatment plant of 140 mL minÀ1). Supplementary hatched trout were exposed to (WWTP). The developed method was applied to determine the 10% and 20% v/v of effluent previously treated with 15 mg LÀ1 of concentrations and distribution of antidepressant in liver, brain, ozone. During exposure, around 15% of mortality by microbiological and muscle tissues. Based on bioaccumulation results, a new infections was observed in mesocosms containing diluted final efficient biomarker (Na/K/serotonin-dependent ATPase activity in effluent. However, no fish mortality was reported in mesocosms neurons synaptosomal membranes) is proposed to track the early containing final effluent treated with ozone. Typically, the Montreal biological effects of SSRIs. WWTP processes 2.7 million m3 of raw sewage daily, yielding a rep- resentative mean effluent discharge of 19.8 m3 sÀ1. Final effluent is discharged directly into the St. Lawrence River. Montreal’s treat- 2. Experimental ment plant uses a physico-chemical process, which involves a grit removal and a physico-chemical treatment by flocculation and 2.1. Chemicals and standards sedimentation to reduce suspended matter and phosphorus content. Dissolved organic carbon (DOC) concentrations and pH values Certified standard (>98% purity grade) fluoxetine (FLU), nor- ranged from 90 to 110 mg LÀ1 and 8.1 to 8.2 respectively (Gagnon fluoxetine (NFLU), paroxetine (PAR), sertraline (SER), and (S)- et al., 2008). Brook trout were dissected in order to obtain brain, citalopram (CIT) were provided by Toronto Research Chemical liver, and filet (muscle) tissues. Because of limited quantities, fish Inc. (North York, Ontario, Canada). Desmethylsertraline (DSER), homogenates (100 mg) composed of pooled tissue (10 alive speci- venlafaxine (VEN), O-desmethylvenlafaxine (DVEN), and bupropi- mens for each tested conditions) were diluted in 1.0 mL of an on-d9 were obtained from Nanjing Jinglong PharmaTech (Nanjing, ice-cold 10 mM HEPES–NaOH (pH 7.4) buffer solution containing China). Amitriptyline (AMI) and nortriptyline (NTRI), HEPES, so- 150 mM NaCl, 1 mM EDTA, 0.1 mM dithiothreitol, and 1 lgmLÀ1 dium chloride, sodium hydroxide, ethylenediaminetetraacetic acid aprotinin. All tissues samples were stored at À80 °C until analysis. (EDTA), dithiothreitol, aprotinin, sucrose, magnesium chloride, and Total proteins were determined using serum bovine albumin for the bioluminescence commercial ATP detection kit based on lucif- calibration (Bradford, 1976).

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx 3

2.2.3. Sample extraction (Thermo Fisher). Mean LOD value of 0.01 ng gÀ1 (variability: 6.7%) Fish homogenate samples were combined with 9 mL of a solu- was estimated for fish tissues extracts when 200 mg of wet tissues tion composed of methanol/0.1 M acetic acid buffer solution pH were spiked with 25 ng of each antidepressant. The higher individ- 4.0 (1:1 v/v) in a 16 Â 150 mm borosilicate glass screw-top conical ual LOD variability (%) was found with brain tissues samples (e.g. tube. To each tube were added 100 lL of a surrogate standard solu- SER, up to 19.3%). However, lower variability was observed with tion in methanol (bupropion-d9: 0.25 mg LÀ1) equivalent to 25 ng. liver tissues (e.g. DSER, 0.5%). Following the same procedure, mean Samples were then shaken vigorously and mixed on a rotary LOQ values were determined by a signal-to-noise ratio (S/N)of extractor (Caframo REAX 2) for 15 min. After extraction, samples 10:1. The LOD and LOQ values are given as Supporting material were centrifuged at room temperature and 320g for 5 min. Follow- in Table S2, Appendix A. Matrix effects were also investigated on ing the solid-phase extraction (SPE) protocol of Lajeunesse et al. all extracted tissues. To prevent possible losses due to extraction, (2008), supernatants were transferred to strong cation exchange the ratio of the spiked extracted matrix (25 ng in 200 mg of tissue) cartridges (Strata-SCX, Phenomenex) preconditioned with 4 mL of minus the ratio of the extracted matrix alone was compared to a methanol and at least 8 mL of the same extraction buffer solution. 25-ng pure non-extracted standard prepared directly in the mobile The dried extracts containing 50 lL of cis-Tramadol13-d3 in meth- phase. For practical purposes, a relative signal response to matrix anol (2.5 mg LÀ1) as internal standard (ISTD) were reconstituted effects has been reported. Values around 1.0 indicate no apprecia- with 0.5 mL of the mobile phase solution made of ammonium ble signal suppression or enhancement: results clearly depict lim- bicarbonate (10 mM) pH 7.8/acetonitrile (1:1 v/v). ited matrix effects on the overall ion signal of each tissue (Fig. S1, Appendix A). The highest ion suppression occurred within brain tissue samples (À12% versus pure non-extracted standard solu- 2.2.4. Na/K-ATPase activity determinations tion). In order to overcome possible matrix interference within Total serotonin content and the activity of Na/K-dependent and tissues, standard addition curves were used for quantitation. Fur- Mg-dependent ATPases in synaptosomal membranes were deter- thermore, before each extraction, a standard surrogate (bupropi- mined in the brain of brook trout based on a previous methodology on-d9) was added to ensure the quality control of the extraction (Kakko et al., 2003). Brain synaptosomes were prepared from the process. Recoveries of 70–120% are normally obtained for standard homogenates as follows: homogenates (a portion was kept for total surrogate, otherwise the overall sequence is rejected. During quan- serotonin) were centrifuged at 3000g for 15 min at 2 °C. The super- tification process, no adjustments were done for antidepressants natant was overlaid on cold 0.8 M sucrose containing 10 mM against the STD surrogate response. The linearity of antidepres- HEPES–NaOH and 1 mM EDTA and centrifuged at 10 000g for sants had been tested across a range of concentrations that was 20 min at 2 °C. The synaptosomes (at the 0.8 M sucrose fraction) previously reported for pharmaceuticals in fish residues (Ramirez were collected from the mitochondria (pellet) and diluted one et al., 2007). Hence, the linearity of standard addition calibration third in distilled water. Na/K-ATPase and Mg-ATPase activities curves yielded a correlation coefficient r2 > 0.995 for a concentra- were determined in synaptosomes by mixing the synaptosomes tion range of 0.50–50 ng per 200 mg of spiked tissue. The mean with 100 lM ATP with or without 50 lM serotonin in 10 mM ratios of two zero points (e.g. extracts alone without any standard HEPES–NaOH buffer at pH 7.4 in the presence of either 140 mM addition) in the calibration curve were used to quantify the ana- NaCl and 5 mM KH2PO4 or 50 mM MgCl2. The reaction mixture lytes in each type of tissue. was incubated for up to 30 min at 30 °C, and the levels of ATP were measured using a bioluminescence commercial ATP detection kit 3.1.2. Application to fish analysis based on a luciferin/luciferase enzymatic system. The effects of Fish that were exposed to ozone- or primary-treated municipal serotonin and selected SSRIs (CIT, AMI, PAR, and FLU) were also effluents were sampled and investigated for antidepressants. The examined in synaptosome from control (unexposed) fish. This pro- occurrence of the antidepressant target molecules in liver, brain, cess involved pre-treating the membranes to increasing concentra- and filet (muscle) tissues was confirmed by observing two multiple tions of PAR and SER for 15 min before taking the ATPase reactions monitoring (MRM) transitions from mass spectrometer measurements. The data were expressed as ATP hydrolysis minÀ1 (except FLU, NFLU), and also by their respective average retention mg proteinsÀ1. Total serotonin levels were also determined using times (t ) ± 0.20 min. The validated protocol enables the sensitive a competitive immunoassay procedure developed in our labora- R detection of the eight studied antidepressants using only 100-mg tory. Details about this protocol are available as Supporting infor- tissue homogenates. The new methodology proved to be a valuable mation in Appendix B. tool in the analytical characterization of antidepressants (e.g. SSRIs, tricyclic) and their metabolites in a variety of biological tissues, 3. Results and discussion especially when the mass of tissues is limited. The relative distribution and concentrations of each detected antidepressant from 3.1. Chemical analysis brook trout exposed to 20% v/v of effluent are represented in Fig. 1. As described in the previous section, brook trout were exposed 3.1.1. LC–MS/MS method validation for 3 months in a continuous flow-through setup at the Montreal Recovery tests were completed on spiked reference brook trout WWTP to 10% and 20% v/v of renewed effluent. The main goal of tissues. Additional validation tests were also performed on two dif- using a continuous flow-through system was to simulate the dilu- ferent freshwater mussel tissues. The percent recovery of each tion factor retrieved in the receiving water of the St. Lawrence Riv- antidepressant (e.g. 25 ng of antidepressant in 200 mg of tissue) er at 0.3 km and 1 km from the effluent outfall. Others studies were was calculated as follows: the ratio of the spiked sample (area ana- also conducted with this system, using 10% and 20% v/v of effluent lyte/area ISTD) divided by the ratio of the standard previously pre- previously treated with 15 mg LÀ1 of ozone (Hébert et al., 2008). As pared by spiking set amounts of analytes in the extracted matrix. shown in Table 1, all studied antidepressants have been detected in Satisfactory results of recovery were obtained for all extracted tis- final effluent waters used for mesocosm exposition. sues (data available in Fig. S1, Appendix A). The extraction procedure was successfully applied to the analy- The developed analytical method enables a sensitive detection sis of all tissue homogenates (Table 2). An extracted ion chromato- of antidepressants in all tissues. Limit of detection (LOD) values gram (EIC) of DSER and SER from LC–MS/MS analysis of brook trout were defined as the minimum detectable amount of analyte giving exposed to 20% v/v of effluent is shown in Fig. 2. The highest a signal-to-noise ratio (S/N) of 3:1 using the Xcalibur 1.2 software concentrations of antidepressant were retrieved in liver tissues

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 4 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx

Detected antidepressants / Extracted tissues 100.0

10.0 ) -1 Liver 1.0 Brain

log (ng g log (ng Muscle

0.1

0.0

Sertraline Paroxetine Fluoxetine Venlafaxine Amitriptyline (S)-Citalopram Norfluoxetine

DesmethylSertraline

Fig. 1. Mean concentrations of antidepressant in 100 mg of tissue extracts from brook trout exposed in mesocosms to 20% v/v of efﬂuent at the Montreal WWTP.

Table 1 Table 2 Mean concentrations (±standard deviation) of antidepressants in 20% v/v effluent Concentrations of antidepressants retrieved in 100 mg of liver, brain, filet tissue mesocosm and estimated bioconcentration factor of antidepressants in brook trout. extracts from brook trout exposed in mesocosms to diluted effluents at the Montreal WWTP. Compounds Concentration Bioconcentration factor (BCF)b (ng LÀ1) Substances Concentrations (ng gÀ1) from 100 mg of wet tissue Liver Brain Muscle 20% v/v effluent (n =4) (n =4) (n =4) (n =4) mesocosma ABCD Desmethylvenlafaxine 55.3 ± 2.7 – – – Liver tissues Venlafaxine 38.5 ± 1.6 18 11 2 Venlafaxine 0.69 ± 0.14 0.16 ± 0.06 0.16 ± 0.02 n/d Paroxetine 0.96 ± 0.08 365 198 – Paroxetine 0.35 ± 0.10 0.15 ± 0.05 0.14 ± 0.03 n/d Citalopram 10.5 ± 0.2 39 17 – S-citalopram 0.41 ± 0.19 n/d 0.06 ± 0.03 n/d Norfluoxetine 0.36 ± 0.04 1750 – – Norfluoxetine 0.63 ± 0.17 0.32 ± 0.10 n/d n/d Nortriptyline 0.54 ± 0.06 – – – Fluoxetine 0.20 ± 0.06 0.11 ± 0.05 n/d n/d Fluoxetine 0.58 ± 0.02 345 138 224 Desmethylsertraline 10.29 ± 0.76 1.17 ± 0.28 0.06 ± 0.02 n/d Desmethylsertraline 0.84 ± 0.08 12 250 2476 500 Amitriptyline 0.29 ± 0.07 n/d n/d n/d Amitriptyline 3.7 ± 0.2 78 – – Sertraline 0.29 ± 0.05 0.15 ± 0.04 0.15 ± 0.08 0.10 ± 0.03 Sertraline 1.1 ± 0.1 264 191 109 Brain tissues a From Lajeunesse et al. (2008). Venlafaxine 0.43 ± 0.10 n/d n/d n/d b À1 À1 BCF: [conc. tissue, ng g ]/[conc. 20% v/v effluent mesocosm, ng L ] Â 1000, Paroxetine 0.19 ± 0.11 n/d 0.13 ± 0.07 n/d À1 based on water density of 1.0 g mL . S-citalopram 0.18 ± 0.11 n/d 0.06 ± 0.03 n/d Norfluoxetine n/d n/d n/d n/d Fluoxetine 0.08 ± 0.02 0.11 ± 0.05 n/d n/d Desmethylsertraline 2.08 ± 0.45 0.78 ± 0.40 0.06 ± 0.02 n/d (0.20–10.29 ng gÀ1) dissected from fish exposed to the less-diluted Amitriptyline n/d n/d n/d n/d primary-treated effluent. Other results revealed that up to Sertraline 0.21 ± 0.08 0.16 ± 0.06 0.04 ± 0.01

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx 5

retrieved for liver tissue extracts: DSER (12 250), NFLU (1750), PAR (350), and FLU (345). However, the distribution and bioaccumulation are less important in hydrophobic tissue (e.g. filet) under experimented pH condition (ionized species seem to be much B(extract + 25 ng spike) more important than nonionized species). Our calculated BCF values for NFLU and FLU were found comparable with those reported by Nakamura et al. (2008) after an exposition of Japanese medaka (Oryzias latipes) at pH 8 during 30 d (NFLU: 3300, FLU: 580). Nev- ertheless, despite its usefulness, the BCF remained somewhat difficult to calculate. As pointed out by Metcalfe et al. (2010), the origins of metabolites detected in fish tissues could be assigned to direct uptake or by in vivo metabolism. Indeed, a metabolism study conducted by Smith et al. (2010) on rainbow trout liver microsomes showed that hepatic FLU metabolism in fish was much less than that seen in mammals and the major mammalian metabolite, NFLU, was possibly not the primary metabolite of FLU in fish. They also suggested that the low level of metabolism could likely be the reason that fish bioaccumulate FLU in vivo. These assump- tions are consistent with our results and could possibly explain why the metabolite NFLU was detected neither in brain nor in filet tissues. However, the physico-chemical properties of the antidepressants and the water pH seem also to play an important role in the molecules distribution (Nakamura et al., 2008; Metcalfe et al., 2010). Therefore, more data and additional information related to the metabolism and the bioaccumulation would be desirable to clarify the possible environmental pathways of antidepressants along their biologically metabolites under given pH conditions.

A(extract) 3.2. Na/K-ATPase activity

In addition to using LC–MS/MS for the quantification of bioaccumulated antidepressants, this paper investigated biomarkers of effects sensitive to this class of pharmaceuticals. The potential toxicological effects of the SSRIs in municipal effluents on brook Fig. 2. Extracted ion chromatogram (EIC) of DSER and SER from LC–MS/MS analysis of brook trout exposed to 20% v/v of effluent: (A) liver tissues extract, and (B) spiked trout were also examined. First, a methodology for evaluating the liver tissues extract. activity of serotonin reuptake was developed and optimized by following the rate of Na/K-dependent ATPase activity in synaptosomal membrane vesicles. As clearly depicted in Fig. 3A, the and SER were detected (0.12–0.42 ng gÀ1). Recently, Schultz et al. addition of 5-hydroxytryptamine or serotonin (10–100 lM) in syn- (2010) have reported similar concentration of DSER (1.8–3.0 ng gÀ1) aptosomal membrane vesicles induced an increase in ATP hydroly- in brain homogenates from native White Suckers species exposed sis, i.e. serotonin stimulation of the ATPase activity to maintain the downstream to outfalls discharging of two effluent-impacted electromotive gradient in synaptosomal membranes. streams located in Colorado and Iowa, USA. Another study com- Pre-incubation with selected SSRIs (CIT, AMI, PAR, and FLU) was pleted in Ontario (Canada) by Metcalfe et al. (2010) reports on the also able to reduce the activity of the ATP-dependent Na/K-ATPase. positive detection of five antidepressants cumulated in whole To verify the validity of the chosen biomarker, Na/K-dependent Fathead Minnows species caged at locations downstream of ATPase, four SSRIs antidepressants were added in synaptosomal WWTPs. Interestingly, during our bioaccumulation study the membrane vesicles to block serotonin transport. As expected, N-desmethyl metabolites were mostly measured at higher concen- Na/K-ATPase activity was decreased by the addition of 100 lMof trations in individual liver tissues than their parent molecules. SSRIs to the membrane preparations (see Fig. 3B). The reuptake Brooks et al. (2005) and Schultz et al. (2010) observed the same of serotonin at the membrane interface occurs with the efflux of trends for some antidepressants in different fish species. Neverthe- K+, which in turn stimulates ATPase to maintain the electroconduc- less, results have shown much lower concentrations of antidepres- tive gradient. According to experimental results, the SSRIs reduced sants in fish exposed to disinfected effluent by ozone treatment Na/K-ATPase activity where PAR and FLU were the most potent (15 mg LÀ1). A pilot study conducted at the Montreal WWTP in ones. In addition, Mg-dependent ATPase activity was also affected 2005 and 2006 demonstrated the effectiveness of ozone disinfec- by these drugs, but to a lesser extent. Among the four antidepres- tion on several classes of pharmaceuticals (Gagnon et al., 2008). In sants tested, PAR was found to be the most potent SSRI for brook fact, a removal rate of 70% was achieved for six of the eight pharma- trout, followed by FLU and AMI. ceuticals detected in the effluent when using an optimal ozone Total serotonin and Na/K-ATPase activity was analyzed in brain concentration of 15–20 mg LÀ1. Therefore, the remarkable effi- synaptosomes from trout exposed to primary-treated effluent be- ciency observed with ozone treatment would certainly decrease fore and after ozone treatment. Total serotonin content remain the bioavailability of antidepressants to fish exposed to wastewater constant in fish tissues (ANOVA p = 0.24; results not shown). How- effluents. ever, the Na/K-dependent ATPase activity was readily inhibited by According to our calculated tissue concentrations, the estimated the municipal effluent before and after ozone treatment (Fig. 4). bioconcentration factors (BCFs) for 20% v/v effluent mesocosm (pH: Ozone treatment seemed to mitigate this effect, but not com- 8.1–8.2) ranged from 2 to 12 250 (Table 1). The highest BCFs were pletely; Na/K-dependent ATPase activity was still inhibited at

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 6 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx

Mean Stand. error 14

Na/K-ATPase activity 8 (ATP hydrolysis / min mg proteins) 6

4 01050100 Added serotonin (µM) A (n=6)

120 Na/K-ATPase 110 Mg-ATPase

100

80 * * * 70 * ATPase activity 60 * 50 (ATP hydrolysis / min mg proteins) 40

20 0 Citalopram Amitryptiline Paroxetine Fluoxetine Drug pre-treatment (100 µM) B (n=6)

Fig. 3. Modulation in vitro of Na/K-dependent and Mg-dependent ATPase activity in the brain tissues by selected SSRIs following (A) added serotonin, and (B) pre-treatment with antidepressants.

20% v/v effluent dilution. This effect is consistent to the presence of The Na/K-dependent ATPase activity was negatively correlated SSRIs in urban effluent. This was further supported by the signifi- with brain tissue concentrations of FLU (r = À0.57; p < 0.03), DSER cant correlation between SSRIs in tissues and Na/K-ATPase in (r = À0.84; p < 0.001), and SER (r = À0.82; p < 0.001). These results synaptosomes. However, this relationship cannot rule out the pos- suggest that the presence of SSRIs in the brain, in part at least, was sibility that other compounds in effluent might have contributed to associated to lower Na/K-ATPase activity in synaptosomal mem- the observed responses. Considering the importance of eventually branes, which is in agreement with the mode of action of these defining a hazard assessment for SSRIs to fish, the reported meth- drugs and the observed responses during the in vitro experiments odology for measuring toxicological effects caused by antidepres- (Fig. 3). In keeping with biomarkers used for pharmaceuticals in sants may represent a promising means of determining whether the environment (Gagné and Blaise, 2004), physiological biomark- these changes are linked with drug content in studied tissues. ers that respond to the mode of action of drugs and address the

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx 7

0,00022 Mean Stand. error

0,00020

0,00018 )

-1 0,00016 * 0,00014

0,00012 *

Na/K ATPase activity 0,00010 (rate ATP hydroysis min * 0,00008

0,00006

0,00004 Control 20 % NT 20 % Oz 10 % NT 10 % Oz Effluent concentration and treatment

Fig. 4. Modulation of nerve synapse Na/K-ATPase activity after diluted effluent exposure, n = 6 (NT: not treated, Oz: ozone treatment at 15 mg LÀ1). toxic consequences are needed in environmental studies. Based on However, these interesting observations cannot rule out other con- the present study, the Na/K-ATPase activity biomarker was shown taminants given the complex nature of municipal effluents. The to respond to these bioaccumulated SSRIs and other compounds combination of tissue loadings and early biological effects provide targeting electrolyte balance in membranes. This highlights the useful information on potential ecotoxicological impacts of dis- need to strengthen the relationships between the presence of sero- charged drugs in urban effluents to aquatic life. tonergic-like compounds and serotonergic effects in complex environmental mixtures. The data presented for the effluent-exposed fish highlighted the evidence that municipal effluents contained Acknowledgements serotonin mimicking compounds (Gagné and Blaise, 2003). Indeed, solid phase C-8 extracts of surface water downstream the same The authors thank the city of Montreal for providing access to municipal effluent of the present study were capable of activating the mesocosms. Sophie Trépanier from Environment Canada is also human serotonin receptors and incubation of female ripe gonads of acknowledged for her meticulous efforts in the experimental setup Elliptio complanata to the filtrated municipal effluent, significantly of the mesocosms. Additional thanks are extended to Keltie Purcell stimulated egg release (Gagné et al., 2004). from Environment Canada for editing the manuscript. This work was supported by the Chemical Management Plan – Health Canada, the St. Lawrence Action Plan and the Canadian Foundation 4. Conclusion for Innovation.

A sensitive and rugged LC–MS/MS method has been developed for the simultaneous analysis of antidepressants along with their Appendix A. Supplementary material respective N-desmethyl metabolites in limited fish tissues extracts samples. The results reported supports the observation that anti- Supplementary data associated with this article can be found, in depressant may accumulate in fish living in an effluent-dominated the online version, at doi:10.1016/j.chemosphere.2010.12.026. stream. In the present study, most drug residues partitioned in liver and brain tissues depicting the possibility of distribution in References specific tissues. The higher concentrations of antidepressant were observed in liver tissues when the brook trout were exposed to Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of 20% v/v of effluent. Results have shown much lower concentrations microgram quantities of protein utilizing the principle of protein-dye binding. of antidepressants in fish exposed to disinfected effluent by ozone Anal. Biochem. 72, 248–254. treatment (15 mg LÀ1). Brooks, B.W., Chambliss, C.K., Stanley, J.K., Ramirez, A., 2005. Determination of Tested SSRIs were shown to reduce the Na/K-ATPase activity in select antidepressants in fish from an effluent-dominated stream. Environ. Toxicol. Chem. 24, 464–469. the presence of serotonin in synaptosomal membranes: PAR and Calisto, V., Esteves, V.I., 2009. Psychiatric pharmaceuticals in the environment. FLU being the most potent ones. Brain membrane Na/K-ATPase Chemosphere 77, 1257–1274. activity was also reduced in fish exposed to a primary-treated Carlsson, C., Johansson, A.K., Alvan, G., Bergman, K., Kuhler, T., 2006. Are pharmaceuticals potent environmental pollutant? Part I: environmental risk effluent before and after ozone treatments showing a similar effect assessment of selected active pharmaceutical ingredients. Sci. Total Environ. (of SSRIs mode of action). This was further supported by the signif- 364, 67–87. icant and negative correlation of brain Na/K-ATPase with brain tis- Chu, S., Metcalfe, C.D., 2007. Analysis of paroxetine, fluoxetine and norfluoxetine in fish tissues using pressurized liquid extraction, mixed mode solid phase sue levels of FLU, DSER and SER which suggests that these drugs extraction cleanup and liquid chromatography–tandem mass spectrometry. J. might have contributed, at least in part, to the observed effects. Chromatogr. A 1163, 112–118.

Please cite this article in press as: Lajeunesse, A., et al. Distribution of antidepressants and their metabolites in brook trout exposed to municipal wastewaters before and after ozone treatment – Evidence of biological effects. Chemosphere (2011), doi:10.1016/j.chemosphere.2010.12.026 8 A. Lajeunesse et al. / Chemosphere xxx (2011) xxx–xxx

Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in Lajeunesse, A., Gagnon, C., Sauvé, S., 2008. Determination of basic antidepressants the environment: agents of subtle change? Environ. Health Perspect. 107, 907– and their N-desmethyl metabolites in raw sewage and wastewater using solid- 938. phase extraction and liquid chromatography–tandem mass spectrometry. Anal. DeVane, C.L., 1999. Metabolism and pharmacokinetics of selective serotonin Chem. 80, 5325–5333. reuptake inhibitors. Cell. Mol. Neurobiol. 19, 443–465. Lishman, L., Smyth, S.A., Sarafin, K., Kleywegt, S., Toito, J., Peart, T., Lee, B., Servos, M., Fong, P.P., 1998. Zebra mussel spawning is induced in low concentrations of Beland, M., Seto, P., 2006. Occurrence and reductions of pharmaceuticals and putative serotonin reuptake inhibitors. Biol. Bull. 194, 143–149. personal care products and estrogens by municipal wastewater treatment Fong, P.P., 2001. Antidepressants in aquatic organisms: a wide range of effects. In: plants in Ontario, Canada. Sci. Total Environ. 367, 544–558. Daughton, C.G., Jones-Lepp, T.L. (Eds.), Pharmaceuticals and Personal Care Metcalfe, C.D., Chu, S., Judt, C., Li, H., Oakes, K.D., Servos, M.R., Andrews, D.M., 2010. Products in the Environment: Scientific and Regulatory Issue. ACS Symposium Antidepressants and their metabolites in municipal wastewater, and Series, Washington DC, USA. downstream exposure in an urban watershed. Environ. Toxicol. Chem. 29, Gagné, F., Blaise, C., 2003. Effects of municipal effluents on serotonin and dopamine 79–89. levels in the freshwater mussel Elliptio complanata. Comp. Biochem. Physiol. C Nakamura, Y., Yamamoto, H., Sekizawa, J., Kondo, T., Hirai, N., Tatarazako, N., 2008. Toxicol. Pharmacol. 136, 117–125. The effects of pH on fluoxetine in Japanese medaka (Oryzias latipes): acute Gagné, F., Blaise, C., 2004. Effects of pharmaceuticals on aquatic biota – a review. toxicity in fish larvae and bioaccumulation in juvenile fish. Chemosphere 70, Curr. Topics Toxicol. 1, 73–86. 865–873. Gagné, F., Fournier, M., Blaise, C., 2004. Serotonergic effects of municipal effluents: Paterson, G., Metcalfe, C.D., 2008. Uptake and depuration of the anti-depressant induced spawning activity in freshwater mussels. Fresenius Environ. Bull. 13, fluoxetine by Japanese medaka (Oryzias latipes). Chemosphere 74, 125–130. 1099–1103. Péry, A.R.R., Gust, M., Vollat, B., Mons, R., Ramil, M., Fink, G., Terne, T., Garric, J., 2008. Gagnon, C., Lajeunesse, A., Cˇejka, P., Gagné, F., Hausler, R., 2008. Degradation of Fluoxetine effects assessment on the life cycle of aquatic invertebrates. selected acidic and neutral pharmaceutical products in a primary-treated Chemosphere 73, 300–304. wastewater by disinfection processes. Ozone Sci. Eng. 30, 387–392. Radjenovic´, J., Jelic´, A., Petrovic´, M., Barceló, D., 2009. Determination of Halling-Sørensen, B., Nors Nielsen, S., Lanzky, P.F., Ingerslev, F., Holten Lützhøft, pharmaceuticals in sewage sludge by pressurized liquid extraction (PLE) H.C., Jergensen, S.E., 1998. Occurrence, fate and effects of pharmaceutical coupled to liquid chromatography–tandem mass spectrometry (LC–MS/MS). substances in the environment – a review. Chemosphere 36, 357–393. Anal. Bioanal. Chem. 393, 1685–1695. Hébert, N., Gagné, F., Cˇejka, P., Bouchard, B., Haussler, R., Cyr, D.G., Blaise, C., Ramirez, A.J., Mottaleb, M.A., Brooks, B.W., Chambliss, C.K., 2007. Analysis of Fournier, M., 2008. Effects of ozone, ultraviolet and peracetic acid disinfection of pharmaceuticals in fish using liquid chromatography–tandem mass spectro- a primary-treated municipal effluent on the immune system of rainbow trout metry. Anal. Chem. 79, 3155–3163. (Oncorhynchus mykiss). Comp. Biochem. Physiol. C Toxicol. Pharmacol. 148, Schultz, M.M., Furlong, E.T., 2008. Trace analysis of antidepressant pharmaceuticals 122–127. and their select degradates in aquatic matrixes by LC/ESI/MS/MS. Anal. Chem. IMS Health Canada, 2006. Antipsychotic/Estimated Pill Consumption, Canadian 80, 1756–1762. CompuScript: Montreal, Quebec, Canada. Schultz, M.M., Furlong, E.T., Kolpin, D.W., Werner, S.L., Schoenfuss, H.L., Barber, L.B., Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2006. Partitioning behaviour of five Blazer, V.S., Norris, D.O., Vajda, A.M., 2010. Antidepressant pharmaceuticals in pharmaceutical compounds to activated sludge and river sediment. Arch. two US effluent-impacted streams: occurrence and fate in water and sediment, Environ. Contam. Toxicol. 50, 297–305. and selective uptake in fish neural tissue. Environ. Sci. Technol. 44, 1918–1925. Kakko, I., Toimela, T., Tahti, H., 2003. The synaptosomal membrane bound ATPase as Smith, E.M., Chu, S., Paterson, G., Metcalfe, C.D., Wilson, J.Y., 2010. Cross-species target for the neurotoxic effects of pyrethroids, permethrin and cypermethrin. comparison of fluoxetine metabolism with fish liver microsomes. Chemosphere Chemosphere 51, 475–480. 79, 26–32. Kanoh, N., Hori, K., Ishigaki, T., Hori, S., 1998. Effect on serotonin on ouabain- Stanley, J.K., Ramirez, A.J., Chambliss, C.K., Brooks, B.W., 2007. Enantiospecific sensitive, K+-dependent, p-nitrophenylphosphatase activity in strial marginal sublethal effects of the antidepressant fluoxetine to a model aquatic vertebrate cells of normal and reserpinized guinea pigs. Histochem. J. 30, 263–266. and invertebrate. Chemosphere 69, 9–16.