Effects of Environmental Steroid Mixtures Are Regulated by Individual Steroid Receptor Signaling

Research Collection

Journal Article

Effects of environmental steroid mixtures are regulated by individual steroid receptor signaling

Author(s): Schmid, Simon; Willi, Raffael A.; Fent, Karl

Publication Date: 2020-09

Permanent Link: https://doi.org/10.3929/ethz-b-000426943

Originally published in: Aquatic Toxicology 226, http://doi.org/10.1016/j.aquatox.2020.105562

Rights / License: Creative Commons Attribution 4.0 International

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library Aquatic Toxicology 226 (2020) 105562

Contents lists available at ScienceDirect

Aquatic Toxicology

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

Eﬀects of environmental steroid mixtures are regulated by individual steroid receptor signaling T

Simon Schmida,1,Raﬀael Alois Willia,1, Karl Fenta,b,* a University of Applied Sciences and Arts Northwestern Switzerland, School of Life Sciences, Hofackerstrasse 30, CH–4132 Muttenz, Switzerland b Swiss Federal Institute of Technology (ETH Zürich), Institute of Biogeochemistry and Pollution Dynamics, Department of Environmental Systems Science, CH–8092 Zürich, Switzerland

ARTICLE INFO ABSTRACT

Keywords: Fish are exposed to steroids of different classes in contaminated waters, but their effects are not sufficiently Zebrafish embryo understood. Here we employed an anti-sense technique using morpholino oligonucleotides to knockdown the Morpholino glucocorticoid receptors (GRs, GRα and GRβ) and androgen receptor (AR) to investigate their role in physio- Steroids logical and transcriptional responses. To this end, zebrafish embryos were exposed to clobetasol propionate Mixture effects (CLO), androstenedione (A4) and mixtures containing different classes of steroids. CLO caused a decrease of spontaneous muscle contraction and increase of heart rate, as well as transcriptional induction of pepck1, fkbp5, sult2st3 and vitellogenin (vtg1) at 24 and/or 48 h post fertilization (hpf). Knockdown of GRs eliminated these effects, while knockdown of AR decreased the ar transcript but caused no expressional changes, except induction of sult2st3 after exposure to A4 at 24 hpf. Exposure to a mixture of 6 steroids comprising progesterone (P4) and three progestins, cyproterone acetate, dienogest, drospirenone, 17β-estradiol (E2) and CLO caused a significant induction of pepck1, sult2st3, vtg1 and per1a. Knockdown of GRs eliminated the physiological effects and the up- regulation of vtg1, sult2st3, pepck1, fkbp5 and per1a. Thus, as with CLO, responses in mixtures were regulated by GRs independently from the presence of other steroids. Exposure to a mixture comprising A4, CLO, E2 and P4 caused induction of vtg1, cyp19b, sult2st3 and fkbp5. Knockdown of AR had no effect, indicating that regulation of these genes occurred by the GRs and estrogen receptor (ER). Our findings show that in early embryos GRs cause vtg1 and sult2st3 induction in addition to known glucocorticoid target genes. Each steroid receptor regulated its own target genes in steroid mixtures independently from other steroids. However, enhanced expressional induction occurred for vtg1 and fkbp5 in steroid mixtures, indicating an interaction/cross-talk between GRs and ER. These findings have importance for the understanding of molecular effects of steroid mixtures.

1. Introduction environments. Adverse effects in fish were first investigated for natural and syn- Fish are exposed to ng/L concentrations of glucocorticoids, andro- thetic estrogens (Brion et al., 2004; Länge et al., 2001; Tyler et al., gens, progestins and estrogens in surface water and their high potency 1998), and subsequently, for androgens (Ankley et al., 2020) and pro- may cause adverse effects in fish. Steroids and their metabolites origi- gestins (Fent, 2015; Hou et al., 2018; Liang et al., 2019; Schmid et al., nate from municipal and hospital wastewater and agricultural runoff, 2020). Recently, glucocorticoids came into the focus for their effects in either from excretion as natural hormones including 17β-estradiol (E2), adult fish (Kugathas et al., 2013; Zhao et al., 2016) and early life stages testosterone, progesterone (P4), cortisol and their metabolites, or from (McNeil et al., 2016; Willi et al., 2019, 2018; Zhao et al., 2018). Effects medical applications as synthetic hormones (Chang et al., 2009; Weizel were studied mainly for single steroids, followed by combination of two et al., 2018; Zhang et al., 2017). Synthetic estrogens and progestins steroid classes (Garcia-Reyero et al., 2009; Liang et al., 2017; Rossier originate mainly from contraceptives and glucocorticoids from various et al., 2016; Siegenthaler et al., 2017) and only recently, for steroid therapeutic applications. Consequently, aquatic organisms are exposed mixtures. Thereby, mostly mixtures of steroids of one class, including to mixtures of several steroid classes in contaminated aquatic estrogens (Brian et al., 2007; Silva et al., 2002), progestins (Schmid

⁎ Corresponding author at: University of Applied Sciences and Arts Northwestern Switzerland, School of Life Sciences, Hofackerstrasse 30, CH–4132 Muttenz, Switzerland. E-mail address: [email protected] (K. Fent). 1 Both contributed equally to the work. https://doi.org/10.1016/j.aquatox.2020.105562 Received 14 February 2020; Received in revised form 9 June 2020; Accepted 1 July 2020 Available online 03 July 2020 0166-445X/ © 2020 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/). S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562 et al., 2020; Zucchi et al., 2014) and glucocorticoids (Willi et al., 2019) glucocorticoids. The same effect occurs in steroid mixtures in- were studied. Only recently, mixtures of different steroid classes were dependently from the presence of other steroids. Furthermore, genes investigated (Thrupp et al., 2018; Willi et al., 2020; Zhang et al., 2017). regulated by more than one steroid receptor will be identified by Steroids exhibit their effects by receptor signaling leading to dif- transcriptional responses in knockdown experiments. Our systematic ferential expression of specific target genes. For most genes, effects of analysis demonstrates GRs mediated effects and assists in under- one steroid class were shown to occur independently in steroid mixtures standing the contribution of receptor signaling in the effects of steroid from the presence of other steroid classes (Willi et al., 2019, 2020). mixtures. However, there were also genes that were regulated by more than one steroid class and underlying receptors (Willi et al., 2019; Faltermann 2. Material and methods et al., 2020). The joint effects of such steroid mixtures were shown to be additive; they led to reduced fecundity in adult fathead minnows 2.1. Chemicals (Thrupp et al., 2018). In a steroid mixture, the contribution of specific sex hormones and Dimethylsulfoxide (DMSO) (purity ≥99.8 %, CAS no. 67-68-5) was glucocorticoids on differential expression of genes can be identified and purchased from Carl Roth GmbH, Karlsruhe, Germany. The steroids delineated by the abundance of selected target gene transcripts regu- 17β-estradiol (E2) (purity ≥98 %, CAS no. 50-28-2), 4-androstene-317- lated by the specific steroids, as shown in embryos (Willi et al., 2019, dione (A4) (analytical standard, CAS no. 63-05-8), clobetasol propio- 2020) and adult fish (Faltermann et al., 2020). Furthermore, effects of nate (CLO) (purity ≥98 %, CAS no. 25122-46-7), cyproterone acetate glucocorticoids can be identified by determining physiological end- (CPA) (purity ≥98 %, CAS no. 472-51-0), dienogest (DNG) (purity ≥98 points via decrease of spontaneous muscle contraction, increase of heart %, CAS no. 65928-58-7), drospirenone (DRO) (purity ≥98 %, CAS no. rate and accelerated hatching (Chen et al., 2016; Zhao et al., 2016; Willi 67392-87-4) and progesterone (P4) (purity ≥99 %, CAS no. 57-83-0) et al., 2018). were purchased from Sigma Aldrich, Fluka AG, Buchs, Switzerland. The A specific receptor can be made non-functional by knockdown of its steroids were dissolved in DMSO and diluted to the desired exposure transcript or protein (besides genetic knockout). By loss of function concentration. Maximal final DMSO concentration in steroid exposures studies, the contribution of a specific steroid receptor in a steroid was 0.01 %. mixture can be deciphered. Morpholino oligonucleotides (MOs) are a straightforward tool to examine the role of hormone receptors in the 2.2. Experimental design reaction to exogenous steroid hormone exposure. A limitation of em- ploying MOs is the fact that they can only be used in very early embryos For egg collection, one female and one male adult zebrafish (4 due to dilution of MOs in the course of embryonic development. months old) from our in-house zebrafish culture were separated in fi α β Zebra sh has two glucocorticoid receptors (GRs), GR and GR . breeding tanks in the evening prior to the experiments. When the lights fi α β Speci c target genes are regulated by GR , while GR may act as in- went on in the morning of egg collection, the barriers in the tanks were α hibitor of GR (Chatzopoulou et al., 2015). GR signaling is important removed, zebrafish were left to spawn for 15 min and eggs immediately for embryonic growth, hatching and development of somites and removed for exposure or for morpholino injection (1–2 cell stage em- muscles (Nesan et al., 2012; Pikulkaew et al., 2011) and heart (Wilson bryos). α et al., 2015). GR knockdown was associated with reduced transcrip- We performed nine separate experiments, in which transcriptional fi tional abundance of speci c genes (Nesan et al., 2012; Pikulkaew et al., expression of selected target genes compared to the solvent control was fi 2011). There is only one androgen receptor (AR) in zebra sh (Gorelick determined at 24 h post fertilization (hpf) and 48 hpf. When the glu- et al., 2008; Hossain et al., 2008). In early embryos at 24 hpf, AR is cocorticoid CLO was present, physiological effects were assessed at 24 expressed in olfactory placodes, midbrain cells and presumptive pro- and 48 hpf. In total, five single steroid exposures and four steroid nephros (Gorelick et al., 2008). The ar transcript is deposited mater- mixture exposures both without and with MOs were performed nally in embryos and signs of expression could be detected at 24 hpf (Table 1). All steroid exposures included a solvent control containing (Hossain et al., 2008). 0.01 % DMSO. Every exposure group consisted of a set of up to 100 In our present study, we investigate the role of GRs and AR in early embryos and every experiment was performed independently five embryos in response to exogenous exposure to single steroids and mixtures. We used splice-blocking MOs to knockdown both zygotic GRs Table 1 and AR, thereby blocking the splicing of their pre-mRNAs in early Performed experiments and exposure concentrations, each repeated in- embryos. These loss of function experiments allowed us to decipher the dependently five times. Each experiment included a solvent control (0.01 % role of GRs and AR in mediating physiological and transcriptional ef- DMSO). First, single steroids were analyzed followed by mixture experiments. fects after exogenous exposure to single steroids and steroid mixtures. Experiment Steroids Concentration (μg/ Used morpholinos To assess the effects of knockdown of GRs and AR, respectively, we L) determined the mRNA abundance of selected target genes that are up- regulated by these receptors. The choice of target genes was based on I, single steroid CLO 10 no II, single steroid CLO 10 COMO, GRMO previous studies, where glucocorticoid-regulated genes (Willi et al., III, single steroid E2 10 no 2018; Zhao et al., 2018; Schmid and Fent, 2020), androgen-regulated IV, steroid mixture E2, CPA, DNG, 10 each no genes (Fent et al., 2018) and estrogen-regulated genes (Siegenthaler A DRO, P4, CLO et al., 2017) were described. In the case of glucocorticoid-regulated V, steroid mixture E2, CPA, DNG, 10 each COMO, GRMO genes, we choose genes associated with gluconeogenesis, development, A DRO, P4, CLO VI, single steroid A4 10 no immune system regulation and circadian rhythm regulation. Similarly, VII, single steroid A4 10 COMO, ARMO androgen and estrogen responsive genes were selected. VIII, steroid A4, CLO, P4, E2 10 each no In our study, we addressed the following questions: How are phy- mixture B siological and expressional responses of steroids affected by knockdown IX, steroid mixture A4, CLO, P4, E2 10 each COMO, ARMO B of GRs and AR? Which gene transcripts are regulated by GRs and AR, and which genes are regulated by more than one receptor? Can the joint Abbreviations: ARMO, androgen receptor (AR) morpholino; COMO, control effects in mixtures be assigned to GRs or AR signaling or both? We morpholino; GRMO, GR morpholino; CLO, clobetasol propionate; E2, 17β-es- hypothesize that GR knockdown eliminates glucocorticoid-regulated tradiol; CPA, cyproterone acetate; DNG, dienogest; DRO, drospirenone; P4, physiological and transcriptional effects upon exposure to progesterone; A4, androstenedione.

2 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562 times, thus each exposure consisted of five replicates. During exposure, MOs against both GR isoforms and AR, respectively, and a standard embryos were kept in an incubator at 27 ± 1 °C under a 14:10 light- control MO that targets a human beta globin intron mutation were dark cycle until sampling at 24 h post fertilization (hpf) and 48 hpf. used. MO oligonucleotide were described previously for GR and control The design of the experiments was such that first, the response to (Chatzopoulou et al., 2015) and AR (Peal et al., 2011) with sequences single steroid exposures was assessed, and second, the effect of MO given in Table S1. MO were resuspended with 300 μL sterile water receptor knockdown to evaluate the receptor contribution. We focused resulting in a 1 mM stock solution according to the manufacturer’s on changes in physiological endpoints and transcriptional alterations recommendation and further diluted to a working solution with 0.05 % that are expected to be altered by endogenous hormone signaling dis- phenol red in reconstituted water to a final concentration of 0.05 mM ruption. In the single steroid exposures, responses to CLO, E2 and A4 at for the GRMO (0.18 mM for the AR blocking MO). Before injections, the concentration of 10 μg/L were assessed. The single steroid experiments working solution was heated up to 65 °C for 5 min and snap cooled on served as basis for the subsequent steroid mixture experiments. ice. Two steroid mixtures consisting of six and four steroids, respec- Bo-glass capillaries with filament (length: 100 mm, outer diameter: tively, of different steroid classes were analyzed, each steroid in con- 1 mm, wall thickness: 0.21 mm; Hilgenberg GmbH, Malsfeld, Germany) centrations of 10 μg/L. Mixture A consisted of six steroids, of which four were pulled into micropipettes on a P-97 micropipette puller (Sutter were progestins, namely E2, CPA, DNG, DRO, P4 and CLO (Table 1). Instruments, Novato, USA). Micropipettes were then backfilled with 3 While the focus here was on progestins, in mixture B, the focus was on μL MO working solution. The micropipettes were clipped with a ra- the joint action of steroids of four different classes. Thereby, re- zorblade and calibrated by injecting a drop of working solution into a presentative steroids, one estrogen (E2), androgen (A4), progesterone drop of mineral oil (Sigma-Aldrich, Switzerland). The drop was then (P4) and one glucocorticoid (CLO) were mixed at equal concentrations. measured using a microscope calibration slide and the volume was First, the effects of the mixture exposure were analyzed to assess the calculated (Formula: 4/3*r3*π). Approximately 1 nL of working solu- joint activity of these environmental steroids. This was then followed by tion was injected into zebrafish 1–2 cell stage embryos (43 pg and 1.5 determining the effects of MO receptor knockdown. ng MO per embryo for the GRMO and ARMO, respectively) using a Single steroids and steroid mixtures were selected based on their Femtojet 4i Microinjector (Eppendorf, Switzerland). presence in the environment. CLO was chosen due to its high activity (Willi et al., 2018), A4 as prominent androgen in the environment (Fent 2.4. Physiological endpoints et al., 2018), P4 due to its natural origin and high prevalence in aquatic systems (Zhang et al., 2017) and the synthetic progestins due to their Glucocorticoids, but not E2 and androstenedione, affected sponta- use in contraception and presence in the environment as a new gen- neous muscle contraction and heartbeat (heart rate) at the exposure eration progestin (Rossier et al., 2016; Schmid et al., 2020; Zucchi et al., concentration as these were regulated by glucocorticoid signaling (Willi 2013). The exposure concentration of 10 μg/L was chosen on the et al., 2018). In exposures, where the glucocorticoid CLO was present ground that it induces uneqivocal effects based on these previous stu- (single CLO exposure, mixture A), end points such as spontaneous dies. Effects are reported for nominal concentrations, as real con- muscle contraction at 24 hpf and heartbeat was determined at 48 hpf as centrations are close to nominal as demonstrated by previous studies previously (Willi and Fent, 2018; Zhao et al., 2016). In mixture B, using identical exposures (Willi et al., 2018). Only E2 and P4 con- physiological endpoints were not assessed. centrations are lower than nominal due to degradation (Schmid et al., 2020; Willi et al., 2020). 2.5. Transcription analysis MOs are oligomers of 25 bases that are targeted via complementary base pairing to the specific mRNA leading to knockdown of the trans- RNA extraction and RT-qPCR of target genes were performed as lation of the gene, and thereby, allow to identify its function. This anti- previously described (Willi et al., 2019; Schmid and Fent, 2020). sense knockdown technique uses injection of MOs into a 1–8 cell stage Briefly, total RNA was extracted from embryos at 24 and 48 hpf using embryo. The MO binds complementary to the mRNA sequence and 500 μL TRIreagent (Sigma Aldrich, Fluka AG, Buchs, Switzerland) fol- specifically blocks, depending on the type of MO, both maternal and lowing homogenization of embryos by passing them repeatedly through zygotic transcripts and blocks translation. MOs are either designed from a syringe. For RT-qPCR, 1 μg RNA was reverse transcribed using a re- cDNA of expressed (and spliced) mRNA, and often, the 25 bases sur- action mixture consisting of M-MLV (Promega, Switzerland), random rounding the start codon are used. Alternatively, splice-blocking MOs primers (Promega, Switzerland) and dNTP (Sigma-Aldrich, Switzer- can be used, as done in our experiments. While translation blocking land). qPCR was conducted by BIO-RAD CFX96 and BIO-Rad CFX MOs cause knockdown of maternal and zygotic mRNA, splice-blocking Connect Real-Time PCR Detection Systems (BIO-RAD, Switzerland) MOs block zygotic mRNA only. using FastStart Universal SYBR Green Master (Rox) (Roche Diagnostics, In our MO experiments, embryos were injected with either a control Basel, Switzerland). Selected target genes and primer sequences are morpholino (COMO), or GR morpholino (GRMO) or AR morpholino based on our previous studies (Fent et al., 2017; Willi et al., 2018). For (ARMO) and sequences are given in the supplementary material in normalization, β-actin was used due to its high stability upon chemical Table S1. The COMO was used to control for injection e ffects. The and hormonal treatment as previously described (Zhao et al., 2016). GRMO was a splice-blocking MO that knocked down zygotic mRNA of Table S2 gives primer sequences. both isoforms of the GR, thus both GRα and GRβ. MOs were injected in We selected 15 different target genes for assessing the effects of the embryos exposed to single steroids and mixtures and the respective steroids and the morpholino treatment. These genes are previously solvent controls (SC). The following exposures were made: solvent shown to be regulated by either glucocorticoids (Willi et al., 2018; Zhao control (COMO + SC and GRMO + SC), CLO (COMO + CLO and et al., 2018), androgens (Fent et al., 2018) or estrogens (Siegenthaler GRMO + CLO), A4 (COMO + A4 and ARMO + A4), and embryos et al., 2017). Their encoded proteins and functions are listed in Table 2. exposed to the steroid mixtures A (COMO + Mixture A and GRMO + Mixture A) and steroid mixture B (COMO + Mixture B and ARMO + 2.6. Data processing and statistics Mixture B) (Table 1). We calculated the relative gene expression using the 2-ΔΔCt method 2.3. Morpholino microinjections (Ct = threshold cycle) (Livak and Schmittgen, 2001). The abundance of mRNA of target genes in COMO injected embryos exposed to the sol- We used MO modified antisense oligonucleotides that were custom- vent control were considered as baseline for the 2-ΔΔCt method. In synthesized by Gene Tools, LCC (Oregon, USA). Two splice blocking experiments without MO injections, the solvent control group was

3 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Table 2 Assessed genes, their encoded proteins and functions.

Gene Encoded protein Function and associated pathways1

ar androgen receptor androgen signaling cyp19b aromatase, cytochrome P450, family 19, subfamily A, polypeptide b steroid hormone biosynthesis, ovarian steroidogenesis cyp2k22 cytochrome P450, family 2, subfamily k, polypeptide 22 androgen responsive, unknown function esr1 estrogen receptor estrogen signaling f2 coagulation factor II blood coagulation cascade, coagulation factor fkbp5 FK506 binding protein 5 glucocorticoid, estrogen signaling pathway fga ﬁbrinogen alpha chain blood coagulation cascade, coagulation factor grα glucocorticoid receptor alpha glucocorticoid signaling g6pca glucose-6-phosphatase gluconeogenesis irg1l immune-responsive gene 1-like metabolic pathway, tyrosine metabolism, xenobiotic metabolism mmp-13 matrix metalloproteinase-13 glucocorticoid responsive, collagenase 3, breakdown of extracellular collagen pepck1 phosphoenolpyruvate carboxykinase gluconeogenesis per1a period 1a circadian rhythm regulation sult2st3 sulfotransferase phase II metabolism vtg1 vitellogenin yolk sac protein precursor

1 Sources: KEGG pathways, http://www.genome.jp/kegg/ and Willi et al. (2018, 2019), Schmid and Fent (2020). considered as baseline. For graphical illustration and statistical ana- GR regulated genes lack responses when embryos are treated with lysis, we used GraphPad Prism 8.00 for Windows (GraphPad Software, GRMO after exposure to glucocorticoids as shown in Fig. 2B. Treatment La Jolla, CA, USA). with GRMO eliminated GR mediated induction of most genes known to Data of all 5 replicates were pooled for statistical analysis. be induced by glucocorticoids (Willi et al., 2018; Zhao et al., 2016). Transcriptional as well as physiological data were analyzed for statis- Induction of pepck1 and g6pca involved in glucose metabolism was tical significance by unpaired t-test if normal distribution was given. If abolished at 24 and 48 hpf. Target genes involved in immune system normal distribution of the data was not given, Mann-Whitney test was regulation, including fkbp5 showed a significant abolishment of in- performed. Thereby, the effect of the morpholino treatment was as- duction, which was lower with the irg1l transcript (Fig. 2B). The sessed in comparison to the control morpholino treatment. elimination of vtg1 induction occurred at both exposure times under GRMO treatment but was significant at 24 hpf. This was also the case for the abundance of the key regulating circadian rhythm gene tran- 3. Results script per1a, although it was statistically not significant compared to CLO exposure and COMO treatment (Fig. 2B). Up-regulation of sult2st3 3.1. Effects of GR knockdown on clobetasol propionate exposure encoding the metabolism enzyme sulfotransferase, fga, encoding fi- brinogen, a protein in the blood coagulation cascade, and g6pca were As a basis for the loss of function receptor studies, we first assessed also abolished (Fig. 2B). The significant elimination of fkbp5 and pepck1 transcriptional and physiological effects of single steroids at 24 and 48 induction and the elimination trend for sult2st3, irg1l, fga, per1a and hpf, as previously, effects of sex steroids and glucocorticoids exposure g6pca indicate that these genes are regulated by the GRs in early em- were assessed at later embryonic stages of 96–144 hpf (Fent et al., bryos. 2018; Rossier et al., 2016; Schmid et al., 2020; Willi et al., 2018; Zhao et al., 2018). The selected target genes are regulated by glucocorticoid signaling, including genes involved in glucose metabolism, immune 3.2. Effects of GR knockdown on steroid mixture A exposure consisting of system regulation and metabolism (Willi et al., 2020, 2018). Ad- CLO + E2 + P4 + CPA + DNG + DRO ditionally, steroid hormone receptors were included to evaluate the expression of the receptors by MO treatment. Fig. 1 shows that gluco- As a basis for the mixture study, first, single steroid exposure ex- corticoid target genes pepck1 and fkbp5 were induced at 24 and 48 hpf. periments have been performed. Furthermore, there are no comparable Furthermore, sult2st3 was induced at 48 hpf and vtg1 at 24 hpf, which is studies on the effects of E2 on the selected target genes in zebrafish unexpected as vtg1 has been known as estrogen-regulated gene. embryos at the exposure times of 24 and 48 h. First, we assessed Fig. 2A shows that exposure to CLO caused a significant decrease in transcriptional effects of E2. Although an induction of vtg1 occurred at spontaneous muscle contraction at 24 hpf. The comparison between 24 hpf, it was statistically not significant (Fig. 3). However, E2 exposure solvent control and CLO exposed embryos, both treated with COMO, led to a significant decrease of the estrogen receptor (esr1) and coa- confirms the significant decrease described previously without MO gulation factor 2 (f2) transcripts at both time points and a down-reg- treatment (Willi et al., 2018). The increase in heart rate at 48 hpf upon ulation of the aromatase transcript (cyp19b) at 24 hpf. CLO exposure from average 158.5–174.5 beats per minute was statis- Subsequently, effects of mixture A comprising CLO, E2, P4 and the tically not significant. The lack of significance was probably due to one new generation progestins CPA, DNG and DRO were determined at 24 sample that showed a lower heart rate than other samples. Previously, a and 48 hpf. Fig. 4 shows that glucocorticoid-regulated genes pepck1 and significant increase in heart rate occurred (Willi et al., 2018). In con- per1 were significantly induced at 24 and 48 hpf. A very strong in- trast to the treatments with COMO, treatment with GRMO eliminated duction occurred for the vtg1 transcript at both exposure times (Fig. 4). these physiological effects; muscle contraction was not decreased and The transcript sult2st3 showed a significant induction at 48 hpf, while heart rate not increased (Fig. 2A). This demonstrates their regulation by the abundance of the cyp19b and cyp2k22 transcript was unaltered GRs. (Fig. 4). In comparison to E2 exposure alone (Fig. 3), these data suggest The abundance of the GRα transcript was decreased when embryos that the induction of sult2st3 in mixture A is not based on the activity of were treated with GRMO, both in embryos exposed to the solvent E2 but on CLO and the progestins. The strong induction of vtg1 in control (GRMO + SC) and to CLO (GRMO + CLO) compared to em- mixture A seems not to be related to exposure to E2 alone but ad- bryos treated with COMO (Fig. 2B). The GRα transcript was sig- ditionally to CLO that also induced vtg1 (Fig. 2B). nificantly down-regulated by the GRMO treatment, which is the basis Knockdown of GRs by GRMO caused significant alterations of for the lack of physiological effects. physiological and transcriptional responses after exposure to mixture A.

4 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 1. Alteration of gene transcripts after exposure to 10 μg/L clobetasol propionate (CLO) relative to solvent control at 24 and 48 hpf. Shown are transcriptional abundances as fold change relative to solvent control for glucorti- coid regulated genes, encoding phosphoenolpyruvate carboxykinase (pepck1), which is involved in gluconeogenesis, and FK506 binding protein 5 (fkbp5), a member of the immunophilin protein family, which plays a role in immunoregulation and cellular processes (protein folding and trafficking) and a co-chaperone in the heat shock protein 90 and steroid complex, and sult2st3 encoding the phase II enzyme sulfotransferase. Furthermore, the transcript vtg1 encoding vitellogenin, which is a known estrogen-regulated protein, was assessed. Statistically significant differences of CLO exposed embryos compared to the solvent control are indicated with asterisks: *p≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Data represent mean ± SEM of 5 replicates.

The decrease in spontaneous muscle contraction and increase in heart effective in the knockdown of the ar transcript. When embryos were rate were abolished after exposure to mixture A when embryos were exposed to A4 and treated with ARMO, putative target genes of ar treated with GRMO but not when treated with COMO (Fig. 5A). As this signaling, including sult2st3, cyp2k22 and cyp19b were also unaltered as response is identical as in the CLO exposure (Fig. 2A), this indicates that with A4 exposure alone. However, sult2st3 was upregulated at 24 hpf. the physiological effects induced by mixture A were based on the glu- In contrast to glucocorticoids, A4 did not alter transcripts at this short cocorticoid CLO and completely abolished when GR signaling was in- exposuretime, and therefore, the effect of AR silencing could not be hibited. detected. Treatment with GRMO significantly decreased the abundance of the Exposure to mixture B, which contained an estrogen (E2), androgen GRα transcript in the solvent control and steroid mixture A exposed (A4), progestin (P4) and glucocorticoid (CLO) led to induction of embryos, which demonstrates the down-regulation of the GRα tran- cyp19b and sult2st3 at 48 hpf (Fig. 7A). Extremely strong induction script at 24 and 48 hpf (Fig. 5B). Treatment of embryos with GRMO occurred for fkbp5 and vtg1 at both exposure times, the latter similar in significantly abolished the induction of GR-regulated target genes amplitude as in mixture A. pepck1 and fkbp5 after exposure to the steroid mixture A (GRMO + Treatment with ARMO was effective in knocking down the ar Mixture A). The effect was very strong for fkbp5 (Fig. 5B). Functional transcript in both the solvent control and mixture B exposed embryos silencing of GRs had a very similar effect as in the CLO exposure. (Fig. 7B). We also tested the role of AR in the regulation of five tran- Treatment of embryos with GRMO also abolished the induction of vtg1 scripts in this steroid mixture. Treatment of embryos with ARMO had at 24 and 48 hpf after exposure to mixture A (GRMO + Mixture A) no effect on the abundance of the cyp2k22, vtg1 and fkbp5 transcripts, (Fig. 5B). This also holds for the sult2st3 transcript at 48 hpf. Further- indicating that they are not regulated by the AR in early embryos. This more, GRMO treatment tended to abolish the induction of the per1a and was also the case for the sult2st3 and cyp19b transcript, although at 24 fga transcripts, while cyp19b expression was not altered (Fig. 5B). These hpf, treatment with ARMO led to slight induction. These data indicate data indicate that the expression of pepck1, fkbp5, sult2st3 and per1a that the genes cyp2k22, vtg1, cyp19b, sult2st3 and fkbp5 were regulated was regulated by GRs in mixture A, similarly as with CLO alone. by other receptors than the AR at these early embryonic stages. Abolishment of vtg1 was also found with CLO alone, suggesting that vtg1 was regulated by GRs in early embryos. This is novel as vtg1 was 4. Discussion previously thought to be regulated solely by the estrogen receptor. In our study, we aimed to investigate the role of GRs and AR in 3.3. Effects of AR knockdown on steroid mixture B exposure consisting of mediating physiological and transcriptional responses in fish embryos E2 + A4 + P4 + CLO exposed to single steroids and mixtures. Our study gives novel insights into the action of GRs and AR in transcriptional regulation of genes at Exposure of embryos to A4 did not alter transcripts of sult2st3, early embryonic stages. The findings advance current knowledge on the cyp2k22, cyp19b and vtg1 at 24 and 48 hpf (Fig. 6A). Treatment of regulation of physiological and expressional effects of CLO (Willi et al., embryos with a MO that knocked down the AR (ARMO) led to a sig- 2018), as well as that of mixtures of glucocorticoids (Willi et al., 2019) nificant decrease of the ar transcript in solvent controls and A4 exposed and steroids of different classes (Schmid et al., 2020; Willi et al., 2020). embryos (Fig. 6B). This demonstrates that treatment with ARMO was We found that GRs mediate the assessed physiological and

5 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 2. Physiological and transcriptional alterations of genes after exposure to clobetasol propionate (CLO) and treatment with control morpholino (COMO) and glucocorticoid receptor morpholino (GRMO) at 24 and 48 hpf. A, spontaneous muscle contractions at 24 hpf and heart rate at 48 hpf. B, transcriptional alterations of genes as fold changes of transcriptional abundance relative to solvent controls for glucocorticoid receptor grα, pepck1, fkbp5, sult2st3, irg1l, mmp-13, fga, per1a, g6pca and vtg1. Table 2 gives names and functions of encoded proteins. Statistically significant differences between embryos treated with COMO and GRMO (Figure A and B) of solvent controls and CLO exposed embryos, and between the solvent control and CLO exposed embryos treated with COMO (Figure A) are indicated with asterisks: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001, ns, not significant. Data represent mean ± SEM of 5 replicates.

6 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 3. Alteration of gene transcript abundance after exposure to 10 μg/L 17β-estradiol relative to solvent control at 24 and 48 hpf. Shown are transcriptional abundance as fold change relative to solvent control for estrogen receptor (esr1), vitellogenin (vtg1), sulfotransferase (sult2st3), coagulation factor II (f2) and aromatase (cyp19b). Statistically significant differences compared to the solvent control are indicated with asterisks: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Data represent mean ± SEM of 5 replicates. transcriptional effects. Down-regulation of GR mRNAs abolished ex- that might also be reduced. Although GRβ also regulates expression of pressional responses in target genes, including vtg1. In a mixture of four genes, GRβ acts as an inhibitor of transcriptional activity of GRα de- steroids classes, GR knockdown eliminated expression of target genes, pending on the expression level of GRα (Chatzopoulou et al., 2015). thus GRs regulated genes independently from the presence of other Increased activation of GRα by dexamethasone regulated the expression steroids, suggesting no cross-talk between receptors for regulation of of a number of genes and some of them assessed in our study are these genes. However, expression of vtg1 and fkbp5 were enhanced regulated by either GRα or both GRs (Chatzopoulou et al., 2015). At when CLO and E2 were present together in a mixture, which suggests this early embryonic stage, transcripts of GRs are partly maternal but de interaction and cross-talk of GRs and estrogen receptor (ER) for reg- novo embryonic synthesis starts very early; it is widely expressed at 10 ulation of this gene. hpf and later (Nesan and Vijayan, 2012). Our MOs disrupted splicing of Knockdown of steroid receptors by MOs is a straightforward tech- both GR isoforms and AR mRNA, thus de novo synthesized embryonic nique to investigate the role of steroid receptors in the regulation of GR and AR mRNAs were targeted. Consequently, zygotic and not ma- different effects. However, due to dilution of MOs in the course of de- ternal GR signaling was responsible for the observed physiological and velopment, such experiments are limited to very early stages in zeb- expressional effects. rafish development. Previously, the role of GRs in mediating effects of Previously, the role of GRα was investigated in the development of environmental steroids was derived from expressional alterations after zebrafish. Disruption of glucocorticoid signaling by GRα knockdown exposure to glucocorticoids. The present study advances current caused smaller hearts, reduced cardiac function and reduced mRNA knowledge by direct demonstration of GR and AR signaling via loss of levels of specific genes in zebrafish embryos (Wilson et al., 2015). GRα function experiments. knockdown also displayed delayed somitogenesis and tail morphogenesis, and reduced embryo size and survival after hatching (Nesan and Vijayan, 2012). This was associated with reduced transcriptional 4.1. Role of glucocorticoid receptors in physiological effects abundance of specific myogenic genes and showed that GRα signaling is important for zebrafish muscle development. Embryonic exposure to an Zebrafish expresses two GR isoforms. The zebrafish GR gene gives extremely high concentration of dexamethasone and hydrocortisone rise to GRα and GRβ as a result of alternative splicing (Chatzopoulou (100 mg/L) caused changes in craniofacial morphogenesis, suggesting et al., 2015). In our study, we employed a GRMO that knocked down that head cartilage formation is sensitive to GR signaling (Hillegass both GR isoforms. The GRMO targets the splice donor site of exon 2 of et al., 2008). In our experiments, however, we did not notice gross the GR gene, causing retention of intron 2 in both GRα and GRβ mRNA developmental defects or changes. (Chatzopoulou et al., 2015). Thus, our GRMO did not not distinguish An advancement of current knowledge (Willi et al., 2020, 2018) lies between the isoforms, and thus, we assume that both GR isoforms in the fact that our present findings directly demonstrate the im- contributed to and were responsible for the responses. Our loss of portance of GRs in heart rate regulation. Knockdown of GRs completely function experiments showed that indeed GRα was down-regulated by abolished the increase in heart rate caused by CLO and steroid mixture GRMO (Figs. 2B and 5 B) but we did not determine the GRβ transcript

7 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 4. Transcriptional alterations of genes at 24 hpf and 48 hpf after exposure to 10 μg/L of mixture A comprising clobetasol propionate, 17β-estradiol, progesterone and the progestins cyproterone acetate, dienogest, and drospirenone. Shown are alterations of gene transcripts as fold changes of transcriptional abundance relative to solvent controls for pepck1, sult2st3, per1a, vtg1, cyp19b and cyp2k22. Statistically significant differences compared to the solvent control are indicated with asterisks: ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Data represent mean ± SEM of 5 replicates. exposure. Previous studies showed that other receptors are also of in- hpf (Fig. 1). Treatment with MO eliminated the induction, which de- fluence. MO mediated knockdown of brain aromatase affected heart monstrates that GRs regulated this gene at this early embryonic stage rate at 48 hpf with an increase at low and decrease at high con- (Fig. 2B). Of the eight vitellogenin (vtg) genes in zebrafish (Yilmaz et al., centrations after E2 exposure (Ulhaq and Kishida, 2018), which sug- 2018), here we analyzed vtg1. Previously, vtg was thought to be solely gested that heart rate is under the control of serotonin signaling and regulated by estrogens. Induction of vtg1 by ethinyl estradiol was shown regulated by E2. Moreover, heart rate regulation through thyroid hor- to occur at 96 and 144 hpf, but not at 48 hpf (Rossier et al., 2016; mone (T3) has been shown to act downstream of the G protein-coupled Siegenthaler et al., 2017). In our E2 exposure, no significant alteration estrogen receptor and suggested that it regulates embryonic heart rate of the vtg1 transcript occurred at 24 and 48 hpf (Fig. S1), suggesting by altering maternal estrogen and embryonic T3 levels (Romano et al., that estrogens do not activate estrogen receptors (ERs) in early em- 2017). However, the importance of GRs as demonstrated in our study is bryonic stages up to 48 hpf. Thus, GR signaling but not ER signaling plausible as GRα signaling is of major importance for heart develop- occurs in these early stages. We assume that at stages later than 48 hpf, ment (Wilson et al., 2015), and reduced GRα signaling altered structure GRs become silenced and ER signaling is responsible for vtg1 regulation. and function of the embryonic heart. Estrogen response element(s) (ERE) occur in the promotor region of the Glucocorticoids play a significant role in development, growth and vtg1 gene (Bouter et al., 2010; Gorelick and Halpern, 2011; Mushirobira hatching of zebrafish. As expression of GRs occur already at 24 hpf, et al., 2018). Based on our findings, we postulate an additional presence glucocorticoids have the potential to influence development of other of glucocorticoid responsive elements in the promotor. This is also tissues in addition to the heart. GRα signaling is also important for based on the fact that previously, it was shown that vtg3 is up-regulated zebrafish muscle development (Nesan et al., 2012). Treatment of em- and vtg6 down-regulated by GRβ in zebrafish embryos (Chatzopoulou bryos by MO in our study abolished spontaneous muscle contractions, et al., 2015). suggesting that this reaction is mediated by GRs. In a previous study, The responsiveness of the vtg1 gene suggests that GR signaling but larvae treated with GRMO showed alteration in swimming behavior not ER signaling occurred in early embryonic stages. We hypothesize (reduction in velocity and distance travelled) (Wilson et al., 2016). that the differential activation of vtg1 by CLO and E2 in different de- Thus, disruption of GR signaling impairs zebrafish muscle development velopmental stages may be based on epigenetic processes, among and function, suggesting its role in proper muscular function, either others. Methylation of cytosin/guanosin dinucleotides (CpGs) are as- directly or indirectly. sociated with epigenetic silencing of transcription. This may account for the silencing of ER regulated genes including vtg1 in early embryos. The non-responsiveness to estrogen exposure due to epigenetic processes 4.2. Role of glucocorticoid receptors and estrogen receptors in expression of would explain the lack of vtg1 induction up to 48 hpf. Epigenetic the vitellogenin gene changes occur in the promotor region of steroid responsive genes, as shown for another ER responsive gene (Kangaspeska et al., 2008). DNA fi Our study con rms up-regulation of target genes demonstrated after methylation and demethylation occur in the promoter of a gene acti- fi CLO exposure (Willi et al., 2018). A novel and unexpected nding was vated by estrogens in human cells (Métivier et al., 2008). Thus, the the up-regulation of vtg1 by CLO in early embryos, particularly at 24

8 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 5. Physiological and transcriptional alterations of genes at 24 hpf and 48 hpf after exposure to 10 μg/L of mixture A comprising clobetasol propionate, 17β- estradiol, progesterone and the progestins cyproterone acetate, dienogest, and drospirenone and treatment with control morpholino (COMO) and glucocorticoid receptor morpholino (GRMO). A, spontaneous muscle contractions and heart rate. B, transcriptional alterations of genes as fold changes of transcriptional abundance of mixture A relative to solvent controls (SC) for glucocorticoid receptor grα, pepck1, fkbp5, sult2st3, per1a, g6pca, vtg1, cyp19b and fga. Table 2 gives names and functions of encoded proteins. Statistically significant differences between COMO and GRMO treated embryos in solvent controls and in embryos exposed to mixture A (Figure A and B), and between solvent control and mixture A exposed embryos treated with COMO (Figure A) are indicated with asterisks: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Data represent mean ± SEM of 5 replicates. methylation status of promotor CpGs, for instance in steroid hormone induction occurred at 24 and 48 hpf in steroid mixture A and B that responsive elements in the vtg1 promotor region, is a critical event in contained E2 and CLO (Figs. 4, 5 and 7). The response is much more transcriptional control. The vtg1 gene may be silenced for ER (but not pronounced than with single steroids and points to an enhanced re- GRs) regulation by epigenetic processes in the early embryo. sponse in the presence of both steroids. This observation supports si- In comparison to the single steroid CLO, an extremely high vtg1 milar findings at later development stages for vtg1 and sult2st3 (Willi

9 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 6. Abundance of gene transcripts after exposure to 10 μg/L androstenedione (A4) relative to solvent control at 24 and 48 hpf and treatment with androgen receptor morpholino (ARMO). A, exposure to A4. B, exposure to A4 and treatment with control morpholino (COMO) and ARMO. Shown are transcriptional abundance as fold change relative to solvent control (SC) for sulfotransferase (sult2st3), cyp2k22, aromatase (cyp19b), vitellogenin (vtg1) (Figure A) and ar (Figure B). Statistically signiﬁcant diﬀerences between COMO and ARMO treated embryos of the solvent control and A4 exposed embryos are indicated with asterisks: * p ≤ 0.05; ** p ≤ 0.01. Data represent mean ± SEM of 5 replicates.

et al., 2020). Enhanced responses may be related to GR activation and 4.3. Role of glucocorticoid receptors in expressional changes of steroid cross-talk with additional receptors, such as ERs. mixtures

The transcriptional effects of steroid mixtures confirmed previous findings, including induction of glucocorticoid target genes sult2st3,

10 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Fig. 7. Abundance of transcripts at 24 hpf and 48 hpf after exposure to 10 μg/L of mixture B comprising androstenedione, clobetasol propionate, 17β-estradiol and progesterone, and treatment with control morpholino (COMO) and androgen receptor morpholino (ARMO). A, exposure to mixture B. B, exposure to mixture B and treatment with control morpholino (COMO) and androgen receptor morpholino (ARMO). Shown are transcriptional alterations of genes as fold changes of transcriptional abundance relative to solvent control (SC) for androgen receptor (ar), sult2st3, cyp19b, vtg1, cyp2k22 and fkbp5. Statistically signiﬁcant diﬀerences between embryos exposed to mixture B compared to the solvent control (Figure A) and between COMO and ARMO treated embryos in the solvent control and mixture B exposure (Figure B) are indicated with asterisks: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001; **** p ≤ 0.0001. Data represent mean ± SEM of 5 replicates.

11 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562 pepck1 and per1a in mixtures containing five, eight and 13 different GR signaling. class steroids (Willi et al., 2020). Our present study also confirms that in steroid mixtures containing different steroid classes, glucocorticoid 5. Conclusions target genes were regulated independently from the presence of other steroids. However, an interactive effect between GRs and ERs is sug- Our findings advance current knowledge about effects of environ- gested for a single gene, vtg1, when embryos were exposed to CLO and mental steroid mixtures in fish that may exhibited at ng/L concentra- E2 at the same time as an enhanced transcript abundance occurred in tions. MO mediated knockdown of GRs highlighted the importance of the mixture. Each steroid class activated its own receptor and regulated these receptors for the physiological and expressional effects exhibited its target genes but cross-talk between receptors and endocrine path- in early embryos after exposure to glucocorticoids. Not only sponta- ways may also occur. This was the case for vtg1, where interaction neous muscle contraction and heart rate are regulated by GRs but also a between different receptors are suggested for gene activation. Our number of glucocorticoid-regulated genes. Surprisingly, vtg1 was also previous studies suggested that additional genes, including cyp19b, regulated by GRs, a gene that was thought to be exclusively regulated sult2st3 and cyp2k22 are regulated by more than one steroid receptor at by estrogens. In steroid mixtures, GRs regulated their target genes in- later embryonic stages (Schmid et al., 2020; Willi et al., 2020). Fur- dependently from the presence of additional steroids. However, when thermore, when embryos were exposed to mixtures of several gluco- CLO and E2 occurred together in mixtures, expression of vtg1 was en- corticoids, transcriptional abundance of glucocorticoid regulated genes hanced, which points to a cross-talk between GRs and ERs. A similar was additive (Willi et al., 2019). Thus, our gene expression approach effect occurred for fkbp5 that was extremely up-regulated in mixture B. allows to determine the activity of each steroid receptor for regulation Our results support previous findings that in mixtures of different of its target genes. steroid classes, the joint activity is mainly reflected by the activity of In our GR knockdown experiments, we specifically identified GR individual steroids that interact with their own receptors, with the regulated genes, including pepck1, fkbp5, per1a, vtg1 and sult2st3 exception of vtg1 and fkbp5 that were enhanced in mixtures. Our study (Fig. 4). Their induction was eliminated when embryos were treated not only highlights the role of GRs in the effects of environmental with GRMO, which indicates that these genes were solely regulated by glucocorticoids but sheds new light on the regulation of vtg1 and the GRs in this embryonic phase. While pepck1 and fkbp5 are known to be interactive responses in steroid mixtures. regulated by GRs, regulation of vtg1 and sult2st3 by GRs, is a novel finding. The vtg1 gene is also regulated by ERs at later embryonic stages CRediT authorship contribution statement (Rossier et al., 2016; Schmid et al., 2020) and at high E2 concentration (Hao et al., 2013), while AR regulates sult2st3 at 120 hpf. Our study Simon Schmid: Data curation, Formal analysis, Investigation, shows that cyp19b is not regulated by GRs but up-regulated by estro- Methodology, Visualization. Raffael Alois Willi: Data curation, Formal gens (Willi et al., 2020) but also by androgens, probably by an indirect analysis, Investigation, Methodology, Visualization. Karl Fent: Funding mechanism via metabolism to estrogens (Fent et al., 2018). This gene acquisition, Conceptualization, Project administration, Validation, contains an ERE in its promoter (Menuet et al., 2005). Steroid mixture Viszualzation, Supervision, Writing - original draft, Writing - review & studies also suggested that cyp19b, sult2st3 and cyp2k22 (Schmid et al., editing. 2020; Willi et al., 2020) are regulated by more than one steroid receptor at 120 hpf. Declaration of Competing Interest The extremely high induction of the glucocorticoid-regulated gene encoding FK506 binding protein 5 (fkbp5) after exposure to mixture B The authors declare that they have no known competing interests or (Fig. 7) has not been observed in previous experiments with other personal relationships that could have appeared to influence the work steroid mixtures containing glucocorticoids and other steroid classes reported in this paper. (Willi et al., 2020), mixtures of progestins and CLO (Schmid et al., fi 2020), nor with CLO alone (Fig. 2B). Thus, the speci c steroid mixture Acknowledgement B is responsible for this extremely high induction. This leads to the conclusion that the composition of steroids in the mixture matters, and The study was funded by the Swiss National Science Foundation the combination of A4, CLO, P4 and E2 enhances transcription of this (contract no. 310030_169205 to K.F.). gene, which may be the result of receptor cross-talks. Appendix A. Supplementary data 4.4. Role of androgen receptor Supplementary material related to this article can be found, in the While the importance of GRs for responses of early embryos to online version, at doi:https://doi.org/10.1016/j.aquatox.2020.105562. exogenous steroid exposure is clearly demonstrated in our study, it also ff shows a very di erent role for the AR. It is known that the AR plays an References important role in male sexual differentiation and development. However, it is not important for the responses to A4 and steroid mix- Ankley, G.T., Blackwell, B.R., Cavallin, J.E., Doering, J.A., Feifarek, D.J., Jensen, K.M., tures at 24 and 48 hpf as assessed in our study. In zebrafish embryos, Kahl, M.D., LaLone, C.A., Poole, S.T., Randolph, E.C., Saari, T.W., Villeneuve, D.L., ff maternally derived and newly synthesized ar transcripts occur (Gorelick 2020. Adverse outcome pathway network-based assessment of the interactive e ects of an androgen receptor agonist and an aromatase inhibitor on fish endocrine func- et al., 2008; Hossain et al., 2008). Our data show that treatment of tion. Environ. Toxicol. Chem. 39, 913–922. https://doi.org/10.1002/etc.4668. n/a. embryos by ARMO caused down-regulation of the ar transcript. How- Bouter, A., Buisine, N., Le Grand, A., Mouchel, N., Chesnel, F., Le Goff, C., Le Tilly, V., Wolff, J., Sire, O., 2010. Control of vitellogenin genes expression by sequences de- ever, exposure to the androgen A4 did not alter the abundance of target rived from transposable elements in rainbow trout. Biochim. Biophys. Acta - Gene gene transcripts at this very early stage of 24 and 48 hpf. AR signaling Regul. Mech. 1799, 546–554. https://doi.org/10.1016/j.bbagrm.2010.07.003. seems to become active only in later stages, as induction of sult2st3 and Brian, J.V., Harris, C.A., Scholze, M., Kortenkamp, A., Booy, P., Lamoree, M., Pojana, G., Jonkers, N., Marcomini, A., Sumpter, J.P., 2007. Evidence of estrogenic mixture ef- cyp19b occurred at 120 hpf (Fent et al., 2018; Schmid et al., 2020). fects on the reproductive performance of fish. Environ. Sci. Technol. 41, 337–344. Therefore, the functional loss of AR had no consequences in our loss of https://doi.org/10.1021/es0617439. Brion, F., Tyler, C.R., Palazzi, X., Laillet, B., Porcher, J.M., Garric, J., Flammarion, P., function experiments. While the down-regulation of ar mRNA was 2004. Impacts of 17β-estradiol, including environmentally relevant concentrations, confirmed, no effects on gene expression occurred after exposure to A4 on reproduction after exposure during embryo-larval-, juvenile- and adult-life stages and the steroid mixture. AR signaling was therefore not detectable after in zebrafish (Danio rerio). Aquat. Toxicol. 68, 193–217. https://doi.org/10.1016/j. aquatox.2004.01.022. exposure to the androgen A4 at early embryonic stages, in contrast to

12 S. Schmid, et al. Aquatic Toxicology 226 (2020) 105562

Chang, H., Wan, Y., Hu, J., 2009. Determination and source apportionment of five classes Nesan, D., Kamkar, M., Burrows, J., Scott, I.C., Marsden, M., Vijayan, M.M., 2012. of steroid hormones in urban rivers. Environ. Sci. Technol. 43, 7691–7698. https:// Glucocorticoid receptor signaling is essential for mesoderm formation and muscle doi.org/10.1021/es803653j. development in zebrafish. Endocrinology 153, 1288–1300. https://doi.org/10.1210/ Chatzopoulou, A., Roy, U., Meijer, A.H., Alia, A., Spaink, H.P., Schaaf, M.J.M., 2015. en.2011-1559. Transcriptional and metabolic effects of glucocorticoid receptor α and β signaling in Peal, D.S., Mills, R.W., Lynch, S.N., Mosley, J.M., Lim, E., Ellinor, P.T., January, C.T., zebrafish. Endocrinology 156. https://doi.org/10.1210/en.2014-1941. Peterson, R.T., Milan, D.J., 2011. Novel chemical suppressors of long QT syndrome Chen, Q., Jia, A., Snyder, S.A., Gong, Z., Lam, S.H., 2016. Glucocorticoid activity detected identified by an in vivo functional screen. Circulation 123, 23–30. by in vivo zebrafish assay and in vitro glucocorticoid receptor bioassay at environ- Pikulkaew, S., Benato, F., Celeghin, A., Zucal, C., Skobo, T., Colombo, L., Valle, L.D., mental relevant concentrations. Chemosphere 144, 1162–1169. https://doi.org/10. 2011. The knockdown of maternal glucocorticoid receptor mRNA alters embryo de- 2016/j.chemosphere.2015.09.089. velopment in zebrafish. Developmental Dynamics 240. Wiley-Liss, pp. 874–889. Faltermann, S., Hettich, T., Küng, N., Fent, K., 2020. Effects of the glucocorticoid clo- Romano, S.N., Edwards, H.E., Souder, J.P., Ryan, K.J., Cui, X., Gorelick, D.A., 2017. G betasol propionate and its mixture with cortisol and different class steroids in adult protein-coupled estrogen receptor regulates embryonic heart rate in zebrafish. PLoS female zebrafish. Aquatic Toxicology 218, 105372. https://doi.org/10.1016/j. Genet. https://doi.org/10.1371/journal.pgen.1007069. aquatox.2019.105372. Rossier, N.M., Chew, G., Zhang, K., Riva, F., Fent, K., 2016. Activity of binary mixtures of Fent, K., 2015. Progestins as endocrine disrupters in aquatic ecosystems: concentrations, drospirenone with progesterone and 17α-ethinylestradiol in vitro and in vivo. Aquat. effects and risk assessment. Environ. Int. 84, 115–130. https://doi.org/10.1016/j. Toxicol. 174, 109–122. https://doi.org/10.1016/j.aquatox.2016.02.005. envint.2015.06.012. Schmid, S., Fent, K., 2020. 17β-Estradiol and the glucocorticoid clobetasol propionate Fent, K., Siegenthaler, P.F., Schmid, A.A., 2018. Transcriptional effects of androstene- affect the blood coagulation cascade in zebrafish. Environ. Pollut. 259, 113808. dione and 17α-hydroxyprogesterone in zebrafish embryos. Aquat. Toxicol. 202, 1–5. https://doi.org/10.1016/j.envpol.2019.113808. https://doi.org/10.1016/j.aquatox.2018.06.012. Schmid, S., Willi, R.A., Salgueiro-González, N., Fent, K., 2020. Effects of new generation Garcia-Reyero, N., Villeneuve, D.L., Kroll, K.J., Liu, L., Orlando, E.F., Watanabe, K.H., progestins, including as mixtures and in combination with other classes of steroid Sepúlveda, M.S., Ankley, G.T., Denslow, N.D., 2009. Expression signatures for a hormones, on zebrafish early life stages. Sci. Total Environ. 709, 136262. https://doi. model androgen and Antiandrogen in the fathead minnow (Pimephales promelas) org/10.1016/j.scitotenv.2019.136262. ovary. Environ. Sci. Technol. 43, 2614. Siegenthaler, P.F., Bain, P., Riva, F., Fent, K., 2017. Effects of antiandrogenic progestins, Gorelick, D.A., Halpern, M.E., 2011. Visualization of estrogen receptor transcriptional chlormadinone and cyproterone acetate, and the estrogen 17α-ethinylestradiol (EE2), activation in zebrafish. Endocrinology 152, 2690–2703. https://doi.org/10.1210/en. and their mixtures: transactivation with human and rainbowfish hormone receptors 2010-1257. and transcriptional effects in zebrafish (Danio rerio). Aquat. Toxicol. 182. https://doi. Gorelick, D.A., Watson, W., Halpern, M.E., 2008. Androgen receptor gene expression in org/10.1016/j.aquatox.2016.11.001. the developing and adult zebrafish brain. Dev. Dyn. 237, 2987–2995. https://doi. Silva, E., Rajapakse, N., Kortenkamp, A., 2002. Something from “Nothing” − eight weak org/10.1002/dvdy.21700. estrogenic chemicals combined at concentrations below NOECs produce significant Hao, R., Bondesson, M., Singh, A.V., Riu, A., McCollum, C.W., Knudsen, T.B., Gorelick, mixture effects. Environ. Sci. Technol. 36, 1751–1756. https://doi.org/10.1021/ D.A., Gustafsson, J.Å., 2013. Identification of estrogen target genes during zebrafish es0101227. embryonic development through transcriptomic analysis. PLoS One 8, 1–18. https:// Thrupp, T.J., Runnalls, T.J., Scholze, M., Kugathas, S., Kortenkamp, A., Sumpter, J.P., doi.org/10.1371/journal.pone.0079020. 2018. The consequences of exposure to mixtures of chemicals: something from Hillegass, J.M., Villano, C.M., Cooper, K.R., White, L.A., 2008. Glucocorticoids alter ‘nothing’ and ‘a lot from a little’ when fish are exposed to steroid hormones. Sci. Total craniofacial development and increase expression and activity of matrix metallo- Environ. 619–620. https://doi.org/10.1016/j.scitotenv.2017.11.081. proteinases in developing zebrafish (Danio rerio). Toxicol. Sci. 102, 413–424. https:// Tyler, C.R., Jobling, S., Sumpter, J.P., 1998. Endocrine disruption in wildlife: a critical doi.org/10.1093/toxsci/kfn010. review of the evidence. Crit. Rev. Toxicol. 28, 319–361. https://doi.org/10.1080/ Hossain, M.S., Larsson, A., Scherbak, N., Olsson, P., Orban, L., 2008. Zebrafish androgen 10408449891344236. receptor: isolation, molecular, and biochemical Characterization. Biol. Reprod. 78, Ulhaq, Z.S., Kishida, M., 2018. Brain aromatase modulates serotonergic neuron by reg- 361–369. https://doi.org/10.1095/biolreprod.107.062018. ulating serotonin levels in zebrafish embryos and larvae. Front. Endocrinol. 9, 230. Hou, L., Chen, H., Tian, C., Shi, W.J., Liang, Y., Wu, R., Fang, X., Zhang, C., Liang, Y., Xie, https://doi.org/10.3389/fendo.2018.00230. L., 2018. The progestin norethindrone affects sex differentiation and alters tran- Weizel, A., Schlüsener, M.P., Dierkes, G., Ternes, T.A., 2018. Occurrence of glucocorti- scriptional profiles of genes along the hypothalamic–pituitary–gonadal and hy- coids, mineralocorticoids, and progestogens in various treated wastewater, rivers, pothalamic–pituitary–adrenal axes in juvenile zebrafish Dario rerio. Aquat. Toxicol. and streams. Environ. Sci. Technol. 52. https://doi.org/10.1021/acs.est.7b06147. 201. https://doi.org/10.1016/j.aquatox.2018.05.019. Willi, R.A., Fent, K., 2018. Interaction of environmental steroids with organic anion Kangaspeska, S., Stride, B., Métivier, R., Polycarpou-Schwarz, M., Ibberson, D., transporting polypeptide (Oatp1d1) in zebrafish (Danio rerio). Environ. Toxicol. Carmouche, R.P., Benes, V., Gannon, F., Reid, G., 2008. Transient cyclical methyla- Chem. 37, 2670–2676. https://doi.org/10.1002/etc.4231. tion of promoter DNA. Nature 452, 112–115. https://doi.org/10.1038/nature06640. Willi, R.A., Faltermann, S., Hettich, T., Fent, K., 2018. Active glucocorticoids have a range Kugathas, S., Runnalls, T.J., Sumpter, J.P., 2013. Metabolic and reproductive effects of of important adverse developmental and physiological effects on developing zebra- relatively low concentrations of beclomethasone dipropionate, a synthetic gluco- fish embryos. Environ. Sci. Technol. 52, 877–885. https://doi.org/10.1021/acs.est. corticoid, on fathead minnows. Environ. Sci. Technol. https://doi.org/10.1021/ 7b06057. es4019332. Willi, R.A., Salgueiro-González, N., Carcasio, G., Fent, K., 2019. Glucocorticoid mixtures Länge, R., Hutchinson, T.H., Croudace, C.P., Siegmund, F., Schweinfurth, H., Hampe, P., of fluticasone propionate, triamcinolone acetonide and clobetasol propionate induce Panter, G.H., Sumpter, J.P., 2001. Effects of the synthetic estrogen 17α-ethinyles- additive effects in zebrafish embryos. J. Hazard. Mater. 374, 101–109. https://doi. tradiol on the life-cycle of the fathead minnow (Pimephales promelas). Environ. org/10.1016/j.jhazmat.2019.04.023. Toxicol. Chem. 20, 1216–1227. https://doi.org/10.1002/etc.5620200610. Willi, R.A., Castiglioni, S., Salgueiro-González, N., Furia, N., Mastroianni, S., Faltermann, Liang, Y.Q., Huang, G.Y., Zhao, J.L., Shi, W.J., Hu, L.X., Tian, F., Liu, S.S., Jiang, Y.X., S., Fent, K., 2020. Physiological and transcriptional effects of mixtures of environ- Ying, G.G., 2017. Transcriptional alterations induced by binary mixtures of ethiny- mental estrogens, androgens, progestins, and glucocorticoids in zebrafish. Environ. lestradiol and norgestrel during the early development of zebrafish (Danio rerio). Sci. Technol. 54, 1092–1101. https://doi.org/10.1021/acs.est.9b05834. Comp. Biochem. Physiol. Part – C Toxicol. Pharmacol. https://doi.org/10.1016/j. Wilson, K.S., Baily, J., Tucker, C.S., Matrone, G., Vass, S., Moran, C., Chapman, K.E., cbpc.2017.02.005. Mullins, J.J., Kenyon, C., Hadoke, P.W.F., Denvir, M.A., 2015. Early-life perturba- Liang, Y.Q., Huang, G.Y., Zhen, Z., Tian, F., Hou, L., Lin, Z., Ying, G.G., 2019. The effects tions in glucocorticoid activity impacts on the structure, function and molecular of binary mixtures of estradiol and progesterone on transcriptional expression pro- composition of the adult zebrafish (Danio rerio) heart. Mol. Cell. Endocrinol. 414. files of genes involved in hypothalamic-pituitary-gonadal axis and circadian rhythm https://doi.org/10.1016/j.mce.2015.07.025. signaling in embryonic zebrafish (Danio rerio). Ecotoxicol. Environ. Saf. https://doi. Wilson, K.S., Tucker, C.S., Al-Dujaili, E.A.S., Holmes, M.C., Hadoke, P.W.F., Kenyon, C.J., org/10.1016/j.ecoenv.2019.03.021. Denvir, M.A., 2016. Early-life glucocorticoids programme behaviour and metabolism Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using real- in adulthood in zebrafish. J. Endocrinol. 230. https://doi.org/10.1530/joe-15-0376. time quantitative PCR and the 2-ΔΔCT method. Methods. https://doi.org/10.1006/ Yilmaz, O., Patinote, A., Nguyen, T., Bobe, J., 2018. Multiple vitellogenins in zebrafish meth.2001.1262. (Danio rerio): quantitative inventory of genes, transcripts and proteins, and relation to McNeil, P.L., Nebot, C., Sloman, K.A., 2016. Physiological and behavioral effects of ex- egg quality. Fish Physiol. Biochem. 44, 1509–1525. https://doi.org/10.1007/s10695- posure to environmentally relevant concentrations of prednisolone during zebrafish 018-0524-y. (Danio rerio) embryogenesis. Environ. Sci. Technol. 50. https://doi.org/10.1021/acs. Zhang, K., Zhao, Y., Fent, K., 2017. Occurrence and ecotoxicological effects of free, est.6b00276. conjugated and halogenated steroids of 17α-hydroxypregnanolone and pregnanediol Menuet, A., Pellegrini, E., Brion, F., Gueguen, M.-M., Anglade, I., Pakdel, F., Kah, O., in Swiss wastewater and surface water. Environ. Sci. Technol. acs.est, 7b01231. 2005. Expression and estrogen-dependent regulation of the zebrafish brain aromatase https://doi.org/10.1021/acs.est.7b01231. gene. J. Comp. Neurol. 485, 304–320. https://doi.org/10.1002/cne.20497. Zhao, Y., Zhang, K., Fent, K., 2016. Corticosteroid fludrocortisone acetate targets multiple Métivier, R., Gallais, R., Tiffoche, C., Le Péron, C., Jurkowska, R.Z., Carmouche, R.P., end points in zebrafish (Danio rerio) at low concentrations. Environ. Sci. Technol. 50, Ibberson, D., Barath, P., Demay, F., Reid, G., Benes, V., Jeltsch, A., Gannon, F., 10245–10254. https://doi.org/10.1021/acs.est.6b03436. Salbert, G., 2008. Cyclical DNA methylation of a transcriptionally active promoter. Zhao, Y., Zhang, K., Fent, K., 2018. Regulation of zebrafish (Danio rerio) locomotor be- Nature 452, 45–50. https://doi.org/10.1038/nature06544. havior and circadian rhythm network by environmental steroid hormones. Environ. Mushirobira, Y., Nishimiya, O., Nagata, J., Todo, T., Hara, A., Reading, B.J., Hiramatsu, Pollut. https://doi.org/10.1016/j.envpol.2017.09.057. N., 2018. Molecular cloning of vitellogenin gene promoters and in vitro and in vivo Zucchi, S., Castiglioni, S., Fent, K., 2013. Progesterone alters global transcription profiles transcription profiles following estradiol-17β administration in the cutthroat trout. at environmental concentrations in brain and ovary of female zebra fish (Danio rerio). Gen. Comp. Endocrinol. 267, 157–166. https://doi.org/10.1016/j.ygcen.2018.06. Environ. Sci. Technol. 47. https://doi.org/10.1021/es403800y. 017. Zucchi, S., Mirbahai, L., Castiglioni, S., Fent, K., 2014. Transcriptional and physiological Nesan, D., Vijayan, M.M., 2012. Embryo exposure to elevated cortisol level leads to responses induced by binary mixtures of drospirenone and progesterone in zebrafish cardiac performance dysfunction in zebrafish. Mol. Cell. Endocrinol. https://doi.org/ (Danio rerio). Environ. Sci. Technol. 48, 3523–3531. https://doi.org/10.1021/ 10.1016/j.mce.2012.07.010. es405584f.