Corticosteroid Fludrocortisone Acetate Targets Multiple End Points in

Sep 12, 2016 - Synthetic corticosteroids may pose an environmental risk to fish. Here, we describe multiend point responses of adult zebrafish (8 mont...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/est

Corticosteroid Fludrocortisone Acetate Targets Multiple End Points in Zebrafish (Danio rerio) at Low Concentrations Yanbin Zhao,† Kun Zhang,† and Karl Fent*,†,‡ †

School of Life Sciences, University of Applied Sciences and Arts Northwestern Switzerland, Gründenstrasse 40, CH-4132 Muttenz, Switzerland ‡ Institute of Biogeochemistry and Pollution Dynamics, Department of Environmental System Sciences, Swiss Federal Institute of Technology (ETH Zürich), CH-8092 Zürich, Switzerland S Supporting Information *

ABSTRACT: Synthetic corticosteroids may pose an environmental risk to fish. Here, we describe multiend point responses of adult zebrafish (8 months old) upon 21-day exposure to a commonly prescribed corticosteroid, fludrocortisone acetate (FLU), at concentrations between 0.006 and 42 μg/L. No remarkable reproductive impacts were observed, while physiological effects, including plasma glucose level and blood leukocyte numbers were significant altered even at 42 ng/L. Ovary parameters and transcriptional analysis of hypothalamic−pituitary−gonadal−liver axis revealed negligible effects. Significant alterations of the circadian rhythm network were observed in the zebrafish brain. Transcripts of several biomarker genes, including per1a and nr1d1, displayed strong transcriptional changes, which occurred at environmental relevant concentrations of 6 and 42 ng/L FLU. Importantly, the development and behavior of F1 embryos were significant changed. Heartbeat, hatching success and swimming behavior of F1 embryos were all increased even at 6 and 42 ng/L. All effects were further confirmed by exposure of eleuthero-embryos. Significant transcriptional changes of biomarker genes involved in gluconeogenesis, immune response and circadian rhythm in eleuthero-embryos confirmed the observations in adult fish. Hatching success, heartbeat, and swimming activity were increased at 81 ng/L and higher, as with F1 embryos. These results provide novel insights into the understanding of potential environmental risks of corticosteroids.



glucose metabolism in aquatic animals, including fish. In fathead minnows, glucocorticoid exposures resulted in a decrease of fecundity and adverse reproductive effects at relatively high concentrations of 50 and 500 μg/L of dexamethasone.7 On the other hand, prednisolone and beclomethasone dipropionate affected physiological end points including plasma glucose concentrations and percentage of blood total leukocytes in fathead minnows at much lower concentrations of 0.1 and 1 μg/L.8,9 Transcriptional analysis in recent reports further revealed sensitive molecular targets, of which, some transcripts involved in gluconeogenesis or immune response, like pepck, baiap2, and tpp2, were disrupted by glucocorticoids even at environmental relevant concentrations in zebrafish (Danio rerio) and rainbow trout (Oncorhynchus mykiss).13,14 Synthetic corticosteroids are highly prescribed, including fludrocortisone acetate (FLU), which has an estimated log P of 1.67.15 FLU displays moderate glucocorticoid potency and greater mineralocorticoid potency. Therapeutically, it has been

INTRODUCTION Synthetic glucocorticoids and mineralocorticoids are pharmaceuticals prescribed in human and veterinary medicine that enter the aquatic environment by wastewater and agricultural runoff. As do natural corticoids, they occur in surface and ground waters in the low ng/L range.1−3 In animal farm runoff and influent/effluent of wastewater treatment plants, higher concentrations of up to hundreds of ng/L occur3−6 Adverse outcomes in reproduction and physiology have been reported recently in fish species, such as fathead minnows (Pimephales promelas).7−9 Besides aquatic wildlife, humans may also be exposed unintentionally to corticosteroids via contaminated seafood.10 Synthetic corticosteroids employed as anti-inflammatory drugs mimic the role of endogenous cortisol and aldosterone and exert effects primarily via binding to nuclear hormone receptors, the glucocorticoid receptor (GR) and mineralocorticoid receptor (MR). Endogenous corticosteroids are involved in a wide range of physiological processes, including stress response, immune response, glucose metabolism, regulation of inflammation, and behavior in mammals and teleosts.11,12 Therefore, the widespread occurrence of synthetic corticosteroids in aquatic ecosystems raises concerns about their potential adverse effects on immune and stress response, and © 2016 American Chemical Society

Received: Revised: Accepted: Published: 10245

July 8, 2016 August 28, 2016 August 29, 2016 September 12, 2016 DOI: 10.1021/acs.est.6b03436 Environ. Sci. Technol. 2016, 50, 10245−10254

Article

Environmental Science & Technology

flow through system; in this proportion, DMSO displayed negligible effects on the adult zebrafish and embryo development in response to different steroid hormones as described previously.18,21 Low concentrations of FLU were selected to reflect environmentally realistic doses,1,17 and high concentrations were chosen as pharmacologically relevant, based on the reproductive and physiological effects reported for dexamethasone, prednisolone, and beclomethasone dipropionate.7−9 The experiment was conducted according to the OECD Test Guideline (TG) 229/230 with slight modification. The detailed procedure was described previously.18,19 In brief, after a five-day acclimatization, the experiment started with a pre-exposure period of 14 days to establish the baseline rate of fecundity for each tank (and spawning groups), followed by 2 days of equilibration when chemical-dosing started, and finally, 21 days of FLU exposure as the OECD test guideline recommended. The whole experiment was performed by employing a flowthrough system, which ensured a complete change of the reconstituted water every 12 h. Temperature (27 ± 1 °C), pH value (6.7−7.2), dissolved oxygen concentration (>70%), nitrate (normally ≤10 mg/L), and nitrite (normally at 0 mg/ L) were continuously measured and ensured water quality. The photoperiod was 14:10 h light/dark. During the whole exposure period, mortality and any abnormalities in appearance of fish were recorded as the OECD TG recommended. No compound related effects occurred. Fish were fed twice daily with TetraMin flakes and a combination of frozen brine shrimps (A. salina), white mosquito larvae, and Daphnia magna. Eggs were collected and counted during the whole experimental period. At the last 5 days of exposure, eggs were collected at about 9 a.m. (ZT2) each day, transferred to Petri dishes with wellaerated reconstituted fish water, and examined under a stereomicroscope (Zeiss, DV4) to determine fertilization success. About 50−100 fertilized embryos were randomly selected from each tank and transferred to new Petri dishes with appropriate reconstituted fish water. Petri dishes were then placed into the fish egg incubator (Flohr Instruments, Netherlands) with constant temperature (27 °C), air humidity (50%), and photoperiod (14:10 h light/dark). Every 24 h, dead embryos were removed and water was completely changed. Contraction rate of embryos at 24 h, heartbeat at 36 h, hatching success at 24, 48, 72, 96, and 120 h, as well as swimming behavior at 120 h were measured for determination of transgenerational effects in the F1 generation. At the end of exposure, fish were anesthetized by KoiMed Sleep (1.5−3 mL/L water). Before dissection, three females and three males from each replicate (n = 9 for each gender of each treatment) were randomly selected and measured for wet weight (mg) and length (cm), which was used to calculate the condition factor. Two females and two males from each replicate were then dissected immediately. Brain (whole brain including pituitary), liver and gonads of two fish were pooled, transferred to RNAlater and stored at −80 °C for subsequent RNA extraction. Pooling was necessary due to the small tissue sizes and varying extraction efficiencies. Before pooling, ovaries of each fish were weighed in order to assess the gonadosomatic index (GSI = gonad weight (g)/body weight (g) × 100). In addition, blood samples were collected from two females and two males (all anesthetized) of each replicate by tail ablation. Plasma glucose levels and numbers of different types of white blood cells were determined as described below.

used in the treatment of cerebral salt wasting syndrome, postural tachycardia syndrome and low blood pressure.16 Despite its frequent use, information about its environmental occurrence is very limited and the ecotoxicological implications are unknown. Residues of FLU were detected up to 36 ng/L in municipal wastewater influents and at 14 ng/L in surface waters in Switzerland, and up to 15 ng/L in wastewater influents and at 8 ng/L in rivers in the Czech Republic.1,17 In hospital wastewater, up to 82 ng/L were noted.1 Of 24 targeted corticosteroids, FLU occurred at highest concentrations in surface water samples in Switzerland.17 Despite environmental relevance, little is known about the bioactivity of FLU, and there is no information on the fate in the environment, biotransformation, bioaccumulation, or the ecotoxicological implications in aquatic wildlife. In our previous studies, we focused on circadian rhythm regulation in zebrafish (Danio rerio) by environmental progestins. We observed remarkable alterations on the expression of several key circadian rhythm genes in response to progestins, such as per1a and nr1d2a.18−20 Our subsequent investigations revealed similar effects on circadian rhythm regulation by corticosteroids. Compared to dexamethasone and betamethasone with high glucocorticoid potencies, the alterations caused by FLU were the most prominent (data not published). Thus, we aimed at evaluating whether corticosteroids, including FLU, also result in alterations on the circadian rhythm network. Corticosteroids were also reported to disrupt endocrine homeostasis in fish. Beclomethasone dipropionate induced male secondary sexual characters with a decrease in plasma vitellogenin (vtg) concentrations and related gene expression in female fathead minnows at concentrations down to 0.1 μg/L.9 Dexamethasone also displayed capabilities to decrease fish plasma estradiol (17βE2) and down-regulate vtg gene expressions.7 Whether the endocrine system of teleost is a common target for corticosteroids, like FLU, is still not clear. On this basis, the aim of present study was to evaluate unknown effects of FLU in adult zebrafish by assessing multiple important end points, including reproductive, physiological (gluconeogenesis and immune response), behavioral and molecular reactions (endocrine regulation and circadian rhythm). Concomitantly to breeding pairs of adults, the development and behavior of F1 embryos were evaluated. Additionally, development of embryos, behavior and transcriptional responses of target genes were assessed in eleutheroembryos to better understand the effects of FLU. Our data show that FLU exhibits important adverse effects, particularly in developing embryos, at environmental concentrations.



MATERIALS AND METHODS Chemicals and Maintenance of Zebrafish. This information can be found in the Supporting Information (SI). Experimental Design. Adult Zebrafish Exposure. Adult zebrafish (8 months old) were selected from the 300 L culture tank and randomly placed into 10 L stainless steel tanks in wellaerated reconstituted water. The temperature was controlled by automatic water-bath heating device and was constantly at 27 ± 1 °C during the whole experiment. The experimental setup consisted of solvent control (0.01% DMSO) and increasing FLU concentrations of nominal 0.01, 0.1, 1, 10, and 100 μg/L. Each treatment consisted of three replicates, each consisting of 5 females and 5 males as breeding pairs. The 0.01% DMSO was employed as solvent control due to the practical constraints of 10246

DOI: 10.1021/acs.est.6b03436 Environ. Sci. Technol. 2016, 50, 10245−10254

Article

Environmental Science & Technology

use of a ViewPoint behavior recording system, together with the associated movement tracking and analysis software (Zebrabox, ViewPoint Life Sciences, France). For each treatment, 16 eleuthero-embryos were selected randomly (four eleuthero-embryos from each of the four replicates), and then gently placed into a 48-well cell culture plate (Corning Costar). After an initial 5 min acclimation period within the zebrabox instrument under visible light conditions, swimming behavior activity was recorded for the next 10 min. Light intensity was controlled by the ViewPoint behavior recording system to keep the uniform illumination distribution on 48-well plate and was constantly during the behavior measurements. Movement thresholds were set at 5 mm/s for “small movements” and 10 mm/s for “large movements” that are characteristic of a startle/escape behavior. RNA Isolation and qRT-PCR Analysis. RNA isolation, first strand cDNA synthesis and the relative quantitation in real time RT-PCR were performed as described previously18−20 with slight modifications. In brief, total RNA was extracted from the different tissues of adult zebrafish using TRIzol reagent (LuBioScience, Switzerland) and from a pool of 15 zebrafish eleuthero-embryos by use of RNeasy Mini Kit (Qiagen, Basel, Switzerland). The samples were then treated with RNase-free DNase (Qiagen, Basel, Switzerland) to purify the RNA from DNA contamination. RNA concentrations and qualities were analyzed using a NanoDrop 1000 spectrophotometer (Nanodrop Technologies Inc. Wilmington DE, U.S.) and then stored at −80 °C for subsequent RT-qPCR analysis. The first-strand cDNA synthesis was performed by MMLV (Promega, Switzerland) in the presence of random hexamers (Roche, Switzerland) and dNTP (Sigma−Aldrich, Switzerland). RT-PCR was conducted on BIO-RAD CFX96 RealTime PCR Detection System (BIO-RAD, Switzerland) using SYBR Green Fluorescence (Roche Diagnostics, Basel, Switzerland) as recommended by the manufacture’s guidelines. A twostep real-time PCR profile was used: enzyme activation step at 95 °C (10 min), and 40 cycles of 95 °C (30 s), 58−62 °C (60 s) depending on the target transcript, followed by a melting curve analysis post run (65−95 °C). For gene expression, the reference gene, β-actin, was selected as housekeeping gene for normalization, because it showed high gene expression stability in adult zebrafish tissues and embryos in response to different steroid hormones.20,26,27 In the present study, the stability of β-actin expression was also demonstrated; the transcripts displayed very little variations in different treatments, tissues and during early development (Figure S1). For primers design, the online software, Primer-BLAST (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) was employed. The intron/exon boundary-spanning primers were preference to minimize the genomic DNA contamination. Melting curves were analyzed to ensure that only a single product was amplified. Primer details are presented in the Supporting Information (Table S4). Threshold cycle (CT) values were recorded in the linear phase of amplification and the data were analyzed using the delta−delta CT method of relative quantification.22,28 Data Analysis and Statistics. The significance of differences between the solvent control and FLU exposed adult fish or embryos in transcript levels, egg production, embryo development parameters, condition factors, GSI, glucose levels and number of leukocytes were analyzed by one-way analysis of variance (ANOVA) followed by Tukey posthoc test (95% confidence interval). Before running the ANOVA, data was

Plasma vitellogenin was not analyzed due to the limited blood volume obtained. Gonadal histopathology was not performed due to the negligible effects on gonadal weight, GSI and HPG-L axis gene expressions. Considering that the sampling duration is a crucial factor that can result in artifacts in the transcriptional responses due to the endogenous circadian oscillations of genes,22 a team of co-workers (nine people) restricted the amount of sampling time within 2 h. The processing of fish sampling was following the order: control group, low concentrations to high concentrations. Embryo Exposure. A separate embryo exposure experiment was performed by use of the procedure as previously described for several progestins.18,23 In brief, at 2−3 h post fertilization (hpf), 100 blastula-stage embryos per replicate (four replicates for each treatment) were randomly placed in 150 mL covered glass beakers containing 100 mL of reconstituted fish water at 27 ± 1 °C. The experiment consisted of four FLU dose groups with increasing concentrations of nominal 0.1, 1, 10, and 100 μg/L and a solvent control group. A 24 h semistatic procedure was applied. Every 24 h, lethal and sublethal effects were evaluated, and dead embryos were removed. Water was completely changed every day with the new reconstituted fish water with appropriate FLU concentrations. In embryos and eleuthero-embryos, respectively, contraction rate, heartbeat, hatching success, and swimming behavior were measured as described for the F1 embryos of the adult fish exposure. At 120 hpf, 15 eleuthero-embryos were pooled and stored in RNAlater for further molecular analysis. Chemical Analysis. The analytical methods as described previously23 were employed to determine FLU concentrations in exposure waters of the adult and embryonic experiments. The analysis was performed by use of solid phase extraction (SPE) and liquid chromatography-tandem mass spectrometry (HPLC-MS-MS). Recoveries in the investigated matrix were 89.5 ± 3.7%. The limits of quantification were calculated directly from samples as the values giving a signal-to-noise ratio (S/N) of 10 and were 0.6 ng/L. Detailed information about water sampling, analytical procedures used for chemical analysis is provided in the Supporting Information (SI Text and Tables S1−S3). Plasma Glucose Measurement and Leukocytes Counting. Blood samples were collected using capillary tubes from two anesthetized females and two anesthetized males of each replicate of exposure by tail ablation (n = 6 females and 6 males per treatment). Plasma glucose concentrations were then determined by use of a TESTA med kit, GlucoCheck Advance (Trend Pharma GmbH, Saarbrücken, Germany; detection range from 20 mg/dL to 600 mg/dL) according to the manufacturer’s instructions. The cut edge of the fish tail was used to make a blood film on a microscopic slide. Blood cell staining and leukocytes counting were performed based on the method described previously24,25 with modification. In brief, blood smear was stained with Wright stain solution for 2 min, then incubated with the same volume of distilled water for 2 min, and rinsed with distilled water until the edges showed faintly pinkish-red. Then the slide was stained with Giemsa stain solution for 1 min, rinsed with distilled water and airdried. Different blood cell counting was then performed using the microscope Olympus BX41 (Olympus Corporation, Tokyo, Japan) by recording 500−1000 total blood cells per fish using Olympus DP21 digital camera and software. Analysis of Swimming Behavior. Swimming behavior (activity) of eleuthero-embryos was performed at 120 hpf by 10247

DOI: 10.1021/acs.est.6b03436 Environ. Sci. Technol. 2016, 50, 10245−10254

Article

Environmental Science & Technology

Figure 1. Reproductive effects and embryonic development of F1 generation. (A) Average number of eggs per female per day during the 3 weeks of FLU exposure given in measured concentrations in μg/L. Each bar represents mean value ± SD of three replicates per treatment. (B) Condition factors for females (left) and males (right) (n = 9 per treatment). (C) Heartbeat rate of F1 embryos at 36 h (n = 24−36 per treatment). (D) Hatching success of F1 eleuthero-embryos at 72 h (n = 4−6 per treatment). (E) Swimming behavior (activity) of F1 eleuthero-embryos at 120 h. Left: representative swimming behavior at different FLU concentrations recorded in 1 min and represented by Viewpoint. Right: bar graphs representing average values per minute (n = 16). Asterisks: treatments that are significantly different from solvent control (*p < 0.05).

checked in future experiments in order to evaluate and control it. In the embryo exposure experiment, FLU concentrations were measured during the 120 h exposure period by collecting two samples each at the beginning (0 h; G1 and G3) and after 24 h prior to water renewal (G2 and G4), whereby the replicates were pooled. The mean measured concentrations were 0.081, 0.683, 6.52, and 60.21 μg/L, which were 19%−40% lower than nominal (Table S3). The concentrations were slightly lower than nominal at 0 h and further decreased during the 24 h exposure. This decrease of nominal concentrations is consistent with the observations in our previous studies with progestins and estrogens.18,23 Reproductive Effects and F1 Eleuthero-Embryo Development. During the 14 days pre-exposure and 21 days FLU exposure period, the egg production was consistent and similar across all the groups (ranging from 208 to 371 for pre-exposure and 200 to 349 eggs for exposure per female in total), showing no obvious alterations (Figures 1A, S2). Exposure to 41.95 μg/L FLU caused a slight but not significant decrease in body weight and condition factor in females. In males, the condition factors significant decreased at 0.48 and 41.95 μg/L FLU (Figures 1B, S3). This was possibly related to the imposed stress by FLU exposure that may have affected fish fitness, metabolism and immune competence. A similar effect was reported after exposure to 500 μg/L dexamethasone for 29 days with a significant reduction in weight and length of fathead minnow fry. The same dexamethasone concentration decreased the egg production in adult fathead minnow.7 This difference compared to our study was possibly due to the lower concentration used in our study and different properties and activities of the compounds. The fertilization rate was not adversely affected by FLU. It was about 80% in all the treatments with only a small decrease

tested for the normality and homogeneity of variances using Shapiro-Wilk and Levene’s tests, respectively. Where data were not normally distributed, log-transformation was performed to allow the parametric testing. Nonparametric Kruskal−Wallis test were used where necessary. Hierarchical clustering map was constructed by use of MultiExperimental Viewer v4.9 (DanaFarber Cancer Institute, Boston, MA). Data for gene expressions were graphically illustrated and statistically analyzed by GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA). Results are given as mean ± standard deviation (S.D.). Differences were considered as significant at p < 0.05.



RESULTS AND DISCUSSION Chemical Analysis. In the adult zebrafish experiment, FLU concentrations were measured during the whole exposure period at day 1, 7, 14, and 21. Mean FLU concentrations in the different dose groups were 0.006, 0.042, 0.480, 4.90, and 41.95 μg/L, respectively, which were 39%−58% lower than nominal (Table S2). The concentrations did not significantly vary within the same dose group, thus the delivered concentrations were stable during the whole course of the experiment. In our previous studies with progestins, a similar decrease in nominal concentrations was found under similar exposure procedures.18,19 As previously observed, a series of factors could be responsible for this decrease as for example, adsorption to flowthrough system (plastic tubes), fish metabolism, particles and debris in the exposure experiment. Besides, partial (bio)transformation of FLU to fludrocortisone also occurred and was found to be in the range of 22−49% at high doses of FLU. However, we did not account for the fludrocortisone levels (thus our exposure concentrations relate to FLU only), although fludrocortisone probably had the same biological activity as FLU. This transformation effect should be carefully 10248

DOI: 10.1021/acs.est.6b03436 Environ. Sci. Technol. 2016, 50, 10245−10254

Article

Environmental Science & Technology

Figure 2. Plasma glucose levels, immune responses, and related transcript levels in liver of zebrafish after FLU exposure (μg/L). (A) Plasma glucose levels in control and FLU treated groups of zebrafish (n = 6). (B) Blood leukocyte counts in control and treated groups of zebrafish (n = 6). (C) Frequency analysis of zebrafish blood cells. Each bar from top to bottom indicates: monocytes (purple); granulocytes (green); lymphocytes + thrombocytes (red) and erythrocytes (blue). Data are shown as mean values of six fish. (D) Hierarchical clustering of gene transcriptional responses in female (F_) and male (M_) zebrafish liver. Genes from top to bottom are g6pca, tpp2, pepck1, socs3, Na/K ATPase, il17a, pepck2, slco2a1, pxr, and slc5a1.

storage of FLU in eggs, which became bioavailable during embryonic development. Similarly, we found a transgenerational effect for progesterone and drospirenone.19 Thus far, studies focused on direct exposures of embryos, such as with prednisolone that affected physiological and behavioral end points. The heartbeat and hatching success showed similar increases at concentrations as low as 0.1 μg/L.31 This is consistent with our findings on FLU, indicating that corticosteroids could impose stress on embryonic development of fish at very low concentrations. Plasma Glucose Regulation and Immune Response. Plasma glucose concentrations and immune system are the primary targets of corticosteroids for their pharmaceutical use in human and veterinary medicine. Corticosteroids were also reported to regulate plasma glucose levels and cause immune responses in fish species.8,9,33 In our 21 days exposure to FLU, plasma glucose levels significantly increased at 42 ng/L in females, and at 0.48 and 41.95 μg/L in males, while no dosedependent response was observed for both females and males (Figure 2A). This phenomenon was possibly due to the imposed stress on adult zebrafish by high doses of FLU exposure. Similarly, total numbers of blood leukocytes were slightly but significantly decreased at concentrations of 0.042, 4.9, and 41.95 μg/L in females, and at 4.9 μg/L in males (Figure 2B). These results are consistent with previous reports, in which the increase on glucose levels and reduction of leukocyte counts were observed simultaneously at corticosteroid concentrations of 1 μg/L or lower.8,9 Our detailed analysis revealed that the proportion of different leukocytes changed. Lymphocyte + thrombocyte counts were significantly decreased in response to high doses of FLU, with up to 43% and 36% decrease in females and males, respectively. Granulocyte counts showed slight but not significant decreases. No alterations were observed for monocytes, which held steady

in the lowest FLU exposure (Figure S4). In F1 eleutheroembryos, the frequency of spontaneous muscle contractions at 24 hpf showed a slight but not significant decrease at several FLU concentrations (Figure S4). However, heartbeat rates at 36 hpf embryos were remarkable induced between 4.6% to 5.1%, even at the lowest concentrations of 6 ng/L FLU (Figure 1C, S4). Corticosteroid signaling can affect the structure and function of the heart in zebrafish embryos.29 Injection of cortisol into zebrafish embryos causes heart deformities and suppression of essential cardiac genes.30 Though no remarkable physiological defects were observed for embryo heart development in our study, the present data suggest that cardiogenesis of embryos would be disrupted by FLU exposure at environmental relevant concentrations. FLU also accelerated hatching in F1 embryos. At 72 hpf, acceleration of hatching success was significant at exposures of F0 fish to 4.9 and 41.95 μg/L FLU. Hatching was already about 98%−99%, while in controls, it was only 84% (Figure 1D). At 48 hpf, 96 hpf and 120 hpf hatching success was not different (Figure S4). This increase in 72 hpf hatching success agrees with the previous reports, in which, dexamethasone and prednisolone also resulted in precocious hatching in zebrafish embryos, and the cause is more likely to be related to the changes in the release of proteolytic enzymes.31,32 The swimming behavior (activity) of 120 hpf F1 embryos displayed a significant increase at 42 ng/ L, 480 ng/L, and 41.95 μg/L of FLU, while no monotonic dose response curve was observed (Figure 1E, S5). This was possibly due to the imposed stress on adult zebrafish by high doses of FLU exposure that affected the fitness of adult fish and consequently development and behavior of F1 embryos. Our results suggest an unexpected transgenerational effect of FLU, affecting F1 embryos of exposed parents. Reasons could be that stress-related physiology was transferred in F0 fish to their gametes by an unknown process, or via accumulation and 10249

DOI: 10.1021/acs.est.6b03436 Environ. Sci. Technol. 2016, 50, 10245−10254

Article

Environmental Science & Technology

Figure 3. Endocrine responses to FLU exposure (μg/L). (A) Ovary weight and gonadal somatic index (GSI) in female zebrafish exposed to FLU. (B) Heat map depicts alteration of gene transcripts of hypothalamic-pituitary−gonadal (HPG)-liver axis genes in zebrafish exposed to FLU in a color scheme: green represents significant down-regulation, red significant up-regulation, and yellow no significant alteration. (C) Relative transcript levels of hsd17β3 in gonads compared to solvent control. Data are shown as mean ± SD of three biological replicates. *p-value