Science of the Total Environment 482–483 (2014) 125–136

Contents lists available at ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Benzotriazole UV-stabilizers and benzotriazole: Antiandrogenic activityin vitro and activation of aryl hydrocarbon receptor pathway in zebrafisheleuthero-embryos

Karl Fent a,b,⁎, Geraldine Chew a, Jun Li a, Elena Gomez c

a University of Applied Sciences and Arts Northwestern Switzerland, School of Life Sciences, Gründenstrasse 40, CH-4132 Muttenz, Switzerlandb ETH Zurich, Institute of Biogeochemistry and Pollutant Dynamics, Universitätsstrasse 16, CH-8092 Zürich, Switzerlandc UMR Hydrosciences — Université Montpellier 1, DSESP — Faculté de Pharmacie, BP 1449115, Av. Charles Flahault, F-34093 Montpellier Cedex 05, France

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Effects of benzotriazole UV-stabilizersand benzotriazoles are little known.

• UV-P and 1HBT show antiandrogenicactivity in the YAS.

• A targeted transcription analysis with 26genes is performed in zebrafish embryos.

• UV-P andUV-326are of theAHRpathwayin zebrafish embryos.

⁎ Corresponding author at: Gründenstrasse 40, CH-413E-mail address: karl.fent@fhnw.ch (K. Fent).

http://dx.doi.org/10.1016/j.scitotenv.2014.02.1090048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 February 2014Received in revised form 22 February 2014Accepted 23 February 2014Available online 15 March 2014

Keywords:Benzotriazole UV stabilizers1H-benzotriazole2-(2-Hydroxy-5-methylphenyl)benzotriazole (UV-P)2-(3-t-Butyl-2-hydroxy-5-methylphenyl)-5-chlorobenzotriazole (UV-326)ZebrafishTranscriptomics

BenzotriazoleUV-stabilizers (BUVs) are applied inmaterials for protection against UV-irradiation. They arewide-ly used, bioaccumulate and share structural similarities to benzotriazole. Benzotriazole (1HBT) finds applicationas corrosion inhibitor in dishwashing detergents, antifreeze (vehicles) and aircraft de-icing agent. BUVs and1HBT are persistent and ubiquitous in the aquatic environment, but there is little understanding of the ecotoxi-cological implications. Here, we comparatively analyze the hormonal activity in vitro and effects in zebrafisheleuthero-embryos in vivo. 2-(2-Hydroxy-5-methylphenyl)benzotriazole (UV-P), 2-(3-t-butyl-2-hydroxy-5-methylphenyl)-5-chlorobenzotriazole (UV-326), UV-327, UV-328, UV-329 and UV-320 showed no estrogenicity(YES assay) and androgenicity (YAS assay). However, UV-P and 1HBT showed significant antiandrogenic activity.We assessed the transcription profiles of up to 26 genes associated with different toxicological pathways inzebrafish eleuthero-embryos to elucidate potential modes of action of UV-P, UV-326 and 1HBT. Embryos wereexperimentally exposed for 144 hpf to three measured concentrations of 15.8, 70.8, and 690 μg/L UV-P, 7.5,31.7, and 84.3 μg/L UV-326 and 7.9, 97.3 and 1197.3 μg/L 1HBT. Among the 26 transcripts, the induction of thearyl hydrocarbon receptor (AHR) pathway by UV-P and UV-326 was the most significant finding. UV-P led todose-related induction of AHR1, ARNT2 and cyp1a1, as well as of phase II enzymes glutathione-S-transferase(gstp1) and ugt1a. UV-326 led to a significant induction of cyp1a1 and AHR2, but down-regulation of gstp1 at84 μg/L. Only little transcriptional alterations occurred in genes related to apoptosis, oxidative stress, hormonereceptors, and steroidogenesis including aromatase. 1HBT led to only a few expressional changes at 1197 μg/L.Our data lead to the conclusion that UV-P andUV-326 activate the AHR-pathway,whereas 1HBT shows only little

2 Muttenz, Switzerland. Tel.: +41 61 467 45 71.

126 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

transcriptional alterations. It should be noted, however, that effects have been observed at concentration muchhigher than those occurring in the environment. Forthcoming studies should show whether the observedantiandrogenic activities and transcriptional changes translate into physiological effects .

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Analytical chemical trace analysis increasingly demonstrates the ubiq-uitous presence of anthropogenic chemicals in environmental media andbiota. Many of them are persistent and accumulate in the food chain andoccur as contaminants along the aquatic and terrestrial food-chain. In con-trast to the demonstration of the environmental occurrence, often very lit-tle is known about the potential adverse effects, modes of action andtoxicological profile of these contaminants. This temptedus to focuson fre-quently used personal care products and industrial chemicals that are per-sistent, tend to bioaccumulate in the food-chain, but of which theecotoxicological implications are little known.

Benzotriazole UV-stabilizers (BUVs) are widely used to protect mate-rials against harming UV-radiation. Many different (2-hydroxyphenyl)benzotriazole derivatives are in use and they constitute one of the mostimportant families of UV-adsorbents and are used in large volumes(Nakata et al., 2010). BUVs have wide applications in various consumerand industrial products, including buildingmaterials, automobile compo-nents, waxes, paints, adhesive agents, and many more to preventyellowing and degradation of materials and products (Parks, 2009).BUVs have a phenolic group bound to the benzotriazole moiety and areoften referred to as Tinuvin (Table 1). They absorb the full spectrum ofUV-light (UV-A and UV-B) from 280 to 400 nm. An important propertyof BUVs is their high log Kow values, which range from 6.58 to 8.28(Nakata et al., 2009). This is of concern as this may result in high bioaccu-mulation in biota (Nakata et al., 2009, 2010).

In the eastern U.S.A. (Narragansett Bay, RI), BUVs (trade nameTinuvin) have been produced from 1961 to 1985 (Reddy, 2009), whichresulted in the contamination of the local marine system (Reddy et al.,2000). Subsequently, BUVs including 2-(2-hydroxy-5-methylphenyl)benzotriazole (UV-P), 2-(3-tert-butyl-2-hydroxy-5-methylphenyl)-5-chlorobenzotriazole (UV-326), 5-chloro-2-(3,5-di-tert-butyl-2-hydro-xyphenyl)benzotriazole (UV-327), and 2-(3,5-di-tert-amyl-2-hydro-xyphenyl)-benzotriazole (UV-328) have been detected in householddust (Wang et al., 2013). A series of BUVswere found inmunicipalwaste-water and sewage sludge (Ruan et al., 2012). They were also detected inrivers (Liu et al., 2014), seawater (Montesdeoca-Esponda et al., 2012),sediment (Zhang et al., 2011) and biota (Reddy et al., 2000; Nakataet al., 2009). BUVswere detected inmussels from several Asian countriesand the U.S.A. (Nakata et al., 2012). In Japan, UV-320, UV-326, UV-327,and UV-328 were detected in sediments and marine organisms at con-centrations of several hundred ng/g (lipid weight) (Nakata et al., 2009).UV-P concentrations in Chinese rivers were 8 ng/L (Liu et al., 2014) and3–4 ng/L in seawater in Spain, where UV-360 (4–5 ng/L)was also detect-ed (Montesdeoca-Esponda et al., 2012).

In contrast to knowledge about widespread contamination, there islittle understanding of toxicological and ecotoxicological implications.BUVs have a low acute mammalian toxicity and moderate chronictoxicity, primarily in the liver and kidneys. In rats, repeated dose exper-iments with 1-(2′-hydroxy-3′,5′-di-ter-butylphenyl)benzotriazole(UV-320, CAS no. 3848-71-7) showed toxicity to the liver, heart,blood, kidney and thyroid. Males were more susceptible than females.The no observed adverse effect level was less than 0.5 mg/kg/day inmales (Hirata-Koizumi et al., 2007).

To date, the ecotoxicological profile andpotential effects of these com-pounds are only marginally known. Environmental concern arises due tothe widespread occurrence, high lipophilicity and persistence of BUVs.Some of them accumulate in the aquatic food chain (Nakata et al., 2009,2012). Bioconcentration factors of UV-320 and UV-327 are relativelyhigh in carp, ranging from 1380 to 9000 (Nakata et al., 2009). Among

eight BUVs, UV-328was predominantly foundwith amean concentrationof 34 ng/g (lipid weight) in fish in the Philippines (Kim et al., 2011a).UV-327 was most frequently detected in several marine animals withapparent bioaccumulation factors up to 14,000 infish. InAsia-Pacificwa-ters,mussels contain BUVs , predominantlyUV-326 andUV-328, occurringin concentrations up to 1.5 μg/g lipid weight (Nakata et al., 2012). The ac-cumulation of BUVs in the food chain is also foundby other UV-filters fromcosmetics including octocrylene (Blüthgen et al., 2014) and 2-ethyl-hexyl-4-trimethoxycinnamate (EHMC) (Fent et al., 2010). UV filters orig-inating from cosmetic and material protection have been detected inmany countries and aquatic ecosystems (Fent et al., 2008; Díaz-Cruzet al., 2012; Bachelot et al., 2012) andmay negatively affect reproductionof fish (Weisbrod et al., 2007).

BUVswere shown to have a low acute toxicity in Daphnia (Kim et al.,2011b) and algae (Seeland et al., 2012). However, chronic effects ofBUVs and potential toxicological modes of action remain elusive. In re-combinant yeast, no estrogenic activity has been detected for severalof these compounds (Moroshi et al., 2005). The interference with hor-mone receptors in vivo andwith other toxicological or physiological tar-gets in fish is unknown at present.

Benzotriazoles including 1H-benzotriazole (1HBT) are a high pro-duction volume chemicals (annual production of 10,000 tons in theU.S.A.) with broad applications in industry and household (Hart et al.,2004). They are used as anticorrosive agents in dishwasher detergentsas well as in deicing of aircrafts and antifreeze fluids in motor vehicles(Hart et al., 2004; Janna et al., 2011). As other benzotriazoles 1HBT iswater soluble having a low Kow of 1.23 (Hart et al., 2004) and persistentin the environment. Consequently, benzotriazoles are widespread inwastewater and surface water (Voutsa et al., 2006; Giger et al., 2006)as well as in runoff from airfields and associated surface and groundwater (Corsi et al., 2003). In treated wastewater up to 8 μg/L 1HBTwas detected (Voutsa et al., 2006; Giger et al., 2006; Janna et al., 2011;van Leerdam et al., 2009), being between 1 and 10 μg/L across Europe(Reemtsma et al., 2006; Janna et al., 2011). In surface water, 1HBT con-centrations are in the range of 0.2–3.4 μg/L (Giger et al., 2006;Weiss andReemtsma, 2005) with 0.5 μg/L on average in Europe (Reemtsma et al.,2006). In rivers across Europe up to 8 μg/L with an average of 0.49 μg/Lhas been reported (Loos et al., 2009), and in groundwater, up to 1 μg/Lwas found (Loos et al., 2010). Furthermore, benzotriazoles including1HBT and benzothiazole derivatives are also present in human urineoriginating from seven countries (Asimakopoulos et al., 2013). This in-dicates human exposure.

In fathead minnows 1HBT showed a LC50 value of 65 mg/L (Pillardet al., 2001). Up to 1 mg/L no toxicity has been noted (Harris et al.,2007). Similar to BUVs , very little is known on chronic toxicity and onthe toxicological modes of action. 1HBT showed an antiestrogenicactivity in vitro but no hormonal activity has been found in vivo as de-termined by vitellogenin (VTG) protein expression in fatheadminnows(Harris et al., 2007). However, induction of VTG was noted in marinemedaka (He et al., 2012). At present, the ecotoxicological data basis isnot sufficient (Janna et al., 2011).

Despite high production,widespread environmental occurrence andpersistence of BUVs and 1HBT, knowledge on their potential effects onhumans and the environment are limited. The current knowledge gapon ecotoxicological consequences of BUVs and 1HBT including theirchronic toxicity does not allow a comprehensive environmental riskcharacterization. Moreover, the modes of action of these compoundsare unknown. Consequently, in this study we analyzed some of theecotoxicological effects of this structurally related group of compounds.The aims of our present study were (1) to examine, in a comparative

Table 1Chemicals, structures and octanol–water partition coefficients (log Kow) of benzotriazoleUV-stabilizers and 1H-benzotriazole.

Chemicals CAS no. Structure Molecular weight log Kow

UV-P 2440-22-4 225.25 4.2a

UV-326 3896-11-5 315.80 N6a

1HBT 95-14-7 119.12 1.44a

UV-327 3864-99-1 357.88 7.81b

UV-328 25973-55-1 351.49 8.28b

UV-329 3147-75-9 323.43 6.2c

UV-320 3846-71-7 323.44 7.21b

Surrogate standard for analyticsChD12 1719-03-5 240 5.80d

Abbreviations: UV-P: 2-(2-hydroxy-5-methylphenyl)benzotriazole (C13H11N3O); UV-326:2-tert-butyl-6-(5-chloro-2H-benzotriazol-2-yl)-4-methylphenol (C17H18ClN3O); 1HBT: 1H-benzotriazole (C6H5N3); UV-327: 2,4-di-tert-butyl-6-(5-chloro-2H-benzotriazol-2-yl)phenol(C20H24ClN3O); UV-328: 2-(2H-benzotriazol-2-yl)-4,6-di-tert-pentylphenol (C22H29N3O);UV-329: 2-(2H-benzotriazol-2-yl)-4-(1,1,3,3-tetramethylbutyl)phenol (C20H25N3O);UV-320: 2-(3,5-di-tert-butyl-2-hydroxyphenyl)-2H-benzotriazole (C20H25N3O).Data sources: aSigma Aldrich; bBommanna G. Loganathan, Paul Kwan-Sing Lam (2011Global Contamination Trends of Persistent Organic Chemicals). cU.S. EnvironmentalProtection Agency Hazard Characterization Document, September (2009); dPredicteddata cited from chemspider.com, which is generated using U.S. Environmental ProtectionAgency's EPI Suite™ (KOWWIN v1.67 estimate).

127K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

manner, the transcriptional effects of benzotriazole UV-stabilizers UV-P,UV-326, and 1HBT on zebrafish eleuthero-embryos in order to elucidatepotential modes of action; (2) to elucidate effects on the endocrine sys-tem in vitro and in vivo as well as on the arylhydrocarbon receptor(AHR)-mediated pathway. Important environmental chemicals includ-ing polychlorinated dibenzodioxins, polychlorinated biphenyls andpolyaromatic hydrocarbons are activators of this pathway, ultimatelyleading to significant adverse effects, including cancerogenicity, endo-crine activity and developmental effects (Billiard et al., 2006; Safeet al., 2013). In our present study, we followed the hypothesis that un-known modes of action of environmental chemicals can be identifiedby a targeted gene expression approach. In fact, we identified the activa-tion of the AHR pathway as an important effect of UV-P and UV-326.

2. Materials and methods

2.1. Chemicals

UV-stabilizers 2-(2-hydroxy-5-methylphenyl)benzotriazole (UV-P,CAS no. 2440-22-4), 2-t-butyl-6-(5-chloro-2H-benzotriazol-2-yl)-4-methylphenol (UV-326, CAS no. 3896-11-5), 2,4-di-t-butyl-6-(5-chloro-2-H-benzotriazol-2-yl)phenol (UV-327, CAS no. 3864-99-1), 2-(2H-benzotriazol-2-yl)-4,6-di-t-pentylphenol (UV-328, CAS no. 25973-55-1),and 2-(2H-benzotriazol-2-yl)-4-(1,1,3,3-tetramethylbutyl)phenol (UV-329, CAS no. 3147-75-9) were purchased from Sigma-Aldrich (Buchs,Switzerland). 2-(3,5-Di-t-butyl-2-hydroxyphenyl)-2H-benzotriazole(UV-320, CAS no. 3846-71-7) was purchased from Dr. EhrenstorferGmbH (Augsburg, Germany).

Chrysene-d12 (ChD12)was used as surrogate standard. All structuresand physicochemical properties are given in Table 1. Analytical-gradeheptane, ethanol and sodium sulfate (Na2SO4) anhydrous were suppliedby Carlo Erba (Val de Reuil, France). Ultrapure water used for blanks wasobtained with a Milli-Q system (Simplicity UV, Millipore).

1H-benzotriazole (1HBT purity N 98%) and formic acid werepurchased from Sigma-Aldrich (Buchs, Switzerland). The surrogate stan-dard 5,6-dimethyl-1H-benzotriazolemonohydrate (purity 99%)was pur-chased from Sigma-Aldrich and used as the surrogate standard for 1HBT.Methanol and acetone were from J.T. Baker (Deventer, The Netherlands).Heptane was from J.T. Baker and Carlo Erba (Val de Reuil, France).

2.2. In vitro assays (YES and YAS)

To investigate the in vitro hormonal activity of selected BUVs and1HBT, a recombinant yeast assay expressing the human estrogen (hERαin the yeast estrogen screen YES) and androgen receptor (yeast androgenscreen YAS) was used. The YES assay was carried out as described byRoutledge and Sumpter (1996), and the YAS assay as describedby Sohoni and Sumpter (1998). Estradiol (E2), dihydrotestosterone(DHT), and flutamide (FLU) were used as standards for the appropriateassays. E2, DHT and FLU were serially diluted from 4.99 × 10−10 to2.44 × 10−13 M, 5.00 × 10−8 to 2.44 × 10−11 M and 1.12 × 10−4 to5.44 × 10−8 M. Agonistic activities of chemicals used as standardswere assessed first to determine the concentration used to assessthe antagonistic (anti-androgenic) activities. Serial dilution of allchemicals was prepared and chemical concentrations used in the agonis-tic/antagonistic YES and YASwere as follows: UV-P 2.20 × 10−4 to 2.62 ×10−11 M, UV-326 1.60 × 10−4 to 1.91 × 10−11 M, 1HBT 8.39 × 10−3 to1.00 × 10−9 M, UV-327 1.52 × 10−4 to 1.81 ×10−11 M, UV-328 1.49 ×10−4 to 1.78 × 10−11 M, UV-329 1.61 × 10−4 to 1.92 × 10−11 Mand UV-320 1.43 × 10−4 to 1.71 × 10−11 M. An agonist concentrationof 2 × 10−9 M DHT was added to the medium to produce a sub-maximal response (65%) in the YAS antagonist assay. Dose response curvesfor target chemicals were always run in parallel with standard curves.

The assay development was followed by reading absorbance at 540and 620 nm (YES and YAS). All chemicals were tested in duplicate andtriplicate, respectively, and each experiment was carried out two to

three times. Solvent controls (ethanol) and blanks (medium only)were included on each 96-well plate. Further details are given in theSupplementary material.

2.3. Chemical analysis

Chemical analysis of BUVs and 1HBT has been performed to assessthe concentration of the chemicals in the different zebrafish exposure ex-periments. Concentrations were measured in water samples collected atthe beginning (0 h) and before water renewal (24 h) from exposuretanks in the fish embryo exposure study. Replicate water samples fromthe solvent control, UV-P, UV-326 and 1HBT were analyzed. Watersamples were regularly taken from exposure tanks at 0 h and 24 h andfrozen at −20 °C in carefully rinsed, acetone-cleaned and dried (at300 °C) water bottles. Prior to chemical analysis, water was thawedand extracted thereafter (UV-P, UV-326) or used directly (1HBT) foranalysis.

128 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

2.3.1. UV-P and UV-326Extraction ofwater samples has been performed by liquid–liquid ex-

traction using heptane. Extracts were filtered and analyzed by GC–MS/MS. Details of the sample preparation, extraction, analysis and analyticalperformance, method validation, recoveries and detection limits aredescribed in the Supplementary material.

2.3.2. 1H-benzotriazoleAn ultra performance liquid chromatography method (HPLC) was

established to determine the concentration of 1HBT in exposurewater. The analytical procedure was developed in the laboratorybased on the method of Zhao et al. (2010). Briefly, solid-phase extrac-tionwas performed and eluates were analyzed by HPLC. Details of sam-ple pretreatment, solid-phase extraction, instrumental analysis, qualitycontrol and quantification are described in the Supplementary material.

2.4. In vivo exposures of zebrafish embryos

Two embryo studies were conducted with different concentrationsof UV-P andUV-326, respectively, and one studywith1HBT. For all com-pounds three exposure concentrations and controlswere included. Eachdose group consisted of 5 replicates consisting of 80 fertilized eggs.Freshly fertilized eggs were exposed 2–4 h post fertilization (hpf) for144 h, and eleuthero-embryos were then sampled for gene expressionanalysis. Eggs were placed in glass jars containing 100 mL of exposurewater at 27 °C and 80% air humidity. Water was replaced every 24 h.Embryoswere inspected daily formortality, and hatchingwas recorded.

Stock solutions of UV-P and UV-326 of 105, 106 and 107 μg/L weremade in 10 mL DMSO. 10 μL of these stock solutions was then addedto 100 mL reconstituted water to reach final exposure concentrationsof 10, 100 and 1000 μg/L for UV-P, and 10, 100 and 500 μg/L forUV-326. The concentration of DMSO was 0.01% in all replicates in bothstudies. 1HBT is water soluble, and concentrations of 10, 100 and1000 μg/L were made by appropriately dissolving appropriate amountsof the stock solution of 1 g/L. Fishmedium (water) was used in the con-trol group.

For analytical chemical determination of exposure concentrations,water samples were taken three times randomly during the exposurestudy, every time at the beginning of exposure (0 h) and before water re-newal after 24 h. For each group, the exposure water of all replicates wascollected in the same bottle (nearly 500mL) and stored at−20 °C (UV-Pand UV-326 samples) and 4 °C (1HBT samples) until analysis.

After 144 h, the total RNA of control and exposed eleuthero-embryos(30 embryos pooled for each control and dose group replicate) was ex-tracted using the RNeasy Mini Kit (Qiagen, Basel, Switzerland). RNAwas further purified by treatment with RNase free DNase set(Qiagen, Basel, Switzerland) to remove DNA contamination and tosubsequently remove DNase and divalent cations from the samples.

-14 -12 -10 -8 -6 -4 -20.5

1.0

1.5

2.0

2.5E2

UVPUV3261HBT

blank ETOHblank assaymedia

A

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

Fig. 1. Activity of benzotriazole UV-stabilizers UV-P, UV-326 and 1HBT (A) and UV-327, UV-32817b-estradiol (E2) shows estrogenic activity. Mean of two to three independent experiments.

RNA concentrations and quality were analyzed using a NanoDrop1000 spectrophotometer (NanoDrop Technologies Inc., Wilmington,DE, USA); the purity of every RNA sample was between 1.8 and 2.0(260 nm/280 nm ratio).

A total of 1 μg of RNA template was reverse-transcribed in thepresence of random hexamers (Roche Diagnostics, Basel, Switzerland)and deoxynucleoside triphosphate (Sigma-Aldrich, Buchs, Switzerland).The reaction mixture was incubated for 5 min at 70 °C, and then for 1 hat 37 °C. The reaction was stopped by heating at 95 °C for 5 min.

The cDNA was used to perform SYBR-PCR based on SYBR-GreenFluorescence (FastStart Universal SYBR Green Master, Roche Diagnos-tics, Basel, Switzerland). Gene-specific primers for 22 genes weresynthetized according to literature data and primer sequences are listedin Table 2. The ribosomal protein L13a (RpL13a) was selected as house-keeping gene for normalization, as expression is stable during the expo-sure time (Zucchi et al., 2011). Themajority of the selected genes belongto the endocrine pathways including hormone receptors and enzymesimportant for steroidogenesis (cyp19b, 17bhsd3). Other genes investi-gated are important for oxidative stress (catalase), apoptosis (p53),phase I and phase II metabolism and associated arylhydrocarbonreceptor-mediated pathway (cyp1a1, ARNT2, AHR1, AHR2, gstp1, ugt)(Table 2). Expression data of RT-qPCR were analyzed using the ΔΔCtmethod (Livak and Schmittgen, 2001).

2.5. Statistics

The activities in the YES and YAS assays and alteration of gene tran-scripts of pooled replicates (30 eleuthero-embryos per replicate) ana-lyzed by RT-qPCR were graphically illustrated with GraphPad Prism 5(GraphPad Software, San Diego, CA, USA). The significance of differ-ences in transcript levels was analyzed by an unpaired t-test (for onlyhigh concentration samples) and ANOVA for all concentrations. Resultsare given as mean ± standard deviation of the mean. Differences wereconsidered significant at p ≤ 0.05.

3. Results

3.1. Hormonal activity in vitro

As UV-filters used in cosmetic and material protection showedhormonal activities in vitro (Kunz and Fent, 2006) and in vivo in fish(Weisbrod et al., 2007; Christen et al., 2011), we determined the estro-genic, androgenic and antiandrogenic activities of benzotriazoleUV-stabilizers and 1HBT. UV-326, UV-327, UV-328, UV-329 andUV-320 showed no estrogenic activity at concentrations of 0.0001 to1 g/L (Fig. 1). Only UV-P showed a slight but not significant estrogenicactivity at higher concentrations (Fig. 1). Furthermore, 1HBT showed

-14 -12 -10 -8 -6 -40.5

1.0

1.5

2.0

2.5E2blank ETOHblank assaymediaUV327UV328UV329UV320

B

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

, UV-329 and UV-320 (B) in the recombinant yeast estrogen screen (YES). Positive control

-12 -10 -8 -6 -4 -20.5

1.0

1.5

2.0DHTblank ETOHblank assaymediaUVPUV3261HBT

A

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

-12 -10 -8 -6 -40.5

1.0

1.5

2.0

2.5

3.0

DHTblank ETOHblank assaymediaUV327UV328UV329UV320

B

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

Fig. 2. Activity of benzotriazole UV-stabilizers UV-P, UV-326 and 1HBT (A) and UV-327, UV-328, UV-329, and UV-320 (B) in the recombinant androgenic yeast screen (YAS). Positivecontrol dihydrotestosterone (DHT) shows androgenic activity. Mean of two to three independent experiments.

Table 2Primer sequences of genes used for qRT-PCR.

Target gene Genebank number Sense primer (5′–3′) Antisense primer (5′–3′) Annealing temperature (°C)

Hormone receptors and genes involved in steroidogenesisesr1a NM 152959 TGAGCAACAAAGGAATGGAG GTGGGTGTAGATGGAGGGTTT 60arb NM 001083123 CACTACGGAGCCCTCACTTGCGGA GCCCTGAACTGCTCCGACCTC 60vtg1c AY034146 AGCTGCTGAGAGGCTTGTTA GTCCAGGATTTCCCTCAGT 57.517bhsd3d NM 200364.1 TTCACGGCTGAGGAGTTTG GGACCCAGGTAGGAATGG 58cyp19be AF183908 CGACAGGCCATCAATAACA CGTCCACAGACAGCTCATC 58.5cyp19ae AF226620 CTGAAAGGGCTCAGGACAA TGGTCGATGGTGTCTGATG 58.5hsdb3d AY279108 GCAACTCTGGTTTTCCACACTG CAGCAGGAGCCGTGTAGCTT 60hsd11b2f,g NM 212720 TGCTGCTGGCTGTACTTCAC TGCATCCAACTTCTTTGCTG 60

Oxidative stressCatalaseh AF170069 CTGTTCATGTTTATTCACG CACACTTCTATAAGAGAATG 60

Apoptosisp53i AF365873 GGGCAATCAGCGAGCAAA ACTGACCTTCCTGAGTCTCCA 60casp9j NM152884 AAATACATAGCAAGGCAACC CACAGGGAATCAAGAAAGG 58bax1k NM 131562.2 ACAGGGATGCTGAAGTGACC GAAAAGCGCCACAACTCTTC 60p21k NM 001002717 GGAAAGATGACCGATGAGGA GACGCACCTTGTCCAATTTT 60

Lipid metabolism/fatty acid metabolismPPARαl NM 001161333 GATTCAAATCTTGCCGTGGT TCGTCGCTGAGAGACTGAGA 60

Thyroid hormone receptorTRαl NM 131396 CAATGTACCATTTCGCGTTG GCTCCTGCTCTGTGTTTTCC 60

Metabolismcyp1a1l AF210727 CCTGGGCGGTTGTCTATCTA TGAGGAATGGTGAAGGGAAG 60ARNT2m AF219989 GAATGGTCTCGGTCCGTCA AGCTGGTCACCTGCAGTCTT 60AHRn NM 131028 TAG ACA GCG ATA TAC AGC AG TCTCTCCAACACCATTCATG 60AHR2n HQ116647 ACGGTGAAGCTCTCCCATA AGTAGGTTTCTCTGGCCAC 60gstp1o NM 131734 CTAGGAGCAGCTTTGAAACGCAC CGTTGTTGGAGAAATGTTGTACCGACG 60gstp2o NM 001020513 CACTCTCACATACTTCGCTATC AATATTTTCAAATGGTTTGAACTC 60sult1st7p NM 001139481 TCGCAACTTACCCCAAAGCA GATAGGCAGGGAGGATTGGC 60ugt1aq GU299096 AGCTACCCGAAGCCAAAGCCAG AGCCCTAACCTTAGGATGGCCCAA 60ugt1bq GU299110 ACCCACTGACGAAGGAGGTGGAG CGCCACAACACCCTCTGTGGAA 60ugt5a1q GU299146 CTTACTTGTAGAATGGCTCCCT CTTTGCTGTCCCTTTTGCTTC 60ugt5a4q GU299149 GTCAGGTTGGCTGCAATGAC CCACGGTGATACTGTGACCC 60

a Martyniuk et al. (2007).b Hossain et al. (2008).c Hoffmann et al. (2006).d Hoffmann et al. (2008).e Arukwe et al. (2008).f Fuzzen et al. (2011).g Alsop and Vijayan (2008).h Gerhard et al. (2000).i Cawkwell et al. (1994)j Deng et al. (2009).k Choi et al. (2010).l Liu et al. (2013).m Tanguay et al. (1999).n Andreasen et al. (2002a,b).o Suzuki et al. (2005).p Liu et al. (2008).q Christen and Fent (2014 ).

129K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

-8 -6 -4 -20.5

1.0

1.5

2.0

2.5

FLUblank ETOHblank assaymediaUVPUV3261HBT

A

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

-8 -6 -40.5

1.0

1.5

2.0

2.5

FLUblank ETOHblank assaymediaUV327UV328UV329UV320

B

Concentration log(M)

Cor

rect

ed a

bsor

banc

e at

540

nm

Fig. 3. Antagonistic activity of benzotriazole UV-stabilizers UV-P, UV-326 and 1HBT (A) and UV-327, UV328, UV-329 and UV-320 (B) in the recombinant androgenic yeast screen (YAS).Positive control flutamide shows clear anti-androgenic activity as well as UV-P and 1HBT. Mean of three independent experiments.

130 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

no estrogenic activity, confirming previous data (Moroshi et al., 2005;Harris et al., 2007).

No significant androgenic activity occurred for UV-P, UV-326, UV-327, UV-328, UV-329 and UV-320 and 1HBT (Fig. 2). Interestingly, aclear antiandrogenic activity was observed for UV-P and 1HBT (Fig. 3).UV-P showed even a higher activity than the model antiandrogenflutamide. These data indicate no estrogenic and androgenic activitiesof the benzotriazole UV-stabilizers. However, UV-P and 1HBT showeda clear antagonistic interaction with the androgen receptor in the YAS.The tested concentrations were below the cytotoxic concentration,thus demonstrating an antiandrogenic activity. An exception was1HBT at the highest concentrations showing toxicity to the yeast cells.

3.2. Zebrafish embryo exposures

3.2.1. Measured exposure concentrationsIn solvent controls, no UV-P and UV-326 were detected. In all

exposure groups, UV-P and UV-326 concentrations were lower thannominal. At the lowest concentration, UV-P was constant during expo-sure with a mean (±S.D.) concentration of 16 ± 1 μg/L. In contrast,concentrations decreased from 0 h to 24 h at mid and high doses.UV-P concentrations decreased from 82 ± 9 to 60 ± 21 μg/L and from950±190 to 430±170 μg/L, respectively (Table 3).Mean exposure con-centrations between 0 h and 24 h (±SD ) were 16± 1, 71± 19 and 690± 328 μg/L. Similarly, concentrations of UV-326 were constant between0 h and 24 h at the lowest dose with a mean (±SD ) of 8 ± 1 μg/L. Con-centrations in the mid and high doses decreased between 0 h and 24 hfrom 33 ± 11 (0 h) to 30 ± 5 μg/L (24 h) and from 113 ± 96 (0 h) to

Table 3Mean concentrations (±standard deviation SD) of UV-P, UV-326 and 1H-benzotriazole inwater samples collected after 0 h and 24 h.

Measured concentrations

0 h 24 h Mean conc. ± SD

UV-P levels of exposure (nominal)10 μg/L 17* 15 ± 1 16 ± 1100 μg/L 82 ± 9 60 ± 21 71 ± 191000 μg/L 950 ± 190 430 ± 170 690 ± 328

UV-326 levels of exposure (nominal)10 μg/L 8** 7 ± 1 8 ± 1100 μg/L 33 ± 11 30 ± 5 32 ± 8500 μg/L 113 ± 96 56 ± 21 84 ± 41

1HBT levels of exposure (nominal)10 μg/L 7 ± 4 8 ± 1 8 ± 3100 μg/L 98 ± 6 96 ± 1 97 ± 41000 μg/L 1197 ± 6 1198 ± 14 1197 ± 10

Sample number is n = 3, otherwise *n = 1, **n = 2.

56 ± 21 μg/L (24 h) (Table 3). Mean concentrations (±SD ) were 8 ±1 , 32 ± 8 and 84 ± 41 μg/L. These results suggest sorption of UV-Pand UV-326 to glass walls, debris and embryos as well as instability ofUV-326 in water.

In 1HBT exposures,measured concentrations of all dose groupswereclose to nominal. No 1HBT was detected in the water control samples.Concentrations of water-soluble 1HBT did not significantly differbetween 0 h and 24 h (Table 3), demonstrating a constant exposurelevel. This is in accordance with the high resistance of 1HBT to biodeg-radation (Breedveld et al., 2003).

3.2.2. Gross toxicological observationsIn all exposures, no compound-related increase in mortality of

embryos was noted. Both UV-stabilizers and 1HBT did not affecthatching time or hatching rate as compared to controls (data not shown).

3.2.3. Expressional changes of target genes

3.2.3.1. UV-P. Alterations in the expression of 26 genes belonging todifferent pathways were investigated in eleuthero-embryos exposedfor 144 h to 690 μg/L UV-P and compared to the solvent control.Employing a target gene expression approach, we analyzed genesbelonging to different pathways including hormonal receptors andpathways (esr1, ar, vtg1, TRa, PPARa), steroidogenesis (17bhsd3,cyp19a, cyp19b, hsdb3, hsd11b2), AHR pathway (AHR1, AHR2, ARNT2,cyp1a1), phase II metabolism (gstp1, gstp2, sult1st7, ugt1a, ugt1b,ugt5a1, ugt5a4), oxidative stress (catalase), apoptosis and DNA damageresponse (p53, bax, p21, caspase 9). In addition, prominently altered tran-scripts associated with the AHR pathway were investigated at mid andlow doses, including AHR1, AHR2, ARNT2 cyp1a1, gstp1 and ugt1a. Thedata depicted in Fig. 4 demonstrate a significant dose-related inductionof cyp1a1. In addition, genes belonging to the AHR signaling pathways,AHR1 and ARNT2, were significantly induced at 690 μg/L. This indicatesan activation of the AHR pathway. Transcripts of phase II enzymes gstp1and ugt1awere also significantly induced at 960 μg/L (Fig. 5).

In addition, a significant transcriptional down-regulation of hormonereceptor genes PPARa, TRa and hsdb3 occurred (Fig. 6). All other 12 genesbelonging to different pathways showed no significant alterations inzebrafish eleuthero-embryos after the 144 h exposure to 690 μg/L UV-P.

3.2.3.2. UV326.We investigated whether UV-326 also activates the AHRpathway. Data shown in Fig. 7 indicate a significant induction of cyp1a1at 84 μg/L. In contrast to UV-P, an induction of the AHR2 transcriptoccurred at 84 μg/L, but no induction of AHR1 and ARNT2. It should benoted, however, that the exposure concentrations were lower than intheUV-P exposures. In addition, gstp1 but not gstp2 showed a significanttranscriptional down-regulation (Fig. 8). Of the additional 11 genetranscripts belonging to different pathways none of them showed

cyp1a1

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L-1

0

1

2

3

4

5

6

**

***

***

A

Fo

ld C

han

ge

(lo

g2)

ARNT2

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L-1

0

1

2

3

4

5

6

******

B

Fo

ld C

han

ge

(lo

g2)

AHR1

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L

-1

0

1

2

3

4

5

6

***

C

Fo

ld C

han

ge

(lo

g2)

AHR2

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L

-1

0

1

2

3

4

5

6D

Fo

ld C

han

ge

(lo

g2)

Fig. 4. Relative transcriptional expression of genes of the AHR pathway. Phase I metabolism related gene cyp1a1 (A), ARNT2 (B), AHR1 (C) and AHR2 (D) in zebrafish eleuthero-embryosafter exposure to 16, 71 and690 μg/L UV-P. Relative transcript abundancewas quantifiedby real-time reverse transcription PCR (qRT-PCR); the fold changes (log2) in cyp1a1, ARNT2,AHR1and AHR2 abundance as compared to control values were determined using the 2−ΔΔCT method. Results are given as the mean value ± standard deviation (n = 5 replicates). Asterisksindicate significantly higher expression than solvent control (**p b 0.001), and (***p b 0.0001). bMDL, below minimal detection limit.

131K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

significant transcriptional alterations at the highest dose (Fig. 9). Someof the transcripts were barely detectable.

3.2.3.3. 1H-benzotriazole. We analyzed whether 1HBT also activates theAHR pathway. Based on the cyp1a1 transcript no significant activationwas noted, although the ARNT2 transcript showed a significant induction(Fig. 10). Of the additional 10 different gene transcripts only phase IIenzymes gstp1 and ugt1a showed a significant down-regulation.

4. Discussion

Basedonour investigationson theoccurrence (Fent et al., 2010) andef-fects (Kunz and Fent, 2006; Fent et al., 2008; Christen et al., 2011; Blüthgenet al., 2012, 2014) of UV-filters employed in cosmetic andmaterial protec-tion, our objective was to focus on additional chemicals, such as benzotri-azole UV-stabilizers used for UV-protection of materials. Very little isknown on their potential ecotoxicological hazards and risks. In addition,we focused on a structurally related chemical, 1HBT, which has other ap-plications butwhich is known to be persistent in the aquatic environment.Themodes of action and the toxicological profile of this compound are alsonot well known (Harris et al., 2007; Janna et al., 2011).

These chemicals have a relatively low acute toxicity to aquatic biota(Seeland et al., 2012), but their chronic toxicity is only little known(Harris et al., 2007). However, benzotriazole UV-filters and 1HBT arepersistent in the aquatic environment and lipophilic benzotriazoleUV-stabilizers accumulate in aquatic organisms (Reddy et al., 2000;Nakata et al., 2009, 2012). Their potential toxicity and ecotoxicologicalimplications need to be investigated. As the mammalian toxicity is

also not well studied, we had to choose an open concept to identify un-known effects by following a targeted gene expression approach in ourinvestigation. Our present study sheds new lights on the transcriptionaleffects of these compounds in zebrafish early life stages.

4.1. Endocrine effects

On the basis of previously reported in vitro data (Gomez et al., 2005;Kunz and Fent, 2006) and in vivo effects of UV-filters in fish (Kunz andFent, 2006; Blüthgen et al., 2012; Zucchi et al., 2011; Christen et al.,2011), we have investigated potential hormonal activities of BUVsin vitro in the YES and YAS, as well as in zebrafish in vivo bytranscriptomic analysis. 1HBT was previously shown to have anantiestrogenic activity in the YES, but interaction to the androgen receptorwas not analyzed. No estrogenic activity was found in fathead minnows(Harris et al., 2007), whereas estrogenicity was reported in medaka (Heet al., 2012). Our study indicates no estrogenic activity, thus confirmingprevious data on BUVs (Moroshi et al., 2005) and 1HBT (Harris et al.,2007). Furthermore, we found no androgenic agonistic activity of benzo-triazole UV-stabilizers and 1HBT. However, we showed for the first timean antiandrogenic activity of UV-P and 1HBT. The activity was evenstronger than the known antiandrogenic pharmaceutical flutamide. Thisindicates that the six benzotriazole UV-stabilizers analyzed in ourstudy have no agonistic activity on the estrogen and androgen recep-tors. However, UV-P and 1HBT showed antagonistic activity on theandrogen receptor (Fig. 3). The consequences of this in vitro activityshould be analyzed in forthcoming studies with adult fish. Thesecompounds may eventually lead to inhibitory action of androgen

gstp1

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L

-1

0

1

2

3

4

5

6

***

A

Fo

ld C

han

ge

(lo

g2)

sult1st7

< M

DL

690

µg/L

-1

0

1

2

3

4

5

6B

Fo

ld C

han

ge

(lo

g2)

ugt1a

< M

DL

16 µ

g/L

71 µ

g/L

690

µg/L

-1

0

1

2

3

4

5

6

***

C

Fo

ld C

han

ge

(lo

g2)

Fig. 5. Relative transcriptional expression of phase II metabolism related genes gstp1 (A), sult1st7 (B), and ugt1a (C) in zebrafish eleuthero-embryos after exposure to 16, 71 and 690 μg/LUV-P. Relative transcript abundancewas quantified by real-time reverse transcription PCR (qRT-PCR); the fold changes (log2) in gstp1, sult1st7, and ugt1a abundance as compared to con-trol valueswere determined using the 2−ΔΔCTmethod. Results are given as themean value± standard deviation (n=5 replicates). Asterisks indicate significantly higher expression thansolvent control (***p b 0.0001). bMDL, below minimal detection limit.

< M

DL

690

µg/L

-2

-1

0

1

2

*

*

*

PPARaTRa

ar

esrvtg1

17hsdb3

cyp19acyp19b

hsdb3hsd11b2

Hormone receptor genes:

baxp21p53caspase9catalase

Apoptosis/oxidative stress genes:

Fo

ld C

han

ge

(lo

g2)

Fig. 6. Relative transcriptional expression of hormone receptor genes PPARa, TRa, ar,esr1, vitellogenin vtg1, genes of steroidogenesis 17hsdb3, cyp19a, cyp19b, hsdb3 andhsd11b2 and apoptosis and oxidative stress related genes bax, p21, p53, caspase 9, and cata-lase in zebrafish eleuthero-embryos after exposure to 690 μg/L UV-P. Relative transcriptabundance was quantified by real-time reverse transcription PCR (qRT-PCR); the foldchanges (log2) in all gene abundances as compared to control values were determinedusing the 2−ΔΔCT method. Results are given as the mean value ± standard deviation (n =7 replicates). Asterisks indicate significantly higher expression than solvent control (*pb 0.05). bMDL, below minimal detection limit.

132 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

receptor mediated activities, including endocrine disrupting effectson spermatogenesis and male fertility.

To confirm the in vitro activity, we analyzed the effect of UV-P, UV-326 and 1HBT on the expression of hormonal receptors of zebrafisheleuthero-embryos. R elative transcription levels of genes related tothe endocrine system and steroidogenesis are low andwithin biologicalvariability of controls . Thus, the lack of estrogenic and androgenicactivities in vitro is confirmed by the in vivo data. One exception is thedown-regulation of the PPARa and TRa transcripts (Fig. 6) , which maysuggest an interaction of UV-P to fatty acidmetabolism and thyroid hor-mone system. No transcriptional alterations of hormone system relatedgene transcripts were detected for UV-326 and 1HBT. In conclusion, ourresults suggest that only little endocrine effects on zebrafish earlylife stages are indicated based on transcriptional data. However, physi-ological effects of benzotriazole UV-stabilizers and 1HBT should beevaluated in forthcoming experiments, in particular as the in vitrodata point to an antiandrogenic activity of UV-P and 1HBT.

4.2. Transcriptional effects in zebrafish eleuthero-embryos

The present study focused on the transcription profile of selectedtranscripts in zebrafish eleuthero-embryos. We employed a targetedgene expression approach to evaluate many different toxicologicalpathways. We have analyzed up to 26 gene transcripts belonging to di-verse and important toxicological pathways to elucidate the unknownpotential modes of action of these compounds. Therefore, here we pro-videfirst insights into transcriptomics effects of UV-P, UV-326 and 1HBTon various biological processes.

cyp1a1

< M

DL

8 µg

/L

32 µ

g/L

84 µ

g/L-1

0

1

2

3

4

5

6

***

A

Fo

ld C

han

ge

(lo

g2)

ARNT2

< M

DL

8 µg

/L

32 µ

g/L

84 µ

g/L-1

0

1

2

3

4

5

6B

Fo

ld C

han

ge

(lo

g2)

AHR1

< M

DL

84 µ

g/L -1

0

1

2

3

4

5

6C

Fo

ld c

han

ge

(lo

g2)

AHR2

< M

DL

84 µ

g/L-1

0

1

2

3

4

5

6

**

D

Fo

ld c

han

ge

(lo

g2)

Fig. 7. Relative transcriptional expression of genes of the AHR pathway. Phase I metabolism related gene cyp1a1 (A), ARNT2 (B), AHR1 (C) and AHR2 (D) in zebrafish eleuthero-embryosafter exposure to 8, 32 and84 μg/L UV-326. Relative transcript abundancewas quantifiedby real-time reverse transcription PCR (qRT-PCR); the fold changes (log2) in cyp1a1, ARNT2,AHR1and AHR2 abundance as compared to control values were determined using the 2−ΔΔCT method. Results are given as the mean value ± standard deviation (n = 5 replicates). Asterisksindicate significantly higher expression than solvent control (**p b 0.001), and (***p b 0.0001). bMDL, below minimal detection limit.

133K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

The activation of the AHR pathway by UV-P (Fig. 4) and UV-326(Fig. 7) was most significant. Target genes cyp1a1, ARNT2 and AHR1showed a dose-related induction, whereas AHR2 was unaffected by UV-P . Zebrafish have three AHR (AHR1a, AHR1b, AHR2) (Evans et al.,2005), of which AHR2 is necessary for dioxin or polycyclic aromatic hy-drocarbon toxicity (Prasch et al., 2003; Billiard et al., 2006). ZebrafishAHR2 is more closely related to the mammalian AHR than AHR1(Andreasen et al., 2002a,b). Activation of this pathway can, but mustnot necessarily, induce the transcript of AHR2, as this depends on the ac-tivator and time frame. Therefore, despite the lack of transcriptional AHR2

gstp1

< M

DL

84 µ

g/L

-5

-4

-3

-2

-1

0

1

2A

***Fo

ld c

han

ge

(lo

g2)

Fig. 8. Relative transcriptional expression of phase II related genes gstp1 (A), and gstp2 (B) in zebwas quantified by real-time reverse transcription PCR (qRT-PCR); the fold changes (log2) in gstmethod. Results are given as the mean value ± standard deviation (n = 5 replicates). Asterisk

induction UV-P is indicated to act as an activator of the AHR pathway asdemonstrated by the dose-related induction of ARNT2 and cyp1a1.

The activation of the AHR pathway was less pronounced withUV-326, showing a significant induction of AHR2 and cyp1a1 at thehighest dose of 84 μg/L. The relative magnitude of induction was lowerthan a similar concentration of UV-P (71 μg/L), thus UV-326 acts as aweaker AHR-agonist. In contrast, 1HBT did not induce the AHR pathway.

In addition, transcripts of phase II enzymes, gstp1 and ugt1a, aresignificantly induced at 690 μg/L UV-P, similar to the phase I enzymetranscript cyp1a1. We have shown that ugt1a is related to the AHR

gstp2

< M

DL

84 µ

g/L

-5

-4

-3

-2

-1

0

1

2B

Fo

ld c

han

ge

(lo

g2)

rafish eleuthero-embryos after exposure to 84 μg/L UV-326. Relative transcript abundancep1, and gstp2 abundance as compared to control values were determined using the 2−ΔΔCT

s indicate significantly higher expression than solvent control (***p b 0.0001).

< M

DL

84 µ

g/L

-2

-1

0

1

2

ugt5a1ugt5a4

ugt1augt1b

caspase9

catalasep53

17bhsd3

TRaPPARa

cyp19b

Fo

ld c

han

ge

(lo

g2)

Fig. 9. Relative transcriptional expression of additional genes caspase9, 17bhsd3, catalase,ugt5a1, ugt5a4, ugt1a, ugt1b, p53, TRa, PPARa, and cyp19b in zebrafish eleuthero-embryosafter exposure to 84 μg/L UV-326. Relative transcript abundance was quantified by real-time reverse transcription PCR (qRT-PCR); the fold changes (log2) in all gene abundanceas compared to control values were determined using the 2−ΔΔCT method. Results aregiven as the mean value ± standard error of the mean (SEM) (n = 5 replicates). bMDL,below minimal detection limit.

134 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

pathway, as benzo(a)pyrene led to transcriptional induction of cyp1a1and ugt1a (Christen and Fent, 2014). In contrast, UV-326 and 1HBTdid not result in the induction of these transcripts. To the contrary,UV-326 led to down-regulation of the gstp1 transcript, and 1HBT ofgstp1 and ugt1 at the highest concentration.

The AHR is a ligand-activated transcription factor that forms aheterodimer with the ARNT protein and activates gene expressionthrough interactions with dioxin response elements in the promoterregion of genes. Activation of the AHR pathway has several critical im-plications, as prominently seen with polychlorinated dibenzodioxins,polychlorinated biphenyls and polyaromatic hydrocarbons. In mam-mals increased expression of AHR results in cancer (Safe et al., 2013).In zebrafish, the AHR pathway has an important role in mediatingchemical-induced developmental toxicity (Prasch et al., 2003; Billiardet al., 2006). Inducers of AHR also lead to the increased expression of

ND

1197

µg/

L

-3

-2

-1

0

1

2

ar

esr1vtg1

PPARaTRa

gstp1ARNT2cyp1a

ugt5a1ugt1a

*

*

*

Fo

ld C

han

ge

(lo

g2)

Fig. 10. Relative transcriptional expression of cyp1a1, ARNT2, gstp1, ugt5a1, ugt1a, esr1,vtg1, ar, PPARa and TRa in zebrafish eleuthero-embryos after exposure to 1197 μg/L1H-benzotriazole. Relative transcript abundance was quantified by real-time reversetranscription PCR (qRT-PCR); the fold changes (log2) in all gene abundance as comparedto control values were determined using the 2−ΔΔCT method. Results are given as themean value ± standard error of the mean (SEM) (n = 5 replicates). Asterisks indicatesignificantly higher expression than control (*p b 0.05). ND, not significantly detectedtranscripts.

genes that catalyze the metabolism and conjugation of xenobiotics,such as Cyp1a1 . Therefore, there is a risk that UV-P and UV-326 mayultimately result in metabolic imbalance, developmental toxicity andeven tumor promotion.

Taken together our data indicate that UV-P and UV-326 interferewith phase I metabolism. In addition the modulation of phase IIenzymes is indicated for these compounds and 1HBT to a minor extent.Forthcoming studies should show whether additional BUVs interferewith expression of phase I and II enzyme transcripts and whetherthese transcriptional alterations translate to effects on the enzyme orphysiological levels .

The search for additionally affected pathways did not indicatefurther significant effects of these compounds. An exemption is thedown-regulation of the caspase 9, which is involved in apoptosis athigh UV-P concentrations. Furthermore, genes involved in oxidativestress did not show an alteration on the transcriptional level.

5. Conclusions

The in vitro analysis indicates neither estrogenic, nor androgenicactivity of the benzotriazole UV-stabilizers tested in a wide range ofconcentrations (from 10−7 to 1 g/L). However, UV-P and 1HBT showsignificant antiandrogenic activity at higher concentrations. Furtherstudies should explore the associated consequence in vivo. Exposureof zebrafish eleuthero-embryos to UV-P, UV-326 and 1HBT confirmedthe lack of estrogenic and androgenic activities. However, the tran-scripts of TRa and PPARa were down-regulated, suggesting a potentialinterference with the thyroid hormone system and fatty acid metabo-lism, respectively. Our data on 1HBT are in linewith previously reportedlack of vitellogenin induction in fathead minnows (Harris et al., 2007).

The benzotriazole UV-stabilizer UV-P, and to a lesser extent UV-326,alter transcriptional expression of cyp1a1 and associated ARNT2 inzebrafish eleuthero-embryos. Our data indicate that UV-P and UV-326exposure leads to the activation of the AHR pathway and modulationof transcripts of phase I and phase II enzymes. Consequently, these com-pounds may lead to metabolic imbalance and developmental toxicity.Further research should show, whether additional UV-stabilizers sharethis activity. In addition research is needed with a focus on the conse-quences of these alterations on the enzyme and physiological levels.

It should be noted that environmental levels of benzotriazoleUV-stabilizers and 1HBT are much lower than effect concentrations inour present study. Therefore, we conclude that the effects found hereare probably only relevant at high concentrations, but chronic effectsmay occur after a longer period of exposure.

Conflict of interest statement

The authors declare that there are no conflicts of interest.

Acknowledgments

We thankMarina Groz, University of Montpellier 1, for assistance onUV-P and UV-326 analysis and the Chair Emerging Pollutants,Université Montpellier 1. The YES and YAS was kindly provided byPetra Kunz, Ecotox Center, Dübendorf, Switzerland. The study wassupported in part by the Federal Office of the Environment, Bern,Switzerland.

Appendix A. Data

Detailed results of chemical analysis of UV-P, UV-326 and 1HBT(water, chemical property test and quality control) and details of theYES and YAS. Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.scitotenv.2014.02.109.

135K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

References

Alsop D, Vijayan MM. Development of the corticosteroid stress axis and receptor expres-sion in zebrafish. Am J Physiol Regul Integr Comp Physiol 2008;294(3):R711–9.

Andreasen EA, Hahn ME, Heidemann W, Peterson RE, Tanguay RL. The zebrafish (Daniorerio) aryl hydrocarbon receptor type 1 is a novel vertebrate receptor. Mol Pharmacol2002a;62:234–49.

Andreasen EA, Spitsbergen JM, Tanguay RL, Stegeman JJ, Heideman W, Peterson RE.Tissue-specific expression of AHR2, ARNT2, and CYP1A in zebrafish embryos andlarvae: effects of developmental stage and 2,3,7,8-tetrachlorodibenzo-p-dioxinexposure. Toxicol Sci 2002b;419:403–19.

Arukwe A, Nordtug T, Kortner TM, Mortensen AS, Brakstad OG. Modulation of steroido-genesis and xenobiotic biotransformation responses in zebrafish (Danio rerio)exposed to water-soluble fraction of crude oil. Environ Res 2008;107:362–70.

Asimakopoulos AG, Wang L, Thomaidis NS, Kannan K. Benzotriazoles and benzothiazolesin human urine from several countries: a perspective on occurrence, biotransforma-tion and human exposure. Environ Int 2013;59:274–81.

Bachelot M, Li Z, Munaron D, Le Gall P, Casellas C, Fenet H, et al. Organic UV-filter concen-trations in marine mussels from French coastal regions. Sci Total Environ 2012;420:273–9.

Billiard SM, Timme-Laragy AR, Wassenberg DM, Cockman C, Di Giulio RT. The role of theAHR pathway in mediating synergistic developmental toxicity of polycyclic aromatichydrocarbons to zebrafish. Toxicol Sci 2006;92:526–36.

Blüthgen N, Zucchi S, Fent K. Effects of the UV-filter benzophenone-3 (oxybenzone) at lowconcentrations in zebrafish (Danio rerio). Toxicol Appl Pharmacol 2012;263:184–94.

Blüthgen N, Meili N, Chew G, Odermatt A, Fent K. Accumulation and profile of theUV-filter octocrylene in adult and embryonic zebrafish. Sci Total Environ 2014;476–477:207–17.

Breedveld GD, Roseth R, Sparrevik M, Hartnik T, Hem LJ. Persistence of the de-icing addi-tive benzotriazole at an abandoned airport. Water Air Soil Pollut Focus 2003;3:91–101.

Cawkwell L, Lewis FA, Quirke I. Frequency of allele loss of DCC, p53, RBI, WTI, NFI, NM23and APC/MCC in colorectal cancer assayed by fluorescentmultiplex polymerase chainreaction. Brit J Cancer 1994;70:813–8.

Choi JE, Kim S, Ahn JH, Youn P, Kang JS, Park K, et al. Induction of oxidative stress andapoptosis by silver nanoparticles in the liver of adult zebrafish. Aquat Toxicol 2010;100(2):151–9.

Christen V, Fent K. Tissue-, sex- and developmental-specific transcription profiles of eightUDP-glucuronosyltransferase in zebrafish (Danio rerio) and regulation by activator ofaryl hydrocarbon receptor. Aquat Toxicol 2014. http://dx.doi.org/10.1016/j.aquatox.2014.02.019. [in press].

Christen V, Zucchi S, Fent K. Effects of the UV-filter 2-ethyl-hexyl-4-trimethoxycinnamate(EHMC) on expression of genes involved in hormonal pathways in fathead minnows(Pimephales promelas) and link to vitellogenin induction and histology. Aquat Toxicol2011;102:167–76.

Corsi SR, Zitomer DH, Field JA, Cancilla DA. Nonylphenolethoxylates and other additives inaircraft de-icers, anti-icers, and waters receiving airport runoff. Environ Sci Technol2003;37:4031–7.

Deng J, Yu L, Liu C, Yu K, Shi X, Yeung LWY, et al. Hexabromocyclododecane-induceddevelopmental toxicity and apoptosis in zebrafish embryos. Aquat Toxicol 2009;93(1):29–36.

Díaz-Cruz MS, Gago-Ferrero P, Llorca M, Barceló D. Analysis of UV-filters in tap water andother clean waters in Spain. Anal Bioanal Chem 2012;402:2325–33.

Evans BR, Karchner SI, Franks DG, Hahn ME. Duplicate aryl hydrocarbon receptor repressorgenes (ahrr1 and ahrr2) in the zebrafish Danio rerio: structure, function, evolution, andAHR-dependent regulation in vivo. Arch Biochem Biophys 2005;441:151–67.

Fent K, Kunz PY, Gomez E. UV filters in the aquatic environment induce hormonal effectsand affect fertility and reproduction in fish. Chimia 2008;62:368–75.

Fent K, Zenker A, Rapp M. Widespread occurrence of estrogenic UV-filters in aquaticecosystems in Switzerland. Environ Pollut 2010;158:1817–24.

Fuzzen MLM, Bernier NJ, Van Der Kraak G. Differential effects of 17β-estradiol and11-ketotestosterone on the endocrine stress response in zebrafish (Danio rerio).Gen Comp Endocrinol 2011;170:365–73.

Gerhard GS, Kauffman EJ, Grundy MA. Molecular cloning and sequence analysis of theDanio rerio catalase gene. Comp Biochem Physiol 2000;127:447–57.

Giger W, Schaffner C, Kohler HE. Benzotriazole and tolyltriazole as aquatic contaminants.1. Input and occurrence in rivers and lakes. Environ Sci Technol 2006;40:7186–92.

Gomez E, Pillon A, Fenet H, Rosain D, Duchesne MJ, Nicolas JC, et al. Estrogenic activity ofcosmetic components in reporter cell lines: parabens, UV screens, and musks.J Toxicol Environ Health A 2005;68:239–51.

Harris CA, Routledge EJ, Schaffner C, Brian JV, Giger W, Sumpter JP. Benzotriazole isantiestrogenic in vitro but not in vivo. Environ Toxicol Chem 2007;26:2367–72.

Hart DS, Davis LC, Erickson LE, Callender TM. Sorption and partitioning parameters ofbenzotriazole compounds. Microchem J 2004;77:9e17.

He T, Liang B, Liu W, Shin PKS, Wu RSS. Estrogenic potential of benzotriazole on marinemedaka (Oryzias melastigma). Ecotoxicol Environ Saf 2012;80:327–32.

Hirata-Koizumi M,Watari N, Mukai D, Imai T, Hirose A, Kamata E, et al. A 28-day repeateddose toxicity study of ultraviolet absorber 2-(2′-hydroxy-3′,5′-di-tert-butylphenyl)benzotriazole in rats. Drug Chem Toxicol 2007;30:327–41.

Hoffmann JL, Torontali SP, Thomason RG, Lee DM, Brill JL, Price BB, et al. Hepatic gene ex-pression profiling using genechips in zebrafish exposed to 17alpha-ethynylestradiol.Aquat Toxicol 2006;79:233–46.

Hoffmann JL, Thomason RG, Lee DM, Brill JL, Price BB, Carr GJ, et al. Hepatic gene expressionprofiling using GeneChips in zebrafish exposed to 17α-methyldihydrotestosterone.Aquat Toxicol 2008;87:69–80.

Hossain MS, Larsson A, Scherbak N, Olsson PE, Orban L. Zebrafish androgen receptor:isolation, molecular, and biochemical characterization. Biol Reprod 2008;78(2):361–9.

Janna H, Scrimshaw MD, Williams RJ, Chrchley J, Sumpter JP. From dishwasher to tap?Xenobiotic substances bentotriazole and tolyltriazole in the environment. EnvironSci Technol 2011;45:3858–64.

Kim JW, Isobe T, Ramaswamy BR, Chang KH, Amano A, Miller TM, et al. Contaminationand bioaccumulation of benzotriazole ultraviolet stabilizers in fish from Manila Bay,the Philippines using an ultra-fast liquid chromatography–tandem mass spectrome-try. Chemosphere 2011a;85:751–8.

Kim JW, Chang KH, Isobe T, Tanabe S. Acute toxicity of benzotriazole ultraviolet stabilizerson freshwater crustacean (Daphnia pulex). J Toxicol Sci 2011b;36:247–51.

Kunz PY, Fent K. Multiple hormonal activities of UV-filters and comparison of in vivo andin vitro estrogenic activity of ethyl-4-aminobenzoate in fish. Aquat Toxicol 2006;79:305–24.

Liu TA, Bhuiyan S, Snow R, Yasuda S, Yasuda T, Yang YS, et al. Identification and character-ization of two novel cytosolic sulfotransferases, SULT1 ST7 and SULT1 ST8, fromzebrafish. Aquat Toxicol 2008;89(2):94–102.

Liu C, Wang Q, Liang K, Liu J, Zhou B, Zhang X, et al. Effects of tris(1,3-dichloro-2-propyl)phosphate and triphenyl phosphate on receptor-associated mRNA expression inzebrafish embryos/larvae. Aquat Toxicol 2013;128–129:147–57.

Liu R, Ruan T,Wang T, Song S, Guo F, Jiang G. Determination of nine benzotriazole UV sta-bilizers in environmental water samples by automated on-line solid phase extractioncoupled with high-performance liquid chromatography–tandemmass spectrometry.Talanta 2014;120:158–66.

Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quanti-tative PCR and the 2−ΔΔCt method. Methods 2001;25:402–8.

Loos R, Gawlik BM, Locoro G, Rimaviciute E, Contini S, Bidoglio G. EU-wide survey of polarorganic persistent pollutants in European river waters. Environ Pollut 2009;157:561–8.

Loos R, Locoro G, et al. Pan-European survey on the occurrence of selected polar organicpersistent pollutants in ground water. Water Res 2010;44:4115–26.

Martyniuk CJ, Gerrie ER, Popesku JT, Ekker M, Trudeau VL. Microarray analysis in thezebrafish (Danio rerio) liver and telencephalon after exposure to low concentrationof 17alpha-ethinylestradiol. Aquat Toxicol 2007;84:38–49.

Montesdeoca-Esponda S, Sosa-Ferrera Z, Santana-Rodriguez JJ. On-line solid phaseextraction coupled to ultra-performance liquid chromatography with tandem massspectrometry detection for the determination of benzotriazole UV stabilizers incoastal marine and wastewater samples. Anal Bioanal Chem 2012;403:867–76.

Moroshi K, Yamamoto H, Kamata R, Shiraishi F, Koda T, Morita M. Estrogenic activity of 37components of commercial sunscreen lotions evaluated by in vitro assays. Toxicol inVitro 2005;19:457–69.

Nakata H, Murata S, Filatreau J. Occurrence and concentrations of benzotrizole UVstabilizers in marine organisms and sediments from the Ariake Sea, Japan. EnvironSci Technol 2009;43:6920–6.

Nakata H, Shinohara R, Murata S, Watanabe M. Detection of benzotriazole UV stabilizersin the blubber of marine mammals by gas chromatography-high resolution massspectrometry (GC–HRMS). J Environ Monit 2010;12:2088–92.

Nakata H, Shinohara RI, Nakazawa Y, Isobe T, Sudarynto A, Subramanian A, et al. Asia-Pacific mussel watch for emerging pollutants: distribution of synthetic musk and benzo-triazole UV stabilizers in Asian and US coastal waters. Mar Pollut Bull 2012;64:2211–8.

Parks N. UV stabilizing chemicals contaminating Japans marine environment. Environ SciTechnol 2009;43:6896–7.

Pillard DA, Cornell JS, Dufresne DL, Hernandez MT. Toxicity of benzotriazole and benzotri-azole derivatives to three aquatic species. Water Res 2001;35:557–60.

Prasch AL, Teraoka H, Carney SA, Dong W, Hiraga T, Stegeman JJ, et al. Aryl hydrocarbonreceptor 2 mediates 2,3,7,8-tetrachlorodibenzo-p-dioxin developmental toxicity inzebrafish. Toxicol Sci 2003;76:138–50.

Reddy CM. Comment on “Occurrence and concentrations of benzotrizole UV stablilizers inmarine organisms and sediments from the Ariake Sea, Japan”. Environ Sci Technol2009;43:7998.

Reddy CM, Quinn JG, King JK. Free and bound benzotriazoles in marine and freshwatersediments. Environ Sci Technol 2000;34:973–9.

Reemtsma T, Weiss S, Mueller J, Petrovic M, Gonzalez S, Barcelo D, et al. Polar pollutantsentry into the water cycle by municipal wastewater: a European perspective. EnvironSci Technol 2006;40:5451–8.

Routledge EJ, Sumpter JP. Estrogenic activity of surfactants and some of their degradationproducts assessed using a recombinant yeast screen. Environ Toxicol Chem 1996;15:241–8.

Ruan T, Liu R, Fu Q, Wang T, et al. Concentrations and composition profiles of benzotri-azole UV stabilizers in municipal sewage sludge in China. Environ Sci Technol2012;46:2071–9.

Safe S, Lee SO, Jin UH. Role of the aryl hydrocarbon receptor in carcinogenesis and poten-tial as a drug target. Toxicol Sci 2013;135:1–16.

Seeland A, Oetken M, Kiss A, Fries E, Oehlmann J. Acute and chronic toxicity ofbenzotriazoles to aquatic organisms. Environ Sci Pollut Res 2012;19:1781–90.

Sohoni P, Sumpter JP. Several environmental oestrogens are also anti-androgens.J Endocrinol 1998;158:327–39.

Suzuki T, Takagi Y, Osanai H, Li L, Takeuchi M, Katoh Y, et al. Pi class glutathioneS-transferase genes are regulated by Nrf2 through an evolutionarily conservedregulatory element in zebrafish. Biochem J 2005;73:65–73.

Tanguay RL, Abnet CC, Heideman W, Peterson RE. Cloning and characterization of thezebrafish (Danio rerio) aryl hydrocarbon receptor 1. Biochim Biophys Acta 1999;1444:35–48.

Van Leerdam JA, Hogenboom AC, van der Kooi MME, de Voogt P. Determination of polar1H-benzotriazoles and benzothiazoles in water by solid-phase extraction and liquidchromatography LTQ FT Orbitrap mass spectrometry. Int J Mass Spectr 2009;282:99–107.

136 K. Fent et al. / Science of the Total Environment 482–483 (2014) 125–136

Voutsa D, Hartmann P, Schaffner C, Giger W. Benzotriazoles, alkylphenols, and bisphenolA inmunicipal wastewaters and in the Glatt River, Switzerland. Environ Sci Pollut ResInt 2006;13:333–41.

Wang L, Asimakopoulos AG, Moon HB, Nakata H, Kannan K. Benzotriazole, benzothiazole,and benzophenone compounds in indoor dust from the United States and East Asiancountries. Environ Sci Technol 2013;47:4752–9.

Weisbrod CJ, Kunz PY, Zenker AK, Fent K. Effects of the UV-filter benzophenone-2 onreproduction in fish. Toxicol Appl Pharmacol 2007;225:255–66.

Weiss S, Reemtsma T. Determination of benzotriazole corrosion inhibitors from aqueousenvironmental samples by liquid chromatography-electrospray ionization-tandemmass spectrometry. Anal Chem 2005;77:7415–20.

Zhao X, Xu B, Wu F, Zhao, Liao H. Benzotriazole removal from water by Zn–Al–O binarymetal oxide adsorbent: behavior, kinetics and mechanism. J Hazard Mater 2010;184:147–55.

Zhang Z, Ren N, Li YF, Kunisue T, Gao D, Kannan K. Determination of benzotriazole andbenzophenone UV filters in sediment and sewage sludge. Environ Sci Technol2011;45:3909–16.

Zucchi S, Blüthgen N, Ieronimo A, Fent K. The UV-absorber benzophenone-4 alterstranscripts of genes involved in hormonal pathways in zebrafish (Danio rerio)eleuthero-embryos and adult males. Toxicol Appl Pharmacol 2011;250:137–46.