Early-life exposure to 17β-estradiol and 4-nonylphenol impacts the growth hormone/insulin-like growth-factor system and estrogen receptors in Mozambique tilapia, Oreochromis mossambicus

Abstract

It is widely recognized that endocrine disrupting chemicals (EDCs) released into the environment through anthropogenic activities can have short-term impacts on physiological and behavioral processes and/or sustained or delayed long-term developmental effects on aquatic organisms. While numerous studies have characterized the effects of EDCs on temperate fishes, less is known on the effects of EDCs on the growth and reproductive physiology of tropical species. To determine the long-term effects of early-life exposure to common estrogenic chemicals, we exposed Mozambique tilapia (Oreochromis mossambicus) yolk-sac fry to 17β-estradiol (E2) and nonylphenol (NP) and subsequently characterized the expression of genes involved in growth and reproduction in adults. Fry were exposed to waterborne E2 (0.1 and 1.0 μg/L) and NP (10 and 100 μg/L) for 21 days. After the exposure period, juveniles were reared for an additional 112 days until males were sampled. Gonadosomatic index was elevated in fish exposed to E2 (0.1 μg/L) while hepatosomatic index was decreased by exposure to NP (100 μg/L). Exposure to E2 (0.1 μg/L) induced hepatic growth hormone receptor (ghr) mRNA expression. The high concentration of E2 (1.0 μg/L), and both concentrations of NP, increased hepatic insulin-like growth-factor 1 (igf1) expression; E2 and NP did not affect hepatic igf2 and pituitary growth hormone (gh) levels. Both E2 (1.0 μg/L) and NP (10 μg/L) induced hepatic igf binding protein 1b (igfbp1b) levels while only NP (100 μg/L) induced hepatic igfbp2b levels. By contrast, hepatic igfbp6b was reduced in fish exposed to E2 (1.0 μg/L). There were no effects of E2 or NP on hepatic igfbp4 and igfbp5a expression. Although the expression of three vitellogenin transcripts was not affected, E2 and NP stimulated hepatic estrogen receptor (erα and erβ) mRNA expression. We conclude that tilapia exposed to E2 and NP as yolk-sac fry exhibit subsequent changes in the endocrine systems that control growth and reproduction during later life stages.

1. Introduction

Particular compounds released into the environment through anthropogenic activities impact the endocrine systems of vertebrates, including fishes (Colborn et al., 1996). These compounds, known as endocrine disrupting chemicals (EDCs), include hormones, pharmaceuticals, pesticides, plasticizers, and naturally occurring compounds. Fish are employed as indicator species for environmental pollution in aquatic systems because they are among the first animals exposed to waterborne chemicals. The adverse activities of EDCs in fish are known to include impacts on fertility, sexual maturation, somatic growth, and circulating hormone levels. Moreover, EDCs can activate stress responses and induce cellular damage, effects that may increase the incidence of disease and mortality (Ankley et al., 2009; Bernanke and Kohler, 2009; Bhandari et al., 2015; Breves et al., 2018; Celino et al., 2009; Jones et al., 2000; Lerner et al., 2007a, b).

Many EDCs act as agonists or antagonists of estrogen receptors (Er) (cf. Ankley et al., 2009). Among the most pervasive EDCs in the aquatic environment are 17β-estradiol (E2) and nonylphenol (NP) (Aris et al., 2014; Giger et al., 1984; Xu et al., 2014). E2 is one of the most common feminizing compounds found in sewage effluent discharged into rivers (Desbrow et al., 1998). Nonylphenol ethoxylates (NPEs) are widely used as surfactants in industrial processes and products, including cleaners, detergents, and plastics. As in the case of E2, NPEs are also discharged through domestic and industrial wastewater (Mao et al., 2012; Servos et al., 2003). NPEs are degraded into NP, which persists in the environment (Ahel et al., 1993). Free NP is presumed to be widely distributed in surface waters (Ekelund et al., 1990; Ekelund et al., 1993) with concentrations ranging from approximately 30 to 30,000 ng/L in Guangzhou riverine waters in China, the Seine estuary in France, and the European river basin in Spain (Brix et al., 2010; Cailleaud et al., 2007; Peng et al., 2008). NP accumulates in various aquatic organisms at concentrations ranging from 0.68–160 ng/g tissue weight (Vethaak et al., 2005; Zhou et al., 2019). NP exerts feminizing effects in mice (Hernandez et al., 2006), reduces fecundity and fertility in Japanese medaka (Oryzias latipes) (Ishibashi et al., 2006; Kang et al., 2003), reduces semen volume in rainbow trout (Oncorhynchus mykiss) (Lahnsteiner et al., 2005), and diminishes plasma testosterone in male carp (Cyprinus carpio) (Amaninejad et al., 2018). Moreover, the presence of NP and NPEs in the environment was linked to a low male:female sex ratio in wild Nile tilapia (Oreochromis niloticus) (Chen et al., 2014). Most studies reporting on the effects of E2 and NP on growth and reproduction in fishes have been conducted with temperate species (Harries et al., 2000; Filby et al., 2006; Goetz et al., 2009; Duffy et al., 2014; Breves et al., 2018).

Given its importance to worldwide aquaculture (FAO, 2005), the Mozambique tilapia (Oreochromis mossambicus) is one of the most thoroughly studied tropical fishes with respect to how environmental conditions impact growth and reproductive endocrinology (Davis et al., 2009a; Davis et al., 2009b; Gaigher and Krause, 1983; Kiilerich et al., 2011; Moorman et al., 2016; Kajimura et al., 2005). Tilapia are widely distributed in tropical areas where they are cultured for human consumption. They inhabit regions where agricultural, municipal, and industrial waters are discharged and are therefore exposed to persistent environmental EDCs based on the contaminants detected in their tissues (Authman et al., 2008; Babu and Ozbay, 2013; Hemmatinezhad et al., 2017; Osman et al., 2012).

The endocrine system of fishes mediates the effects of environmental stimuli, including contaminants, on growth and reproduction. The growth hormone (Gh)/insulin-like growth-factor (Igf) system plays a major role in regulating the growth and development of vertebrates, including teleosts (Duan et al., 2010; Reindl and Sheridan, 2012). Upon binding to the Gh receptor (Ghr), Gh stimulates the release of Igf1 which has growth-promoting actions in target tissues (Butler and Le Roith, 2001; Duan, 1998; Fan et al., 2009; Le Roith et al., 2001; Le Roith and Roberts, 2003). Igfs interact with a family of binding proteins, known as Igf binding proteins (Igfbps), which influence their availability and activities (Duan and Xu, 2005; Duan et al., 2010; Rajaram et al., 1997) and teleost fishes possess an expanded suite of Igfbps (Allard and Duan, 2018). Steroid hormone receptors mediate target-tissue responsiveness to the actions of steroid hormones, in addition to compounds that mimic hormone actions (Park et al., 2007; Gross and Yee, 2002). The production of vitellogenin (Vtg), a precursor of egg yolk protein produced by the liver of female oviparous animals (Denslow, 1999; Hiramatsu et al., 2005), is stimulated by activation of ers and Ers (Bowman et al., 2002; Flouriot et al., 1996; Jalabert, 2005; Nelson and Habibi, 2010). These same Ers are the pathway in which estrogenic EDCs interfere with normal estrogen signaling.(Shanle and Xu, 2011). Hence, Vtg/vtg and Ers/ers are often used as indicators of estrogenic EDC exposure (Jones et al., 2000; Matozzo et al., 2008; Leet et al., 2011; Park et al., 2007). Plasma Vtg has been detected in male white sucker (Catostomus commersoni) and rainbow trout inhabiting waters contaminated by sewage effluent (Purdom et al., 1994; Vajda et al., 2008). Estrogenic EDCs such as E2, 17α-ethinylestradiol (EE2), diethylstilbestrol, and NP induce vtg expression and plasma Vtg levels in male fish (Davis et al., 2007; Davis et al., 2009b; Folmar et al., 2000; Hemmer et al., 2001). In male tilapia, injection of E2 induces Vtg production while concurrently suppressing the Gh/Igf system (Davis et al., 2007; Davis et al., 2008). Moreover, in juvenile Atlantic salmon, vtg transcripts were induced, while hepatic er, ghr, igfs, and several igfbps were reduced following waterborne exposure to EE2 and NP (Breves et al., 2018). Filby et al. (2006) previously showed that male fathead minnows (Pimephales promelas) exposed to E2 exhibited reduced hepatic igf1 expression levels. Here, we examined the long-term effects of waterborne exposure to NP and E2 (as positive control) in Mozambique tilapia. We reared tilapia fry in water containing E2 (0.1 and 1.0 μg/L) and NP (10 and 100 μg/L) for 21 days, and then measured the expression of pituitary gh, and hepatic ghr, igfs, igfbps, vtgs, and ers after an additional 112 days of growth and development.

2. Materials and methods

2.1. Animals

Mozambique tilapia yolk-sac fry were obtained from broodstock tanks maintained in dechlorinated city water at the Hawai‘i Institute of Marine Biology (Kāne‘ohe, HI). Fry were initially reared in 7-L conical tanks supplied with filtered dechlorinated city water for 2-3 days. Fry were then distributed to 5-L aerated flow-through tanks (33 fry/tank) supplied by 19-L header tanks and reared for an additional 5-10 days until yolk absorption was ~90% complete. Header and fish-holding tanks were lined with modified polytetrafluoroethylene (MPTFE) (Welch Fluorocarbon, Inc., Dover, NH). Two replicate tanks were used for each treatment and subsequent rearing. Water temperature was maintained at ~26-28 °C under a 12L: 12D photoperiod. After yolk absorption, fry were fed crushed trout chow pellets (Skretting, Tooele, UT) twice daily for the remainder of the experiment. Fecal material and uneaten food were siphoned out from the tanks before 60-70% of the water volume was changed daily. All housing and experimental procedures were conducted in accordance with the principles and procedures approved by the Institutional Animal Care and Use Committee of the University of Hawai‘i.

2.2. Chemical exposures and rearing

Fry (0.029 ± 0.001 g body weight) were exposed to E2 (0.1 and 1.0 μg/L) and NP (10 and 100 μg/L) in fresh water for 21 days. The duration of the exposures was chosen following previous studies aimed at identifying responses to chronic EDC exposures (Woltering, 1984; Lerner et al., 2007a, 2007b). The range of nominal concentrations of E2 used in this study were above those typically found in the environment inasmuch as they were included as a positive control. The concentrations of E2 and NP were based on those employed in previous studies where Atlantic salmon yolk-sac larvae, fry, juvenile, and smolts were subjected to the aqueous exposure of these chemicals; the concentrations of the chemicals in the water have been previously validated (Breves et al., 2018; Duffy et al., 2014; Lerner et al., 2007a, b). E2 and NP were purchased from Sigma-Aldrich (St. Louis, MO) and Acros Organics (Fair Lawn, NJ). All chemicals were solubilized in ethanol and then added to fresh water at a final concentration of 0.0001% ethanol to minimize solvent toxicity. Control treatments received solvent only. Header tanks were covered and refilled every two days with filtered and dechlorinated city water with or without the experimental chemicals. The flow-rate from the header tanks to the rearing tanks averaged 0.2 L/h. The fry were maintained in 5-L MPTFE-lined tanks until the end of the exposures (see Fig. 1 for the experimental setup) at which time the juveniles were transferred to 19-L aerated flow-through MPTFE-lined containers and reared for an additional 112 days until sampling. MPTFE was used to prevent the leaching of chemicals from the plastic containers used as tanks and header buckets. This approach has been previously used in a similar study employing coho salmon (Harding et al, 2016). Bodyweight (BW) and total length (TL) were measured every two weeks throughout the duration of the experiment.

Fig. 1.

Schematic illustration of header and experimental tanks. Nineteen-liter capacity header tanks supplied water either containing solvent only (control), 0.1 μg/L or 1.0 μg/L 17β -estradiol (E2), or 10 μg/L or 100 μg/L nonylphenol (NP) to fish tanks. Both header and fish tanks were lined with modified polytetrafluoroethylene. Black squares = aeration, white rectangles = pipes.

At the end of the experiment, male tilapia were netted and anesthetized with 2-phenoxyethanol (0.3ml/L; Sigma-Aldrich) and BW and TL were measured. Anesthetized fish were euthanized by rapid decapitation. Testes and liver were removed and weighed for calculation of gonadosomatic index (GSI; (gonad weight/BW)*100) and hepatosomatic index (HSI; (liver weight/BW)* 100). Condition factor (CF) was calculated as CF = (BW/TL³)*100. Liver and pituitary were collected, immediately snap-frozen in liquid nitrogen, and stored at −80 °C until RNA extraction.

2.3. Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from liver and pituitary using TRI Reagent (MRC, Cincinnati, OH) according to the manufacturer’s protocols. The concentration and purity of extracted RNA were assessed using a NanoDrop (NanoDrop One, Thermo Scientific). Total RNA (100-500 ng) was reverse-transcribed using a High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA). The mRNA levels of reference and target genes were determined by the relative quantification method using a StepOnePlus real-time PCR system (Thermo Fisher Scientific). The qRT-PCR reaction mix (15 μL) contained Power SYBR Green PCR Master Mix (Thermo Fisher Scientific), 200 nM of forward and reverse primers, and 1 μl cDNA. Dilution of experimental cDNA from liver ranged from 10− to 175-fold. PCR cycling parameters were: 2 minutes at 50 °C, 10 minutes at 95 °C followed by 40 cycles at 95 °C for 15 seconds and 60 °C for 1 minute. All qRT-PCR primers have been previously described; PCR efficiencies are reported in Table 1. Since hepatic elongation factor 1α (ef1α), β-actin, and 18s levels varied across treatments, the geometric mean ($\sqrt[3]{x_1\ast x_2\ast x_3}$; where x= quantity of each reference gene) of these reference genes was used to normalize target genes in the liver, ef1α levels were used to normalize pituitary gh expression after verification that ef1α expression did not vary across treatments. Data are expressed as a fold-change relative to control.

Table 1.

Lis of primers used in qPCR assays.

Gene	Forward primer	Reverse primer	R2	% efficiency	Accession Number	Reference
	Sequence (5'-3')	Sequence (5'-3')
18s	GCTACCACATCCAAGGAAGGC	TTCGTCACTACCTCCCCGAGT	0.999	89.1	AF497908	Magdeldin et al., 2007
ef1α	AGCAAGTACTACGTGACCATCATTG	AGTCAGCCTGGGAGGTACCA	0.999	95.1	AB075952	Breves et al., 2010
β-actin	CTCTTCCAGCCTTCCTTCCT	ACAGGTCCTTACGGATGTCG	0.998	96.5	FN673689	Tipsmark et al., 2011
ghr	CACACCTCGATCTGGACATATTACA	CGGTTGGACAATGTCATTAACAA	0.997	96.8	EF452496	Pierce et al., 2007
igf1	CTGCTTCCAAAGCTGTGAGCT	GATCGAGAAATCTTGGGAGTCTTG	0.999	92.3	AF033796	Kajimura et al., 2004
igf2	GCTTTTATTTCAGTAGGCCAACCA	CACAGCTACAGAAAAGACACTCCTCTA	0.997	90.1	AH006117	Davis et al., 2008
igfbp1b	CCTTCCCTTTGATCACCAAG	GTGTGACATGGACCCTGTTG	0.997	90.8	XM_003438121	Breves et al., 2014
igfbp2b	CCGACTTCCCTTTACAGCAG	TCAGTCCCATGCACCTCATA	0.998	92.2	XM_0054504847	Breves et al., 2014
igfbp4	ATCCCCATACCCAACTGTGA	TGATCCACACACCAGCATTT	0.999	85.6	XM_003454633	Breves et al., 2014
igfbp5a	AACTGGACGGGATCATCATTCAG	GCACTGTTTGCGTTTGAAGA	0.999	106.7	XM_003443250.2	Breves et al., 2014
igfbp6b	TCCTACCTGCAGAGGAAAGC	CGCAGCTCAGAGTGTAGACG	0.975	96.5	XM_003441337	Breves et al., 2014
gh	TTACATCATCAGCCCGATCG	AGATCGACAGCAGCTTCAGGA	0.999	94.3	AF033806	Magdeldin et al., 2007
vtga	GAATGTGAATGGGCTGGAAATAC	TTTGTTTGATCTGGATGTCAGCTT	0.999	90.3	EF408235	Davis et al., 2007
vtgb	AAGTTGCAGACTGGATGAAAGGA	GCGGTACTCGTCTCCGACAT	1.000	97.7	EF408236	Davis et al., 2007
vtgc	GGACCTTGCAGAACCCAAAG	CATCGTTTCTTGCCAGTTCCA	0.998	94.9	EF408237	Davis et al., 2007
erα	GGCTCAGCAGCAGTCAAGAA	TGCCTTGAGGTCCTGAACTG	0.990	87.0	AM284390	Park et al., 2007
erβ	ACCTTCCGGCAGCAGTACAC	TCCAACATCTCCAGCAACAG	0.994	108.0	AM284391	Park et al., 2007

2.4. Statistical analysis

Group comparisons were performed by one-way ANOVA followed by Fisher’s protected LSD test. In order to meet assumptions of normality (assessed by Kolmogorov-Smirnov), individual values were log-transformed when necessary prior to statistical analysis. Pearson correlation coefficients were used to describe the relationships between gh, ghr, igf1, and igf2 mRNA levels. Statistical calculations were performed using Prism 8.0 (GraphPad, La Jolla, CA). Significance for all tests was set at P < 0.05.

3. Results

3.1. Effects of E2 and NP on physical characteristics

BW was significantly higher in fish exposed to 10 μg/L NP compared with controls (Fig. 2A) whereas TL and CF were unaffected by E2 and NP (Fig. 2B, C). The high concentration of NP (100 μg/L) reduced HSI relative to controls (Fig. 2D). GSI was elevated in fish exposed to E2 at 0.1 μg/L (Fig. 2E).

Fig. 2.

Body weight (BW) (A), total length (TL) (B), condition factor (CF) (C), hepatosomatic index (HSI) (D), and gonadosomatic index (GSI) (E) of Mozambique tilapia adults 112 days after 21-day exposure as fry to water containing 0 (control), 0.1 μg/L or 10 μg/L 17β-estradiol (E2) or 10 μg/L or 100 μg/L nonylphenol (NP). Values are means ± SEM (n = 5-23; BW, TL, and CF; n = 5-12, HSI and GSI). Asterisk indicates significant difference between treatment and control group (One-way ANOVA; Fisher’s protected LSD; P < 0.05).

3.2. Effects of E2 and NP on ghr, igf1, igf2, and gh gene expression

Hepatic ghr levels were >2-fold higher in fish exposed to E2 (0.1 μg/L) compared with controls; E2 at 1 μg/L and NP at both concentrations did not impact ghr levels (Fig. 3A). By contrast, hepatic igf1 was elevated >2-fold in fish exposed to the high concentration (1 μg/L) of E2 and >1-fold to 2-fold to both concentrations of NP (Fig. 3B). Hepatic igf2 levels were highly variable and were not impacted by E2 and NP (Fig. 3C). Pituitary gh levels in the E2 and NP treatments were not different from control levels (Fig. 3D). Hepatic ghr was significantly correlated with igf1 (r²=0.33) and igf2 (r²=0.22).

Fig. 3.

Hepatic ghr (A), igf1 (B), igf2 (C), and pituitary gh (D) mRNA levels in Mozambique tilapia adults 112 days after 21-day exposure as fry to water containing 0 (control), 0.1 μg/L or 10 μg/L 17β-estradiol (E2) or 10 μg/L or 100 μg/L nonylphenol (NP). mRNA levels are presented as fold-change relative to the control group. Values are means ± SEM (n = 5-12). Asterisk indicates significant difference between treatment and control group (One-way ANOVA; Fisher’s protected LSD; P < 0.05).

3.3. Effects of E2 and NP on igfbp gene expression

Exposures to E2 and NP at 1 and 10 μg/L, respectively, induced hepatic igfbp1b levels by >2-fold from controls (Fig. 4A). Fish exposed to NP (100 μg/L) exhibited elevated igfbp2b levels compared with controls (Fig. 4B). igfbp4 and igfbp5a levels were unaffected by E2 and NP exposures (Fig. 4C, D) whereas igfbp6b was the only igfbp transcript suppressed by E2 (Fig. 4E).

Fig. 4.

Hepatic igfbp1b (A), igfbp2b (B), igfbp4 (C), igfbp5a (D), and igfbp6b (E) mRNA levels in Mozambique tilapia adults 112 days after 21-day exposure as fry to water containing 0 (control), 0.1 μg/L or 10 μg/L 17β-estradiol (E2) or 10 μg/L or 100 μg/L nonylphenol (NP). mRNA levels are presented as fold-change relative to the control group. Values are means ± SEM (n = 5-12). Asterisk indicates significant difference between treatment and control group (One-way ANOVA; Fisher’s protected LSD; P < 0.05).

3.4. Effects of E2 and NP on vtg and er gene expression

Although all three vtg transcripts had a tendency to be elevated in fish exposed to E2, no significant effects of E2 or NP were detected (Fig. 5A-C). On the other hand, erα levels were increased following exposures to all tested concentrations of E2 and NP (Fig. 5D); erβ levels were stimulated by E2 and NP at 0.1 and 100 μg/L, respectively (Fig. 5E).

Fig. 5.

Hepatic vtga (A), vtgb (B), vtgc (C), erα (D), and erβ (E) mRNA levels in Mozambique tilapia adults 112 days after 21-day exposure as fry to water containing 0 (control), 0.1 μg/L or 10 μg/L 17β-estradiol (E2) or 10 μg/L or 100 μg/L nonylphenol (NP). mRNA levels are presented as fold-change relative to the control group. Values are means ± SEM (n = 5-12). Asterisk indicates significant difference between treatment and control group (One-way ANOVA; Fisher’s protected LSD; P < 0.05).

4. Discussion

The objective of this study was to determine whether Mozambique tilapia exposed to estrogenic chemicals as fry exhibit long-term responses that impact physiological systems underlying growth and reproduction as adults. Mature tilapia exhibit concentration-dependent responses to estrogenic compounds commonly found in sewage effluents, such as E2, o,p'-DDE (dichlorodiphenyl dichloroethene), heptachlor, and NP (Davis et al., 2007; Davis et al., 2008; Davis et al., 2009b). Little is known, however, on whether exposure to estrogenic compounds during early-life stages may impart long-term physiological effects on adult tilapia. To our knowledge, this is the first study demonstrating that early-life exposure to estrogenic EDCs affects the Gh/Igf system and er expression in adult tilapia. At very low levels, E2 and NP still elicited physiological responses in exposed individuals.

In previous studies, BW was significantly reduced in fry and adult male fish after exposure to estrogenic compounds such as EE2 and NP (Meredith et al., 1999; Breves et al., 2018). In the current study, however, the BW of adults exposed to NP (10 μg/L) as fry was actually greater than controls. This may be attributed to a capacity for the somatotropic axis to compensate for poor growth during early life-stages (Bertram et al., 1993; Chambers et al, 1998; Gagliano and McCormick, 2007; Segers et al., 2012), in this case when the EDCs were present. Compensatory growth following periods of suppressed growth, such as food restriction, occurs in several teleosts, including striped bass (Morone saxatilis), Atlantic halibut (Hippoglossus hippoglossus), channel catfish (Ictaluruspunctatus), rainbow trout, and hybrid tilapia (O. mossambicus × O. niloticus) (Gaylord and Gatlin 2000; Heide et al., 2006; Montserrat et al., 2007; Picha et al., 2008). In the current study, we observed that HSI was lower in fish exposed to NP (100 μg/L) as fry. Since the liver is a major site for metabolism, detoxification, and vitellogenesis, there are a variety of factors that likely contributed to this response (Roberts, 2012; Asem-Hiablie et al., 2013). HSI is naturally elevated during reproductive periods as a result of increased protein synthesis and Vtg production (Jia et al., 2019). In salmon fry and smolts, E2, EE2, and NP elevated HSI (Lerner et al., 2012; Duffy et al., 2014). While elevations in HSI have been correlated with the occurrence of xenobiotics in polluted zones (Karels et al., 1998; Billiard and Khan, 2003), in some cases, such as in Mozambique tilapia, African catfish (Clarias gariepinus), spotted pim (Pimelodus maculatus), and Japanese medaka, lower HSIs were observed in fish exposed to sewage effluents and agricultural runoffs (Ma et al., 2005; Asem-Hiablie et al., 2013; Sadekarpawar and Parikh, 2013; Araújo et al., 2018). Further investigation is needed to determine the mechanisms that underlie reductions in HSI following EDCs exposures.

GSI has been extensively used as an indicator of sexual maturation as well as a biomarker for exposure of aquatic organisms to estrogenic EDCs. Several laboratory studies have reported that exposure to estrogenic chemicals inhibits testicular development (Gimeno et al., 1997; Komen et al., 1989; Christiansen et al., 1998). Field studies have also documented a correlation between estrogenic compounds and lower GSI in exposed male fish (Andersson et al., 1988; Harries et al, 1997; Kukkonen et al., 1999; Hassanin et al., 2002). In the current study, we observed that fry exposed to E2 (0.1 μg/L) exhibited elevated GSI as juvenile males. Jobling et al. (1996) found that the inhibitory effects of estrogenic compounds on sexually maturing rainbow trout were not evident in mature or regressing fish. In juvenile male salmon, a relatively low concentration of E2 (2 μg/L) increased GSI (Lerner et al., 2007a). Furthermore, E2 also plays a role in male gonad development. Although high concentrations of E2 could inhibit testicular development in some fishes such as Japanese eel (Anguilla japonica) and three-spot wrasse (Halichoeres trimaculatus), low concentrations of E2 were found to induce spermatogonial stem cell renewal and spermatogonial proliferation, suggesting a modulatory role of E2 in normal testicular development (Miura et al., 1999, 2003; Kobayashi et al., 2011). These previous studies, therefore, may explain the increase in GSI in fish exposed to the low concentration of E2. The other concentrations of E2 and NP employed in this study were seemingly insufficient to affect GSI, especially given that fish were exposed to these concentrations as fry. Taken together with previous findings, our results suggest that the impact of estrogenic chemicals on GSI is dependent on both concentration and timing of exposure.

Our results indicate that the somatotropic axis of adult tilapia was impacted by early-life exposure to the tested EDCs. Both stimulatory and inhibitory effects of estrogenic chemicals on gh transcript levels and Gh synthesis were previously reported in teleosts (Elango et al., 2006; Holloway and Leatherland, 1997; Shved et al., 2007, 2008; Zou et al., 1997). In juvenile Atlantic salmon, E2 and NP injection had no effect on pituitary transcript levels of gh (Yadetie and Male, 2002), a pattern that is consistent with our current observations. Diminished hepatic ghr gene expression was associated with reductions in Gh binding capacity, circulating Igf1 levels, and igf1 expression following exposure to estrogenic compounds in salmonids (Breves et al., 2018; Hanson et al., 2017; Lerner et al., 2012; Norbeck and Sheridan, 2011). By contrast, we observed elevations in hepatic ghr and igf1 gene expression in addition to positive correlations between ghr and both igf1 and igf2, following exposure to estrogenic compounds. As discussed above, these patterns may be associated with a compensatory growth response. In this instance, the Gh/Igf system is seemingly ‘activated’ following the withdrawal of estrogenic chemicals. During restricted feeding, for example, the catabolic state preceding compensatory growth is characterized by depressed levels of hepatic ghr, igf1, and plasma Igf1 (Gray et al., 1992; Duan et al., 1995; Pierce et al., 2005; Norbeck et al., 2007; Picha et al., 2008). Upon re-feeding, a rapid increase in specific growth rate and hepatic ghr, igf1, and igf2 expression occurs (Picha et al., 2008). Although there was no clear inhibition of pituitary gh expression, we found a negative correlation between pituitary gh and hepatic igf1, a possible indication of feedback regulation of Gh by Igf1 (Reinecke, 2010). No correlation was observed, however, between hepatic igf2 and pituitary gh. Moreover, the differing responses by hepatic igf1 and igf2 to estrogenic EDCs observed in this study were similar to patterns in male Mozambique tilapia injected with E2 (Davis et al., 2008), in which hepatic igf2 was not affected. In mammals, Igf2 is mainly associated with fetal growth and development (Constancia et al., 2002; Daughaday and Rotwein, 1989). In teleosts, however, some studies suggest that Igf2 is also an important factor in adult growth (Pierce et al., 2011; Reindl and Sheridan, 2012). The varying responses of igf2 to estrogenic compound may be due to differences among species and tissue sensitivity.

Igfbps are key modulators of Igf activity (Duan and Xu, 2005). Only a few studies in mammals and fishes have described how steroid hormones regulate Igfbps (Duan et al., 2010; Garcia de la Serrana et al., 2017; Rajaram et al., 1997; Reindl and Sheridan, 2012), and fewer yet have examined the long-term effects of EDC exposure on igfbps during early developmental stages (Breves et al., 2018). In teleosts, it is generally accepted that the liver is a major site for Igfbp synthesis and secretion (Shimizu and Dickhoff, 2017; Zhou et al., 2008). In Mozambique tilapia, igfbp1b, -2b, -5a, -4, and -6b are expressed in the liver (Breves et al., 2014). In vertebrates, Igfbp1 plays a highly conserved role as a negative regulator of somatic growth by restricting Igf1 from binding to its receptor (Kajimura et al., 2005; Kamei et al., 2008). Knockdown of Igfbp1 in zebrafish, for instance, alleviates hypoxia-induced retardation of growth, while its overexpression causes growth and developmental retardation (Kajimura et al., 2005). In chinook (O. tshawytscha) and Atlantic salmon, Igfbp1b paralogs are important modulators of Igf signaling in response to nutrient availability (Hevrøy et al., 2011; Shimizu et al., 2005, 2006, 2009). Therefore, the increase in igfbp1b transcript levels after E2 and NP exposures may provide a mechanism for EDCs to impact growth. In other words, by enhancing igfbp1b, a negative regulator of growth, estrogenic EDCs inhibit somatic growth. Alternatively, the increase in igf1 mRNA levels (and plasma Igf1) and other factors within the Gh/Igf system following EDC exposures may counterbalance the igfbp1b response. Previous studies in mature male Mozambique tilapia (Riley et al., 2004) and striped bass (Morone saxatilis) (Fukazawa et al., 1995) reported that E2 stimulated the release of putative Igfbp1s from hepatocytes. In Atlantic salmon fry and smolts, however, estrogenic compounds inhibited igfbp1b (Breves et al., 2018). Diets supplemented with E2 also inhibited hepatic igfbp1b1 expression (along with hepatic igf1 and igf2) in rainbow trout (Cleveland and Weber, 2016). These findings support the notion that responses to estrogenic compounds are both species and life stage-dependent.

In salmonids, Igfbp2 paralogs are major carriers of plasma Igf1 (Shimizu and Dickhoff, 2017). In mammals and teleosts, varying patterns exist on the regulation of igfbp2 gene expression. For example, overexpression of Igfbp2 in mouse embryos reduces growth rates, which was proposed to be related to a reduction in Igf availability (Hoeflich et al., 1999). On the other hand, in salmonids, circulating Igfbp2 increase in response to Gh (Garcia de la Serrana and Macqueen, 2018; Shimizu et al., 1999; 2003). In tilapia, igfbp2b expression increased with an increased plasma Igf1 induced by Gh injection (Breves et al., 2014). Moreover, igfbp2b expression was increased by treatment of NP but decreased by treatment of E2 in salmon smolts (Breves et al., 2018). By contrast, in rats, hepatic igfbp2 expression is induced by E2 (Hoeflich et al., 2014; Ricciarelli et al., 1991). In rainbow trout, igfbp2 expression in ovarian follicles is also increased by E2 treatment (Kamangar et al., 2006). Hence, the observed increase in igfbp2b levels following exposure to NP, and its tendency to increase after E2 exposure may be either associated with the increase in igf1 levels or modulated by E2 and E2 analogues. Further studies, however, are needed to assess whether there are direct actions of Igf1 and E2 on hepatic igfbp2b expression. In Atlantic salmon, igfbp4, -5a, and -6b regulate the binding of Igfs to its receptor in the tissues where they are produced (Breves et al., 2017; Cleveland and Weber, 2015; Macqueen et al., 2013). Unlike patterns observed in Atlantic salmon (Breves et al., 2018), we found no significant effect of estrogenic EDCs on igfbp4 and -5a. igfbp6 on the other hand, was significantly decreased following E2 exposure. In teleosts, Igfbp6 inhibits Igf-signaling that supports growth and development (Wang et al., 2009). While the decrease in igfbp6b may be a residual effect from the earlier exposure to E2, additional work should address whether E2 and NP act directly on the liver to regulate igfbp6.

To assess the effects of the tested EDCs on estrogenic biomarkers in males, we measured hepatic vtg and er transcripts. While a trend in all vtg transcripts was observed, no significant effects of E2 or NP were detected. Alternatively, both erα and erβ were stimulated by E2 and NP. In previous studies, vtg and er (α and β) were induced in liver and testis after injection of E2 and other estrogenic compounds in mature male Mozambique tilapia (Davis et al., 2009b). A concurrent increase in vtg and erα expression was also observed in Atlantic salmon embryos, yolk-sac fry, feeding fry, and smolts in response to E2, EE2, and NP (Duffy et al., 2014; Breves et al., 2018). The lack of effects on vtg in the current study may be linked to the time of EDC exposures and life stage. Indeed, it is noteworthy that even after 112 days since E2 and NP exposures, both ers were still elevated. This elevation suggests that males may be more sensitive to E2 and similar chemicals after a previous EDC exposure. Increased sensitivity to estrogenic compounds through enhanced expression of ers may render males more susceptible to further detrimental effects on their reproductive development.

5. Conclusion

Our current findings indicate that early aqueous exposure to estrogenic EDCs exerts long lasting effects on the somatotropic axis of tilapia, a central mediator of adaptive patterns of growth and development throughout the life cycle in vertebrates. Thus, an improved understanding of how EDCs impact the endocrine systems controlling growth and reproduction attest to the importance of fish as sentinels for assessing the health of the aquatic ecosystem. Moreover, studies such as this one shall be instrumental in optimizing culture practices for tropical fishes in environments where EDCs are pervasive. Nonetheless, future work that include female tilapia is needed to characterize the long-term effects of estrogenic EDCs in both sexes. Moreover, additional analyses, such as histological examination of testicular tissue would further shed light on the long-term effects of early exposure to estrogenic EDCs on testicular development. Future investigations should also seek to determine the effects of these estrogenic chemicals on the indices of reproductive capacity such as spawning efficiency, fertilization success, and viability of embryos.

Highlights.

• This study focused on long-term effects of early exposure to estrogenic EDCs on hepatic igf, igfbps, and er in Mozambique tilapia.

• Hepatic ghr and igf1 were stimulated by EDC exposure.

• EDCs stimulated igfbp1b and igfbp2b, but diminished igfbp6b.

• erα and erβ were stimulated by EDCs.