Organophosphate esters cause thyroid dysfunction via multiple signaling pathways in zebrafish brain.

Organophosphate esters (OPEs) are widespread in various environmental media, and can disrupt thyroid endocrine signaling pathways. Mechanisms by which OPEs disrupt thyroid hormone (TH) signal transduction are not fully understood. Here, we present in vivo-in vitro-in silico evidence establishing OPEs as environmental THs competitively entering the brain to inhibit growth of zebrafish via multiple signaling pathways. OPEs can bind to transthyretin (TTR) and thyroxine-binding globulin, thereby affecting the transport of TH in the blood, and to the brain by TTR through the blood-brain barrier. When GH3 cells were exposed to OPEs, cell proliferation was significantly inhibited given that OPEs are competitive inhibitors of TH. Cresyl diphenyl phosphate was shown to be an effective antagonist of TH. Chronic exposure to OPEs significantly inhibited the growth of zebrafish by interfering with thyroperoxidase and thyroglobulin to inhibit TH synthesis. Based on comparisons of modulations of gene expression with the Gene Ontology and Kyoto Encyclopedia of Genes and Genomes databases, signaling pathways related to thyroid endocrine functions, such as receptor-ligand binding and regulation of hormone levels, were identified as being affected by exposure to OPEs. Effects were also associated with the biosynthesis and metabolism of lipids, and neuroactive ligand-receptor interactions. These findings provide a comprehensive understanding of the mechanisms by which OPEs disrupt thyroid pathways in zebrafish.

Introduction

Because fires can cause significant loss of life and damage to property globally, the inclusion of flame retardants in various potentially flammable products is required in many jurisdictions [1]. Owing to the phasing out and regulation of brominated flame retardants (BFRs), in recent years organophosphate flame retardants (OPFRs) have instead been widely used worldwide as replacements [2,3]. It has been reported that the annual consumption of OPFRs has increased significantly over the last few decades [4,5]. In China, the production of OPFRs has increased at an average, annual rate of 15% [6]. In the U.S.A., it was projected that approximately 50,000 tons of OPFRs would be manufactured per year by 2020 [7]. Because they are often not covalently bound into materials, organophosphate esters (OPEs) can be released into the environment, resulting in their widespread detection in multiple environmental media, including abiotic and biotic compartments [6,[8], [9], [10], [11], [12], [13]] and [14]. OPEs have also been reported to be prone to long-range atmospheric transport away from their sources and can move toward the poles via the “grasshopper effect,” which is caused by differential cooling [15,16]. This behavior, eventually results in the global contamination of remote and oceanic environments far from industrial and urban areas where OPEs are manufactured and used [[15], [16], [17], [18]]. Due to this transport in the environment, there is a need for study of the potential exposure of wildlife and humans to OPEs and associated adverse effects.

Recently, multiple mechanisms behind the toxicity of OPEs to aquatic organisms have been reviewed and discussed [19]. Conclusions of that review indicated that neurotoxicity, cardiotoxicity, hepatotoxicity, endocrine disruption including of sex hormones and thyroid hormone (TH), and adverse effects on reproduction and development have been observed, both in vitro and in vivo [[20], [21], [22], [23], [24], [25]]. Previous results suggested that, in theory, some new chemicals should cause less acute toxicity, but might be more toxic owing to chronic exposure [26]. Furthermore, exposure to OPEs was predicted to disrupt endocrine pathways at environmentally relevant concentrations, which were lower than those causing other toxic effects [27] and [19, [28], [29], [30], [31], [32]]. Disruption of endocrine functions was also found to be related to changes in the fitness of fish populations [[33], [34], [35]]. Generally, decreases in biodiversity are preceded by changes at the molecular and physiological levels of organization, which result in changes in the functions of tissues and organs, followed by integrative effects on traits such as scope for growth and reproduction, which can be measured based on body mass and length [36]. TH is associated with regulation of the growth of organisms [37] and understanding the mechanisms and adverse outcome pathways of TH can thus be critical to predicting changes in populations [38,39]. Compared with toxicity with direct links to the estrogen, androgen, and steroidogenesis (EAS) modalities, information on the indirect effects of OPEs on the hypothalamic–pituitary–thyroid (HPT) axis of fish and subsequent effects on reproduction remains limited.

In fish, the thyroid modality of toxic action can involve the disruption of multiple targets of thyroid signals and functions along the thyroid hormone-mediated, signal transduction cascade [19,[40], [41], [42]]. Results of recent studies have demonstrated that tris(2-chloroisopropyl) phosphate (TCIPP) can function as an antagonist by inhibiting the binding of 17β-estradiol to estrogen receptor [27], but also reduce functioning of the thyroid gland and increase the activities of hepatic T4-outer ring deiodinase enzymes [22]. When zebrafish embryos/larvae were exposed to tris(1,3-dichloro-2-propyl) phosphate (TDCIPP) and triphenyl phosphate (TPHP), the synthesis of TH via enzymes (tshβ, slc5a5, and tg), metabolism (dio1), transport (ttr), elimination (ugt1ab), and thyroid development (hhex, nkx2.1, and pax8) were affected, which could disrupt central regulation through signaling pathways and synthesis to increase the TH levels in blood plasma [25,29,37]. Aryl, alkyl, and chlorinated OPEs were all considered as potential disruptors of TH and antagonists of thyroid hormone receptor β (TRβ) [43]. However, research into the acute toxicity of OPEs to zebrafish embryos/larvae suggested that aryl OPEs, including cresyl diphenyl phosphate (CDP) and TPHP, were more likely to cause heart malformations, pericardial edema, and cardiac-looping blocking than alkyl and chlorinated OPEs, which was mainly attributed to their hydrophobicity and potential bioaccumulation [21]. Moreover, TPHP has been suggested to be an effective inhibitor of acetylcholinesterase (AChE) activity to induce developmental neurotoxicity in zebrafish [44,45]. To date, research has focused on the mechanisms by which OPEs impact TH by affecting the amounts of hormones produced, while information on the effects on the molecular mechanisms and signaling pathways of TH has remained limited. Thus, an important aspect to be considered is whether there is a specific biomarker, such as a target protein, target molecule, or specific OPE, for the effects of OPEs on thyroid function.

In this study, three typical OPEs were selected to explore the disruption of TH endocrine functions in vivo/in vitro using zebrafish and GH3 cells, and in silico using molecular docking. Results from the combination of these methods elucidated the primary molecular mechanisms at the molecular, gene expression, cell, protein, and individual levels. Modes of transport of OPEs in the blood were also investigated using molecular docking simulation. In addition, GH3 cell assays were used to explore how OPEs competitively enter cells. To elucidate the primary molecular mechanisms behind the disruption of TH-signaling pathways, after chronic exposure, the expression of key genes, hormones, and proteins in target organs was investigated. Specific aims in this study were to: (1) investigate the correlations between TH biosynthesis and OPE transportation in vivo with precursor proteins or transporters; (2) discuss molecular modes of action regarding TH endocrine disruption by OPEs; and (3) identify a specific biomarker for the disruption of thyroid function in zebrafish.

Materials and methods

Chemicals and reagents

Tris(isopropyl) phosphate (TIPP) (CAS 513-02-0), CDP (CAS 26444-49-5), TCIPP (CAS 13674-84-5), and dimethyl sulfoxide (DMSO) were purchased from J&K Scientific, Ltd. (Beijing, China). The internal standard [tris(1,3-dichloroisopropyl) phosphate-D15, TDCIPP-d15] and phosphate-buffered saline (PBS) were purchased from Toronto Research Chemicals, Inc. (Toronto, ON, Canada), and Coolaber (Beijing, China), respectively. All organic solvents were of HPLC-grade purity and were obtained from Thermo Fisher Scientific (Waltham, MA, USA). RPMI-1640 culture medium was purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China). Charcoal-dextran-treated fetal bovine serum (CD-FBS) and DMEM/F-12 were obtained from HyClone (Logan, UT, USA). MTS kit used for measuring cell proliferation was purchased from Bestbio (Shanghai, China).

Maintenance of zebrafish and exposure to OPEs

Zebrafish were maintained in continuous-flow tanks filled with carbon-filtered water under a photoperiod of 16:8 h light/dark for 25 ± 2 °C. After acclimation, 1-month-old zebrafish were exposed to TIPP, CDP, or TCIPP following OECD test guidelines protocol no. 229 with minor modifications (OECD 229). Briefly, zebrafish were exposed individually in a semi-static exposure system to 10, 50, or 100 μg L−1 of each OPE or blank control for 28 days. A carrier of 0.01% DMSO (v/v) was used and exposure was performed in triplicate. Fifteen zebrafish were exposed in 3 L of dechlorinated carbon-filtered water filtered via a 0.45 μm filter membrane. During 28 dexposure, half of the exposure solution was replaced daily with freshly prepared filtered water, containing a corresponding concentration of TIPP, CDP, or TCIPP. Exposure solutions were sampled before and after renewing these solutions, 4 mL of which were filtered via a 0.45 μm Minisart syringe filter (Sartorius, Göttingen, Germany), followed by storage at −20 °C until analysis. Detailed protocols used for the identification and quantification of TIPP, TCIPP, and CDP in exposure solutions and biota are provided in the Supporting Information (SⅠ) (Tables S1 and S2 and Fig. S1), using a high-performance liquid chromatograph interfaced with a mass spectrometer (HPLC-MS/MS, Nexera-X2/8040; Shimadzu Co. Ltd., Kyoto, Japan) with a Waters C18 column (50 mm × 2.1 mm, 3.5 mm). The mass and snout-to-vent length of each zebrafish were measured before and after the exposure. Blood was collected from the caudal vein using glass capillary tubes to quantify thyroxin (T4). Brain and soma samples were collected for the quantification of T4, thyroperoxidase (TPO), and thyroglobulin (TG). Transcriptome sequencing analysis was performed and the protein levels of thyroxine-binding globulin (TBG), transthyretin (TTR), and Na+/I− symporter (NIS) were quantified. Detailed protocols are provided in the Supporting Information Section SⅣ.

Competitive inhibition assay and Western blot on GH3 cell lines

Rat pituitary tumor GH3 cells were purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China), and cultured in Ham's F–12K culture medium supplemented with 15% horse serum (HS), 2.5% fetal bovine serum (FBS), and 1% penicillin–streptomycin (P/S) at 37 °C in a humidified atmosphere with 5% CO2. Cells were stored in liquid nitrogen with cell freezing medium containing 55% basic medium, 40% FBS, and 5% DMSO.

Based on the proliferation of GH3 cells for 96 h, the concentrations of OPEs and T4 that caused 20% (RIC20) or 50% (RIC50) inhibition of cell proliferation were calculated (Table S3). Proliferation of cells in the T-screen assay for 96 h was carried out by slightly modified versions of previously published methods [43,46,47]. Briefly, cells were cultured for 48 h in phenol red-free, RPMI-1640 culture medium with 5% CD-FBS, and 1% P/S. Cells were collected from culture plates and used in the competitive inhibition assay conducted in a transparent 48-well culture plate. The experimental design included a blank control (0.1% v:v DMSO); exposure to T4, CDP, TIPP, or TCIPP individually; or exposure to combinations of T4 and OPEs: T4+CDP, T4+TIPP, or T4+TCIPP. Concentrations administered for the exposure were set to either RIC20 or RIC50. Each exposure solution was conducted in triplicate and refreshed every 24 h during 96 h of exposure. Absorption measured as optical density (OD490), which was used as a measure of cell proliferation, was determined via a microplate reader at 490 nm after treatment with the MTS kit [43]. The remaining cells were subjected to western blotting to determine the levels of proteins of TBG, TTR, and NIS, with the results normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression. The detailed procedure is provided in Supporting Information Section SⅢ.

Transcriptome sequencing (RNA-seq) in zebrafish brain

Brain samples were taken from each control or treatment group after 28 d. Six brain samples were taken from each replicate exposure and used for transcriptome sequencing related to key signaling and functional processes of the HPT axis. An RNA-seq transcriptome library was prepared by extracting total RNA in TRIzol® Reagent, using the TruSeq™ RNA sample preparation kit from Illumina (San Diego, CA, USA). Before read mapping for these samples, a paired-end RNA-seq sequencing library was sequenced with the Illumina HiSeq xten/NovaSeq 6000 sequencer (2 × 150 bp read length). Differentially expressed genes (DEGs) were analyzed by RSEM (http://deweylab.biostat.wisc.edu/rsem/). Functional enrichments with the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases were carried out using Goatools (https://github.com/tanghaibao/Goatools) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do). Detailed protocols used for the RNA-seq of TIPP, TCIPP, and CDP are provided in Supporting Information Section SⅤ. To validate the results of transcriptome analysis, several DEGs calculated by RSEM were selected to conduct the quantitative real-time PCR (qRT-PCR) for the brain samples, as detailed in Supporting Information Section SⅥ.

Molecular docking simulation

The structures for OPEs and T4 investigated in this study were built and optimized using the software package Discovery Studio Visualizer 4.0 (DS4.0; Accelrys Inc., CA, USA). Molecular docking with TBG (ID: 2RIW) and TTR (ID: 1ICT 3.0 Å) were obtained from the Protein Data Bank (PDB, www.rcsb.org). The molecular docking was performed with the cdocker protocol in DS4.0 based on the results of previous research [48,49]. During the molecular docking, at least 20 random ligand conformations generated in this study were further refined by grid-based simulated annealing in the binding site. Final interaction energy was selected from top-ranking poses derived during the docking analysis.

Statistical analyses

All statistical analyses in this study were performed using IBM SPSS Statistics 22.0 (IBM Corporation, Armonk, NY, USA) and Origin 2018 (OriginLab Corporation, Northampton, MA, USA). First, the normality of parameters was checked using the Shapiro–Wilks test, and the assumption of homogeneity of variance was evaluated using Levene's test. One-way analysis of variance (ANOVA) with least significant difference was used as a multiple range test to measure the significance of differences between the control and OPE-treated groups. The results from at least three independent repetitions were expressed as mean ± standard deviation (SD). In this study, significant differences between the control and treatment groups were defined by a p-value less than 0.05.

Results

Analysis of growth inhibition in zebrafish induced by OPEs

Exposure of zebrafish to TIPP, CDP, or TCIPP inhibited growth and development as determined by body mass and length measurements (Fig. 1a). Exposure to 10 μg CDP L−1 resulted in significantly shorter body length, whereas exposure to other concentrations of CDP did not have significant effects. There was no significant change of body mass in zebrafish exposed to CDP. Exposure to 50 μg L−1 TIPP or TCIPP significantly decreased the body length of zebrafish compared with that in the control. Exposure of zebrafish to TIPP ranging from 10 to 100 μg L−1 resulted in significantly, dose-dependently greater body mass, while for TCIPP, no significant inhibition was observed, except for with 100 μg TCIPP L−1. Impacts on body mass were the major effects of TIPP, while exposure to TCIPP affected body length.

Fig. 1

Inhibition of growth and hormone imbalance of zebrafish exposed to CDP, TIPP, or TCIPP for 28 d a, Body mass and length of zebrafish exposed to OPEs. b, Concentrations of T4 in blood plasma or brain after 28 d exposure. c, Concentrations of TG and TPO in brain of zebrafish. Mean ± standard deviation (SD, n = 3), ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

The concentrations of CDP in tissues of zebrafish exposed to 10, 50, and 100 μg L−1 were dose-dependent and were 5.26 ± 1.90, 21.63 ± 6.42, and 27.43 ± 2.01 ng g−1 wet mass (w/w), respectively (Table S2). Concentrations of TCIPP in tissues of zebrafish were found at 100 μg TCIPP L−1, namely, 20.99 ± 7.31 ng g−1 w/w. However, upon exposure to TIPP, TIPP was consistently not detected in zebrafish tissues.

Analysis of contents of T4, TG, and TPO in various tissues of zebrafish

Contents of T4, TG, and TPO in blood plasma, brain, and somatic tissues (muscle) were measured by ELISA kits to analyze the correlation between hormone levels and growth of zebrafish after exposure to OPEs for 28 d. Exposure of zebrafish to OPEs for 28 d affected the concentrations of T4 in blood plasma, brain, and somatic tissues (muscle) (Fig. 1b and S4a). No significant changes in the contents of T4 were observed in the brain of zebrafish exposed to TCIPP (Fig. 1b). Compared with controls, exposure to CDP caused significantly, dose-dependent, lower contents of T4 in the brain, while exposure to TIPP caused dose-dependent effects at greater contents of T4 (p < 0.05 only at 10 μg L−1), which were all less than that of the control. The only exception was 100 μg L−1. Exposure to CDP and TIPP caused dose-dependent lower total contents of T4 in the blood plasma (Fig. 1b). Compared with the control, the lower content of T4 in the blood plasma was significant at 100 μg CDP L−1, and the greater content was significant at 10 μg TIPP L−1. Contents of T4 were significantly lower in zebrafish exposed to 10 μg TCIPP L−1, while no significant effects on T4 were observed in zebrafish exposed to other concentrations of TCIPP. Exposure to CDP, TIPP, or TCIPP, except for 50 μg CDP L−1 and 100 μg TIPP L−1, did not significantly affect the contents of T4 in somatic tissue (Fig. S4a). Ranges of the contents of T4 in the brain and somatic tissue of zebrafish exposed to CDP, TIPP, or TCIPP accounted for 4.2–9.3% and 3.6–9.6% of the total plasma concentrations of T4, respectively. Mean values of the percentages of T4 in the brain and somatic tissues were 6.7 ± 1.2% and 5.9 ± 2.2%, which were less than one-tenth of the total contents of T4 in blood plasma. Contents of T4 in the brain were greater than those in somatic tissue.

Exposure of zebrafish to CDP, TIPP, or TCIPP caused dose-dependent changes in the contents of TPO and TG which are key proteins in the synthesis of TH in the brain (Fig. 1c). Compared with the control, exposure of zebrafish to CDP caused significantly, dose-dependently lower TG in the brain, but significantly, dose-dependently greater contents of TPO upon exposure to 100 μg CDP L−1. Zebrafish exposed to TIPP had significantly lower contents of TPO in the brain relative to the controls, while the contents of TG were significantly lower than those of controls in zebrafish exposed to the greatest concentration (100 μg TIPP L−1). During chronic exposure to TCIPP, contents of TG in the brain increased in a dose-dependent manner, and that upon exposure to the greatest concentration (100 μg TCIPP L−1) was significantly increased. Contents of TPO in the brain were lower than those of the controls in zebrafish exposed to TCIPP. For somatic tissue (muscle), contents of TG in fish exposed to CDP or TIPP were greater than those in controls in a dose-dependent manner (Fig. S4b). Contents of TPO upon exposure to CDP were reduced in a dose-dependent manner, while they increased in a dose-dependent manner in zebrafish exposed to TIPP (Fig. S4b).

Transcriptome analysis of thyroid disruption caused by OPEs in brain

RNA-seq was used to investigate the key signaling pathways of thyroid dysfunction induced by OPEs in zebrafish. According to the hierarchical cluster analysis of 91 thyroid-related genes or 35 co-expressed genes related to thyroid function, the patterns of expression of genes in fish exposed to TIPP or TCIPP were similar, which differed from the pattern in the brain of fish exposed to CDP (Figs. S9 and S10). The patterns of expression of all genes in unexposed controls also differed from those of fish exposed to OPEs. Twenty-four DEGs related to thyroid function were selected from 9793 DEGs in all samples, which were annotated for the Danio rerio genome, and used for hierarchical cluster analysis and functional enrichment analyses of both GO databases (Fig. 2, Fig. 3, Tables S20 and S21). The results of the functional annotation analysis for thyroid-related genes identified certain GO terms, with the genes being particularly involved in biological processes (e.g., cellular process and biological regulation), cellular components (e.g., cell parts and organelle), and molecular functions (e.g., binding) in zebrafish (Fig. 3a, S11a, and S12a). Based on classified statistics of KEGG signaling pathway, environmental information processing pathways (e.g., signaling molecules and interaction, and signal transduction) were the main KEGG pathways related to the effects of nerve conduction on the thyroid (Figs. S11c and S12c). The results of functional enrichment analyses, using GO and KEGG, demonstrated significant changes of multiple signal pathways related to thyroid effects among all OPE treatments, including binding to TH receptor (THR), thyroid hormone generation, thyroid hormone metabolic process, thyroid hormone transmembrane transporter activity, thyroid gland development, neuroactive ligand–receptor interaction, and insulin signaling pathway (p-adjust ≤ 0.05) (Fig. 3b, S11b and d, and S12b and d, Table S6). Regulation of lipid biosynthetic process (p-adjust < 0.05) could be induced by thyroid-related genes after exposure to OPEs (Fig. 3b, S11b, S12b, and S13). Thyroxine 5′-deiodinase activity, which plays an important role in thyroid hormone generation, was also significantly enriched in the GO analysis (p-adjust <0.05) (Figs. S11b and S12b). Based on the correlation analysis of gene and protein interactions (Figs. S13 and S14), disruption of thyroid function by OPEs might be associated with certain signaling pathways related to the nervous system and lipid biosynthesis.

Fig. 2

Hierarchical clustering analysis of DEGs related to thyroid effects (24 genes) selected from 9793 DEGs in all samples exposed to CDP, TIPP, and TCIPP in zebrafish brain. Colors in the figure indicate normalized expression of genes in each sample. Red represents overexpression of genes and blue represents under-expression. To the left is a tree diagram of gene clustering and the module diagram of sub-clustering, and to the right is the names of genes. The names of the samples are shown at the bottom.

Fig. 3

Functional enrichment analysis of DEGs related to thyroid effects (24 genes) in samples exposed to CDP, TIPP, or TCIPP. a, Functional annotation analysis based on the GO database of 24 DEGs related to thyroid effects in all brain samples exposed to CDP, TIPP, or TCIPP. b, Top 39 GO terms of enrichment from the GO database of DEGs (p-adjust < 0.5). Enrichment factor represents the ratio of the number of genes enriched in the GO term and the total number of annotated genes. Size of dots indicates the numbers of genes associated with the GO term, while colors of dots correspond to p-adjusted ranges.

Quantitative analysis of TBG, TTR, and NIS levels based on RNA-seq results

To verify that transmembrane transport and blood transportation are key signaling pathways, as determined by transcriptional analysis, TTR, TBG, and NIS in the brain were measured by ELISA. Compared with the levels in the control, concentrations of TTR in fish exposed to CDP increased in a dose-dependent manner (Fig. 4a). Exposure to 50 or 100 μg CDP L−1 resulted in significantly greater concentrations of TTR. Exposure to TIPP or TCIPP resulted in significantly (p < 0.01) greater concentrations of TTR in the brain after 28-d exposure. Concentrations of TBG in the brain were not significantly altered by exposure to CDP, TIPP, or TCIPP compared with those of the unexposed control (Fig. 4b). Concentrations of NIS in the brain were generally greater in individuals exposed to CDP, TIPP, or TCIPP than those in brains of unexposed controls (Fig. 4c), especially exposure to 10 or 50 μg CDP L−1, 100 μg TIPP L−1, and all concentrations of TCIPP.

Fig. 4

Amounts of TTR, TBG, and NIS proteins in brain after 28 d of exposure of zebrafish to CDP, TIPP, or TCIPP. a, Changes in concentrations of TTR in the brain of zebrafish exposed to CDP, TIPP, or TCIPP. b, Changes in concentrations of TBG in brain. c, Changes in concentrations of NIS in brain. Mean ± standard deviation (SD, n = 3), ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Competitive inhibition and Western blot analysis of T4 or OPEs

To explore the effects of OPEs on cell proliferation, a competitive inhibition assay was performed on GH3 cells. Based on acute toxicity, exposure to T4 promoted the proliferation of cells, while exposure to OPEs inhibited such proliferation (Table S3). Upon analysis of competitive correlations among OPEs and T4 based on colorimetric responses of the MTS kit (Fig. 5), with RIC20 as the endpoint, OD490 values for cells exposed to individual OPEs were not significantly different from those of cells exposed simultaneously to OPEs in the presence of T4 (T4+OPEs). OD490 values upon exposure to T4 individually were significantly greater than those of T4+OPEs (p < 0.01). However, when the endpoint was the RIC50, exposure to T4+CDP resulted in significantly greater proliferation of GH3 cells than exposure to either T4 or CDP alone (p < 0.01). Exposure to T4+TIPP can significantly inhibit the proliferation of GH3 cells compared with that of cells exposed to T4 or TIPP alone (p < 0.01). Simultaneously, exposure to T4+TCIPP resulted in significantly greater proliferation than exposure to TCIPP alone (p < 0.05), while it resulted in significantly less proliferation than that by cells exposed solely to T4 (p < 0.001). In addition, expression of TTR protein was not detected in GH3 cells (Fig. S2). This might have been due to the occurrence and distribution of TTR protein in cells. In the case of the RIC20, compared with GH3 cells exposed to T4 alone and blank control, concentrations of NIS were significantly decreased upon exposure to TIPP, TCIPP, or T4+OPEs, while there was no significant difference at the RIC50. Likewise, at the RIC50, there was significantly more TBG in GH3 cells exposed to TIPP, TCIPP, and OPEs with T4, compared with cells exposed to T4 alone, while there were no significant differences at the RIC20.

Fig. 5

Proliferation of GH3 cells assessed using the MTS kit after exposure to the RIC20 and RIC50 of CDP, TIPP, and TCIPP for 96 h a, Changes in the proliferation of GH3 cells exposed to RIC20. b, Changes in the proliferation of GH3 cells exposed to RIC50. Mean ± standard deviation (SD, n = 3), ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

Analysis of interaction of OPEs with TBG and TTR

Based on the molecular docking simulation, the interactions between OPEs and TBG or TTR were analyzed and compared with the interactions between T4 and TBG or TTR. Amino acid binding sites involved in the interaction between OPEs or T4 and TBG were Asn A 273, Arg B 381, and Lys A 270, while those involved in interactions between OPEs or T4 and TTR were Ser B 117, Ser D 117, Lys D 15, and Lys B 15 (Table 1, Figs. S16 and S17). Based on “cdocker” energy, binding affinities of OPEs to TBG were in the following order, TNBP > TDCIPP > TCIPP > TIPP > T4 > CDP > TPHP, while those to TTR were TNBP > TCIPP > TDCIPP > TIPP > T4 > CDP > TPHP (Table 1). Affinities of binding between TIPP or TCIPP and TBG or TTR were greater than those of T4 and CDP. When interacting with TBG, Arg B 381 was the common amino acid binding site of T4 or aryl-OPEs (CDP and TPHP), while Asn A 273 and Lys A 270 were specific ones of T4 and CDP, respectively. In addition, Lys D 15 was the common amino acid binding site of T4 or TCIPP interacting with TTR. However, Lys B 15 was the specific amino acid binding site for CDP and TPHP when interacting with TTR, and Ser B 117 and Ser D 117 were the only binding sites for interaction between T4 and TTR. Chlorinated OPEs might interact with TBG via van der Waals or hydrophobic interaction, while alkyl OPEs might interact with TBG and TTR via van der Waals or hydrophobic interaction.

Table 1

The binding affinity of OPEs for TBG and TTR using molecular docking.

Chemicals	CAS	TBG		TTR
		Cdocker Energy kcal mol−1	Interaction	Cdocker Energy kcal mol−1	Interaction
T4	300-30-1	−47.0479	ASN A: 273a, ARG B: 381 b	−42.6893	SER B: 117a, SER D: 117a, LYS D: 15a
TIPP	513-02-0	−48.634	–	−50.7945	–
CDP	26444-49-5	−38.183	ARG B: 381 b, LYS A: 270 b	−40.128	LYS B: 15 b
TCIPP	13674-84-5	−54.5853	–	−55.6138	LYS D: 15a
TNBP	126-73-8	−60.4225	–	−57.6073	–
TPHP	115-86-6	−34.1994	ARG B: 381 b	−36.3499	LYS B: 15 b
TDCIPP	13674-87-8	−56.755	–	−53.7351	LYS D: 15a

Note: a: Hydrogen bond interaction; b: π–cation interaction; —: There is no amino acid binding site between OPEs and TBG or TTR, while van der Waals interaction and hydrophobic interaction mainly occur.

Discussion

Disruption of the TH hormone function by exogenous substances that adversely affect the HPT axis can play significant roles in vertebrate development and homeostasis related to growth and energy metabolism [50,51]. In this study, when 1-month-old zebrafish were chronically exposed to OPEs for 28 d, thyroid function was disrupted at the molecular level (in silico), and concentrations of hormones, transcription of RNA, and concentrations of proteins were altered in vivo and in vitro, which resulted in the inhibition of zebrafish growth.

Significant accumulation of CDP rather than TIPP and TCIPP in zebrafish

A growing number of studies have validated that the three OPEs are present in fish as well as in ambient surface water [14, 32, 52]. In the present study, the differential distribution of concentrations of OPEs in tissues of zebrafish was identified and suggested to be due to the greater hydrophobicity and greater bioconcentration factor (BCF) of CDP than those of TIPP and TCIPP. These results were consistent with the following: (1) During long-term exposure, the accumulation of OPEs in aquatic organisms was proportional to the concentrations to which they were exposed. (2) Greater bioconcentrations of CDP in tissues of zebrafish were observed, while there was little bioconcentration of TIPP in such tissues. In fact, in contrast to alkyl OPEs and chlorinated OPEs, TPHP was generally detected in three different tissues of Chinese rare minnow and was found at greater concentrations in male gonad and liver [52]. Based on these results and those of previous studies, differences in toxic potencies caused by the same concentrations of OPEs might be related to differences in detection methods, exposure environments, and pathways of exposure [19,53]. Thus, internal doses, as determined by the concentrations accumulated in tissues of aquatic organisms among exposure media, might better reflect bioavailability and the actual concentration of toxicity effects than the traditional measure of exposure dose, namely, the nominal ambient concentration. Briefly, the concentrations of CDP accumulated in tissues of zebrafish were greater than those of other OPEs, and CDP also caused observable adverse effects, which supported the hypothesis that the greater potency of CDP was due to its greater accumulation.

OPEs regarded as environmental hormones with TH effect in zebrafish

TPO is the primary enzyme involved in the synthesis of TH, while TG is the prohormone and storage form of TH; both are markers of the physiological function of the thyroid gland [54]. Thus, changes in the content of T4 are closely related to the contents of the two proteins TPO and TG. Exposure to environmentally relevant concentrations of CDP, TIPP, or TCIPP was shown to result in changes in the concentrations of T4 in blood plasma, brain, and somatic tissue, as well as changes in the amounts of TPO and TG in brain and muscle. Results of previous studies demonstrated that the exposure of American kestrels (Falco sparverius) to environmentally relevant concentrations of OPEs for 21 d might decrease thyroid gland activity and increase hepatic deiodinase activity, eventually influencing concentrations of free T4 or T3 in blood plasma [22]. The results of another study demonstrated that exposure to TPHP can disrupt central regulation and pathways of synthesis, metabolism, transport, and elimination of TH, thus increasing the concentrations of TH in embryos/larvae of zebrafish [37]. It has been reported that OPEs could cause thyroid endocrine disruption of not only adult zebrafish, but also their offspring via the maternal transfer of OPEs [25]. Additionally, in epidemiological studies, TPO and TG have been considered to be key indicator proteins to be monitored to interpret thyroid endocrine disorders [54]. In previous studies of thyroid diseases, thyrotropin receptor, TPO, TG, NIS, and pendrin proteins were identified as antigens that resulted in an autoimmune pathogenic response, which also included molecular mechanisms and signaling pathways associated with thyroid diseases [[54], [55], [56], [57]]. TPO and TG could be identified by CD8-positive T cells, and induce Hashimoto's thyroiditis leading to the destruction of thyroid tissue, which was considered to be a manifestation of clinical disease [58]. Further research has found that TPO and TG are targets of IgE or IgG autoantibodies during chronic, spontaneous urticaria or hypothyroidism [59]. Chronic exposure of zebrafish to OPEs for 1 month resulted in positive correlations between the concentrations of TPO or TG in the brain and the concentrations of T4 in blood plasma, especially upon exposure to CDP. Similarly, the ratio of concentrations of T4 in brain to the total concentration of T4 in blood plasma was greater than that in somatic muscle. These results indicated that exposure to OPEs might affect the concentrations of TPO and TG in the brain, thereby disturbing the synthesis and secretion of TH, whose concentrations were closely related to the development of the brain. Together, this information suggests that the brain plays a vital role in the thyroid endocrine system. The results of several studies have shown that the developing brain tends to be more sensitive to xenobiotics, such as OPEs, and changes in concentrations of TH can significantly affect normal brain development, even adversely affecting the behavior of the offspring of exposed individuals [60,61]. Briefly, the results of the proteins analysis suggested that exposure to OPEs could affect the levels of TPO and TG in brain, resulting in anomalous changes in the concentrations of T4 in blood plasma and brain tissue. This in turn resulted in slower growth of zebrafish owing to lower concentrations of T4 in blood plasma.

Thyroid dysfunction induced by OPEs via multiple signaling pathways including nerve conduction and lipid regulation

Functional enrichment analysis of DEGs by GO and KEGG showed that the effects of growth inhibition of thyroid function in zebrafish were due to interactions with and modulation of multiple signaling pathways. Chronic exposure of zebrafish to OPEs affected significant signaling pathways by altering the binding of TH to the THR, development of the thyroid gland, synthesis and metabolism of TH, thyroid hormone transmembrane transporter activity, and neuroactive ligand–receptor interaction, which eventually resulted in effects on the thyroid in zebrafish that were in the stage of rapid development (Fig. 6 and Table S6). Receptor–ligand binding was the key signaling pathway of the endocrine-disrupting effects induced by chemicals. Therefore, the binding of TH to the THR was considered to be a primary potential target of thyroid disruption induced by OPEs [43]. It was found that OPEs could competitively bind to the membrane THR with TH to enter the cells, inducing thyroid endocrine disruption [49,62]. The results of previous RNA-seq demonstrated that OPEs can affect certain signaling pathways related to the synthesis and metabolism of TH, causing thyroid dysplasia, which resulted in lower concentrations of T4 and caused abnormal development in the early life stages of zebrafish [29,37]. For example, in this study, OPEs could affect the content of T4 through the changes in contents of TPO and TG. It was found that lipid biosynthetic and lipid metabolic processes played important roles in the growth and development of organisms, in which cga (encoding a glycoprotein hormone) might be a key gene (Fig. 6 and S13) [63,64]. The results of this study showed that exposure to OPEs can cause significant enrichment of neuroactive ligand–receptor interactions and alter the mitogen-activated protein kinase (MAPK) signaling pathway. Exposure to β-diketone antibiotics (DKAs) caused alterations in 149 KEGG-annotated metabolic pathways in F1 zebrafish, of which the primary affected pathway was the MAPK signaling pathway, followed by neuroactive ligand–receptor interactions [65]. Similarly, exposure of Daphnia magna to tributyltin significantly affected the neuroactive ligand–receptor interaction signaling pathways involved in reproduction and development [66]. Previous studies confirmed that thyroid disruption was linked to neurodevelopmental effects [25,61,67]. In brief, the findings suggested that the growth inhibition observed in zebrafish exposed to OPEs might be due to the mode of action of thyroid endocrine disruption, including via neurodevelopment and lipid regulation.

Fig. 6

Schematic representation of the speculated mechanisms of toxicity of three OPEs on the disruption of functions of the thyroid axis in zebrafish brain. (Top) Exposure of zebrafish to OPEs affected the brain. (Down) In brain, OPEs competed with T4 to bind to TBG or TTR (especially TTR) for transport, and entered cells through transmembrane proteins affected by NIS. The subsequent interaction of nuclear THR with OPEs or T4 induced abnormal expression of some thyroid-related genes, such as slc16a10, ttr, slc5a5, tpo, dio2, cga, thrβ, nkx2.4b, and ncoa2, which could induce multiple signaling pathways related to thyroid function. For example, the abnormal expression of tpo can affect levels of TPO protein in cells, which together with TG caused abnormal changes in TH contents. In addition, expression of ttr can affect levels of TTR protein. Due to these signaling pathways, OPEs may cause thyroid dysfunction in zebrafish brain.

TTR is a likely biomarker of thyroid disruption in zebrafish induced by OPEs

As the primary carrier proteins for T4 in blood plasma, TBG and TTR play crucial roles in transporting TH to organs and tissues. NIS mainly expressed in the basolateral membrane of thyroid cells was reported to be important for mediating the transmembrane transport of iodine [68]. In blood plasma, most T4 (70%) was found to bind to TBG with greater affinity than to TTR, to which only 10–15% of T4 bound, yet less than 1% of T4 was present in blood [69,70]. These results suggested that TBG plays a role in binding to and transporting T4 in blood plasma. However, the results involving these proteins in the brain have demonstrated that the concentrations of TTR and NIS were significantly affected by OPEs, but the amounts of TBG were not changed. These findings indicated that TTR was important for the transport of TH or exogenous sources of natural hormones (e.g., OPEs) into the brain of zebrafish. It was also reported that T4 could localize in the brain and cerebrospinal fluid by passing through the blood–brain barrier after combining with TTR [[71], [72], [73]]. Simultaneously, greater amounts of NIS could increase the transmembrane transport activity of iodine, which might subsequently result in greater concentrations of TH [68]. Based on the Western blot results of GH3 cells, rather than TTR and TBG, NIS might play an important role in the generation of T4. In addition, based on the results of transcriptome analysis, TH transmembrane transporter activity might be related to the concentrations of NIS (Fig. 6). However, the changes of T4 in this study were not completely correlated with the concentrations of NIS, which suggested that the concentrations of T4 in zebrafish were affected by multiple factors rather than the intracellular concentrations of iodine (Fig. 1, Fig. 4). Furthermore, the results of molecular docking simulation showed that molecular docking sites of CDP to TBG and TTR were similar to those for T4, while TCIPP would preferentially bind to TTR rather than CDP or T4, and form more stable bonds with TTR than TIPP. The results of several studies have indicated that molecular docking sites of T4 on TBG included hydrogen bond interaction at Asn 273, π–cation interaction at Arg 381, and salt bridge interaction at Lys 270, while those of binding of T4 to TTR involved hydrogen bond interaction with Lys 15 and Ser 117 [70,74,75]. However, the binding of CDP by hydrogen bonding had greater cdocker energy than that of T4, so it decreased the levels of free T4 in plasma or brain tissues. TIPP and TCIPP showed higher binding affinities to TBG and TTR via van der Waals interaction and hydrophobic interaction than those of T4, which increased the levels of free T4 in plasma or brain tissues. It was thus confirmed that TH transmembrane transporter activity and TH transport may be important signaling pathways of thyroid dysfunction induced by OPEs in zebrafish, suggesting that TBG or TTR could transport OPEs in blood plasma and NIS might be the predominant transmembrane protein affecting intracellular TH biosynthesis (Fig. 6). In particular, the combination of OPEs and TTR can be transported to the brain through the blood–brain barrier, thus affecting brain development and altering homeostasis, which suggests that TTR might be a biomarker for thyroid disruption in zebrafish.

GH3 cell proliferation effectively inhibited by OPEs

TH-disrupting chemicals, including BFRs, OPFRs, and polybrominated diphenyl ethers (PBDEs), have been classified as another class of endocrine-disrupting chemicals other than environmental estrogens [37,43,[76], [77], [78], [79], [80]]. Based on results of the T-screen assay and Western blot analysis, the proliferation of GH3 cells was inhibited by OPEs compared with the level in blank control cells or cells exposed only to T4. The inhibitory effects of OPEs obtained using RIC20 and RIC50 calculated by acute assay of GH3 cells were in the following order: CDP (RIC20 = 2.4 nM, RIC50 = 0.38 μM) > TCIPP (RIC20 = 1.4 μM, RIC50 = 90.3 μM) > TIPP (RIC20 = 1.4 μM, RIC50 = 1.8 mM). The results of Western blot analysis of NIS and TBG in cells exposed to RIC20 and RIC50 concentrations further confirmed that OPEs, as antagonists, might compete with T4 to inhibit GH3 cell proliferation. It was also found that adverse effects on GH3 cells induced by TIPP and TCIPP co-exposed with T4 at the RIC50 were significantly different from those of CDP under the same conditions. T4-TIPP and T4-TCIPP significantly inhibited cell proliferation, while T4-CDP significantly promoted cell proliferation compared with the levels in the blank and T4 groups. The reason for these effects might be the competitive inhibition process of TIPP, TCIPP, or CDP with T4 (Fig. 6), indicating that the superiority of TIPP and TCIPP led to a decrease of the T4 effect, and the inferiority of CDP led to an increase of the T4 effect, as revealed via molecular docking simulation and ELISA. Moreover, the expression levels of key genes related to thyroid effects in zebrafish induced by TIPP and TCIPP differed from those of CDP (Fig. S9). Notably, the differences in the modes of action of OPEs in this study were consistent with previous research on nuclear/membrane THR, suggesting that the patterns of GH3 cell proliferation induced by alkyl and chlorine OPEs differed from those of aryl OPEs [62]. Based on the results of previous studies, the present results of the competitive inhibition assay demonstrated that OPEs and T4 might be competitive in binding to membrane receptor integrin αvβ3 to enter GH3 cells (Fig. 6) [49]. Therefore, TIPP and TCIPP may play a dominant role in the competition with T4 and be the main T4 inhibitors. Although CDP was an inhibitor of cell proliferation, co-exposure with T4 can significantly promote cell proliferation at a high concentration of CDP.

Conclusions

In conclusion, the mechanism of toxicity and adverse outcome pathway of thyroid disruption of zebrafish exposed to OPEs were clarified in this study. The results of this study showed that OPEs could be typical environmental endocrine disruptors affecting TH in aquatic organisms. OPEs mainly bound to TTR (aryl OPEs: hydrogen bonding, alkyl and chlorine OPEs: non-hydrogen bonding) were transported to the target organ (the brain) via the blood–brain barrier. OPEs competed with T4 to bind to integrin αvβ3 to enter cells. Zebrafish growth was inhibited by OPEs via affecting TH endocrine function accompanied by neurodevelopment and lipid regulation, including thyroid hormone receptor binding, thyroid hormone generation, thyroid hormone transmembrane transporter activity, biosynthesis and metabolism of lipids, and neuroactive ligand–receptor interaction. However, the correlations between TH endocrine function and neurodevelopment or lipid regulation require further elucidation, and the molecular modes of action between lipid regulation and the change of body mass of zebrafish induced by OPEs also remain largely unclear. Given that the mechanism behind the toxicity of TH endocrine disruption induced by OPEs was elaborated in zebrafish, this study not only provides a basis for understanding the mechanism by which OPEs threaten aquatic organisms, but also provides a valuable theoretical basis for deriving water quality criteria and ecological risk assessment of OPEs.

Declaration of competing interest

The authors declare that there were no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.