Derivation and Evaluation of Putative Adverse Outcome Pathways for The

SCWRRP #0996

TOXICOLOGICAL SCIENCES, 156(2), 2017, 344–361

doi: 10.1093/toxsci/kfw257 Advance Access Publication Date: February 13, 2017 Research article

Derivation and Evaluation of Putative Adverse Outcome Pathways for the Effects of Cyclooxygenase Inhibitors on Reproductive Processes in Female Fish Dalma Martinovic-Weigelt,* ,1 Alvine C. Mehinto,†,‡ Gerald T. Ankley,§ Jason P. Berninger,§ Timothy W. Collette,¶ John M. Davis,¶ Nancy D. Denslow,† Elizabeth J. Durhan,§ Evan Eid,§ Drew R. Ekman,¶ Kathleen M. Jensen,§ Michael D. Kahl,§ Carlie A. LaLone,§ Quincy Teng,¶ and Daniel L. Villeneuve§

*Biology Department, University of St Thomas, Saint Paul, Minnesota 55105; †Departments of Physiological Sciences and Biochemistry and Molecular Biology, University of Florida, Gainesville, Florida 32611; ‡Southern California Coastal Water Research Project, Costa Mesa, California 92626; §U.S. Environmental Protection Agency, National Health and Environmental Effects Research Laboratory, Mid-Continent Ecology Division, Duluth, Minnesota 55804; and ¶U.S. Environmental Protection Agency, National Exposure Research Laboratory, Ecosystems Research Division, Athens, Georgia 30605

1To whom correspondence should be addressed at University of St. Thomas, 2115 Summit Avenue, Mail OWS 390, St Paul, MN 55105. Tel: +1 651-962-5233. Fax: +1 651-962-2501. E-mail: [email protected].

ABSTRACT Cyclooxygenase (COX) inhibitors are ubiquitous in aquatic systems and have been detected in fish tissues. The exposure of fish to these pharmaceuticals is concerning because COX inhibitors disrupt the synthesis of prostaglandins (PGs), which modulate a variety of essential biological functions, including reproduction. In this study, we investigated the effects of well- characterized mammalian COX inhibitors on female fathead minnow reproductive health. Fish (n ¼ 8) were exposed for 96 h to water containing indomethacin (IN; 100 mg/l), ibuprofen (IB; 200 mg/l) or celecoxib (CX; 20 mg/l), and evaluated for effects on liver metabolome and ovarian gene expression. Metabolomic profiles of IN, IB and CX were not significantly different from control or one another. Exposure to IB and CX resulted in differential expression of comparable numbers of genes (IB ¼ 433, CX ¼ 545). In contrast, 2558 genes were differentially expressed in IN-treated fish. Functional analyses (canonical pathway and gene set enrichment) indicated extensive effects of IN on PG synthesis pathway, oocyte meiosis, and several other processes consistent with physiological roles of PGs. Transcriptomic data were congruent with PG data; IN-reduced plasma PG F2a concentration, whereas IB and CX did not. Five putative AOPs were developed linking the assumed molecular initiating event of COX inhibition, with PG reduction and the adverse outcome of reproductive failure via reduction of: (1) ovulation, (2) reproductive behaviors mediated by exogenous or endogenous PGs, and (3) oocyte maturation in fish. These pathways were developed using, in part, empirical data from the present study and other publicly available data.

Key words: cyclooxygenase inhibition; prostaglandins; reproduction; ﬁsh; adverse outcome pathway; nonsteroidal anti- inﬂammatory drugs.

344 MARTINOVIC´ -WEIGELT ET AL. | 345

Cyclooxygenase (COX; also known as prostaglandin- effects of short-term, waterborne, exposure to IN and IB (nonse- endoperoxide synthase; PTGS) catalyzes conversion of arachi- lective mammalian COX-inhibitors), and celecoxib (CX; predomi- donic acid into endoperoxide prostaglandin G2, a precursor for nantly inhibits the COX-2 isoform) in adult teleost fish synthesis of bioactive PGs (Simmons et al., 2004). PGs modulate (Pimephales promelas—fathead minnow). We assessed effects on variety of essential biological functions (including reproduction, ovarian transcriptomic and hepatic metabolomic profiles, ovar- immune function and vasodilation; Gomez-Abellan et al., 2016; ian COX activity, and circulating PGF2a,17b-estradiol (E2), and vi- Simmons et al., 2004; Takahashi et al., 2010). There are multiple tellogenin (VTG; egg yolk protein precursor). We hypothesized COX isoforms in vertebrates (Havird et al., 2008); the COX-1 iso- that exposure to mammalian COX-inhibitors would impact COX- form is typically expressed constitutively and is involved in activity, circulating PGs and gene sets associated with PG synthe- maintenance of homeostatic functions, while COX-2 is largely sis and ovulation. Conversely, because the most sensitive targets inducible (eg, during discrete stages of gamete maturation, by of COX-inhibitors are COX activity and PGs, we did not anticipate inflammation; Simmons et al., 2004). significant effects on vitellogenesis and associated measures Inhibition of COX activity and PG synthesis is of special con- (plasma E2 and VTG). In vitro steroidogenesis assays indicate that cern in fish because COX inhibitors are ubiquitous in fish tissues IN has no impacts on E2 production (Aggregated Computational (eg, Brozinski et al., 2012; Corcoran et al., 2010) and aquatic sys- Toxicology Resource (ACToR; Judson et al., 2012; last accessed May tems (typically nonsteroidal anti-inflammatory [NSAIDs] drugs 14, 2016), and that it inhibits COX activity and PGE2 synthesis at present in the low lg/l range, with maximal concentrations concentrations that are at least 10–30 times lower than concen- nearing 100 lg/l in treated wastewater effluents; Corcoran et al., trations that initiate next most sensitive molecular targets (ie, in- 2010; Kone? et al., 2013), and because PGs play an important role terleukins, peroxisome proliferator-activated receptors). in fish reproduction. In teleosts, PGs have been shown to medi- In addition to generating empirical data specific to the action ate ovulation (eg, Takahashi et al., 2013), and there is some evi- of COX-inhibitors in fish, we utilized publically available, high- dence indicating their involvement in oocyte maturation (Lister throughput bioactivity data (ACToR), curated depositories of and Van Der Kraak, 2008; Sorbera et al., 2001). Importance of PGs chemical-gene interactions (Comparative Toxicogenomics for fish ovulation has been conclusively demonstrated by stud- Database [CTD]; Davis et al., 2015), and sequence similarity eval- ies that inhibited ovulation by impairing either PG synthesis uation tool (Sequence Alignment to Predict Across Species (using the mammalian COX inhibitor IN; Lister and Van Der Susceptibility; SeqAPASS; LaLone et al., 2016) to compare our Kraak, 2008; Patino et al., 2003; Stacey and Pandey, 1975), or both empirical, fish-related findings to mammalian data, and to de- PG synthesis and signaling (Fujimori et al., 2011). PGs also play velop adverse outcome pathway (AOP) for COX inhibition. an instrumental role in modulating fish reproductive behavior Development of AOPs has potential to support and enhance the and synchronization of reproduction between sexes. For in- use of mechanistic data in regulatory decision-making (Becker stance, endogenous PGs elicit spawning behavior in female fish et al., 2015; Villeneuve et al. 2014a,b). It involves organizing (eg, Kobayashi et al., 2002), while exogenous PGs (excreted by knowledge about mechanistic relationships between initial ovulating females) act as priming pheromones that facilitate chemical-biological interactions (molecular initiating events; physiological and behavioral reproductive readiness in males MIEs), intermediary key events (KEs), and adverse outcomes (Moore et al., 2002; Stacey and Sorensen, 2009). More specifically, (AOs) relevant to risk assessment (Ankley et al., 2010). prostaglandin F2a (PGF2a) and its metabolites released into water by ovulating females facilitate synchronization of reproductive readiness by increasing luteinizing hormone, 17,20b-dihydroxy- MATERIALS AND METHODS 4-pregnen-3-one, testosterone, and 11-ketotestosterone, express- Chemical selection and identification of exposure concentrations. Because ible sperm and reproductive behaviors in receiving male fish interactions of mammalian COX inhibitors (many of which are (Kobayashi et al.,2002; Lado et al.,2013; Moore et al.,2002; Stacey NSAIDs pharmaceuticals) with fish COX are not well character- and Sorensen, 2009). ized, we utilized mammalian mechanistic data to guide chemi- Several studies have indicated that exposure to mammalian cal selection. We selected NSAIDs with varying COX inhibition COX-inhibitors can adversely impact fish reproduction (eg, mechanisms with regard to selectivity and time-dependence of Flippin et al., 2007; Yokota et al., 2015). Medaka ovulation assays (in binding to COX. IB is a non-selective, reversible competitive vitro) indicated that diclofenac sodium, ketoprofen, salicylic acid, inhibitor; it inhibits COX-1 and COX-2 activity immediately after mefenamic acid, and acetylsalicylic acid exert anti-ovulatory ac- addition, and is eliminated from the COX active site rapidly (ie, tivity, which correlated with data published on the inhibitory ac- is not time-dependent; Selinsky et al., 2001). IN is also a non- tivity on human COX (Yokota et al., 2015). Flippin et al. (2007) selective, reversible competitive inhibitor, but unlike IB it typi- observed that a 6-week exposure of adult medaka to waterborne cally requires several minutes to bind, and takes hours to disso- ibuprofen (IB; 100 lg/l) led to a decreased number of spawning ciate from the active site. CX, demonstrates distinct, 3-step events. Lister and Van Der Kraak (2008) noted that exposure to in- inhibitory mechanisms typical of diarylheterocyclic inhibitors; domethacin (IN; 100 lg/l) reduced cumulative egg output in zebra- CX’s selectivity for COX-2 is achieved by a slow irreversible fish upon 16-d waterborne exposure. Although the effects on egg binding step which has been attributed to strong interactions output have been noted, molecular, cellular and organismal with the sulfonamide (or sulfone) group that fits into the side- events underlying the observed reproductive impairments have pocket of COX-2 (Selinsky et al., 2001; Walker et al., 2001). not been fully elucidated. For example, it remains unclear Water exposure concentrations were selected based on a whether mammalian COX-inhibitors can alter fish COX-activity combination of factors. The main objective was to initiate a bio- (Knight and Van Der Kraak, 2015). Consequently, it has been pos- logical response based upon COX inhibition, without causing tulated that their adverse reproductive effects on fish may be me- acute or overt toxicity due, eg, to narcosis or other reactive path- diated via modes of action independent of COX-inhibition (eg, ways. This could be particularly challenging with IB and IN, Knight and Van Der Kraak, 2015). where margins of safety (for mammals) are among the smallest To characterize mechanisms by which mammalian COX- observed among all pharmaceuticals (Berninger and Brooks, inhibitors may impact reproduction of female fish, we evaluated 2010). Because fish toxicity data for these COX inhibitors were 346 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

either not available (ie, CX), or limited to a handful of studies 16:8 h light:dark photoperiod, and fed adult brine shrimp (San and/or endpoints, we selected exposure concentrations for this Francisco Bay Brand, Newark, California) twice a day. Control study based on both in vivo results (where possible) and human (UV-treated Lake Superior) or chemical enriched water was therapeutic concentrations (“read-across” approach; eg, Huggett delivered to exposure tanks at a flow rate of 45 ml/min (com- et al., 2003). Although the “read-across” approach makes several plete tank renewal was achieved every 3.5 h). Chemical stocks, a assumptions that may affect accuracy of its toxicity estimate water blank, chemical-spiked water samples, and a duplicate (eg, the assumption is that there is a target conservation and sample from each one of the exposure tanks (C, IB, IN, and CX) that the interaction between drug and target will result in phar- were collected daily for confirmatory measurement of chemical macological response before a toxicological one; Rand-Weaver concentrations. et al., 2013), it can provide an informative starting point for After the 96 h of exposure, fish were removed from the test exposure concentration selection in absence of in vivo data for tanks, and anesthetized with tricaine methane sulfonate species of interest. To predict an effective dose for fish, the (100 mg/l buffered with 200 mg NaHCO3/l). Ovulation status was human peak plasma concentration (Cmax) and the log octanol- determined by presence/absence of expressible eggs upon apply- water partition coefficient (logKow) of the test chemicals were ing gentle pressure on abdomen. Blood was collected from the used to determine an aqueous effect threshold (AqET)—the con- caudal vein of the fish with a heparinized microcapillary tube centration likely to result in an effect based on the pharmaceut- and centrifuged to obtain plasma. Plasma was stored in micro- ical’s therapeutic mechanism of action. Approaches developed centrifuge tubes at 80 C for VTG, E2, and PG analyses. Ovaries by Huggett et al. (2003) and modified by Berninger et al. (2011) and livers were removed, divided equally amongst 3 microcentri- were used for these calculations and predicted AqET concentra- fuge tubes, cryopreserved in liquid nitrogen, and stored at 80 C tions of 2.8, 257, and 1.4 mg/l for CX, IB, and IN, respectively. for metabolomics, transcriptomics and COX-activity analyses. Because all 3 compounds are weak bases, the equations were modified following published approaches (eg, Berninger et al., Chemical analyses. Analysis of IN, IB, and CX in water was per- 2011; Supplementary Data) resulting in logD adjusted AqET val- formed by liquid chromatography. Sample aliquots of 50 mlwere ues of 2.8, 5902, and 142.6 mg/l for CX, IB, and IN, respectively. directly injected onto an Agilent 1100 Series HPLC with a Zorbax For IB and IN, concentrations where in vivo reproductive effects SB-C18 (2.1 75 mm, 3.5 mm) column. Separations were achieved had been observed (100 mg/l; Flippin et al., 2007; Lister and Van at 200 ml/min using isocratic elution of 56% acetonitrile in 0.1% Der Kraak, 2008) were also considered in exposure concentra- formic acid. Retention times (min) were 3.95, 4.40, and 5.0 for IN, tion selection. For IB, the predicted AqET and in vivo study IB, and CX, respectively. Concentrations of CX were determined responses matched fairly well (Flippin et al. (2007) observed by liquid chromatography-mass spectrometry (LC-MS) using neg- decreased frequency of egg production at 100 mg/l IB), but given ative polarity atmospheric pressure photoionization (APPI). the short time course (4 days) of our experiment, and much Toluene was supplied as a dopant at 50 ml/min for the APPI higher logD adjusted AqET, a higher concentration (200 mg/l IB) source. The MS was operated in SIM (selective ion monitoring) was selected. For IN, the in vivo effect concentrations fell mode at 380 and 381 m/z. Source parameters were optimized by between the AqET and the adjusted AqET, thus a concentration flow injection analysis and were as follows: fragmentor voltage of 100 mg/l was identified as likely to elicit COX inhibition during 100, gas temperature 350C, vaporizer temperature 325C, drying the 4 days exposure period; this is in agreement with in vivo gas 10 l/min, nebulizer pressure 60 PSIG, capillary voltage 4000 v, studies that observed reduced egg output by IN at 100 mg/l (eg, gain 14. IB and IN were both determined using a diode array Lister and Van Der Kraak, 2008). For CX, a final value of 20 mg/l detector at 220 nm. Quantification of all compounds was per- was ultimately selected, given its relatively larger margin of formed by an external standard method. Concentrations of safety and our desire to be above a COX inhibition threshold. No matrix spikes averaged 94.0 6 0.05% (mean 6 SD, n ¼ 10), 101 6 in vivo fish reproduction data were available to further evaluate 0.04% (n ¼ 10), and 101 6 0.04% (n ¼ 10) for CX, IB, and IN, respec- selected CX concentration. tively. Duplicate analysis agreement was 96 6 0.05% for CX (n ¼ 5) and >99% for both IB (n ¼ 5) and IN (n ¼ 5). Animals. All laboratory procedures involving animals were reviewed and approved by the local Animal Care and Use Ovarian COX activity and plasma E2, VTG, and PG analyses. For COX Committee in accordance with Animal Welfare Act regulations activity measurement, ovarian tissue collected from C, IB, IN, and Interagency Research Animal Committee guidelines. Adult, and CX fish, was extracted following the procedure described by sexually mature fathead minnows (age 7 months, females) were Lister and Van Der Kraak (2008). Ovary pieces were homogen- obtained from an on-site culture facility at the U.S. ized with 0.1 M Tris-HCl buffer containing 1 mM EDTA (pH 7.8) Environmental Protection Agency (EPA) lab in Duluth, Minnesota. at a ratio of 10 ml/mg tissue. The homogenates were centrifuged at 10 000 g for 15 min at 4 C. The supernatant was removed, Chemical exposures. Each chemical super-stock was prepared by stored on wet ice and analyzed immediately. Ovarian total COX dissolving “neat” chemical in UV-treated Lake Superior water enzyme activity was measured and quantified using fluorescent (IB 10 mg/l; IN 1 mg/l and CX 1 mg/l) without use of carrier sol- activity assay kit designed to detect combined COX-1 and COX-2 vents. Stocks were sonicated for 2 h at 35 C to assure that activities (Cayman Chemicals, Ann Arbor, Michigan, USA). chemicals were dissolved. Just prior to onset of chemical expo- PGF2a concentrations were measured using an enzyme- sure, these stocks were further diluted with UV-treated Lake linked immunosorbent assay (ELISA) per manufacturer’s Superior water to achieve the target concentrations of 200 mg/l instructions (Cayman Chemicals, Ann Arbor, Michigan, USA). (IB), 100 mg/l (IN), and 20 mg/l (CX). Plasma VTG concentrations were determined using an ELISA One day prior to the onset of exposure, 8 adult female fish with a polyconal antibody to fathead minnow VTG, and purified were placed in each of 8 exposure aquaria (20-l aquaria). fathead minnow VTG as a standard (Ankley et al., 2001). Plasma Chemical treatments (Control (C), IB, IN, or CX) were randomly E2 concentrations were measured by radioimmunoassay assigned, and each chemical exposure was conducted in dupli- (Jensen et al., 2001). Statistical significance of effects of chemical cate. Throughout the exposure the fish were kept at 25 C under treatment on COX activity, E2, PGF2a, and VTG was evaluated MARTINOVIC´ -WEIGELT ET AL. | 347

using 2-way analysis of variance (ANOVA) (with ovulation sta- Bioanalyzer (Agilent, Wilmington, Delaware, USA) and quantity tus, and chemical treatment as factors) followed by the was determined using a Nano-Drop ND-1000 spectrophotome- Dunnett’s post hoc test. The differences were considered signifi- ter (Nanodrop Technologies, Wilmington, Delaware). Total RNA cant if P < .05. All data were evaluated for normality (normal was stored at 80 C until analyzed with oligonucleotide micro- probability of residuals was examined) and homogeneity of var- arrays. Microarray analysis was conducted at the University of iance (Lavene’s test). Because these requirements were violated Florida (Gainesville, Florida, USA) using a 15K feature fathead in some cases, the data was also analyzed using non- minnow microarray (Gene Expression Omnibus, Accession plat- parametric Kruskal-Wallis ANOVA (with chemical treatment as form number GPL9248) designed in Dr Nancy Denslow’s labora- a factor). Because the outcomes of the non-parametric test were tory (University of Florida, Gainesville, Florida, USA) and congruent with those of 2-way ANOVA, which is robust and can manufactured by Agilent Technologies (Palo Alto, California, tolerate some departure from the requirements of normality USA). RNA samples of 500 ng were used for cDNA synthesis; and homoscedasticity, we report the results of 2-way ANOVA cRNA labeling, amplification, and hybridization were performed for COX, E2, VTG, and PG analyses. following the manufacturer’s kits and protocols (Quick Amp Labeling Kit; One-Color Microarray-Based Gene Expression Metabolomics. Individual liver samples were extracted in a 96- Analysis Protocol v 5.7; Agilent Technologies). An Agilent well plate format using a dual-phase extraction process to pro- Technologies G2505B scanner was used to scan microarray file changes in relative abundances of hepatic endogenous images at 5-mm resolution and data were extracted using metabolites (Ekman et al., 2012). Polar extracts were vacuum Agilent Feature Extraction software version 9.5 (Agilent dried and then reconstituted in 200 ml of 0.1 M sodium phos- Technologies). Consistent with the “Minimum Information phate buffered deuterium oxide (pH 7.4) that contained 50 mM About a Microarray Experiment Standards” (Brazma et al., 2001), sodium 3-(trimethylsilyl) propionate-2,2,3,3-d4 (TSP). Samples text versions of the Agilent raw data from this study have been were analyzed with 1H-NMR spectroscopy based on an auto- deposited in the Gene Expression Omnibus (GSE72976). mated push-through direct injection technique (Teng et al., Microarray data were imported into JMP Genomics version 2012), using acquisition parameters reported in the 4.1 (SAS Institute Inc., Cary, North Carolina). Raw intensity data Supplementary Data. were subjected to a logarithm base-2 transformation followed NMR data analysis. The spectra were zero-filled, line broad- by median normalization. Genes differentially transcribed ened at 0.3Hz, and Fourier transformed (ACD/1D NMR Manager, (DEGs) between controls and fish exposed to either IB, IN, or CX Advanced Chemistry Development). Using an automated rou- were identified by 1-way ANOVA (P < .05). The Database for tine, the spectra were phase- and baseline-corrected, referenced Annotation, Visualization and Integrated Discovery (DAVID ver- to TSP, and binned at a width of 0.005 ppm. We excluded several sion 6.7; Huang et al., 2009) was used to identify statistical regions of binned data to eliminate a residual water peak (4.82– enrichment of groups of genes that were populating pathways 4.88 ppm), residual methanol peak (3.33–3.37 ppm), and a large in the Kyoto Encyclopedia of Genes and Genomes (KEGG) resonance from betaine that exhibited high leverage and varia- (Tanabe and Kanehisa, 2012). Human homologs and human- bility (3.26-3.29 ppm). Remaining bins were then normalized to specific pathways were used for these analyses. Human homo- unit total integrated intensity. We used principal component logs were identified using the NCBI Human Reference Sequence analysis (PCA) to compare metabolite profiles across the 4 expo- database; e-value < 1e-4, and identity match >75% were used as sure classes (SIMCA-13.0, Umetrics Inc.), as this statistical a criteria for homolog identification. Only genes represented on framework has detected biological and metabolomic responses the microarray platform were used as background genes for to a wide variety of contaminant exposures in other studies (eg, KEGG analyses. Enrichment was considered significant if Davis et al., 2013, Ekman et al., 2011, Viant et al., 2006). Cross- Expression Analysis Systematic Explorer (EASE) score < 0.1 (a validation was used to determine the number of PCA compo- recommended default cutoff); EASE score is a more conservative nents and cumulative Q2 (Q2cum) was used to assess model equivalent of 1-tail Fisher Exact probability value (Hosack et al., over-fitting. Outliers were identified with a Hotelling’s t2 test at 2003). the 95% CI. A 1-way ANOVA with a post hoc Tukey’s test (R Because the primary purpose of the above microarray- v.2.15.1, http://www.R-project.org) assessed whether average generated data analyses was to generate mechanistic hypothe- PCA score values differed among the exposure classes; a sepa- ses we did not apply stringent multiple adjustment statistics/ rate analysis was run for first and second component scores. To criteria to identify DEGs and associated enriched KEGG identify individual endogenous metabolites that were signifi- Pathways, as we wanted to limit number of false negatives for cantly affected by an exposure, we also generated t-test-filtered these exploratory analyses. Conversely, when conducting gene difference spectra based on the binned and normalized spectra. set enrichment analyses (GSEAs; see below for details), where On a pairwise basis, we subtracted the average spectrum of con- we tested preselected hypotheses (ie, that exposure to COX- trol females from the spectrum of each female within a given inhibitors would lead to enrichment of predefined sets of genes exposure class. We then applied a t-test to each spectral bin to associated with fish reproductive function) we used a more con- assess whether those average differences (ie, exposed vs conservative approach—false discovery rate (FDR) cutoff of 25% was trol) were significantly different. For significantly different spec- used for GSEA (Subramanian et al., 2005). tral bins, we identified metabolite peaks using Chenomx NMR To determine whether and how COX inhibitors interfere Suite 7.0 (Chenomx Inc.). The suitability and interpretation of with sets of gene that modulate fish reproduction (Ingenuity these statistical approaches have been justified and discussed Pathway Analyses (IPA; Ingenuity Systems, Inc., USA) and KEGG elsewhere (Collette et al., 2010; Eriksson et al., 2006). are poorly populated with fish-specific pathways) we conducted GSEA using gene sets corresponding to features represented Transcriptomics. For microarray analyses, total RNA was isolated within a conceptual model of the teleost fish brain-pituitary- from female ovary samples (n ¼ 7–8 per treatment) using an gonadal-hepatic (BPGH) axis (Villeneuve et al., 2012). Briefly, RNeasy Mini RNA purification kit (Qiagen Inc., Valencia, these gene sets contain genes involved in normal BPGH California, USA). RNA quality was assessed with an Agilent 2100 functioning, as well as those indicative of exposure to 348 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

endocrine-active chemicals. The gene sets are organized based mammalian COX enzymes to deliver their therapeutic benefits, on tissue type (ovary, liver, and brain compartments), and key protein sequence and structural conservation was evaluated by functional categories (eg, steroid biosynthesis, oocyte growth, comparing the human PG G/H synthase 1 (COX-1; National vitellogenesis, PG-related; Villeneuve et al., 2012). Center for Biotechnology Information (NCBI) accession For GSEA analyses, the microarray raw data were normalized NP_000953) and the human PG G/H synthase 2 (COX-2; NCBI using Fastlo11 implemented in R (http://www.r-project.org/). accession NP_000954) enzymes across species using SeqAPASS GSEA was conducted using GSEA version 2.0 software (http:// tool (LaLone et al., 2016). Comparisons between the human pro- www.broadinstitute.org/gsea/) following published procedures teins were conducted, initially, by assessing the percent similar- (Subramanian et al., 2005). Normalized gene expression matrix ity between primary amino acid sequences and then at the level files, and files defining the phenotypes (experimental cases) of the functional domains (NCBI Conserved Domain Database, were constructed using custom R code (Villeneuve et al.,2012). cd09816 for both COX-1 and COX-2), which contain the heme These files, along with a .gmt file that included 23 gene sets binding site, substrate binding site, and homodimer interface (Villeneuve et al. 2012), were imported into the GSEA software. All for each sequence. Further, individual amino acid residues GSEA analyses were performed using 1000 permutations, with deemed to be critical for inhibition of COX-2 were evaluated for the permutation type set to phenotype. Gene sets with size <10 conservation across species using NCBI Constraint-Based Local or >500 were excluded from GSEA analyses. To determine signifi- Alignment Tool, COBALT (Rowlinson et al., 2003). Specifically, cantly enriched gene sets binary combinations of phenotypes the human COX-2 sequence (NP_000954) was used as the tem- were considered in each analysis (ie, C vs IB, C vs CX, C vs IN); a plate, identifying residues that aligned with Tyr-348, Try-355, FDR cutoff of 25% was used. Tyr-385, and Ser-530 across taxa (Supplementary Data, COX-2 Finally, to broaden the scope of the analyses, IPA software Individual AA). was used to identify additional enriched canonical pathways, as well as putative chemicals and transcription factors potentially Preliminary evaluation of functional conservation. To evaluate con- responsible for the observed gene expression changes. cordance of the empirical, fish-specific data generated by our study with the published literature, we also queried the CTD Derivation and evaluation of putative AOPs. Putative AOPs (http://ctdbase.org/), which largely curates chemical-gene/ (Villeneuve et al., 2014b) were developed following OECD guide- protein interaction data from empirical mammalian studies lines (OECD, 2013, 2014) as a part of OECD Project 1.29. “Catalog (Davis et al., 2015). We used CTD to extract genes and pathways of putative AOPs that will enhance the utility of U.S. EPA associated with the chemical (IN, CX, or IB) exposure. Last, we ToxCast high-throughput screening (HTS) data for hazard iden- used the ACToR’s Interactive Chemical Safety for Sustainability tification”. Each of these were deposited in the AOP Wiki Dashboard (iCSS; http://actor.epa.gov/dashboard/) to investigate (https://aopwiki.org)—a central repository for AOPs developed whether there were commonalities between IN fish gene as a part of a joint effort between European Commission—DG expression and in vitro mammalian HTS data for IN provided via Joint Research Center and U.S. EPA. Unique numbers have auto- the iCSS Dashboard. This dashboard provides access to results matically been assigned to the AOPs by AOP-Wiki (63, 100, 101, of high throughput chemical screening data generated by 102, and 103); these numbers are subsequently used throughout Toxicity Forecaster (ToxCast) and Toxicity Testing in the 21st when referring to these AOPs. century (Tox21) programs (Kavlock et al., 2012; Schmidt, 2009). Preliminary assessment of the relative level of confidence in Because the article is focused on the impacts of COX- the newly developed putative AOPs was conducted following inhibitors on reproduction-related genes and endpoints, data the weight of evidence considerations outlined by Becker et al. and assessment of non-reproduction related processes are not (2015). We evaluated the biological plausibility of the putative discussed in detail. However, these results, and statistical anal- AOPs; this involved determining whether relationships between yses relevant to other processes are presented in the upstream and downstream KEs are consistent with established Supplementary Data as they may be of interest to the commun- understanding of unperturbed biological pathways. We also ity and may inform the development of other AOPs addressed the essentiality of the KEs in a context of each AOP; (Supplementary Tables 1–3). essentiality is typically evaluated by determining whether downstream KEs/AOs can be prevented if upstream KEs are blocked. Typically, weight of evidence analysis for well- RESULTS characterized AOPs also involves assessment of empirical sup- Mean concentrations (mg/l 6 SD) of chemicals in exposure tanks port (eg, toxicity studies with contaminants that disrupt MIE of were 213 (19.1) for IB, 91.9 (6.6) for IN, and 29.3 (4.9) for CX. There interest) for temporal, dose–response, and incidence concord- was 1 mortality in the IB treatment; it occurred within 1.5 h of ance. Given the putative nature of our AOPs, and the nature of introduction of the fish to the exposure tank. The number of this article (research article), we place emphasis on presenta- ovulating females (defined as females that had expressible tion/interpretation of the empirical data generated, and only eggs) on the morning of tissue collection (day 4 of exposure) incorporate the evaluation of empirical evidence in the discus- was similar across treatments; 56% for control and 50% for IB, sion in a cursory fashion. Based on the collaborative ethos of IN, and CX treatments. the wiki platform for AOP development, it is expected that other investigators will contribute additional data and analyses that Ovarian COX Activity can help support or refute the relationships depicted. Chemical treatment had a significant effect on ovarian COX Nevertheless, major uncertainties, inconsistencies and data activity in IN and CX vs. control fish (2-way ANOVA, Figure 1A). gaps related to the putative AOPs were identified and are There was also a significant effect of ovulation status on COX presented. activity, and an interaction between ovulation status and COX inhibition. However, post hoc tests did not detect significant dif- Evaluation of sequence/structural conservation of MIE. With knowl- ferences between nonovulated controls versus nonovulated IB, edge that the chemicals of interest (IN, IB, CX) act on IN or CX-treated fish and ovulated controls versus ovulated IB, MARTINOVIC´ -WEIGELT ET AL. | 349

FIG. 1. Effects of mammalian COX inhibitors on (A) ovarian COX activity (n ¼ 47), and, (B) circulating prostaglandin-F2a (n ¼ 64) in female fathead minnows (2-way ANOVA followed by the Dunnett’s post hoc test). Speciﬁc treatments signiﬁcantly different (Dunnett’s pos -hoc test, P < .05) from their respective controls (ie, ovulated treated vs ovulated controls, and non-ovulated treated vs non-ovulated controls) are labeled with an asterisk (*).

IN or CX- treated fish. Thus, throughout the article, we refer to reduced concentrations of PGF2a; this effect was observed in COX as not affected by IN, CX, or IB. Both, log transformed and both ovulated and non-ovulated fish (Figure 1B). There was no raw data yielded same results. significant effect of ovulation status, chemical treatment, or their interaction on plasma E2 or VTG concentrations Plasma E2, VTG, PG Analyses (Supplementary Figs. 1A and B). Both ovulation status and chemical treatment had an effect on plasma PGF2a concentrations, but there was no significant Metabolomics interaction between the 2 factors. Plasma PGF2a concentrations PCA indicated that liver metabolomic profiles of females were higher in ovulated females. Only IN exposure significantly exposed to IN, IB, or CX were similar to control fish; exposure 350 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

classes were not significantly different from one another along TABLE 1. Overview of enrichment of predefined sets of genes associ- either PC1 or PC2 (ANOVA PC1: F2, 43 ¼ 0.571, P ¼ .569; ANOVA ated with fish reproductive function—GSEA for microarray data PC2: F2, 43 ¼ 0.871, P ¼ .426). IN fish exhibited the most separa- Gene Set (total number of genes in each gene set) IN IB CX tion from other treatments (along PC2) suggesting that IN had the most extensive, albeit small, effects on the liver metabo- Gonad All (78) C lome (Supplementary Figure 2A). Average difference spectra Gonad All With Neighbors (227) C (exposed minus lab-control) of hepatic metabolites for IN, IB, Gonad Steroid Biosynthesis (11) C and CX females indicated that only a few metabolites changed Gonad PG-related (10) IN relative to the controls (Supplementary Figure 2B). Due to the Gonad Oocyte Maturation (21) paucity of compelling liver metabolome responses, ovarian Gonad Oocyte Growth (10) transcriptomic endpoints and other apical measurements ulti- Gonad Cholesterol Synthesis (12) CX mately served as the primary basis for AOP development. Gonad Minus Cholesterol Synthesis (67) C Nevertheless, while NMR methodologies may not be able to Gonad Ccnb1 Plus Neighbors (106) detect subtle effects, especially not the effects on the less abun- Gonad Acvrl1 Plus Neighbors (12) CC dant metabolites, the lack of effects observed via NMR has some Gonad Oocyte Maturation Plus Neighbors (143) utility for data interpretation—it indicates that the concentra- Gonad Oocyte Growth Plus Neighbors (113) Gonad Cholesterol Uptake Plus Neighbors (15) C tions of COX-inhibitors were likely below those that would Gonad Steroid Biosynthesis Plus Neighbors (12) C cause overt toxicity. This suggests that concentration we used Gonad Ccnb1 Zp3 Plus Non-Redundant Neighbors (118) CX in this study likely minimized initiation of less sensitive, secon- Liver All (38) C dary targets, and predominantly interacted with primary targets Liver Cholesterol Biosynthesis (13) (COX activity/PG synthesis in this case), at least in case of IN. Liver Steroid Metabolism (17) Pituitary All (26) Brain All (19) Transcriptomics Hypothalamic-Pituitary-Gonadal-Hepatic All (138) C When compared with controls, exposure to IB or CX resulted in differential expression of a comparable numbers of genes (IB ¼ Gray shading indicates significant enrichment in the control group compared 433, CX ¼ 545). In contrast, there were 2558 differentially with the treated; black shading indicates significant enrichment in the treated expressed genes (DEGs) in IN-treated fish compared with con- group relative to control; lack of shading indicates no significant enrichment (FDR < 25% was considered statistically significant). C, Control; IN, Indomethacin; IB, trols, indicating more extensive effects. About 28% of IB DEGs Ibuprofen; CX, Celecoxib. and 26% of CX DEGs were also differentially expressed in IN fish (Supplementary Table 1); 24 DEGs were common to all 3 chemical treatments. associated with: (1) ribosomes, ribosomal proteins and transla- Gene set enrichment analyses. Focused investigation of impacts tion, (2) zona pellucida, and (3) rRNA binding. of IN, IB, and CX on gene sets involved in fish BPGH functioning In conclusion, GSEA pointed to a variety of responses indica- (using gene lists from Villeneuve et al., 2012) indicated that mul- tive of effects on the BPGH functioning, but only IN-exposed tiple reproduction-related gene sets were altered (Table 1). The fish exhibited enrichment of gene sets specifically associated “Gonad PG-related” gene set was significantly enriched in IN with PG synthesis (additional GSEA outputs are provided for IN- fish when compared with C (Table 1, Supplementary Table 2). exposed fish in Supplementary Table 2). The genes driving enrichment of the PG gene set were mildly Canonical pathway enrichment. Exposure to IN, IB, and CX (max 1.8-fold versus control), but consistently up-regulated in largely resulted in enrichment of: (1) canonical pathways asso- IN-treated fish (eg, PTGSs 2 and 2a, carbonyl reductase 1, and PG ciated with processes normally modulated by PGs (eg, KEGG D2 and E synthases). Interestingly, only 1 gene contributing to pathways for renin-angiotensin, cardiac contraction, oocyte the PG gene set enrichment was down-regulated (0.85-fold vs meiosis, cell adhesion molecules, nucleotide excision repair; control) in IN fish—phospholipase A2 group IVa. GSEA for IB Iyer et al., 2000; Wennmalm, 2013; Zou et al., 2002), or (2) path- and CX did not reveal any effects on PG-related gene set. ways indicative of known therapeutic effects/side effects of When comparing IB and control fish, enrichment was COX-inhibitors (eg, beneficial uses for treatment of symptoms observed only in controls (Table 1). The genes that contributed of neurodegenerative diseases and oxidative stress; Simmons to core enrichment of gene sets “gonad steroid biosynthesis” et al., 2004; Supplementary Table 2). More specifically, 3 KEGG and “gonad steroid biosynthesis with neighbors” included cyto- pathways were enriched in IB-treated fish: ERbB signaling, cell chrome P450, Family 19, Subfamily A, Polypeptide 1a (cyp19a1a) adhesion molecules and long term depression (Supplementary and cytochrome P450 17a1 (cyp17a1), hydroxy-delta-5-steroid Table 2). Despite the low number of DEGs (545) in CX fish, 7 dehydrogenase, 3 beta steroid delta-isomerase (3bhsd), and 11- KEGG pathways were significantly enriched, including neuroac- beta-hydroxysteroid dehydrogenase-like (11bhsd-like)—all were tive ligand-receptor interaction, renin-angiotensin systems, downregulated in IB fish when compared with control. cell-adhesion molecules and the gonadotropin releasing hor- CX had a somewhat different profile than IB; 2 gene sets were mone (GnRH) signaling pathway (Supplementary Table 2). IN enriched in CX fish (Table 1) and 4 were enriched in controls impacted 8 pathways, including Huntington’s, Alzheimer’s and (including “gonad activin receptor1 plus neighbors”, set, which Parkinson’s disease, ribosome, oxidative phosphorylation, car- was also enriched in controls when compared with IB; Table 1). diac muscle contraction, and oocyte meiosis (Supplementary Genes driving enrichment of the “cholesterol uptake” gene set Table 2). Closer examination of the oocyte meiosis pathway (eg, apolipoprotein Eb [apoeb], steroidogenic acute regulatory enrichment indicates dysregulation by IN was not limited to a protein [star]) were down-regulated in CX-treated fish. Two gene discrete portion of the pathway; several regulatory (androgen sets were enriched/upregulated in CX-exposed fish: (1) “gonad receptor; ar) and signaling molecules (adenylate cyclase [ac], cholesterol synthesis”, and (2) “gonad_ccnb1_zp3_plus_non- protein kinase A [pka], ribosomal protein S6 kinase, 90 kDa, pol- redundant_neighbors”—genes in this category are functionally ypeptide 6 [rsk 1/2], as well as the downstream genes MARTINOVIC´ -WEIGELT ET AL. | 351

FIG. 2. Enrichment of oocyte meiosis pathway in IN-exposed fish. The functional annotation analyses (DAVID) indicated statistically significant enrichment of oocyte meiosis pathway (KEGGs). EASE Score (a more conservative equivalent of 1-tail Fisher Exact probability value used by DAVID) was considered significant at P < .1. DEGs contributing to oocyte meiosis pathway enrichment are labeled with a star adjacent to the gene symbol.

(anaphase-promoting complex-cyclosome [apc/c], budding maturation (Figure 4, https://aopwiki.org/wiki//index.php/ uninhibited by benzimidazoles [bub3], mitotic arrest deficient Aop:102; https://aopkb.org/aopwiki/index.php/Aop:103) were [mad1/2], and separase) involved in oocyte meiosis were developed. impacted by IN-exposure (Figure 2). At this early stage of AOP development, we elected to use Additional canonical pathways associated with reproductive total COX inhibition as the MIE. Currently there is only a weak function were enriched in IN fish (IPA analyses). These included empirical evidence that fish COX activity can be impacted by androgen signaling, estrogen receptor signaling, GnRH signal- human COX-inhibitors (as determined by the indirect assessing, dopamine receptor signaling, peroxisome proliferator- ment of the peroxidase activity of the COX). However, based on activated receptors/retinoid X receptor (PPAR/RXR) activation, a handful of empirical studies in which impacts on fish COX and glucocorticoid receptor and corticotropin releasing hor- activity were observed (eg, Flippin et al., 2007; Galus, 2014), the mone signaling (Supplementary Table 3). similarity of fish and mammalian COX functional domains, and the need to definitively address this gap in knowledge we desig- Derivation of Putative AOPs for COX Inhibition Leading to nated decreased COX activity, and not PG reduction, as the MIE Reproductive Impairment (for further discussion of this decision please see “Evaluation of Because IN was the only chemical treatment that had an effect Sequence/Structural Conservation of MIE” and “Discussion” secon circulating PG, and because those responses were congruent tions). Moreover, we retained COX inhibition as the assumed with the transcriptomic data (altered PG synthesis pathway; MIE because of the high strength of empirical evidence for its Table 1), we used IN as a model COX inhibitor for AOP develop- inhibition by NSAIDs in humans, and because of the nature of ment. Putative AOPs for COX inhibition leading to reproductive the fish studies that reported lack of inhibition. Specifically, impairment via disruption of: (1) ovulation (Figure 3A; https:// most of the fish studies have used COX-activity assays that aopkb.org/aopwiki/index.php/Aop:63), (2) pheromonally- were not optimized for fish (Patel, 2014), were conducted, and mediated synchronization of behavior and reproductive readi- did not employ temporally or concentration-intense designs in ness (Figure 3B; https://aopkb.org/aopwiki/index.php/Aop:101), well-characterized tissues that would allow one to confidently (3) performance of female spawning behavior (Figure 3C; reject this hypothesis. In addition, conceptually one would not https://aopkb.org/aopwiki/index.php/Aop:100), and (4) oocyte necessarily expect to observe effects on MIE and KEs at the 352 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

FIG. 3. Putative adverse outcome pathways for cyclooxygenase inhibition leading to reproductive dysfunction via inhibition of (A) ovulation, (B) female spawning behavior, and (C) pheromone release in reproductively mature female ﬁsh. MIE, molecular initiating event; KE, key event; AO, adverse outcome.

same point in time. Indeed, the expected, temporally concord- performance of female-specific spawning behaviors and subse- ant pattern may be that effects such as that on COX activity is quent decrease in female’s ability to attract spawning mates transient and by day 4 offset by compensatory responses, but (AOP 100; Figure 3B), and (3) in urine it leads to reduced output nonetheless still influencing downstream events (ie, circulating of PG-based pheromones and subsequent decreases in female’s PGs). Furthermore, in many cases highly dynamic molecular ability to attract/secure spawning mates (AOP 101; Figure 3C). scale responses, such as COX activity, will be much better The AOPs defined above share the 2 AOs—reduced reproductive studied/demonstrated in vitro, and thus lack of evidence for success and subsequent declining trajectory of the populations. inhibition of COX with in vivo fish studies is not necessarily Two putative AOPs that involve dysregulation of GnRH and unexpected. subsequent impairment of oocyte maturation (Figure 4) were We did not attempt to resolve which COX isoform was developed in response to extensive transcriptomic effects along impacted as there are typically 3–4 isoforms of COX in fish oocyte meiosis (Figure 2) and GnRH signaling pathways (Ishikawa et al., 2007), and their identities and specific roles in (Supplementary Table 3) in IN-exposed fish, as well as the evi- fish reproduction have not been fully characterized. A handful dence from mammalian literature indicating involvement of of studies that attempted to elucidate differential role of COX-1 PGs in GnRH release and GnRH neuronal function (Clasadonte versus COX-2 in fish reproduction have indicated that both iso- et al., 2011a,b). These 2 AOPs form an AOP network that shares forms have roles in reproduction (zebrafish; Lister and Van Der the MIE and the first 5 KEs describing reduced PG synthesis, Kraak, 2008), and that their roles may be species-specific GnRH and LH secretion, and subsequent reduction of matura- (Fujimori et al., 2011). Overall, COX-2 has been more strongly tion inducing steroids (MIS). Because IN-exposed females exhib- implicated as a driver of ovulation in some teleost species (ie, ited significant alterations in GnRH signaling and medaka; Fujimori et al., 2011; Takahashi et al., 2013). downregulation of nuclear AR expression, as well as dysregula- For the first 3 AOPs (Figs. 3A–C) COX inhibition is identified tion of the androgen signaling pathways associated with both as the MIE which leads to the first KE—reduction of PG synthe- genomic and non-genomic androgen signaling we propose that sis. PGs PGE2 and PGF2a are suggested for the first KE because impacts on oocyte maturation may be associated with these they have been either reduced upon exposure to IN in our study, changes. Both, androgens (in mammals, amphibians and some or because literature has indicated that they serve as the media- fish; Hammes, 2004; Tokumoto et al., 2011) and/or 17a,20b-dihy- tors of ovulation (eg, PGE2- Fujimori et al., 2011; PGF2a—Crespo droxy-4-pregnen-3-one (17,20 bP and its derivatives in fish; et al., 2015), female-specific spawning behavior (eg, PGF2a— Suwa and Yamashita 2007; Tokumoto et al., 2012) have been Juntti et al., 2016) and/or have a pheromonal function in fish (eg, shown to serve as MIS. Several frog and fish studies indicate PGF2a—Stacey and Sorensen, 2009). After KE1 the AOPs diverge: that androgens may play an important role in promoting oocyte (1) in ovarian theca cells KE1 leads to reduced stimulation of PG maturation (most likely via classical nuclear AR agonism; receptors in granulosa cells and thus impairment of ovulation Hammes, 2004; Lutz et al., 2001). For above reasons, COX inhibi- (AOP 63; Figure 3A), (2) in the preoptic area of the brain it leads tion and subsequent reduction in nuclear AR and/or membrane to reduced stimulation of PG receptors, which results in reduced PR signaling (Figure 4) is hypothesized to lead to (1) increased MARTINOVIC´ -WEIGELT ET AL. | 353

FIG. 4. Putative adverse outcome pathways for cyclooxygenase inhibition leading to reproductive dysfunction via interference with of oocyte meiosis via (A) reduced meiotic prophase I/metaphase I transition (reduced germinal vesicle breakdown) due to cyclic adenosine monophosphate increase, and (B) upregulation of spindle assembly checkpoint proteins and aneuploidy in reproductively mature female ﬁsh. MIE, molecular initiating event; KE, key event; AO, adverse outcome; E2, 17b-estradiol; GnRH, gonadotropin releasing hormone; MAD2, mitotic arrest deﬁcient-like 2. intracellular cAMP (supported by the upregulation of adenylate candidates (Supplementary Figs. 3A and B). Of the vertebrate spe- cyclase and protein kinase A mRNA in IN-exposed fish) and cies sharing sequence similarity to the human COX-1 and COX-2, reduced meiotic prophase I/metaphase I transition (measurable 90 and 54 of these protein sequences were identified as ortholog by decreased germinal vesicle breakdown [GVBD] rates; Figure candidates, respectively (Supplementary Data). Because ortholog 4A, AOP 102), or (2) upregulation of spindle assembly checkpoint sequences are more likely to maintain similar function, these proteins (supported by the upregulation of Mad-2 in IN-exposed results provide a line of evidence to indicate that vertebrates and fish), and thus inhibition of anaphase onset of oocyte meiosis some invertebrates preserve a relevant MIE for interaction with and interference with chromosome segregation (Figure 4B, COX inhibitors. AOP 103). Both of these AOPs culminate in reduced reproduc- Evaluation of sequence similarity within regions identified tive success and subsequent declining trajectory of the as the key functional domains for COX-1 indicated 69.7–96.6% populations. similarity among ortholog candidates for species within the taxonomic groups Mammalia, Aves, Crocodylia, Testudines, Evaluation of Sequence/Structural Conservation of MIE Amphibia, Actinopteri, Lepidosauria, Chondrichthyes, and SeqAPASS analyses of cross-species primary amino acid Chondrichthyes (Supplementary Figure 3C). Consistent with the sequence similarity compared with the human COX-1 and COX-2 primary amino acid sequence evaluation, COX-2 maintains con- indicated a substantial degree of conservation across vertebrate servation of the functional domain among both vertebrates and species for both enzymes (Supplementary Figure 3). Sequence invertebrates (ortholog candidates between 47.4 and 97.7% simi- similarity ranging between 53.6 and 15.5% compared with larity; Supplementary Figure 3D). These data provide an addi- human COX-2 was also identified with invertebrates (Bivalvia, tional line of evidence that COX enzymes are conserved among Branchiostomidae, Anthozoa, Malacostraca, Ascidiacea, and vertebrates and some invertebrates, predicting they may simi- Gastropoda). Ten invertebrates were identified as ortholog larly interact with known mammalian COX inhibitors. 354 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

SeqAPASS was also used to align 4 key residues linked to are associated with PG function or therapeutic targets of COX- functional conformations relative to binding of COX-inhibitors. inhibitors. Furthermore, as we have noted before, GSEA was Specifically, site-directed mutagenesis and evaluation of a crys- statistically more conservative than canonical pathway analy- tal structure of murine COX-2, identified amino acid residues ses in terms of limiting FDRs. Unfortunately, our COX activity important for chelation of ligands and direct chemical interac- data did not provide definitive clarification as to whether the tion. Using the human COX-2 as a template protein sequence, MIE of interest was influenced by IB and CX. There are several we aligned all species identified as sharing sequence similarity possible explanations for this. First, basal levels of COX activity at the primary amino acid sequence level, specifically identify- in fish tissues are close to the limit of detection for the COX- ing residues aligning with tyrosine (Tyr) at position 348, Tyr at activity assay, and thus unlikely to provide accurate and/or pre- position 355, Tyr at position 385, and serine (Ser) at position 530. cise interpolation of COX activity (Patel, 2014). Second, the dif- This analysis showed that all 4 structurally and functionally ferences in the COX activity may be difficult to detect because important residues were conserved across the majority of verte- COX activity is spatially and temporally dynamic, especially in brate species (including fish) and invertebrates from Bivalvia, the ovary of asynchronous spawning fish (Lister and Van Der Branchiostomidae, Gastropoda, Anthozoa (Table 2; Kraak, 2008) such as the fathead minnow. Typically, COX and Supplementary Data). The 4 human COX-2 amino acid residues PGs are elevated only shortly before ovulation (eg, Fujimori aligned with 26 fish species (Organism class, Actinopteri). et al., 2011). Third, the homogenization and dilution of the sam- Overall, these data identified high conservation of key COX-2 ples may have impacted the evaluation of COX-activity inhib- residues comparing other vertebrates and certain invertebrates ition—the inhibition is not based on covalent modification so it to humans, further supporting that chemical perturbation of is plausible that tissue handling altered binding of COX- COX-2 as a MIE could be well conserved across a variety of inhibitors. species. Below, we examine extant literature to evaluate the proposition that the exposure concentrations of CX and IB were below Functional Conservation those needed to initiate targets of interest (COX and PG) in fish. To further evaluate the degree to which our proposed AOPs There are no published fish CX studies that could help with (early KEs in particular) might be applicable not only to fish, but evaluation of this proposition. Existing IB studies (Bhandari and higher vertebrates, we compared outcomes of our empirical IN Venables, 2011; Patel, 2014) indicate that the “read-across” pre- fish data with human HTS data and CTD-curated chemical- dicted IB concentration used in this study was within the order gene interaction data (largely mammalian). Comparison of the of magnitude of the effective dose for fish, but did not offer a IN-related empirical fish data with HTS outputs (http://actor. definitive answer. The exposure concentration used in our epa.gov/dashboard/) indicated many similarities between the study (ie, 200 lg/l IB) fell between a concentration that was too human HTS (Supplementary Table 4) and fish datasets. The low (100 lg/l IB) to reach plasma IB level within the human Cmax most sensitive targets of IN in HTS assays were PGs (ie, PGE2 range, and the concentration that was sufficient to do so measured indirectly by assessing PG E2 receptor stimulation; (270 lg/l IB) (fathead minnow, 48 h waterborne exposure; Patel,

EC50 ¼ 0.02 mM) and COX-1 activity (EC50 ¼ 0.07 mM; COX-2 was 2014). It is important to note that a large individual variability in not evaluated). Other sensitive targets of IN included cytokines IB plasma concentrations was observed for the 270 lg/l IB

(IL6 and CSF1; EC50 ¼ 0.6–0.7 mM). At higher concentrations (EC50 treatment—only one individual had IB plasma levels within ¼ 2.8–18 mM), IN resulted in PPAR activation, matrix metallopro- Cmax range, whereas others were outside of the range (2 were teinase upregulation, and alteration of endpoints associated below and 1 was above Cmax; Patel, 2014). Patel (2014) could not with cell adhesion, dopamine transport, and cell cycle. Majority reliably evaluate impacts on COX activity due to low basal levels of the above targets and/or pathways associated with those tar- of COX, but reported decreased gill PG metabolite levels upon gets were also altered in or experimental fish exposed to IN exposure to 370 lg/l IB, but not 9 lg/l IB. PG metabolite levels of (based on the transcriptomic data; Supplementary Tables 1–3). fish exposed to IB concentrations similar to those used in this Evaluation of CTD chemical–gene interaction outputs for IN study (200 lg/l) were not evaluated by Patel (2014). Bhandari and indicated that the following reproduction-related pathways Venables (2011) found that exposure to 50 and 100 lg/L IB that were enriched in IN-exposed fish were also curated in CTD reduced PGE2 in the gills of a bluntnose minnow, Pimephales as being impacted by IN: steroid hormone biosynthesis, PPAR notatus (a species closely related to P. promelas), and based on signaling, GnRH signaling, arachidonic acid metabolism, the gill IB and PGE2 measurements concluded that “read- progesterone-mediated oocyte maturation and oocyte meiosis across” prediction based on human data underestimated sensi- (Supplementary Table 5). tivity of the bluntnose minnow COX to IB. Because of lack of empirical evidence and poor understanding of intra- and interspecific variation in absorption, distribution, metabolism, and DISCUSSION sensitivity of fish COX to IB and CX, it remains uncertain Lack of the effects on COX activity, plasma PGF2a and PG- whether the concentrations of CX and IB used in this study related gene sets (via GSEA) in CX- and IB-exposed fish indicates were sufficient to impact COX and related processes. Therefore, that concentrations of CX and IB may have been below those only IN data (where significant impacts on plasma PGF2a and needed to initiate the MIE of interest (COX) and subsequent PG PG-related gene sets, and reproduction-related canonical path- reductions. However, it is notable that in both, IB and CX fish, ways were observed) were used to inform AOP development. canonical pathway enrichment analyses detected changes in gene expression indicative of exposure to COX-inhibitors (but Putative AOPs—Evidence GSEA analyses did not). This finding is not unexpected, as can- Reduction of PG synthesis, following exposure to COX inhibitors onical analyses and GSEA approaches evaluated different sets has been consistently demonstrated in mammals and is a basis of genes. The PG-related set for GSEA was largely comprised of a for therapeutic use of these compounds (reviewed in Simmons handful of genes involved in PG synthesis, whereas enriched et al., 2004). Fish studies that support this KE relationship are canonical pathways captured a broad variety of processes that largely lacking. Although fish studies have consistently reported MARTINOVIC´ -WEIGELT ET AL. | 355

TABLE 2. Comparison of key amino acid residues across select vertebrate species from different taxonomic groups using human PG G/H synthase 2 as the template sequence

NCBI Accession Protein Name Species Name Classa Y348 Y355 Y385 S530

NP_000954.1 PG G/H synthase 2 precursor Homo sapiens Mammalia YYYS XP_004028113.1 PREDICTED: PG G/H synthase 2 Gorilla gorilla gorilla Mammalia YYYS XP_524999.3 PREDICTED: PG G/H synthase 2 Pan troglodytes Mammalia YYYS XP_006023936.1 PREDICTED: PG G/H synthase 2 isoform X2 Alligator sinensis Crocodylidae YYYS XP_006267636.1 PREDICTED: PG G/H synthase 2 Alligator mississippiensis Crocodylidae YYYS XP_009504608.1 PREDICTED: PG G/H synthase 2 isoform X1 Phalacrocorax carbo Aves YYYS EOB04442.1 PG G/H synthase 2, partial Anas platyrhynchos Aves YYYS XP_003208549.1 PREDICTED: PG G/H synthase 2-like Meleagris gallopavo Aves YYYS NP_001161191.1 PG G/H synthase 2 isoform 1 precursor Gallus gallus Aves YYYS XP_009086668.1 PREDICTED: PG G/H synthase 2 Serinus canaria Aves YYYS XP_002190785.1 PREDICTED: PG G/H synthase 2 Taeniopygia guttata Aves YYYS XP_007056755.1 PREDICTED: PG G/H synthase 2 Chelonia mydas Testudines YYYS XP_005313923.1 PREDICTED: PG G/H synthase 2 Chrysemys picta bellii Testudines YYYS XP_006110924.1 PREDICTED: PG G/H synthase 2 Pelodiscus sinensis Testudines YYYS XP_003223472.1 PREDICTED: PG G/H synthase 2 Anolis carolinensis Lepidosauria YYYS ETE58294.1 PG G/H synthase 2, partial Ophiophagus hannah Lepidosauria YYYS XP_007435833.1 PREDICTED: PG G/H synthase 2 Python bivittatus Lepidosauria YYYS XP_006009311.1 PREDICTED: PG G/H synthase 2 isoform X1 Latimeria chalumnae Lepidosauria YYYS NP_001020675.1 PG G/H synthase 2 precursor D. rerio Actinopteri YYYS XP_003445100.2 PREDICTED: PG G/H synthase 2-like Oreochromis niloticus Actinopteri YYYS AAS21313.2 COX 2 Fundulus heteroclitus Actinopteri YYYS NP_001118139.1 PG G/H synthase 2 precursor Oncorhynchus mykiss Actinopteri YYYS NP_001165867.1 PG G/H synthase 2 precursor O. latipes Actinopteri YYYS NP_001025697.1 PG G/H synthase 2 precursor Xenopus (Silurana) tropicalis Amphibia YYYS NP_001086946.1 PG G/H synthase 2 precursor X. laevis Amphibia YYYS YP_004935976.1 PG G/H synthase 2 Cercopithecine herpesvirus 5 Herpesvirales YYYS AFL03426.1 Rh10 Macacine herpesvirus 3 Herpesvirales YYYS XP_007893749.1 PREDICTED: PG G/H synthase 2 Callorhinchus milii Chondrichthyes YYYS AAL37727.1 AF420317_1 COX Squalus acanthias Chondrichthyes YYYS ACH73268.1 COX-2 Myxine glutinosa Myxinformes YYYS ACP28169.2 COX Crassostrea gigas Bivalvia YYYS XP_002586987.1 hypothetical protein BRAFLDRAFT_129952 Branchiostoma ﬂoridae Branchiostomidae YYYS AHW81484.1 COX -c Branchiostoma belcheri tsingtauense Branchiostomidae YYYS XP_009047579.1 hypothetical protein LOTGIDRAFT_139178, partial Lottia gigantea Gastropoda YYYS XP_005090976.1 PREDICTED: PG G/H synthase 1-like Aplysia californica Gastropoda YYYS AAF93168.1 COX Gersemia fruticosa Anthozoa YYYS AAF93169.1 COX Plexaura homomalla Anthozoa YYYS AHA44500.1 COX Penaeus monodon Malacostraca YYYS ADB65785.1 COX Caprella sp. KV-2010a Malacostraca YYYS ADB65786.1 COX Gammarus sp. KV-2010a Malacostraca YYYS XP_002123273.1 PREDICTED: PG G/H synthase 2-like Ciona intestinalis Ascidiacea Y F YS EFX85708.1 putative COX Daphnia pulex Branchiopoda Y Y Y A EHJ75729.1 putative oxidase/peroxidase Danaus plexippus Insecta W EHF XP_006558125.1 PREDICTED: myeloperoxidase-like isoform X2 Apis mellifera Insecta FWHF XP_003487986.1 PREDICTED: hypothetical protein LOC100740410 Bombus impatiens Insecta FWHF XP_004933948.2 PREDICTED: peroxidase-like Bombyx mori Insecta W EHF XP_005179733.1 PREDICTED: uncharacterized protein LOC101892480 Musca domestica Insecta FWHF EAT44728.2 AAEL003933-PA Aedes aegypti Insecta FWHF XP_003741036.1 PREDICTED: uncharacterized protein LOC100899356 Metaseiulus occidentalis Arachnida L P W F XP_002400188.1 peroxidase, putative Ixodes scapularis Arachnida F P W F AAO33164.1 major ampullate gland peroxidase Nephila senegalensis Arachnida Y S F F CBI34960.3 unnamed protein product Vitis vinifera Rosids W TYA CDX93419.1 BnaA06g06290D Brassica napus Rosids W TYA XP_008360922.1 PREDICTED: alpha-dioxygenase 2-like Malus domestica Rosids W TYA XP_003553937.1 PREDICTED: alpha-dioxygenase 1-like Glycine max Rosids W TYA XP_007210768.1 hypothetical protein PRUPE_ppa020149mg Prunus persica Rosids W TYA BAJ90503.1 predicted protein Hordeum vulgare subsp. vulgare Liliopsida W TYA CDM84254.1 unnamed protein product Triticum aestivum Liliopsida W TYA EAY82977.1 hypothetical protein OsI_38200 Oryza sativa Indica Group Liliopsida W TYA AFW75180.1 hypothetical protein ZEAMMB73_220081 Zea mays Liliopsida W TYA aClass represents the taxonomic group. Table abbreviations: Y ¼ Tyrosine; W ¼ Tryptophan; F ¼ Phenylalanine; L ¼ Leucine; E ¼ Glutamic Acid; P ¼ Proline; S ¼ Serine; T ¼ Threonine; H ¼ Histidine; A ¼ Alanine; Black shading indicates an amino acid residue that shares a similar side chain with the selected residue from the human template sequence. Other shading indicates amino acid residue that does not match the selected residue from the human template sequence. 356 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

dose-dependent decreases in PGs upon exposure to mammalian In addition to impacting reproductive outcomes via inhibit- COX inhibitors (eg, Bhandari and Venables, 2011; Grosser et al., ing ovulation, the reduction of endogenous PGs could lead to 2002; Lister and Van Der Kraak, 2008; Morthorst et al., 2013; Patel reduced performance of spawning behavior in female fish (AOP 2014), empirical evidence that reduced PG synthesis is driven by 100; Figure 3B). The importance of the ovulatory PGF2a surge for COX inhibition in fish is very weak (including this study). Some performance of female-specific spawning behaviors (and thus of the best evidence for the COX inhibition by mammalian COX- synchronization of female sexual activity with egg presence) is inhibitors in fish comes from studies on embryonic zebrafish thoroughly and consistently documented in externally where COX activity was measureable throughout gastrulation fertilizing teleosts (eg, Juntti et al., 2016; Kobayashi et al., 2002; and segmentation, and was inhibited by both COX-1 (SC-560) Sorensen and Goetz, 1993). Essentiality of PGF2a for female and COX-2 specific (DuP-697) inhibitors (Galus, 2014). The lack spawning behavior has been demonstrated in goldfish where of empirical data that support COX inhibition in fish (by mam- administration of IN blocked performance of female spawning malian COX inhibitors) is unexpected given the strong and con- behavior, which could be rescued by injection of PGF2a (Stacey, sistent evidence for reduction of PGs in fish, as well as a high 1976). Furthermore, the studies of cichlid fish with mutant puta- level of similarity of functional domains in teleost and mamma- tive receptor for PGF2a (ptgfr) had shown that PGF2a signaling is lian COX (as indicated by the SeqAPASS analyses). Due to weak necessary and sufficient for initiation of the final stages of evidence for COX inhibition, additional time- and dose- female reproductive behavior (Juntti et al., 2016). intensive studies (including in vitro studies) with a variety of It is also biologically plausible that reduced synthesis and mammalian COX-inhibitors and variety of tissues are war- release of PGF2a (and its metabolites) by females could lead to ranted. Furthermore, optimization of COX-activity assays that reduced olfactory stimulation of males (AOP 101, Figure 3C), and can be used for evaluation of basal COX-activity and/or assays thus impact female’s ability to attract/secure spawning mates, that incorporate induction of inflammatory response may be and to induce appropriate behavioral and physiological effects needed to settle whether mammalian COX inhibitors can in males (ie, both, increases in socio-sexual interaction, and cir- impact fish COX, and subsequent reduction in PGs (that has culating LH and milt volume are normally induced by pheromo- been widely and consistently documented in fish). nal PGs) that ensure interspecific synchronization of spawning Transcriptomic data for IN-exposed females support the readiness (reviewed in Stacey and Sorensen, 2009). Essentiality proposition that mammalian COX inhibitors can impact PG of PG mediated pheromonal stimulation for reproduction in pathways in fish. Upregulation of genes involved in PG synthe- fish, and its disruption by IN has been demonstrated exten- sis (eg, ptges2/2a and PG synthase E) observed in this study, sively (reviewed in Stacey and Sorensen, 2009). likely reflects compensatory mechanisms to the IN-induced In this study, IN induced many gene expression changes reduction of PG synthesis. For example, increased transcription indicative of interference with oocyte maturation. We propose of ptges2 and/or PG synthases has been reported as a compen- that these changes are linked to PG reductions we observed, satory response to COX inhibition by aspirin and CX in mamma- and could lead to: (1) reduced ability of oocyte to overcome lian models (Johnson et al., 1997; Reese et al., 2001). Such cAMP mediated meiotic arrest (AOP 102), and (2) interference compensatory effects combined with the ability of NSAIDs to with spindle assembly checkpoint (AOP 103). In this study IN- retard degradation of COX protein (Kang et al., 2007) could have exposed fish exhibited upregulation of spindle assembly check- also interfered with the ability to detect effects on COX activity point proteins mad2 (15-fold) and bub3. Upregulation of mad2 in fish. mRNA in mammals has been shown to lead to cell cycle arrest It is biologically plausible that a reduction in PGs could lead in meiosis I (Wassmann et al., 2003) and chromosome missegre- to reduced stimulation of PG receptors. Although specific and gation events in meiosis I (which can result in aneuploid meta- distinct roles of PGs and their receptors in mammalian repro- phase II oocytes; Niault et al., 2007). Further evidence for duction are well elucidated (Funk, 2001; Richards et al., 2002; interference with meiosis progression (in IN-exposed fish) was Sugimoto et al., 2015), comparable fish data are lacking. indicated by upregulated apc/c and dysregulation of rsk1 (Sun Collectively, fish studies indicate that there are species-specific and Kim, 2012; Tunquist and Maller, 2003). These findings of cell differences in PGs and PG receptor types that mediate ovulation cycle disruption are in concordance with high throughput in fish. Both, PGFs (eg, in trout; Cetta and Goetz, 1982) and PGEs mammalian data for IN (eg, effects on cell cycle), and mamma- (eg, in medaka; Fujimori et al., 2011) and their respective recep- lian studies which documented effects of NSAIDs on cancer cell tors have been shown to be essential for ovulation in teleosts. cycle via alteration of proteins involved in kinetochore/centro- In medaka, essentiality of PGE2 and its receptor (E type receptor mere assembly (Bieniek et al., 2014) and cell cycle arrest 4; EP4) for ovulation has been demonstrated (Hagiwara et al., (Narayanan et al., 2003). Nevertheless, it remains to be deter- 2014). In vitro exposure of large-sized ovarian follicles to IN mined whether these are COX-inhibition/PG related effects. reduced PGE2 and ovulation, co-administration of PGE2 with IN Unlike their ovulatory/behavioral function, the role of PGs in rescued ovulation, and exposure to EP4 receptor antagonists fish oocyte maturation (AOPs 102 and 103), while previously diminished ovulation (Fujimori et al., 2011). Overall, there is a suggested (eg, Lister and Van Der Kraak, 2008; Sorbera et al., strong body of evidence that PGs and their receptors are indis- 2001), is not well elucidated. Recent in vitro studies of role of PGs pensable for teleost follicle rupture (Takahashi et al., 2013) and in amphibians demonstrated that PGF2a can induce GVBD and that several pharmaceuticals designed to inhibit human COX meiosis resumption up to metaphase II, and that IN administra- enzymes (eg, IN, NS-398, SC-560, diclofenac sodium, mefenamic tion can lead to dose-dependent decrease in arachidonic acid acid) can block/decrease ovulation in fish likely via decreased induced GVBD (Ortiz et al., 2015). Inhibition of in vitro spontane- PG production (Crespo et al., 2015; Joy and Singh, 2013; Lister ous oocyte maturation of full-grown follicles by IN has also and Van Der Kraak, 2008; Patino et al., 2003; Yokota et al., 2015) been observed in fish (Lister and Van Der Kraak, 2008). These and reduced stimulation of PG receptors (Fujimori et al., 2011). in vitro observations, indicate that effects of IN on the genes Dose-dependence and temporal concordance of these KEs have associated with oocyte maturation observed in this study, may been extensively characterized in mammalian models, but rep- not have to be controlled centrally (via HPG axis) as proposed in resent a data gap in fish studies. AOPs 102 and 103. MARTINOVIC´ -WEIGELT ET AL. | 357

Because of evidence for a role of PGs in amphibian and interaction of NF-jB with the aromatase promoter II (Fan et al., mammalian oocyte maturation, and empirical transcriptomics 2005) have also been noted. We observed both downregulation of fish data indicating impacts of IN on oocyte maturation (present aromatase expression (IB fish) and increased expression of ptges2 study) we developed 2 putative AOPs linking COX-inhibition to (IN fish). Overall, transcriptomics data from CX or IB-exposed fish, impairment of oocyte maturation (102 and 103; Figs. 3D and E). supported the proposition that COX-inhibition independent PGs have been shown to stimulate hypothalamic GnRH release mechanisms (PPAR and NFjB) may have been involved. (Ojeda et al., 1975) and GnRH neuron activity (Clasadonte et al., Transcriptomics data also indicated significant impacts on andro- 2011b), and to modulate autoregulation of GnRHR and differen- gen signaling pathways in IN-exposed fish; anti-androgenic tial release of LH and FSH in mammalian models (Naidich et al., effects of NSAIDs via suppression of nuclear AR expression and 2010; Naor et al., 2007). Although these mechanisms have not signaling are well documented in mammalian cancer models (eg, been elucidated in fish, Pati and Habibi (2002) suggested Kashiwagi et al.,2013; Pan et al.,2003). involvement of arachidonic metabolism in fish GnRH signaling. Overall, evidence that supports KE relationships (especially those relating to interplay between GnRH and PGs) outlined in CONCLUSION oocyte-maturation AOPs is extremely limited in fish, but given We have outlined 5 putative AOPs linking COX-inhibition to the extensive effects of IN on expression of genes associated reproduction-related AOs. Although literature and analyses with GnRH signaling and oocyte meiosis, and clearly defined conducted in this study indicate high level of sequence/struc- role of PGs in oocyte maturation in mammals and amphibians, tural/functional conservation of COX across vertebrates, direct these AOPs deserve further attention. empirical evidence for COX inhibition in fish remains weak, and majority of fish studies have not demonstrated the well- Evaluation of Alternative MIE and/or Pleiotropic Potential of established phenomena of COX inhibition by NSAIDs observed NSAIDs in Fish in mammalian models. Further, time- and dose-intensive stud- Another important part of AOP development involves critical ies, with adequately optimized assays, are warranted to conclu- evaluation of potential alternative mechanisms that may sively determine whether mammalian COX inhibitors can explain observed toxicological effects (OECD, 2013). Such alter- induce COX-inhibition in fish and whether this is the mecha- natives may either suggest other types of pathway-based assays nism by which widely documented PG reduction is initiated. needed to adequately detect impacts of this class of compounds The assessment of confidence in the biological plausibility of and/or may identify other branches in the overall AOP network AOPs 63, 100, and 101 indicated strong support. The evidence that must be considered to accurately predict related toxicologi- for essentiality of the KEs portrayed in these 3 pathways is cal outcomes. The majority of the transcriptomic, pathway- ranked as moderate to high—there is a particularly large num- level effects for all 3 chemicals (IN, CX, and IB) were associated ber of studies that established that the downstream KEs por- with processes like muscle contraction, renin-angiotensin sys- trayed in these pathways (eg, ovulation, spawning behavior) do tem, reproduction, inflammation and neurodegenerative dis- not occur if PG synthesis is blocked/attenuated. In terms of bio- eases, which are normally modulated by PGs (Funk, 2001), and/ logical domain of applicability of these 3 AOPs, we have found or constitute known therapeutic uses of COX-inhibitors strong evidence that AOP 63 is applicable to female vertebrates, (Cudaback et al., 2014; Simmons et al., 2004). This observation with biological plausibility for humans, rats, mice and select may seem surprising, as we did not always detect effects of CX amphibians (Xenopus laevis and Rhinella arenarum) and teleost and IB on COX inhibition and/or PGs—the main therapeutic tar- fish (eg, Oryzas latipes, Danio rerio, Salmo sp.) being particularly gets for NSAIDs. There are several potential explanations. First, strong (Crespo et al., 2015; Duffy, 2015; Hester et al., 2010; Lister it is possible that observed transcriptomic effects in IB and CX and Van Der Kraak, 2008; Pall et al., 2001; Yokota et al., 2015). fish were driven by COX inhibition/PG decreases, but that our AOPs 101 and 102 are applicable to female Teleosts, cyprinids, experimental design, in terms of duration and dose, was not and cichlids in particular, where essentiality of PGs for perform- optimal for detection of COX inhibition/PGs. Temporally inten- ance of reproductive behavior, and pheromonal signaling of sive studies with endocrine-active compounds have shown reproductive readiness has been extensively documented (eg, how factors like compensatory response can strongly influence Juntti et al., 2016). There is a strong evidence for essentiality and dose–response time course behaviors, particularly for early KEs conservation of function of GnRH LH MIS oocyte matura- at the molecular and biochemical level that are subject to highly tion cascade captured in AOPs 102 and 103 across female verte- dynamic control (Ankley and Villeneuve, 2015). brates (Bobe et al., 2008; Jalabert et al., 1991). Conversely, Alternatively, the effects of IB and CX on PG-related processes evidence for biological plausibility of KER “reduced hypothala- could have been a result of COX-inhibition independent mecha- mic PGE2 leading to reduced hypothalamic GnRH” outlined in nisms. Such mechanisms have been described for NSAIDs in the oocyte maturation-related AOPs (102–103) is lacking for mammalian systems and the best understood ones include: (1) most vertebrates with the exception of mammals (Clasadonte Inhibition of enzymes involved in arachidonic acid metabolism et al., 2011b; Naor et al., 2007, Naidich et al., 2010). The role of PGs (ie, lipooxygenases; Sagi et al.,2003), (2) Activation of PPARa/Ç in modulation of hypothalamic GnRH, and its involvement in (Lehmann et al.,1997; Romeiro et al.,2008), and (3) Inhibition of regulation of female reproduction and behavior warrants fur- nuclear factor jBcomplex(NF-jB; Kopp and Ghosh, 1994). It is ther study. It is important to note that there is strong in vitro evi- well established that, aside from COX inhibition, anti- dence that PGs play a role in oocyte maturation and activation inflammatory and reproductive effects of NSAIDs may be medi- and pronuclei formation in female mammals and amphibians ated by the activation of PPARs (Jaradat et al.,2001; Komar 2005; indicating that there may be another pathway, separate from Lehmann et al.,1997)orNF-jB(Kopp and Ghosh,1994; Sagi et al., the HPG-mediated one proposed in AOPs 102 and 103, that could 2003). For example, NSAIDs can modulate expression of ptgs2 via explain extensive effects of IN on oocyte meiosis observed in PPARy (Pang et al.,2003)orNF-jB(Tanabe and Tohnai, 2002). this study (Ortiz et al., 2015). There are no studies that system- Decreases in the expression and activity of aromatase via activa- atically and/or concurrently evaluated effects of mammalian tion of PPARa (Toda et al.,2003)and/orbydisruptingthe COX inhibitors on KE portrayed in these pathways (102, 103) in 358 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

mammals or fish. Overall, evidence for temporal, dose– Berninger, J. P., Du, B., Connors, K. A., Eytcheson, S. A., response, and incidence concordance for oocyte maturation Kolkmeier, M. A., Prosser, K. N., Valenti, T. W., Chambliss, C. AOPs (102 and 103) is lacking in both fish and mammals. Our K., and Brooks, B. W. (2011). Effects of the antihistamine di- data also allude to the possibility of alternative/pleiotropic phenhydramine on selected aquatic organisms. Environ. mechanisms through which NSAIDs may elicit biological Toxicol. Chem. 30, 2065–2072. responses, which may ultimately prove adverse. Specifically, Bhandari, K., and Venables, B. (2011). Ibuprofen bioconcentration our data suggest that potential interactions with PPAR, AR, and and prostaglandin E2 levels in the bluntnose minnow steroidogenic enzymes (eg, aromatase) might also need to be Pimephales notatus. Comp. Biochem. Physiol. C 153, 251–257. considered when trying to predict the toxicity of these com- Bieniek, J., Childress, C., Swatski, M. D., and Yang, W. (2014). pounds. Relative sensitivity of these molecular targets to COX-2 inhibitors arrest prostate cancer cell cycle progression NSAIDs (vs COX inhibition) in fish and other non-human mod- by down-regulation of kinetochore/centromere proteins. els, and their potential to lead to apical AOs relevant to ecologi- Prostate 74, 999–1011. cal risk assessment should be elucidated further. Bobe, J., Jalabert, B., and Fostier, A. (2008). Oogenesis: Post- vitellogenic Events Leading to a Fertilizable Oocyte. Fish Reproduction, pp. 1–36. Science Publishers, Enfield. SUPPLEMENTARY DATA Brazma, A., Hingamp, P., Quackenbush, J., Sherlock, G., Spellman, P., Stoeckert, C., Aach, J., Ansorge, W., Ball, C. A., Supplementary data are available at Toxicological Sciences and Causton, H. C. (2001). Minimum information about a mi- online. croarray experiment (MIAME)—toward standards for microarray data. Nat. Genet. 29, 365–371. Brozinski, J., Lahti, M., Meierjohann, A., Oikari, A., and Kronberg, ACKNOWLEDGEMENTS L. (2012). The anti-inflammatory drugs diclofenac, naproxen We thank Dr Kellie Fay and 3 anonymous reviewers for a and ibuprofen are found in the bile of wild fish caught down- thoughtfulpeerreview.Anyuseoftrade,product,orfirmnames stream of a wastewater treatment plant. Environ. Sci. Technol. is for descriptive purposes only and does not imply endorse- 47, 342–348. ment by the U.S. Government. The contents neither constitute, Cetta, F., and Goetz, F. W. (1982). Ovarian and plasma prostaglan- nor necessarily reflect, U.S. Environmental Protection Agency din E and F levels in brook trout (Salvelinus fontinalis) during policy. pituitary-induced ovulation. Biol. Reprod. 27, 1216–1221. Clasadonte, J., Sharif, A., Baroncini, M., and Prevot, V. (2011a). Gliotransmission by prostaglandin E2: a prerequisite for FUNDING GnRH neuronal function?. Front. Endocrinol. 2, 91. Clasadonte, J., Poulain, P., Hanchate, N. K., Corfas, G., Ojeda, S. R., This work was supported by the University of St. Thomas; and Prevot, V. (2011b). Prostaglandin E2 release from astro- U.S. Environmental Protection Agency (U.S. EPA; Office of cytes triggers gonadotropin-releasing hormone (GnRH) neu- Research and Development’s Chemical Safety for ron firing via EP2 receptor activation. Proc. Natl. Acad. Sci. U. S. Sustainability Research Program) and Oak Ridge Institute for A. 108, 16104–16109. Science and Education research fellowship at U.S. EPA to Corcoran, J., Winter, M. J., and Tyler, C. R. (2010). Martinovic-Weigelt. Pharmaceuticals in the aquatic environment: a critical re- view of the evidence for health effects in fish. Crit. Rev. Toxicol. 40, 287–304. REFERENCES Collette, T. W., Teng, Q., Jensen, K. M., Kahl, M. D., Makynen, E. Ankley, G. T., Jensen, K. M., Kahl, M. D., Korte, J. J., and Makynen, A., Durhan, E. J., Villeneuve, D. L., Martinovic-Weigelt, D., E. A. (2001). Description and evaluation of a short-term repro- Ankley, G. T., and., and Ekman, D. R. (2010). Impacts of an duction test with the fathead minnow (Pimephales prome- anti-androgen and an androgen/anti-androgen mixture on las). Environ. Toxicol. Chem. 20, 1276–1290. the metabolite profile of male fathead minnow urine. Ankley, G. T., Bennett, R. S., Erickson, R. J., Hoff, D. J., Hornung, M. Environ. Sci. Technol. 44, 6881–6886. W., Johnson, R. D., Mount, D. R., Nichols, J. W., Russom, C. L., Crespo, D., Goetz, F. W., and Planas, J. V. (2015). Luteinizing hor- Schmieder, P. K., et al. (2010). Adverse outcome pathways: A mone induces ovulation via tumor necrosis factor a-depen- conceptual framework to support ecotoxicology research dent increases in prostaglandin F2a in a nonmammalian and risk assessment. Environ. Toxicol. Chem. 29, 730–741. vertebrate. Sci. Rep. 5, 1420. Ankley, G. T., and Villeneuve, D. L. (2015). Temporal changes in Cudaback,E.,Jorstad,N.L.,Yang,Y.,Montine,T.J.,andKeene,C. biological responses and uncertainty in assessing risks of D. (2014). Therapeutic implications of the prostaglandin path- endocrine-disrupting chemicals: insights from intensive way in Alzheimer’s disease. Biochem. Pharmacol. 88, 565–572. time-course studies with fish. Toxicol. Sci. 144, 259–275. Davis, A. P., Grondin, C. J., Lennon-Hopkins, K., Saraceni- Becker, R. A., Ankley, G. T., Edwards, S. W., Kennedy, S., Linkov, Richards, C., Sciaky, D., King, B. L., Wiegers, T. C., and I., Meek, B., Sachana, M., Segner, H., Van Der Burg, B., Mattingly, C. J. (2015). The comparative toxicogenomics data- Villeneuve, D. L., et al. (2015). Increasing scientific confidence base’s 10th year anniversary: update 2015. Nucleic Acids Res. in adverse outcome pathways: Application of tailored 43, D-914–D-920. bradford-hill considerations for evaluating weight of evi- Davis, J. M., Collette, T. W., Villeneuve, D. L., Cavallin, J. E., Teng, dence. Regul. Toxicol. Pharmacol. 72, 514–537. Q., Jensen, K. M., Kahl, M. D., Mayasich, J. M., Ankley, G. T., Berninger, J. P., and Brooks, B. W. (2010). Leveraging mammalian and Ekman, D. R. (2013). Field-based approach for assessing pharmaceutical toxicology and pharmacology data to predict the impact of treated pulp and paper mill effluent on endoge- chronic fish responses to pharmaceuticals. Toxicol. Lett. 193, nous metabolites of fathead minnows (Pimephales prome- 69–78. las). Environ. Sci. Technol. 47, 10628–10636. MARTINOVIC´ -WEIGELT ET AL. | 359

Duffy, D. M. (2015). Novel contraceptive targets to inhibit ovula- Huang, D. W., Sherman, B. T., Zheng, X., Yang, J., Imamichi, T., tion: the prostaglandin E2 pathway. Hum. Reprod. Update. 21, Stephens, R., and Lempicki, R. A. (2009). Extracting biological 652–670. meaning from large gene lists with DAVID. Curr. Protoc. Ekman, D. R., Villeneuve, D. L., Teng, Q., Ralston-Hooper, K. J., Bioinformatics 13.11. 1–13.11. 13. Martinovic-Weigelt, D., Kahl, M. D., Jensen, K. M., Durhan, E. Huggett, D., Cook, J., Ericson, J., and Williams, R. (2003). A theo- J., Makynen, E. A., Ankley, G. T., et al. (2011). “Use of gene ex- retical model for utilizing mammalian pharmacology and pression, biochemical, and metabolite profiles to enhance safety data to prioritize potential impacts of human pharma- exposure and effects assessment of the model androgen 17 ceuticals to fish. Hum. Ecol. Risk Assess. 9, 1789–1799. beta-trenbolone in fish.” Environ. Toxicol. Chem. 30, 319–329. Ishikawa, T. O., Griffin, K. J., Banerjee, U., and Herschman, H. R. Ekman, D., Hartig, P., Cardon, M., Skelton, D., Teng, Q., Durhan, (2007). The zebrafish genome contains two inducible, func- E., Jensen, K., Kahl, M., Villeneuve, D., and Gray L. Jr. (2012). tional cyclooxygenase-2 genes. Biochem. Bioph. Res. Co 352, Metabolite profiling and a transcriptional activation assay 181–187. provide direct evidence of androgen receptor antagonism by Iyer, S. N., Yamada, K., Diz, D. I., Ferrario, C. M., and Chappell, M. bisphenol A in fish. Environ. Sci. Technol. 46, 9673–9680. C. (2000). Evidence that prostaglandins mediate the antihy- Eriksson, L., Kettaneh-Wold, N., Trygg, J., Wikstro¨m, C., and Wold, pertensive actions of angiotensin-(1-7) during chronic block- S. (2006). Multi-and Megavariate Data Analysis: Part I: Basic ade of the renin-angiotensin system. J. Cardiovasc. Pharmacol. Principles and Applications. 2nd ed. Umetrics, Umea˚, Sweden. 36, 109–117. Fan, W., Yanase, T., Morinaga, H., Mu, Y., Nomura, M., Okabe, T., Jalabert, B. E., Fostier, A. L., Breton, B. E., and Weil, C. L. (1991). Goto, K., Harada, N., and Nawata, H. (2005). Activation of per- Oocyte maturation in vertebrates. Vertebr. Endocrinol. 4, oxisome proliferator-activated receptor-c and retinoid X re- 23–90. ceptor inhibits aromatase transcription via nuclear factor- Jaradat, M. S., Wongsud, B., Phornchirasilp, S., Rangwala, S. M., jB. Endocrinology 146, 85–92. Shams, G., Sutton, M., Romstedt, K. J., Noonan, D. J., and Flippin, J. L., Huggett, D., and Foran, C. M. (2007). Changes in the Feller, D. R. (2001). Activation of peroxisome proliferator- timing of reproduction following chronic exposure to ibupro- activated receptor isoforms and inhibition of prostaglandin fen in Japanese medaka, Oryzias latipes. Aquat. Toxicol. 81, H 2 synthases by ibuprofen, naproxen, and indomethacin. 73–78. Biochem. Pharmacol. 62, 1587–1595. Fujimori, C., Ogiwara, K., Hagiwara, A., Rajapakse, S., Kimura, A., Jensen, K. M., Korte, J. J., Kahl, M. D., Pasha, M. S., and Ankley, G. and Takahashi, T. (2011). Expression of cyclooxygenase-2 T. (2001). Aspects of basic reproductive biology and endocri- and prostaglandin receptor EP4b mRNA in the ovary of the nology in the fathead minnow (Pimephales promelas). Comp. medaka fish, Oryzias latipes: Possible involvement in ovula- Biochem. Physiol. C 128, 127–141. tion. Mol. Cell. Endocrinol. 332, 67–77. Johnson, R. D., Polakoski, K., Everson, W. V., and Nelson, D. M. Funk, C. D. (2001). Prostaglandins and leukotrienes: advances in (1997). Aspirin induces increased expression of both prosta- eicosanoid biology. Science. 294, 1871–1875. glandin H synthase-1 and prostaglandin H synthase-2 in cul- Galus, M. (2014). Chronic Exposure to Environmentally Relevant tured human placental trophoblast. Obstet. Gynecol. 177, 78–85. Pharmaceutical Concentrations Effects Reproductive and Judson, R. S., Martin, M. T., Egeghy, P., Gangwal, S., Reif, D. M., Developmental Physiology in Zebrafsih (Danio rerio). Kothiya, P., Wolf, M., Cathey, T., Transue, T., and Smith, D. Doctoral Dissertation, McMaster University Hamilton. (2012). Aggregating data for computational toxicology appli- Available at: http://hdl.handle.net/11375/16282 cations: The US Environmental Protection Agency (EPA) ag- Gomez-Abell an, V., and Sepulcre, M. P. (2016). “The role of pros- gregated computational toxicology resource (ACToR) system. taglandins in the regulation of fish immunity.” Mol. Immunol. Int. J. Mol. Sci. 13, 1805–1831. 69, 139–145. Joy, K. P., and Singh, V. (2013). Functional interactions between Grosser, T., Yusuff, S., Cheskis, E., Pack, M. A., and FitzGerald, G. vasotocin and prostaglandins during final oocyte maturation A. (2002). Developmental expression of functional cyclooxy- and ovulation in the catfish Heteropneustes fossilis. Gen. genases in zebrafish. Proc. Natl. Acad. Sci. U. S. A. 99, Comp. Endocrinol. 186, 126–135. 8418–8423. Juntti, S. A., Hilliard, A. T., Kent, K. R., Kumar, A., Nguyen, A., Hagiwara, A., Ogiwara, K., Katsu, Y., and Takahashi, T. (2014). Jimenez, M. A., Loveland, J. L., Mourrain, P., and Fernald, R. D. Luteinizing hormone-induced expression of Ptger4b, a pros- (2016). A neural basis for control of cichlid female reproduc- taglandin E2 receptor indispensable for ovulation of the me- tive behavior by prostaglandin F2a. Curr. Biol. 26, 943–949. daka Oryzias latipes, is regulated by a genomic mechanism Kang, Y., Mbonye, U. R., DeLong, C. J., Wada, M., and Smith, W. L. involving nuclear progestin receptor. Biol. Reprod. 90, 126. (2007). Regulation of intracellular cyclooxygenase levels by Hammes, S. R. (2004). Steroids and oocyte maturation—a new gene transcription and protein degradation. Prog. Lipid Res. look at an old story. Mol. Endocrinol. 18, 769–775. 46, 108–125. Havird, J. C., Miyamoto, M. M., Choe, K. P., and Evans, D. H. (2008). Kashiwagi, E., Shiota, M., Yokomizo, A., Itsumi, M., Inokuchi, J., Gene duplications and losses within the cyclooxygenase Uchiumi, T., and Naito, S. (2013). Prostaglandin receptor EP3 family of teleosts and other chordates. Mol. Biol. Evol. 25, mediates growth inhibitory effect of aspirin through andro- 2349–2359. gen receptor and contributes to castration resistance in pros- Hester, K. E., Harper, M. J., and Duffy, D. M. (2010). Oral adminis- tate cancer cells. Endocr. Relat. Cancer 20, 431–441. tration of the cyclooxygenase-2 (COX-2) inhibitor meloxicam Kavlock, R., Chandler, K., Houck, K., Hunter, S., Judson, R., blocks ovulation in non-human primates when adminis- Kleinstreuer, N., Knudsen, T., Martin, M., Padilla, S., and Reif, tered to simulate emergency contraception. Hum. Reprod. 25, D. (2012). Update on EPA’s ToxCast program: Providing high 360–367. throughput decision support tools for chemical risk manage- Hosack, D. A., Dennis, G., Sherman, B. T., Lane, H. C., and ment. Chem. Res. Toxicol. 25, 1287–1302. Lempicki, R. A. (2003). Identifying biological themes within Knight, O. M., and Van Der Kraak, G. (2015). The role of eicosa- lists of genes with EASE. Genome Biol. 4,1. noids in 17a,20b-dihydroxy-4-pregnen-3-one-induced 360 | TOXICOLOGICAL SCIENCES, 2017, Vol. 156, No. 2

ovulation and spawning in Danio rerio. Gen. Comp. Endocrinol. Niault, T., Hached, K., Sotillo, R., Sorger, P. K., Maro, B., Benezra, 213, 50–58. R., and Wassmann, K. (2007). Changing Mad2 levels affects Kobayashi, M., Sorensen, P. W., and Stacey, N. E. (2002). chromosome segregation and spindle assembly checkpoint Hormonal and pheromonal control of spawning behavior in control in female mouse meiosis I. PLoS One 2, e1165. the goldfish. Fish Physiol. Biochem. 26, 71–84. OECD. 2013. Guidance Document on Developing and Assessing Komar, C. M. (2005). Peroxisome proliferator-activated receptors Adverse Outcome Pathways. Series on Testing and (PPARs) and ovarian function–implications for regulating ste- Assessment No. 184. Available at: http://www.oecd.org/offi roidogenesis, differentiation, and tissue remodeling. Reprod. cialdocuments/publicdisplaydocumentpdf/?cote¼env/jm/ Biol. Endocrinol. 3, 41. mono(2013)6&doclanguage¼en Kone?, M., Cologgi, D. L., Lu, W., Smith, D. W., and Ulrich, A. C. OECD. 2014. Users’ Handbook Supplement to the Guidance (2013). Pharmaceuticals in Canadian sewage treatment plant Document for Developing and Assessing AOPs. Available at: effluents and surface waters: occurrence and environmental https://aopkb.org/common/AOP_Handbook.pdf risk assessment. Environ. Technol. Rev. 2, 17–27. Ojeda, S., Harms, P., and McCann, S. (1975). Effect of inhibitors of Kopp, E., and Ghosh, S. (1994). Inhibition of NF-kappa B by so- prostaglandin synthesis on gonadotropin release in the Rat dium salicylate and aspirin. Science 265, 956–959. 1. Endocrinology 97, 843–854. Lado, W. E., Zhang, D., Mennigen, J. A., Zamora, J. M., Popesku, J. Ortiz, M. E., Arias-Torres, A. J., and Zelarayan, L. I. (2015). Role of T., and Trudeau, V. L. (2013). Rapid modulation of gene ex- arachidonic acid cascade in Rhinella arenarum oocyte matu- pression profiles in the telencephalon of male goldfish fol- ration. Zygote 23, 603–614. lowing exposure to waterborne sex pheromones. Gen. Comp. Pall, M., Fride?n, B. E., and Brannstro€ ¨ m, M. (2001). Induction of Endocrinol. 192, 204–213. delayed follicular rupture in the human by the selective LaLone, C. A., Villeneuve, D. L., Lyons, D., Helgen, H. W., COX-2 inhibitor rofecoxib: a randomized double-blind study. Robinson, S. L., Swintek, J. A., Saari, T. W., and Ankley, G. T. Hum. Reprod. 16, 1323–1328. (2016). Sequence Alignment to Predict Across Species Pan, Y., Zhang, J. S., Gazi, M. H., and Young, C. Y. (2003). The cy- Susceptibility (SeqAPASS): A web-based tool for addressing clooxygenase 2-specific nonsteroidal anti-inflammatory the challenges of cross-species extrapolation of chemical drugs celecoxib and nimesulide inhibit androgen receptor toxicity. Toxicol. Sci. 153, 228–245. activity via induction of c-Jun in prostate cancer cells. Cancer Lehmann, J. M., Lenhard, J. M., Oliver, B. B., Ringold, G. M., and Epidemiol. Biomarkers Prev. 12, 769–774. Kliewer, S. A. (1997). Peroxisome proliferator-activated recep- Pang, L., Nie, M., Corbett, L., and Knox, A. J. (2003). tors a and c are activated by indomethacin and other non- Cyclooxygenase-2 expression by nonsteroidal anti- steroidal anti-inflammatory drugs. J. Biol. Chem. 272, inflammatory drugs in human airway smooth muscle cells: 3406–3410. role of peroxisome proliferator-activated receptors. Lister, A., and Van Der Kraak, G. (2008). An investigation into the J. Immunol. 170, 1043–1051. role of prostaglandins in zebrafish oocyte maturation and Patel, A. (2014). The applicability of the “read-across hypothesis” ovulation. Gen. Comp. Endocrinol. 159, 46–57. for assessing the effects of human pharmaceuticals on fish. Lutz, L. B., Cole, L. M., Gupta, M. K., Kwist, K. W., Auchus, R. J., Doctoral Dissertation, Brunel University London. and Hammes, S. R. (2001). Evidence that androgens are the Pati, D., and Habibi, H. R. (2002). Involvement of protein kinase C primary steroids produced by Xenopus laevis ovaries and and arachidonic acid pathways in the gonadotropin- may signal through the classical androgen receptor to pro- releasing hormone regulation of oocyte meiosis and follicu- mote oocyte maturation. Proc. Natl. Acad. Sci. U. S. A. 98, lar steroidogenesis in the goldfish ovary. Biol. Reprod. 66, 13728–13733. 813–822. Moore, A., Olse?n, K. H., Lower, N., and Kindahl, H. (2002). The Patino, R., Yoshizaki, G., Bolamba, D., and Thomas, P. (2003). Role role of F-series prostaglandins as reproductive priming pher- of arachidonic acid and protein kinase C during maturation- omones in the brown trout. J. Fish Biol. 60, 613–624. inducing hormone-dependent meiotic resumption and ovu- Morthorst, J. E., Lister, A., Bjerregaard, P., and Van Der Kraak, G. lation in ovarian follicles of Atlantic croaker. Biol. Reprod. 68, (2013). Ibuprofen reduces zebrafish PGE 2 levels but steroid 516–523. hormone levels and reproductive parameters are not af- Rand-Weaver, M., Margiotta-Casaluci, L., Patel, A., Panter, G. H., fected. Comp. Biochem. Physiol. Part C 157, 251–257. Owen, S. F., and Sumpter, J. P. (2013). The read-across hy- Naidich, M., Shterntal, B., Furman, R., Pawson, A. J., Jabbour, H. pothesis and environmental risk assessment of pharmaceu- N., Morgan, K., Millar, R. P., Jia, J., Tomic, M., and Stojilkovic, ticals. Environ. Sci. Technol. 47, 11384–11395. S. (2010). Elucidation of mechanisms of the reciprocal cross Reese, J., Zhao, X., Ma, W. G., Brown, N., Maziasz, T. J., and Dey, S. talk between gonadotropin-releasing hormone and prosta- K. (2001). Comparative analysis of pharmacologic and/or ge- glandin receptors. Endocrinology 151, 2700–2712. netic disruption of cyclooxygenase-1 and cyclooxygenase-2 Naor, Z., Jabbour, H. N., Naidich, M., Pawson, A. J., Morgan, K., function in female reproduction in mice 1. Endocrinology 142, Battersby, S., Millar, M. R., Brown, P., and Millar, R. P. (2007). 3198–3206. Reciprocal cross talk between gonadotropin-releasing hor- Richards, J. S., Russell, D. L., Ochsner, S., and Espey, L. L. (2002). mone (GnRH) and prostaglandin receptors regulates GnRH Ovulation: new dimensions and new regulators of the receptor expression and differential gonadotropin secretion. inflammatory-like response. Annu. Rev. Physiol. 64, 69–92. Mol. Endocrinol. 21, 524–537. Romeiro, N. C., Sant’Anna, C. M., Lima, L. M., Fraga, C. A., and Narayanan, B. A., Condon, M. S., Bosland, M. C., Narayanan, N. Barreiro, E. J. (2008). NSAIDs revisited: Putative molecular ba- K., and Reddy, B. S. (2003). Suppression of N-methyl-N-nitro- sis of their interactions with peroxisome proliferator- sourea/testosterone-induced rat prostate cancer growth activated gamma receptor (PPARc). Eur. J. Med. Chem. 43, by celecoxib: Effects on cyclooxygenase-2, cell cycle regula- 1918–1925. tion, and apoptosis mechanism(s). Clin. Cancer Res. 9, Rowlinson, S. W., Kiefer, J. R., Prusakiewicz, J. J., Pawlitz, J. L., 3503–3513. Kozak, K. R., Kalgutkar, A. S., Stallings, W. C., Kurumbail, R. MARTINOVIC´ -WEIGELT ET AL. | 361

G., and Marnett, L. J. (2003). A novel mechanism of Teng, Q., Ekman, D. R., Huang, W., and Collette, T. W. (2012). cyclooxygenase-2 inhibition involving interactions with Ser- Push-through direct injection NMR: an optimized automa- 530 and Tyr-385. J. Biol. Chem. 278, 45763–45769. tion method applied to metabolomics. Analyst 137, Sagi, S. A., Weggen, S., Eriksen, J., Golde, T. E., and Koo, E. H. 2226–2232. (2003). The non-cyclooxygenase targets of non-steroidal Toda, K., Okada, T., Miyaura, C., and Saibara, T. (2003). anti-inflammatory drugs, lipoxygenases, peroxisome Fenofibrate, a ligand for PPARa, inhibits aromatase cyto- proliferator-activated receptor, inhibitor of kappa B kinase, chrome P450 expression in the ovary of mouse. J. Lipid Res. and NF kappa B, do not reduce amyloid beta 42 production. J. 44, 265–270. Biol. Chem. 278, 31825–31830. Tokumoto, T., Tokumoto, M., Oshima, T., Shimizuguchi, K., Schmidt, C. W. (2009). TOX 21: new dimensions of toxicity test- Fukuda, T., Sugita, E., Suzuki, M., Sakae, Y., Akiyama, Y., and ing. Environ. Health Perspect. 117, A348–A353. Nakayama, R. (2012). Characterization of multiple membrane Selinsky, B. S., Gupta, K., Sharkey, C. T., and Loll, P. J. (2001). progestin receptor (mPR) subtypes from the goldfish ovary Structural analysis of NSAID binding by prostaglandin H2 and their roles in the induction of oocyte maturation. Gen. synthase: time-dependent and time-independent inhibitors Comp. Endocrinol. 177, 168–176. elicit identical enzyme conformations. Biochemistry 40, Tokumoto, T., Yamaguchi, T., Ii, S., and Tokumoto, M. (2011). In 5172–5180. vivo induction of oocyte maturation and ovulation in zebra- Simmons, D. L., Botting, R. M., and Hla, T. (2004). Cyclooxygenase fish. PLoS One 6, e25206. isozymes: the biology of prostaglandin synthesis and inhibi- Tunquist, B. J., and Maller, J. L. (2003). Under arrest: cytostatic tion. Pharmacol. Rev. 56, 387–437. factor (CSF)-mediated metaphase arrest in vertebrate eggs. Sorbera, L. A., Asturiano, J. F., Carrillo, M., and Zanuy, S. (2001). Genes Dev. 17, 683–710. Effects of polyunsaturated fatty acids and prostaglandins on Viant, M. R., Pincetich, C. A., and Tjeerdema, R. S. (2006). oocyte maturation in a marine teleost, the European sea bass Metabolic effects of dinoseb, diazinon and esfenvalerate in (Dicentrarchus labrax). Biol. Reprod. 64, 382–389. eyed eggs and alevins of Chinook salmon (Oncorhynchus Sorensen, P. W., and Goetz, F. W. (1993). Pheromonal and repro- tshawytscha) determined by H-1 NMR metabolomics. Aquat. ductive function of F prostaglandins and their metabolites in Toxicol. 77, 359–371. teleost fish. J. Lipid Mediat. 6, 385–393. Villeneuve, D. L., Garcia-Reyero, N., Martinovic-Weigelt, D., Li, Z., Stacey, N. E., and Pandey, S. (1975). Effects of indomethacin and Watanabe, K. H., Orlando, E. F., LaLone, C. A., Edwards, S. W., prostaglandins on ovulation of goldfish. Prostaglandins 9, Burgoon, L. D., and Denslow, N. D. (2012). A graphical systems 597–607. model and tissue-specific functional gene sets to aid tran- Stacey, N., and Sorensen, P. (2009). Hormonal pheromones in scriptomic analysis of chemical impacts on the female tele- fish. In Hormones, Brain and Behavior (D. W. Pfaff, A. P. Arnold, ost reproductive axis. Mutat. Res., Genet. Toxicol. Environ. A. M. Etgen, S. E. Fahrbach and R. T. Rubin, Eds), 2nd ed., pp. Mutagen 746, 151–162. 639–681. Academic Press, San Diego, CA. Villeneuve, D. L., Crump, D., Garcia-Reyero, N., Hecker, M., Stacey, N. (1976). Effects of indomethacin and prostaglandins on Hutchinson, T. H., LaLone, C. A., Landesmann, B., Lettieri, T., the spawning behaviour of female goldfish. Prostaglandins 12, Munn, S., and Nepelska, M. (2014a). Adverse outcome path- 113–126. way (AOP) development I: Strategies and principles. Toxicol. Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, Sci. 142, 312–320. B. L., Gillette, M. A., Paulovich, A., Pomeroy, S. L., Golub, T. R., Villeneuve, D. L., Crump, D., Garcia-Reyero, N., Hecker, M., and Lander, E. S. (2005). Gene set enrichment analysis: a Hutchinson, T. H., LaLone, C. A., Landesmann, B., Lettieri, T., knowledge-based approach for interpreting genome-wide ex- Munn, S., Nepelska, M., et al. (2014b). Adverse outcome pression profiles. Proc. Natl. Acad. Sci. U. S. A. 102, 15545–15550. pathway development II: Best practices. Toxicol. Sci. 142, Sugimoto, Y., Inazumi, T., and Tsuchiya, S. (2015). Roles of pros- 321–330. taglandin receptors in female reproduction. J. Biochem. 157, Walker, M. C., Kurumbail, R. G., et al. (2001). “A three-step kinetic 73–80. mechanism for selective inhibition of cyclo-oxygenase-2 by Sun, S. C., and Kim, N. H. (2012). Spindle assembly checkpoint diarylheterocyclic inhibitors.”. Biochem. J. 357, 709–718. and its regulators in meiosis. Hum. Reprod. Update 18, 60–72. Wassmann, K., Niault, T., and Maro, B. (2003). Metaphase I arrest Suwa, K., and Yamashita, M. (2007). Regulatory mechanisms of upon activation of the Mad2-dependent spindle checkpoint oocyte maturation and ovulation. In The Fish Oocyte (P. J. in mouse oocytes. Curr. Biol. 13, 1596–1608. Babin, J. Cerda, and E. Lubzens Eds.), pp. 323–347. Springer, Wennmalm, e´. (2013). Prostaglandins and the heart. Hearts Dordrecht, The Netherlands. Heart-like Organs 2, 41–92. Takahashi, T., Fujimori, C., Hagiwara, A., and Ogiwara, K. (2013). Yokota, H., Eguchi, S., Hasegawa, S., Okada, K., Yamamoto, F., Recent advances in the understanding of teleost medaka Sunagawa, A., Tanaka, M., Yamamoto, R., and Nakano, E. ovulation: the roles of proteases and prostaglandins. Zool. Sci. (2015). Assessment of in vitro antiovulatory activities of non- 30, 239–247. steroidal anti-inflammatory drugs and comparison with Takahashi, T., Igarashi, H., Amita, M., Hara, S., and Kurachi, H. in vivo reproductive toxicities of medaka (Oryzias latipes). (2010). Roles of prostaglandins during oocyte maturation: Environ. Toxicol. doi: 10.1002/tox.22173. lessons from knockout mice. J. Mammal. Ova Res. 27, 11–20. Zou, M., Shi, C., and Cohen, R. A. (2002). High glucose via peroxy- Tanabe, M., and Kanehisa, M. (2012). Using the KEGG database nitrite causes tyrosine nitration and inactivation of prostacy- resource. Curr. Protoc. Bioinformatics 1.12. 1–1.12. 43. clin synthase that is associated with thromboxane/ Tanabe, T., and Tohnai, N. (2002). Cyclooxygenase isozymes and prostaglandin H2 receptor–mediated apoptosis and adhesion their gene structures and expression. Prostagland. Other Lipid molecule expression in cultured human aortic endothelial Mediat. 68, 95–114. cells. Diabetes 51, 198–203.