Linking chemical exposure to lipid homeostasis: A municipal waste water treatment plant influent is obesogenic for zebrafish larvae

1 Linking chemical exposure to lipid homeostasis: a municipal waste water 1 treatment plant influent is obesogenic for zebrafish larvae 2 Mélanie Audrey Barbosa a,b,c, Ricardo Capela, a Jorge Rodolfo a,c, Elza Fonseca a, Rosa 3 Montes d, Ana André a, Ana Capitão a, António Paulo Carvalho a,b, José Benito 4 Quintana d, L. Filipe C. Castro a,b, Miguel Machado Santos a,b,* 5 a CIMAR/CIIMAR, Interdisciplinary Centre of Marine and Environmental Research, 6 University of Porto, Avenida General Norton de Matos s/n, 4450-208, Matosinhos, Portugal 7 b FCUP-Faculty of Sciences, University of Porto, Rua do Campo Alegre, 4169-007, Porto, 8 Portugal 9 c ICBAS - Institute of Biomedical Sciences Abel Salazar, University of Porto, Rua Jorge de 10 Viterbo Ferreira 228, 4050-313, Porto, Portugal 11 d Department of Analytical Chemistry, Nutrition and Food Sciences, IIAA—Institute for 12 Food Analysis and Research, Universidade de Santiago de Compostela, Constantino 13 Candeira S/N, 15782 Santiago de Compostela, Spain 14 * Corresponding author. CIIMAR, Avenida General Norton de Matos s/n, 4450-208 15 Matosinhos, Portugal. Tel: + 351 22 340 1812 16 E-mail address: [email protected] (M.M. Santos) 17 ABSTRACT 18 . Abbreviations: ACC, acetyl-CoA carboxylase; ACOX, acyl-CoA oxidase; C/EBPα, CCAAT/enhancerbinding protein α ; DBD, DNA binding domain; DEHP, (2-ethylhexyl) phthalate; DGAT2, diacylglycerol acyltransferase; DMSO, dimethylsulfoxide; Dre, Danio rerio; EDC, endocrine disrupting chemical; FAS, fatty acid synthase; HFD, high fat diet; HR, hinge region; Hs, Homo sapiens; LBD, ligand binding domain; LSD, least significant difference; LXR, liver X receptor; MoA, mode of action; MSC, mesenchymal stem cells; NR, nuclear receptor; PCR, polymerase chain reaction; PPAR, peroxisome proliferator-activated receptor; RXR, retinoid X receptor; SD, standard diet; SEM, standard error of the mean; SREBP, sterol regulatory element-binding protein; TBT, tributyltin; WAT, white adipose tissue; WWTP, wastewater treatment plant. This is the postprint (accepted manuscript) version of the article published in Ecotoxicology and Environmental Safety: https://doi.org/10.1016/j.ecoenv.2019.109406 © 2019. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ 2 Obesity, a risk factor for the development of type-2 diabetes, hypertension, cardiovascular 19 disease, hepatic steatosis and some cancers, has been ranked in the top 10 health risk in the 20 world by the World Health Organization. Despite the growing body of literature evidencing 21 an association between the obesity epidemic and specific chemical exposure across a wide 22 range of animal taxa, very few studies assessed the effects of chemical mixtures and 23 environmental samples on lipid homeostasis. Additionally, the mode of action of several 24 chemicals reported to alter lipid homeostasis is still poorly understood. Aiming to fill some 25 of these gaps, we combined an in vivo assay with the model species zebrafish (Danio rerio) 26 to screen lipid accumulation and evaluate expression changes of key genes involved in lipid 27 homeostasis, alongside with an in vitro transactivation assay using human and zebrafish 28 nuclear receptors, retinoid X receptor α and peroxisome proliferator-activated receptor γ. 29 Zebrafish larvae were exposed from 4 day post-fertilization until the end of the experiment 30 (day 18), to six different treatments: experimental control, solvent control, tributyltin at 100 31 ng/L Sn and 200 ng/L Sn (positive control), and wastewater treatment plant influent at 32 1.25% and 2.5%. Exposure to tributyltin and to 2.5% influent led to a significant 33 accumulation of lipids, with white adipose tissue deposits concentrating in the perivisceral 34 area. The highest in vitro tested influent concentration (10%) was able to significantly 35 transactivate the human heterodimer PPARγ/RXRα, thus suggesting the presence in the 36 influent of HsPPARγ/RXRα agonists. 37 Our results demonstrate, for the first time, the ability of complex environmental samples 38 from a municipal waste water treatment plant influent to induce lipid accumulation in 39 zebrafish larvae. 40 Keywords: obesity, obesogen, endocrine disruptor, environmental mixture, nuclear 41 receptor. 42 1. INTRODUCTION 43 3 Obesity, the result of a prolonged disturbance of energy homeostasis favoring triglyceride 44 storage and adipocyte hypertrophy, has been reported to be reaching epidemic proportions, 45 affecting developed and developing countries and increasing prevalence among children 46 and animals (Chamorro-García and Blumberg, 2014; Grün and Blumberg, 2009; Landgraf 47 et al., 2017; Newbold et al., 2007; Minchin & Rawls, 2011; Santos et al., 2012). In addition 48 of being a risk factor for type-2 diabetes, hypertension, cardiovascular disease, hepatic 49 steatosis, premature mortality and some cancers, among other medical conditions (Grün and 50 Blumberg, 2009; Landgraf et al., 2017; Minchin and Rawls, 2011; Newbold et al., 2007), 51 obesity has a detrimental impact on economic development costing billions of dollars 52 annually in direct or indirect care (Spencer & Tilbrook, 2011). Consequently, it is ranked in 53 the top 10 health risks in the world by the World Health Organization, with the number of 54 overweight people exceeding the number of undernourished ones worldwide (Newbold et 55 al., 2007). 56 An increasing body of evidence suggest that this condition does not develop solely from a 57 chronic positive energy balance (high caloric intake versus lack of exercise), but is rather a 58 consequence of complex interactions between genetic and environmental factors (Bašić et 59 al., 2012; Capitão et al., 2017; Chamorro-García and Blumberg, 2014; Dimastrogiovanni et 60 al., 2015; Grün and Blumberg, 2009; Lyssimachou et al., 2015; Minchin and Rawls, 2011; 61 Newbold et al., 2007; Santos et al., 2012; Spencer and Tilbrook, 2011). The disruption of 62 energy homeostasis by endocrine disrupting chemicals (EDCs) is an emerging new field 63 and could explain the rapid increase in obesity rates observed in animals living in close 64 proximity to human populations in industrialized societies (Chamorro-García and 65 Blumberg, 2014; Grün and Blumberg, 2009; Klimentidis et al., 2011; Lyssimachou et al., 66 2015; Newbold et al., 2007), which coincided with the massive release of industrial 67 chemicals over the past decades (Newbold et al., 2007; Santos et al., 2012). 68 4 Lipid homeostasis is controlled by complex networks and feedback mechanisms involving 69 several body organs, transcription factors (e.g. nuclear receptors (NR)), enzymes, hormones 70 and nutrients, with a complex cross-talk between different signaling pathways (Capitão et 71 al., 2017, 2018; Moseti et al., 2016). The central nervous system is the coordinator of all 72 these processes, acting as a lipid and hormone sensor and integrating and transforming 73 afferent information from other organs into signals that will ensure necessary adjustments, 74 for example, by changing feeding behavior or energy expenditure (Lyssimachou et al., 75 2015; Santos et al., 2012). The adipose tissue, besides being one of the main tissues for the 76 occurrence of lipogenesis and where the excess of fatty acids are stored in the form of 77 triglycerides (Kersten, 2001; Lyssimachou et al., 2015; Moseti et al., 2016; Santos et al., 78 2012), functions also as an endocrine and immune organ (Bašić et al., 2012; Birsoy et al., 79 2013; Santos et al., 2012). 80 Members of the peroxisome proliferator-activated receptors (PPARs), transcription factors 81 from the NR superfamily play essential roles in adipocyte differentiation and energy 82 homeostasis, forming obligate heterodimers with the retinoid X receptor (RXR) which 83 changes the transcription of target genes upon the binding of a ligand to the ligand binding 84 domain (LBD) (Janesick and Blumberg, 2011; Santos et al., 2018, 2012; Yoon, 2009). Each 85 subtype possesses different ligand specificity, tissue distributions and biological functions 86 (Santos et al., 2012; Yoon, 2009). 87 Widely known as the “Master regulator of adipogenesis”, due to its absolute requirement 88 for the process, PPARγ is mostly expressed in adipocytes and directly induces genes that 89 regulate the adipogenic cascade, glucose uptake, lipid uptake, synthesis and storage (Birsoy 90 et al., 2013; Janesick and Blumberg, 2011; Lyssimachou et al., 2015; Moseti et al., 2016; 91 Santos et al., 2012). As its ligand-binding pocket is large enough to accept a variety of 92 chemical structures, this NR is particularly sensitive to environmental chemicals (Capitão et 93 al., 2018; Janesick and Blumberg, 2011; Lyssimachou et al., 2015) and since the 94 5 RXR:PPARγ heterodimer is permissive, either receptor pocket can be targeted by ligands 95 and mediate its signaling (Grün & Blumberg, 2009). Moreover, PPARs occur in vertebrate 96 and some non-vertebrate groups (Cephalochordates, Urochordates, Molluscs and 97 Echinoderms) and RXRs in most metazoans, indicating a wide taxonomic scope of the 98 inappropriate regulation of these pathways (Lyssimachou et al., 2015; Santos et al., 2012). 99 Moreover, RXR is the heterodimeric partner of several other nuclear receptors that are 100 involved in lipid homeostasis to some extent [e.g. Liver X Receptor (LXR) and Farnesoid X 101 Receptor] and also form permissive heterodimers (Capitão et al., 2017). 102 Obesogens are a class of EDCs capable of interfering with the lipid metabolic pathways 103 either directly, through interaction with NR leading to changes in patterns of gene 104 expression stimulating adipogenesis and fat accumulation, or indirectly by mimicking or 105 blocking the action of natural hormones and regulating appetite and satiety (Grün and 106 Blumberg, 2007; Janesick and Blumberg, 2011; Lyssimachou et al., 2015; Newbold et al., 107 2007). Yet, several other chemicals may affect lipid homeostasis through mechanisms that 108 do not lead to lipid accumulation (Coimbra et al. 2015; Capitão et al., 2017). Natural or 109 xenobiotics, these chemicals are widely used in everyday products (e.g. food packages, 110 medical devices, toys, cosmetics) (Janesick & Blumberg, 2011), from which they might 111 leach promoting their ubiquitous presence in the environment and threatening not only 112 human populations worldwide, but also the ecosystems (Capitão et al., 2017). 113 The exact MoA of some of these chemicals has not yet been determined, as they may act on 114 the several distinct levels aforementioned. Exposure to lipid homeostasis regulators in utero 115 and during early development can alter developmental programming leading to obesity later 116 in life (Bašić et al., 2012; Grün and Blumberg, 2007; Newbold et al., 2007; Santos et al., 117 2012), as evidenced by mesenchymal stem cells (MSC) with fates biased towards the 118 adipogenic lineage following exposure (Chamorro-Garcia & Blumberg, 2014). Other 119 possible MoA include epigenetic modifications which could modify the expression of genes 120 6 (Bašić et al., 2012; Chamorro-Garcia & Blumberg, 2014; Newbold et al., 2007) and lead to 121 an obesogenic effect at a transgenerational level (Chamorro-Garcia et al., 2017; Chamorro-122 García and Blumberg, 2014). Post-translational modifications, reported to occur in several 123 enzymes and transcription factors involved in lipid metabolism (Carnitine 124 palmitoyltransferase I (Guan et al. 2016), ACC (Capitão et al., 2017; Wang et al., 2015), 125 PPARγ (van Beekum et al., 2009), SREBP (Wang and Sul, 2013)), could affect the 126 interactions of NRs with the transcriptional machinery and the processing, activity, stability 127 and degradations of proteins, as well as their affinity for certain substrates (Guan et al., 128 2016; Janesick and Blumberg, 2011; Lempradl et al., 2015; Wang and Sul, 2013). In 129 addition to the conformational change in the receptors and induction of transcriptional 130 activation through the binding of a ligand, nuclear receptors can also be activated or 131 unrepressed through these post-translational modifications in the absence of ligands 132 (Janesick & Blumberg, 2011). 133 Amongst the 1300 chemicals identified as potential EDCs, very few have been evaluated 134 through lipid homeostasis disruption endpoints (Capitão et al., 2017). So far, identified 135 obesogens include members of the organotins, thiazolidinediones, phthalates, bisphenol-A 136 and its polybrominated and polychlorinated derivatives, organochlorine pesticides, 137 organophosphates, carbamates, solvents and heavy metals such as cadmium and lead, as 138 well as estrogenic chemicals and phytoestrogens (Biemann et al., 2014; Grün and 139 Blumberg, 2007; Janer et al., 2007; Janesick and Blumberg, 2011; Jordão et al., 2015, 2016 140 Newbold et al., 2007; Riu et al., 2014, 2011; Santos et al., 2012; Watt and Schlezinger, 141 2015), but in addition to the lack of information regarding their MoA, data concerning their 142 interactions in mixtures, as they are found in the environment, is scarce (Biemann et al. 143 2014; Capitão et al., 2017). 144 The overall aim of this study was to evaluate the effects of complex environmental samples 145 from a waste water treatment plant (WWTP) on lipid accumulation in zebrafish larvae. As a 146 7 proxy to better understand the underlying MoA of this complex samples, primary 147 approaches were also conducted using transactivation assays with PPARγ/RXRα, alongside 148 with the determination of expression of key lipogenic and adipogenic genes in zebrafish 149 larvae. 150 2. METHODOLOGY 151 2.1. In vivo assay 152 2.1.1. Species selection 153 The zebrafish (Danio rerio), a freshwater cyprinid in the Actinopterygii class, is an 154 emerging model for the study of lipid metabolism and metabolic diseases (Hölttä-Vuori et 155 al., 2010; Landgraf et al., 2017). Besides the widely known advantages of this species (e.g. 156 small size, known ideal growth conditions, large number of eggs, short life cycle), it shows 157 a high conservation of genes, NRs and molecular processes regulating energy homeostasis 158 between mammals and teleosts and the ability to develop metabolic diseases also reported 159 in humans (e.g. hepatic steatosis) (Birsoy et al., 2013; Hill et al., 2005; Hölttä-Vuori et al., 160 2010; Landgraf et al., 2017; Lyssimachou et al., 2015). Studying altered gene expression in 161 response to toxic insults by real-time PCR (qPCR) or genomic screens is facilitated since 162 the zebrafish genome is completely sequenced (Hill et al., 2005; Veldman and Lin, 2008). 163 Another advantage is its optical clarity which allows the screening of lipid accumulations 164 using staining techniques with fluorescent dyes in small larvae without dissecting them. 165 2.1.2. Animal breeding 166 All experiments with living fish were conducted in the facilities of BOGA (Biotério dos 167 Organismos Aquáticos – CIIMAR) and have been approved by the CIIMAR ethical 168 committee and by CIIMAR Managing Animal Welfare Body (ORBEA) according to the 169 European Union Directive 2010/63/EU. 170 8 A breeding stock of adult wild-type zebrafish (Danio rerio; Singapore) was kept in a 250L 171 aquarium with dechlorinated aerated water (28 ± 1°C), under a photoperiod of 14:10h 172 (light:dark). The fish were fed with commercial diet Tetramin (Tetra, Germany) twice a 173 day. Nine females and nine males were placed in cages within a 30L aquarium, the cages 174 possessing a net bottom covered with glass marbles, preventing eggs from being consumed 175 by adults. 176 In the following morning, fertilized eggs were collected, cleaned and randomly allocated to 177 different experimental aquaria. 178 2.1.3. Test conditions 179 The embryos obtained by natural mating were raised in 3L aquaria with dechlorinated water 180 (28 ± 1 °C) with a 14L:10D photoperiod. Each aquarium contained 80 eggs. 181 From hatching, 4th day post-fertilization (dpf) until the end of the experiment (18th dpf), the 182 larvae were exposed to six different treatments in duplicate: experimental control, solvent 183 control (0.0002% dimethylsulfoxide, DMSO), tributyltin (TBT) at 100 ng/L Sn and 200 184 ng/L Sn (positive control), and wastewater treatment plant influent at 1.25% and 2.5%. TBT 185 was selected as a positive control given that Lyssimachou et al., (2015) previously 186 demonstrated massive fat accumulation in zebrafish following a life-cycle exposure to 187 environmentally relevant concentrations. TBT test solutions were obtained by successive 188 dilutions of a stock solution containing TBT chloride (96%, Sigma-Aldrich) in DMSO 189 (99.9%, Sigma-Aldrich). All aquaria had the same DMSO concentration with exception of 190 experimental control. The 24-h composite WWTP influent was obtained at the Porto 191 municipal waste water treatment plant facility in 2017, between 3 and 9 of May, 192 encompassing a full week from Wednesday to Tuesday. Samples were collected daily for 193 seven consecutive days, frozen immediately after collection to prevent degradation of the 194 compounds and used within 1 month for the assays, following a similar sampling design of 195 9 that described by González-Mariño et al., (2017) and Gracia-Lor et al., (2017). This WWTP 196 serves a population of approximately 150.000 people (mean flux of 33000 m3/day) 197 (González-Mariño et al., 2017; Gracia-Lor et al., 2017). The samples were characterized 198 upon collection (mean pH= 7,03; T= 21ºC; Ammonium (mg/L) 47,8). A preliminary 199 characterization of the composite influent was performed by Liquid Chromatography 200 Quadrupole Time-of-flight Tandem Mass Spectrometry (LC-QTOF-MS/MS) with a 201 particular focus on putative PPAR/RXR agonists as detailed in the Supporting Information 202 Table S1 (Appendix A). Detected compounds were classified in confirmatory levels as 203 proposed by (Schymanski et al., 2014). 204 The medium was renewed daily, ammonia levels checked weekly using Palintest tablets and 205 Palintest Photometer 7000se (0.3625 ± 0.033 mg/L N and 0.4375 ± 0.0369 mg/L N for the 206 1.25% and 2.5% WWTP influent, respectively, and 0.1425 ± 0.0347 mg/L N for all 207 remaining treatments), while the temperature was monitored daily. Values of mortality were 208 registered daily and the dead larvae/eggs were removed. Total mortality did not differ 209 among groups (data not shown) and was maintained below 20% up to the end of the assays, 210 which is within the normal range of the species in laboratory assays (Coimbra et al., 2015). 211 A light aeration was introduced to the aquaria on the 11th dpf. 212 Throughout the whole experiment, care was taken in order to avoid the suffering and 213 distress of the animals. 214 2.1.4. Feeding protocol 215 From 5 to 14 dpf larvae were fed with a standard diet (SD), Gemma Micro (Skretting), 216 supplemented with cod liver oil (Fagron) to obtain a final lipid content of 25% dry matter. 217 Artemia spp. cysts (Brine Shrimp Eggs; Ocean Nutrition) put to hatch for 16 hours were 218 supplied every other day from 7 to 13 dpf (0.2 grams fed 3 aquaria). From the 15th to the 219 16th dpf, larvae were fed lyophilized chicken egg yolk as a high fat diet (HFD), followed by 220 16 as average fold changes ± SEM (N=7) of the solvent control and experimental control 338 groups, respectively. *p < 0.05 (one-way ANOVA, followed by Fisher LSD’s test). 339 A statistically significant increase (p < 0.05) in fluorescence was detected in larvae 340 chronically exposed to the environmentally-relevant concentrations of 100 and 200 ng/L Sn 341 TBT and 2.5% WWTP influent, with WAT deposits concentrating in the perivisceral area. 342 Whereas the subcutaneous fat accumulation is considered an adaptive process, exceeding 343 the storage capacity of this adipose depot, diverts the excess of fatty acids towards the 344 accumulation in the visceral cavity, muscle and liver, primary locations for correlations 345 with metabolic disorders and increased mortality (Chamorro-Garcia and Blumberg, 2014; 346 Grün and Blumberg, 2009; Lyssimachou et al., 2015). 347 Preliminary assays performed by our group with regular diet (Gemma Micro, 59% 348 protein, 19% oil) did not lead to an increase in fat accumulation in TBT exposed larvae 349 (data not shown). Hence, the assays were performed with a SD supplied with additional 350 lipids (25% of lipid content). Tingaud-Sequeira and colleagues (2011), reared zebrafish 351 larvae on a standard diet and separated them into two groups which were fed for a day with 352 different diets, SD or hard-boiled egg yolk as an HFD, followed by two days of starvation. 353 During starvation, the initial adiposity was recorded and larvae were posteriorly exposed for 354 24h to 50 nM TBT and 1 nM rosiglitazone, with a final adiposity recording. The results 355 indicated that HFD led to a more dramatic increase in lipid adiposity when the fishes were 356 exposed, being more evident in the treatment with rosiglitazone in which, despite the fasting 357 state, an increase in adiposity was obtained in the HFD group, oppositely to the smaller 358 decrease recorded for the SD group when compared to the control. The present study 359 corroborates previous findings of short-term exposures to higher TBT concentrations 360 (Tingaud-Sequeira et al., 2011) and to life-cycle zebrafish exposure to 10 and 50 ng/TBT 361 (Lyssimachou et al., 2015). Similarly, zebrafish larvae exposed for 15 days to 362 tetrabromobisphenol A (TBBPA), tetrachlorobisphenol A, TBBPA-sulfate or TBT, and fed 363 17 with egg yolk, exhibited significant increases in lipid accumulation (Riu et al., 2014). The 364 authors also pointed out that when the larvae were fed with regular diets, instead of HFD, 365 none of the treatments resulted in the staining of larvae, highlighting the impact of the 366 synergy of both calories and chemical exposure in lipid accumulation. 367 As fat build-up results from the balance between lipogenesis and lipolysis (Kersten, 368 2001), this study assessed the impact of TBT and the WWTP influent on both lipogenesis 369 and adipogenesis focusing in particular in PPARγ:RXRα associated pathways 370 (i.e.,PPARγ:RXRα, SREBP1, C/EBPα, FASn, DGAT2) and also fatty acid oxidation 371 (ACOX1). Bodies of zebrafish larvae from the in vivo assay were analyzed by qPCR to 372 determine whether the lipid accumulation verified in vivo was accompanied by changes in 373 levels of mRNA of key transcription factors and enzymes ruling lipid homeostasis. 374 Interestingly, the PCR results (Figure 3) only reveal statistically significant differences 375 in the expression of the transcription factor SREBP1 at 100 ng/L Sn TBT and RXRα 376 alongside DGAT2 at 200 ng/L Sn TBT, environmental concentrations that triggered 377 significant lipid accumulations in vivo. Both the increase of SREBP1 levels at the lowest 378 concentration and decrease of RXRα expression levels at the highest concentration, suggest 379 a mechanism of negative feedback, whereas the overexpression of DGAT2, the enzyme that 380 catalyzes the final reaction in triglyceride’s synthesis (Lyssimachou et al., 2015), suggests 381 that lipid production has not ceased at the highest concentration of TBT. While the lack of 382 altered significant gene expression for 1.25% influent matches the observed lack of lipid 383 accumulation, the similar observation for the 2.5% concentration, which contrastingly 384 induced significant accumulation, could also be due to a negative feedback mechanism. 385 This could explain the 3-fold decrease in gene expression of C/EBPα in larvae exposed to 386 both WWTP influent treatments. 387 18 388 389 Figure 3. In vivo induction of lipogenic and adipogenic genes in zebrafish, following 390 exposure to TBT and WWTP influents. Values normalized to β-actin and relative to an 391 internal calibrator are expressed as the average fold changes ± SEM of the solvent control 392 and experimental control groups for TBT and the WWTP influents, respectively. *p < 0.05 393 (one-way ANOVA, followed by Fisher LSD’s test). 394 Similarly, Watt and Schlezinger (2015) reported lipid accumulations in bone marrow 395 MSCs of male 9-week-old C57BL/6J mice following exposure to PPARγ-agonists 396 rosiglitazone, TBT and triphenyltin, with no statistical significant increases in PPARγ 397 mRNA, although downstream target genes were up-regulated (Watt & Schlezinger, 2015). 398 Whereas PPARγ plays an important role in lipogenesis and adipogenesis in the adipose 399 tissue, the levels of expression in other tissues are low, for example in the liver where the 400 main mediator of the expression of lipogenic genes is SREBP-1 (Kersten, 2001). Thus, it is 401 not surprising that several studies report organ-specific and sex-specific effects of exposure 402 to TBT and other lipid homeostasis regulators in a variety of model animals (Bian et al., 403 19 2017; Chamorro-García et al., 2013; Guan et al., 2016; Lyssimachou et al., 2015; Zhang et 404 al., 2012). 405 Nevertheless, the transcription factor SREBP1, which was overexpressed at 100 ng/L 406 Sn TBT, is an early regulator of adipogenesis and stimulates PPARγ expression for the final 407 adipocyte maturation where they will modulate the expression of adipogenic and lipogenic 408 enzymes, such as ACCα and FASn (Birsoy et al., 2013; Capitão et al., 2017; Kersten, 2001; 409 Lyssimachou et al., 201; Moseti et al., 2016; Santos et al., 2012). 410 3.2. Transactivation assays 411 To further address if the WWTP influent and TBT could transactivate or repress the 412 PPARγ/RXRα signaling pathways, we have performed transactivation assays using human 413 (Hs) and zebrafish (Dre) PPARγ and RXRα receptors, as well as with the RXRα:PPARγ 414 heterodimer. The results indicate that both influent and TBT are able to interact with the 415 receptors in vitro, which might possibly be one of the mechanisms underlying the observed 416 lipid accumulation and disruption of lipid homeostasis in vivo. The organotin TBT was the 417 first obesogen to be associated with RXR and PPARγ signaling (Janesick and Blumberg, 418 2011; Santos et al., 2012). Indeed, our positive control TBT transactivated HsRXRα, 419 PPARγ and heterodimer at 250 nM (Figure 4). Interestingly, the results for the zebrafish 420 receptors (Figure 5) show a different response when compared to HsPPARγ. While 421 DreRXRα activity was significantly induced and PPARγ not affected by TBT, the 422 heterodimer was massively repressed at the tested concentration. 423 20 424 425 Figure 4. In vitro transactivation of Homo sapiens receptors by 250 nM TBT and WWTP 426 influent. The activity of firefly luciferase (Photinus pyralis) was normalized to the activity 427 of Renilla luciferase (Renilla reniformis) and then expressed as average fold changes ± SEM 428 of the solvent control group. *p < 0.05 (one-way ANOVA, followed by Fisher’s LSD test). 429 430 431 21 Figure 5. In vitro transactivation of Danio rerio receptors by 250 nM TBT and WWTP 432 influent. The activity of firefly luciferase (Photinus pyralis) was normalized to the activity 433 of Renilla luciferase (Renilla reniformis) and then expressed as average fold changes ± SEM 434 of the solvent control group. *p < 0.05 (one-way ANOVA, followed by Fisher’s LSD test). 435 Analyzing the amino acid sequences of human and zebrafish receptors, it was found a 436 high percentage of similarities (99%) in both RXRα DNA-binding domain (DBD) and 437 LBD, oppositely to the 94% similarities at the PPARγ DBD and 74% at the LBD (Ouadah-438 Boussouf & Babin, 2016; Zhao et al., 2015), suggesting different ligand-binding affinities 439 for PPARγ. Furthermore, the cysteine in position 285 of HsPPARγ LBD, which is mutated 440 into a tyrosine in the zebrafish (Capitão et al., 2017), has been shown as essential for the 441 activation of this receptor by TBT (Harada et al., 2015). This mutation could underlie the 442 disparity between the transactivation results, as differences in nuclear receptors’ structures 443 may lead to distinct chemical binding and activation outcomes (Capitão et al., 2017). The 444 underlying mechanism of this finding has been recently discussed by Capitão et al., (2018). 445 Indeed, Riu and collaborators have also noted that known HsPPARγ agonists 446 thiazolidinediones (rosiglitazone, pioglitazone, ciglitazone and troglitazone), TBT and a 447 prostaglandin could not be used as positive controls in assays involving DrePPARγ, as they 448 failed to activate the receptor in vitro (Riu et al., 2014). 449 The lack of DrePPARγ transactivation in comparison to RXRα suggest that the TBT’s 450 obesogenic effects reported here might be mediated by RXRα instead. In fact, this chemical 451 was also reported to interact with other permissive heterodimers (LXR:RXRα heterodimer) 452 (Capitão et al., 2017; Le Maire et al., 2009) and it has been suggested that the covalent 453 binding to HsRXRα’s Cys432 would allow a more efficient stabilization of the active 454 receptor conformation, than its non-covalent ionic bond to HsPPARγ’s Cys285 (Harada et 455 al., 2015; Le Maire et al., 2009). Ouadah-Boussouf and Babin (2016) demonstrated that the 456 22 obesogenic effect of TBT on DreRXR:PPARγ was not inhibited by the HsPPARγ 457 antagonist T0070907, but was fully abolished by the HsRXR antagonist UVI3003 and by a 458 combination of HsRXR:RXR and HsRXR:PPARγ antagonists. The findings of Ouadah-459 Boussouf and Babin (2016) support that RXR plays a central role in TBT-modulated 460 obesogenic outcomes and that in vivo obesogenic effects might occur through RXR-461 dependent pathways which do not necessarily involve PPAR isoforms (Ouadah-Boussouf & 462 Babin, 2016). 463 TBT may modulate other NRs involved in lipid homeostasis such as LXR:RXR 464 heterodimer (Capitão et al., 2017; Le Maire et al., 2009). Interestingly, LXR is tightly 465 related to cholesterol homeostasis and regulates the expression of SREBP-1 (Capitão et al., 466 2017), one of the upregulated target genes after exposure to TBT at 100 ng/L Sn. 467 Regardless, unveiling the crystallographic structure of TBT binding to the human RXR and 468 PPARγ has allowed a better understanding of the interaction with these receptors. Similar 469 studies should be performed with other chemicals and nuclear receptors to help elucidate 470 the different species-responses to a particular lipid homeostasis regulator. 471 Regarding the environmental samples, only the highest tested influent concentration (10%) 472 was able to significantly transactivate the heterodimer HsPPARγ/RXRα, thus indicating the 473 presence in the influent of HsPPARγ/RXRα agonists (Figure 4). It is important to have in 474 mind that the tested concentrations of 1.25% and 2.5% were used when zebrafish larvae 475 were chronically exposed from the 4th to the 18th dpf in vivo. Thus, the results are not easily 476 comparable, as transactivation assays comprise a short 24-hour window of exposure, which 477 could explain the lack of transactivation at these lower concentrations. In DrePPARγ a 1.5-478 fold induction was observed for the 10% influent concentrations, although differences did 479 reach significance in comparison with control (p = 0.089). 480 A simultaneous exposure to a “cocktail” of chemicals can potentially lead to additive or 481 even synergistic effects (e.g. by acting on both RXR and PPARγ simultaneously) (Biemann 482 23 et al., 2014; Riu et al., 2011) but also to antagonistic effects. Since several compounds have 483 also been shown to repress the transactivation of nuclear receptors involved in lipid 484 homeostasis, the possibility of antagonistic actions of different chemicals on the nuclear 485 receptors cannot be disregarded (Biemann et al., 2014). Indeed, the chemical 486 characterization of the influent used in this study (Appendix A, Table S1) shows the 487 presence of a variety of chemicals that have been reported to impact lipid homeostasis 488 triggering lipid storage, i.e., (diclofenac (Puhl et al., 2015), ibuprofen (Puhl et al., 2015), 489 dibutyl phthalate (Mjaeed et al., 2017), methyl paraben (Hu et al., 2016), acesulfame (Bian 490 et al., 2017), saccharin (Suez et al., 2015)) or lipid mobilization, i.e., (metformin (Park et 491 al., 2009), piperine (Shah et al., 2011), irbesartan (Parhofer et al., 2007), valsartan (Cole et 492 al., 2010), chenodiol (Chen et al., 2017), hippuric acid (Zhao et al., 2016), fenofibric acid 493 (Alagona, 2010)). Some of the compounds present in the influent have been reported to 494 interact with the PPAR/RXR receptors, either through induction and/or agonism [PPARγ: 495 acesulfame (Simon et al., 2013), saccharin (Simon et al., 2013), irbesartan (Afzal et al., 496 2016), valsartan (Hasan et al., 2014; Storka et al., 2008), diclofenac (Ayoub et al., 2009; 497 Puhl et al., 2015), ibuprofen (Puhl et al., 2015), dibutyl phthalate (Lapinskas et al., 2005; 498 Pereira-Fernandes et al., 2013), metformin (Elia et al., 2011); methyl paraben (Pereira-499 Fernandes et al., 2013); PPARα: metformin (Maida et al., 2011); dibutyl phthalate 500 (Lapinskas et al., 2005), irbesartan (Rong et al., 2010); fenofibric acid (Qiu et al., 2017); 501 Mouse PPARα: triclosan (Wu et al., 2014);] or suppression and/or antagonism [PPARγ: 502 piperine (Park et al., 2012); PPARα: chenodiol (Chen et al., 2017); human PPARα: 503 triclosan (Wu et al., 2014)]. 504 The preliminary mechanistic approach performed here to address the involvement of the 505 PPARγ/ RXRα in the lipid accumulation observed in vivo in the 2.5% influent exposed 506 larvae is not fully conclusive. Even though the chemical analysis of our environmental 507 24 samples confirms the presence of known PPARγ agonists and antagonists, it seems likely 508 that other mechanisms might have also directed the in vivo effects observed. 509 So far, very few studies reported the impact of chemical mixtures and environmental 510 samples on obesity and metabolic endpoints (Adeogun et al., 2016; Biemann et al., 2014; 511 Lyche et al., 2013, 2011; Maisano et al., 2016), which is concerning, as only the exposure 512 to mixtures depict real-case scenarios and the results are not always predictable by the 513 effects of the individual compounds alone (Biemann et al., 2014; Rodrigues et al., 2006; 514 Santos et al., 2006, 2008; Sarria et al., 2011). In addition to organotins, other groups of 515 compounds have been investigated for their variety of MoA on obesity outcomes and lipid 516 homeostasis disruption: primary phthalates, such as (2-ethylhexyl) phthalate (DEHP), 517 activate PPARα, favoring a lipolysis state, whereas their metabolites show preference for 518 PPARγ (Grün & Blumberg, 2009); Halogenated Bisphenol A derivatives have been shown 519 to disrupt the thyroid hormone receptor, estrogen receptors and the PPARγ signaling (Bašić 520 et al., 2012; Riu et al., 2011). As lipogenesis and adipogenesis involve a complex network 521 of transcription factors and enzymes, the disruption of the energy homeostasis can occur at 522 many levels (Bašić et al., 2012). Weighting these different factors together, through the 523 study of complex samples of chemicals, is important to obtain a more realistic assessment 524 of the impact of these mixtures on animal and human health. Interestingly, epidemiological 525 and toxicological studies show a moderate to strong probability of human obesity being 526 associated with exposure to DDE, Di-2-ethylhexylphthalate and BPA (Trasande et al., 527 2015). 528 4. CONCLUSION 529 The present study demonstrates, for the first time, obesogenic responses in zebrafish 530 larvae following exposure to complex environmental samples reported to contain an array 531 of chemicals known to modulate lipid homeostasis. Yet, the mechanistic studies performed 532 here are not fully conclusive. Hence, future studies should expand the screening to 533 25 additional nuclear receptors known to be engaged in vertebrate’s lipid homeostasis and the 534 study of pathway’s other than the ones regulated by PPARγ/RXRα. 535 DECLARATION OF INTEREST 536 Conflicts of interest: none. 537 538 ACKNOWLEDGMENTS 539 This work was supported by the Norte2020 and FEDER (Coral—Sustainable Ocean 540 Exploitation—Norte-01-0145-FEDER-000036), the Spanish Agencia Estatal de 541 Investigación (CTM2017-84763-C3-2-R), the Galician Council of Culture, Education and 542 Universities (ED431C2017/36) and FEDER/ERDF. 543 544 REFERENCES 545 Adeogun, A.O., Ibor, O.R., Regoli, F., Arukwe, A., 2016. Peroxisome proliferator-546 activated receptors and biotransformation responses in relation to condition factor 547 and contaminant burden in tilapia species from Ogun River, Nigeria. Comp. 548 Biochem. Physiol. Part - C Toxicol. Pharmacol. 183–184, 7–19. 549 https://doi.org/10.1016/j.cbpc.2015.12.006 550 Afzal, S., Sattar, M.A., Johns, E.J., Abdulla, M.H., Akhtar, S., Hashmi, F., Abdullah, 551 N.A., 2016. Interaction between irbesartan, peroxisome proliferator-activated 552 receptor (PPAR-γ), and adiponectin in the regulation of blood pressure and renal 553 function in spontaneously hypertensive rats. J. Physiol. Biochem. 72, 593–604. 554 https://doi.org/10.1007/s13105-016-0497-1 555 Alagona, P., 2010. Fenofibric acid: A new fibrate approved for use in combination with 556 statin for the treatment of mixed dyslipidemia. Vasc. Health Risk Manag. 6, 351–557 362. https://doi.org/10.2147/VHRM.S6714 558 André, A., Ruivo, R., Capitão, A., Froufe, E., Páscoa, I., Castro, L.F.C., Santos, M.M., 559 32 “obesogen” tributyltin targets hepatic triglyceride accumulation and the 710 transcriptional regulation of lipid metabolism in the liver and brain of zebrafish. 711 PLoS One 10, 1–22. https://doi.org/10.1371/journal.pone.0143911 712 Maida, A., Lamont, B.J., Cao, X., Drucker, D.J., 2011. Metformin regulates the incretin 713 receptor axis via a pathway dependent on peroxisome proliferator-activated 714 receptor-α in mice. Diabetologia 54, 339–349. https://doi.org/10.1007/s00125-010-715 1937-z 716 Maisano, M., Cappello, T., Oliva, S., Natalotto, A., Giannetto, A., Parrino, V., 717 Battaglia, P., Romeo, T., Salvo, A., Spanò, N., Mauceri, A., 2016. PCB and OCP 718 accumulation and evidence of hepatic alteration in the Atlantic bluefin tuna, T. 719 thynnus, from the Mediterranean Sea. Mar. Environ. Res. 121, 40–48. 720 https://doi.org/10.1016/j.marenvres.2016.03.003 721 Minchin, J.E.N., Rawls, J.F., 2011. In vivo Analysis of White Adipose Tissue in 722 Zebrafish, in: Detrich III, W.H., Westerfield, M., Zon, L.I., (Eds.), The Zebrafish: 723 Disease Models and Chemical Screens. Methods Cell Biol. 105, 63–86. 724 https://doi.org/10.1016/B978-0-12-381320-6.00003-5 725 Mjaeed, K.A., ur Rehman, H., Yousaf, M.S., Zaneb, H., Rabbani, I., Tahir, S.K., 726 Rashid, M.A., 2017. Sub-chronic exposure to low concentration of dibutyl 727 phthalate affects anthropometric parameters and markers of obesity in rats. 728 Environ. Sci. Pollut. Res. 24, 25462–25467. https://doi.org/10.1007/s11356-017-729 9952-y 730 Moseti, D., Regassa, A., Kim, W.K., 2016. Molecular regulation of adipogenesis and 731 potential anti-adipogenic bioactive molecules. Int. J. Mol. Sci. 17, 1–24. 732 https://doi.org/10.3390/ijms17010124 733 Newbold, R.R., Padilla-Banks, E., Snyder, R.J., Phillips, T.M., Jefferson, W.N., 2007. 734 33 Developmental Exposure to Endocrine Disruptors and the Obesity Epidemic. 735 Reprod. Toxicol. 23, 290–296. https://doi.org/10.1016/j.reprotox.2006.12.010 736 Ouadah-Boussouf, N., Babin, P.J., 2016. Pharmacological evaluation of the mechanisms 737 involved in increased adiposity in zebrafish triggered by the environmental 738 contaminant tributyltin. Toxicol. Appl. Pharmacol. 294, 32–42. 739 https://doi.org/10.1016/j.taap.2016.01.014 740 Parhofer, K.G., Münzel, F., Krekler, M., 2007. Effect of the angiotensin receptor 741 blocker irbesartan on metabolic parameters in clinical practice: The DO-IT 742 prospective observational study. Cardiovasc. Diabetol. 6, 1–6. 743 https://doi.org/10.1186/1475-2840-6-36 744 Park, U.H., Jeong, H.S., Jo, E.Y., Park, T., Yoon, S.K., Kim, E.J., Jeong, J.C., Um, S.J., 745 2012. Piperine, a component of black pepper, inhibits adipogenesis by 746 antagonizing PPARγ activity in 3T3-L1 cells. J. Agric. Food Chem. 60, 3853–747 3860. https://doi.org/10.1021/jf204514a 748 Pereira-Fernandes, A., Demaegdt, H., Vandermeiren, K., Hectors, T.L.M., Jorens, P.G., 749 Blust, R., Vanparys, C., 2013. Evaluation of a Screening System for Obesogenic 750 Compounds: Screening of Endocrine Disrupting Compounds and Evaluation of the 751 PPAR Dependency of the Effect. PLoS One 8, 1–17. 752 https://doi.org/10.1371/journal.pone.0077481 753 Puhl, A.C., Milton, F.A., Cvoro, A., Sieglaff, D.H., Campos, J.C.L., Bernardes, A., 754 Filgueira, C.S., Lindemann, J.L., Deng, T., Neves, F.A.R., Polikarpov, I., Webb, 755 P., 2015. Mechanisms of peroxisome proliferator activated receptor γ regulation by 756 non-steroidal anti-inflammatory drugs. Nucl. Recept. Signal. 13, 1–17. 757 https://doi.org/10.1621/nrs.13004 758 Qiu, F., Matlock, G., Chen, Q., Zhou, K., Du, Y., Wang, X., Ma, J.X., 2017. 759 34 Therapeutic effects of PPARα agonist on ocular neovascularization in models 760 recapitulating neovascular age-related macular degeneration. Investig. Ophthalmol. 761 Vis. Sci. 58, 5065–5075. https://doi.org/10.1167/iovs.17-22091 762 Riu, A., Grimaldi, M., le Maire, A., Bey, G., Phillips, K., Boulahtouf, A., Perdu, E., 763 Zalko, D., Bourguet, W., Balaguer, P., 2011. Peroxisome proliferator-activated 764 receptor γ is a target for halogenated analogs of bisphenol A. Environ. Health 765 Perspect. 119, 1227–1232. https://doi.org/10.1289/ehp.1003328 766 Riu, A., McCollum, C.W., Pinto, C.L., Grimaldi, M., Hillenweck, A., Perdu, E., Zalko, 767 D., Bernard, L., Laudet, V., Balaguer, P., Bondesson, M., Gustafsson, J.A., 2014. 768 Halogenated bisphenol-a analogs act as obesogens in zebrafish larvae (Danio 769 rerio). Toxicol. Sci. 139, 48–58. https://doi.org/10.1093/toxsci/kfu036 770 Rodrigues, P., Reis-Henriques, M.A., Campos, J., Santos, M.M., 2006.Urogenital 771 Papilla feminization in male Pomatoschistus minutus from two estuaries in 772 northwestern Iberian Peninsula. Mar. Environ. Res. 62 (1): S258-S262. 773 https://doi.org/10.1016/j.marenvres.2006.04.032 774 Rong, X., Li, Y., Ebihara, K., Zhao, M., Kusakabe, T., Tomita, T., Murray, M., Nakao, 775 K., 2010. Irbesartan treatment up-regulates hepatic expression of PPAR and its 776 target genes in obese Koletsky (fak/fak) rats: A link to amelioration of 777 hypertriglyceridaemia. Br. J. Pharmacol. 160, 1796–1807. 778 https://doi.org/10.1111/j.1476-5381.2010.00835.x 779 Santos, M.M., Micael, J., Carvalho, A.P., Morabito, R., Booy, P., Massanisso, P., 780 Lamoree, M., Reis-Henriques, M.A., 2006. Estrogens counteract the masculinizing 781 effect of tributyltin in zebrafish. Comp Biochem Physiol C 142, 151-155. 782 https://doi.org/10.1016/j.cbpc.2005.11.014 783 Santos, M.M., M.A. Reis-Henriques, R. Guillot, D. Lima, R. Franco-Duarte, I. Mendes, 784 35 S. Queirós, L. Filipe C. Castro, 2008. Anti-androgenic effects of sewage treatment 785 plant effluents in the prosobranch gastropod Nucella lapillus. Comp Biochem 786 Physiol C 148, 87-93. https://doi.org/10.1016/j.cbpc.2008.03.012 787 Santos, M.M., Reis-Henriques, M.A., Castro, L.F.C., 2012. Lipid homeostasis 788 perturbation by organotins: effects on vertebrates and invertebrates, in: Pagliarani, 789 A., Trombetti, F., Ventrella, V. (Eds.), Biochemical and Biological Effects of 790 Organotins. Bentham Science Publishers, Sharjah, 83-96. 791 https://doi.org/10.2174/978160805265311201010083 792 Santos, M.M., Ruivo, R., Capitão, A., Fonseca, E., Castro, L.F.C., 2018. Identifying the 793 gaps: Resources and perspectives on the use of nuclear receptor based-assays to 794 improve hazard assessment of emerging contaminants. J. Hazard. Mater. 358, 508–795 511. https://doi.org/10.1016/j.jhazmat.2018.04.076 796 Sárria, M.P., Santos, M.M., Reis-Henriques, M.A., Vieira, N.M., Monteiro, N.M., 2011. 797 The unpredictable effects of mixtures of androgenic and estrogenic chemicals on 798 fish early life. Environ. Int. 37(2):418-24. 799 https://doi.org/10.1016/j.envint.2010.11.004 800 Simon, B.R., Parlee, S.D., Learman, B.S., Mori, H., Scheller, E.L., Cawthorn, W.P., 801 Ning, X., Gallagher, K., Tyrberg, B., Assadi-Porter, F.M., Evans, C.R., 802 MacDougald, O.A., 2013. Artificial sweeteners stimulate adipogenesis and 803 suppress lipolysis independently of sweet taste receptors. J. Biol. Chem. 288, 804 32475–32489. https://doi.org/10.1074/jbc.M113.514034 805 Schymanski, E.L., Jeon, J., Gulde, R., Fenner, K., Ruff, M., Singer, H.P., Hollender, J., 806 2014. Identifying Small Molecules via High Resolution Mass Spectrometry: 807 Communicating Confidence Environ. Sci. Technol. 48, 2097–2098. 808 https://pubs.acs.org/doi/abs/10.1021/es5002105 809 36 Spencer, S.J., Tilbrook, A., 2011. The glucocorticoid contribution to obesity. Stress 14, 810 233–246. https://doi.org/10.3109/10253890.2010.534831 811 Storka, A., Vojtassakova, E., Mueller, M., Kapiotis, S., Haider, D.G., Jungbauer, A., 812 Wolzt, M., 2008. Angiotensin inhibition stimulates PPARγ and the release of 813 visfatin. Eur. J. Clin. Invest. 38, 820–826. https://doi.org/10.1111/j.1365-814 2362.2008.02025.x 815 Suez, J., Korem, T., Zilberman-Schapira, G., Segal, E., Elinav, E., 2015. Non-caloric 816 artificial sweeteners and the microbiome: Findings and challenges. Gut Microbes 817 6, 149–155. https://doi.org/10.1080/19490976.2015.1017700 818 Tingaud-Sequeira, A., Ouadah, N., Babin, P.J., 2011. Zebrafish obesogenic test: a tool 819 for screening molecules that target adiposity. J. Lipid Res. 52, 1765–1772. 820 https://doi.org/10.1194/jlr.D017012 821 Trasande, L., Zoeller, R.T., Hass, U., Kortenkamp, A., Grandjean, P., Myers, J.P., 822 Digangi, J., Bellanger, M., Hauser, R., Legler, J., Skakkebaek, N.E., Heindel, J.J., 823 2015. Estimating burden and disease costs of exposure to endocrine-disrupting 824 chemicals in the European Union. J. Clin. Endocrinol. Metab. 100, 1245–1255. 825 https://doi.org/10.1210/jc.2014-4324 826 van Beekum, O., Fleskens, V., Kalkhoven, E., 2009. Posttranslational Modifications of 827 PPAR-γ: Fine-tuning the Metabolic Master Regulator. Obesity 17, 213–219. 828 https://doi.org/10.1038/oby.2008.473 829 Veldman, M.B., Lin, S., 2008. Zebrafish as a developmental model organism for 830 pediatric research. Pediatr. Res. 64, 470–476. 831 https://doi.org/10.1203/PDR.0b013e318186e609 832 Wang, Y., Sul, H.S., 2013. Transcriptional regulation of lipogenesis and its contribution 833 to hepatosteatosis. Clin. Lipidol. 8, 165–168. https://doi.org/10.2217/clp.13.11 834 37 Wang, Y., Viscarra, J., Kim, S.J., Sul, H.S., 2015. Transcriptional regulation of hepatic 835 lipogenesis. Nat. Rev. Mol. Cell Biol. 16, 678–689. 836 https://doi.org/10.1038/nrm4074 837 Watt, J., Schlezinger, J.J., 2015. Structurally-diverse, PPARγ-activating environmental 838 toxicants induce adipogenesis and suppress osteogenesis in bone marrow 839 mesenchymal stromal cells. Toxicology 331, 66–77. 840 https://doi.org/10.1016/j.tox.2015.03.006 841 Wu, Y., Wu, Q., Beland, F.A., Ge, P., Manjanatha, M.G., Fang, J.L., 2014. Differential 842 effects of triclosan on the activation of mouse and human peroxisome proliferator-843 activated receptor alpha. Toxicol. Lett. 231, 17–28. 844 https://doi.org/10.1016/j.toxlet.2014.09.001 845 Yoon, M., 2009. The role of PPARα in lipid metabolism and obesity: Focusing on the 846 effects of estrogen on PPARα actions. Pharmacol. Res. 60, 151–159. 847 https://doi.org/10.1016/j.phrs.2009.02.004 848 Zhang, W., Zhang, Y., Zhang, H., Wang, J., Cui, R., Dai, J., 2012. Sex Differences in 849 Transcriptional Expression of FABPs in Zebrafish Liver after Chronic 850 Perfluorononanoic Acid Exposure. Environ. Sci. Technol. 46, 5175–5182. 851 https://doi.org/10.1021/es300147w 852 Zhao, H., Shen, J., Djukovic, D., Daniel-MacDougall, C., Gu, H., Wu, X., Chow, W.H., 853 2016. Metabolomics-identified metabolites associated with body mass index and 854 prospective weight gain among Mexican American women. Obes. Sci. Pract. 2, 855 309–317. https://doi.org/10.1002/osp4.63 856 Zhao, Y., Zhang, K., Giesy, J.P., Hu, J., 2015. Families of Nuclear Receptors in 857 Vertebrate Models: Characteristic and Comparative Toxicological Perspective. Sci. 858 Rep. 5, 8554. https://doi.org/10.1038/srep08554 859 38 860 861 862 863 864 Linking chemical exposure to lipid homeostasis: a municipal waste water treatment plant influent is obesogenic for zebrafish larvae Mélanie Audrey Barbosa a,b,c, Ricardo Capela, Jorge Rodolfo a,c, Elza Fonseca a, Rosa Montes d, Ana André a, Ana Capitão a, António Paulo Carvalho a,c, José Benito Quintana d, L. Filipe C. Castro a,b, Miguel Machado Santos a,b,* ANNEX A LC-QTOF-MS analysis of the influent wastewater The wastewater sample was submitted to a solid-phase extraction (SPE) protocol prior to their analysis by Liquid Chromatography Quadrupole Time-of-flight Tandem Mass Spectrometry (LCQTOF-MS/MS). Thus, 100 mL of the sample were passed through two 150 mg cartridges OASIS WCX and OASIS WAX (Waters, Milford, MA, USA) connected in series. The cartridges were previously conditioned following the manufacturer’s specifications. After sample loading, the cartridges were dried under nitrogen stream for 20 minutes and eluted with 4 consecutive fractions consisting of 1) 10 mL of methanol, 2) 10 mL of methanol containing 5 % of ammonia, 3) 10 mL of methanol containing 2% of formic acid and 4) 10 mL of ethyl acetate. All these fractions were evaporated to dryness and reconstituted in 100 µL of methanol. The final extracts were injected (10 µL) into a LCQTOF-MS 6520 system (Agilent Technologies, Santa Clara, USA). The chromatographic column was a Synergy Fusion-RP 100x2mm, 4µm (Phenomenex, Torrance, CA, USA) and the mobile phases consisted of ultrapure water (A) and methanol (B) both containing 0.1% formic acid. After the injection, the column was maintained at 95% A during 1 min, then, a linear gradient of 15 minutes was programmed from 95% A to 100% B (hold for 5 min) and the system was maintained at initial conditions again for other 10 min before the following injection. The QTOF-MS system worked in the “All Ions” acquisition mode using a collision energy of either 0 or 20 V (as to obtain the pseudomolecular ion and product ions), in both electrospray (ESI) positive and negative modes. Nitrogen was used as nebulizing (40 psi) and drying gas (300 °C, 5 L min–1) in the dual ESI source and also as collision gas in the MS/MS and All Ions experiments. The QTOF instrument was operated in the high resolution 4 GHz mode. This mode provides a full width at half-maximum (FWHM) resolution of ca. 10 600 at m/z 118.0862 and ca. 16 900 at m/z 922.0097. A reference calibration solution, supplied by Agilent, was continuously sprayed in the source during the chromatographic run, providing the required accuracy of mass assignations, according to the manufacturer instructions. The MS data treatment was made with the MassHunter B.07.01 software using the algorithm “Find by Formula” with fragment confirmation against the Agilent Forensic Toxicology, Veterinary Drugs, Pesticides and Water Pollutants Libraries, which all together contain ca. 11,000 chemicals, of which about 3,500 have MS/MS spectra. Tentatively identified compounds were reinjected in MS/MS mode at 10, 20 and 40 V collision energies for further confirmation. Blank samples of the whole analytical process were also processed and chemicals were only reported if they were detected in at least two replicate analysis of the sample and its signal was 5 times higher than the blank. The identified compounds are shown in Table S1, classified in confirmatory levels, according to Schymanski et al., 2014. Only those compounds at a high enough confirmation level (level 2b or higher) are reported. The presence of twelve out of 20 analytes was confirmed through their comparison with reference standards (Level 1), 5 analytes were confirmed by matching their structure with different commercial libraries (Level 2a) and finally 3 compounds were tentatively identified by checking the possible fragmentation pattern (Level 2b). Table S1. Summary of chemicals detected in the wastewater sample. Compound detected Level* Fórmula Rt (min) ESI mode Uses CAS Metformin 1 C4 H11 N5 1.8 + Hypoglicemic 657-24-9 Paracetamol 1 C8 H9 N O2 3.5 + Analgesic 103-90-2 Dibutylphthalate 1 C16 H22 O4 18.9 + Plasticizer 84-74-2 Benzophenone 8 1 C14 H12 O4 16.4 + UV filter 131-53-3 TCPP 1 C9 H18 Cl3 O4 P 17.2 + Flame retardant 13674-84-5 Piperine 2a C17 H19 N O3 17.7 + Food additive 94-62-2 Methyl paraben 1 C8H8O3 14.6 - Antimicrobial agent 99-76-3 Triclosan 1 C12H7Cl3O2 18.5 - Antiseptic 3380-34-5 Netilmicin 2b C21 H41 N5 O7 12.6 + Antibiotic 56391-56-1 Irbesartan 1 C25 H28 N6 O 16.1 + Antihypertensive 138402-11-6 Valsartan 1 C24H29N5O3 16.9 - Antihypertensive 137862-53-4 Diclofenac 1 C14 H11 Cl2 N O2 18.1 + Antinflammatory 15307-86-5 Hippuric acid 2a C9 H9 N O3 4.8 + Bactericide 495-69-2 Fenofibric acid 2a C17 H15 Cl O4 17.9 + Antihyperlipidemic 42017-89-0 Acesulfame 1 C4H5NO4S 2.5 - Sweetener 33665-90-6 4-Hydroxybenzoic acid 2a C7H6O3 11.3 - Chemical reagent / parabens metabolilte 99-96-7 Chenodiol 2a C24H40O4 19.0 - Anticholesterolemic 474-25-9 Carboxyibuprofen 2b C13H16O4 14.8 - Ibuprofen (NSAID) metabolite 15935-54-3 Saccharin 1 C7H5NO3S 3.8 - Sweetener 128-44-9 Canbisol 2b C24 H38 O3 19.3 + Syntetic cannabinoid 56689-43-1 * Confirmation level according to Schymanski et al. (2014).