Direct and indirect parental exposure to endocrine disruptors and elevated temperature influences gene expression across generations in a euryhaline model fish

Introduction

Many chemicals intended for human use, such as pharmaceuticals and pesticides, are found to pollute adjacent waterways and some are also known to disrupt endocrine function (Brander et al., 2013; Kuivila et al., 2012; Weston & Lydy, 2012; Brander et al., 2016a). Endocrine disrupting compounds (EDCs) influence hormone excretion and signaling in exposed organisms. For example, estrogenic EDCs can alter the expression of genes involved in reproduction and development (Scholz & Gutzeit, 2000; Tu et al., 2016; Brander et al., 2016b). These changes in gene expression may result in higher order consequences potentially impacting organismal function (Brander et al., 2013; Brander et al., 2016b). For example, the synthetic estrogen found in oral contraceptives, 17α-ethinylestradiol (EE2), has been shown to reduce egg production, influence sex ratios, and influence development in fishes exposed to environmentally relevant concentrations (Kidd et al., 2007; Vajda et al., 2008; Schwindt et al., 2014; Bhandari, Vom Saal & Tillitt, 2015). Many pesticides, such as the commonly used pyrethroid pesticide, bifenthrin, can produce estrogenic effects at low doses that deviates from their intended mode of action on the sodium channels of presynaptic neurons of target pests (Brander et al., 2013). Like EE2, bifenthrin has been shown to impact fishes at multiple levels of biological organization; e.g., gene expression, leading to impaired development and reproduction (Brander et al., 2016a; Brander et al., 2016b; Tu et al., 2016).

Estuaries are typically highly impacted by pollution because they are heavily used by growing coastal communities (Cloern & Jassby, 2012). Protection of estuaries is vital for maintaining healthy food webs and ecosystem services. The model species, Menidia beryllina (common name: inland silverside), is a small forage fish that is a food source to a number of estuarine taxa, including commercially important species. Menidia are tolerant of a wide range of salinities but are sensitive to chemical stressors, making them an excellent model species to test the effects of anthropogenic stressors that are common in estuaries (Hemmer, Middaugh & Moore, 1990; Brander et al., 2013). Past studies have shown that the expression of genes involved in reproduction, growth and immune function can be altered following EDC exposure in M. beryllina. Menidia species may be especially susceptible to increasing temperatures and pollutants as their sex determination can be influenced by both temperature and EDC exposure (Duffy, McElroy & Conover, 2009; DeCourten & Brander, 2017). Understanding how organisms with sensitive physiology, such as M. beryllina, respond to EDC exposure will help determine the potential risk of these stressors on estuarine ecosystems.

As our planet continues to warm at an unprecedented rate (Stocker et al., 2013), it becomes increasingly important to understand the impacts pollution may have in the context of global climate change. Increases in temperature can inhibit growth, reproduction and early-life stage survival of marine fishes (Munday et al., 2008; Pankhurst & Munday, 2011). Expression of genes along the hypothalamus-pituitary-gonad (HPG) axis can be reduced by exposure to increased temperatures, inhibiting gonadal function (Miranda et al., 2013). The impacts of increased temperature on fishes could lead to changes to overall community structure and trophic dynamics (Jeppesen et al., 2010). Studies suggest that the effects of estrogenic endocrine disruptors may be influenced by other abiotic factors such as ambient temperature (Hemmer, Middaugh & Moore, 1990; Luzio et al., 2016; DeCourten & Brander, 2017). Testing effects of EDCs under different temperature regimes will help determine the risk they pose to wildlife in the context of global climate change. Elevated temperature may exacerbate the effects of EDCs by increasing the uptake and metabolism of these chemicals in ectothermic organisms (Clarke & Fraser, 2004). Additionally, some chemicals such as bifenthrin, which is converted into a metabolite (4-hydroxybifenthrin), have more potent metabolites than their parent compounds (DeGroot & Brander, 2014). In this study, we investigated the effects of two estrogenic EDCs (EE2 and bifenthrin) on genes involved in development and reproduction of M. beryllina, at two temperatures: 22 °C and 28 °C, to understand how rising temperatures may influence the toxicity of these chemicals. The temperatures chosen for this study represent the most extreme possibility from the worst-case scenario temperature increase (RCP 8.5) expected by the year 2100, put forth by the International Panel for Climate Change (IPCC; Stocker et al., 2013). The RCP 8.5 model predicts a surface temperature increase of 1.4–2.6 °C between 2046 and 2065 and a 2.6–4.8 °C 2081–2100 (Stocker et al., 2013). Assuming these warming periods are additive, we might expect a 4–7.4 °C increase over the next century, illustrating the importance for testing a 6 °C temperature increase. This is especially important for organisms with plasticity in development, such as M. beryllina, whose sex determination can be influenced by temperature and EDC exposure independently (Duffy, McElroy & Conover, 2009; DeCourten & Brander, 2017). The 28 °C upper temperature limit was chosen to allow us to test for effects on TSD without reaching the upper breeding limit of M. beryllina (30 °C), at which spawning is curtailed (Hubbs, 1982). This experiment was conducted over three generations to determine effects at different life stages and effects of parental exposure on offspring. Understanding the effects of concurrent stressors across multiple generations on the gene expression of exposed organisms is necessary to determine the potential for long-term population impacts.

Methods

Results

Parental generation

There were no significant differences in expression of genes tested in the parental generation adult ovaries (Fig. S1) or testes (Fig. S2) following experimental exposures. A full list of pairwise comparisons and p-values across all generations is presented in Table S2.

F2 generation

Within the F2 larvae (Fig. 2), we found that GPR30 expression was reduced by the 22 °C bifenthrin and EE2 treatments (two-way ANOVA, Tukey’s HSD, p ≤ 0.01). There were no differences in expression in the other estrogen receptors (ESR1, ESR2, ESR3). The pattern of expression of FSHR observed in the F1 generation was carried over into the F2 generation (Fig. 2), with the 28 °C EE2 treatment significantly reducing the expression of this gene when compared to all other treatments (two-way ANOVA, Tukey’s HSD, p ≤ 0.01). We did not observe any differences in expression of GnRHR, indicating that receptors of hormones secreted by the pituitary responded differently in the F2 generation. The expression of 17β-HSD was significantly altered by the 28 °C bifenthrin treatment when compared to the 22 °C control (two-way ANOVA, Tukey’s HSD, p = 0.0249). Other genes involved in steroidogenesis (3β-HSD and CYP19b) in the F2 generation were not significantly affected by the experimental exposures. As observed in the previous generation, we did not observe differences in expression of IGF2 among the experimental treatments.

Discussion

Many EDCs have been detected in estuarine waterways, including many commonly used pharmaceuticals and pesticides. Bifenthrin is a pyrethroid pesticide that enters watersheds primarily via runoff following application for urban and agricultural pest control (Kuivila et al., 2012). This compound is found in California surface waters at concentrations up to 106 ng/L (Weston & Lydy, 2012) and higher concentrations have been found in sediments throughout the world (Li et al., 2017). EE2 is a common component of oral contraceptives, which enters the waterways via wastewater effluent, at concentrations up to 7.0 ng/L (Adeel et al., 2017; King et al., 2016). While waste water effluents are primarily found in riverine environments, these rivers discharge into downstream estuarine and marine systems, introducing EE2 at low (ng/L) concentrations (Aris, Shamsuddin & Praveena, 2014). Environmentally relevant concentrations of these estrogenic EDCs have been shown to elicit adverse effects in fish including disruption of homeostasis, reproduction and development (Brander et al., 2016a; Brander et al., 2016b; DeCourten & Brander, 2017). Our results support previous findings illustrating the importance of testing for adverse effects of estrogenic EDCs at low doses, which can produce stronger effects than those observed at moderate doses (Brander et al., 2012; Vandenberg et al., 2012).

We found that the effects of EE2 exposure may be influenced by temperature in the F1 larvae, with the 28 °C EE2 treatment significantly altering expression of most genes tested. Many of the genes selected in this study are involved in the hypothalamic-pituitary-gonadal (HPG) axis and are useful in determining the effects of elevated temperature and EDC exposure on development and reproduction. We found that 28 °C EE2 treatment reduced the expression of the aromatase gene, CYP19b, which is primarily expressed in the brain and plays a pivotal role in sex determination by the conversion of androgens into estrogens (Diotel et al., 2010). Opposite of our findings with EE2, previous studies found that CYP19b expression was increased following exposure to environmental androgens or estrogens (Baumann et al., 2014; Filby et al., 2007). We found a similar pattern of expression in ESR2 and ESR3, which are nuclear, ligand-dependent estrogen receptors that can regulate estrogenic responses via direct mechanisms or by binding to estrogen response elements in the promoter regions of some genes (Filby & Tyler, 2005). Our findings in the 28 °C EE2 group are contrary to previous studies that reported an upregulation of estrogen receptors following exposure to estrogenic EDCs (Filby & Tyler, 2005; Filby et al., 2007; Rhee et al., 2011) indicating that these effects may be dependent on temperature. These discrepancies may also be attributed to the length of chemical exposure (0 hpf to 21 dph, including a 7–10 day embryonic incubation), and responses regulated via ligand–receptor feedback loops can be dependent on exposure period (Golshan et al., 2014). There is evidence that genes upregulated during earlier timepoints in early life stage exposure to EDCs can begin to down regulate as the exposure continues past 30 days post fertilization (Shved et al., 2008). Analogous to the expression of nuclear estrogen receptors, we found a reduction in the expression of GPR30, an estrogen receptor coupled with seven transmembrane G proteins that plays a pivotal role in reproductive success including activating oocyte maturation (Maggiolini & Picard, 2010; Pang, Dong & Thomas, 2008). The production of the enzymes 17β-HSD and 3β-HSD, which act as catalysts in the steroidogenesis pathway, is essential in the synthesis pathway of biologically active sex steroids (Marchais-Oberwinkler et al., 2011; Liu et al., 2012). We found that EE2 (28 °C) downregulated the expression of genes coding for the production of these enzymes, corroborating previous findings that production of this class of enzymes can be influenced by exposure to EE2 (Liu et al., 2012; Filby et al., 2007). Similarities between the expression of the above-mentioned genes indicate that the response to EE2 can exacerbated by temperature along the HPG axis following exposure at the larval stage. Changes in the expression of these genes may influence the sexual development and reproductive capacity of exposed fish. At the adult stage, F1 fish in this experiment were observed to have a more female biased sex ratio in the 28 °C EE2 than any other treatment (DeCourten & Brander, 2017). The observed changes in gene expression may explain the skewed sex ratios of the F1 adults in the 28 °C EE2 treatment (DeCourten & Brander, 2017) as sex of M. beryllina is determined in the larval stage (Conover & Fleisher, 1986; Strüssmann et al., 2010). Sex ratios deviating from the evolutionarily stable 50:50 males to females have the potential to result in population declines in M. beryllina (White et al., 2017).

We investigated the response of additional genes in larval tissue essential to reproductive function at later life stages. We found that the 28 °C EE2 treatment downregulated the production of GnRHR, which indicates that less gonadotropin releasing hormone (GnRH) is being produced in the hypothalamus (Levavi-Sivan et al., 2010). GnRH stimulates gonadotropin production in the gonads and is essential for gonad maturation, reproduction, and reproductive behaviors (Jodo, Ando & Urano, 2003). A similar pattern of expression was found with the FSHR gene indicating that EE2 (28 °C) downregulated the production of follicle stimulating hormone (FSH) in the pituitary of the larval fish. FSH levels can vary based on reproductive status of the individual and mediates steroidogenesis, gametogenesis and ovulation in the gonads (Rocha et al., 2009). A previous study using fish exposed to BPA, found an upregulation on GnRHR and FSHR, opposite of our finding (Rhee et al., 2011). However, Rhee et al. (2011) found similar expression patterns across GnRHR, FSHR, ERs, and CYP19a, indicating that these genes respond to EDCs similarly. Our findings align with Rhee et al. (2011) as the downregulation observed in the 28 °C EE2 treatment was similar across all genes we tested, indicating a similar response to EE2 exposure throughout the HPG axis. These results could explain the reduction in egg production by the F1 adults in the 28 °C EE2 treatment in this experiment as multiple genes within the HPG axis were downregulated within the 28 °C EE2 treatment, potentially impacting the development and reproduction (DeCourten & Brander, 2017). However, these changes in gene expression were not observed in the F1 adult gonad tissues, suggesting that gene expression altered immediately after EE2 exposure was not measurable later in life after rearing in clean water. Despite the lack of altered gene expression at the adult stage, we were able to measure changes to higher order processes (e.g., sex ratio, egg production) that could have been mediated through the expression of genes at the larval stage, which highlights the importance of testing across levels of biological organization (Brander, Hecht & Kuivila, 2015).

Previous literature has established that exposure to estrogenic EDCs has been shown to alter the expression of many genes involved in the reproductive and developmental pathways in gonadal tissues (Scholz & Gutzeit, 2000; Filby et al., 2007; Zhang et al., 2008). Likewise, increased temperature has been shown to influence expression of genes such as gonadal aromatase, which is the mechanism through which animals are able to achieve environmental sex determination (Kitano et al., 1999; Uchida et al., 2004). However, we were unable to detect any significant differences in the expression of genes in the parental ovaries (Fig. S1), parental testes (Fig. S2), F1 adult ovaries (Fig. S3), F1 adult testes (Fig. S4) in any treatments tested in this study. This is likely due to high variability of measurements between our samples and the loss of statistical power within our tests resulting from a high number of pairwise comparisons. These discrepancies with existing literature may also be explained by exposing our fish to EDCs at a salinity of 15 ppt because EE2 uptake can be influenced by changes in salinity (Blewett et al., 2013). Additionally, because we were only able to sample fish at one time point in each life stage we may have missed an opportunity to detect changes in gene expression that had occurred earlier on or there may have been unmeasured differences in protein or hormone concentrations that were not represented by gene expression quantifications at the end of the exposure. This may explain why these exposures resulted in phenotypic changes such as decreases in egg production within 48 h of the sampling time (DeCourten & Brander, 2017).

Like the genes involved in the HPG axis, we found that the IGF2 expression was reduced by exposure to EE2 at 28 °C. Production of IGFs are regulated by growth hormones, mediating the uptake of glucose and amino acids, and protein synthesis (Shamblott et al., 1995). A study with fathead minnows (Pimephales promelas) found that bifenthrin downregulated the expression of IGF2, when using higher concentrations than those used in our study (Beggel et al., 2011). This supports our findings that high doses or highly estrogenic EDCs, such as EE2 in our study, can downregulate the production of IGF2, although we did not find a significant difference between low concentrations of bifenthrin and the controls. Previous literature has shown that IGF2 transcription is not influenced by temperature (Gabillard et al., 2003), despite the growth of ectothermic organisms being heavily influenced by elevated temperatures (Moyano et al., 2016). Our results support the findings that temperature alone does not influence IGF2 transcription. However, we found that EE2 exposure, in combination with elevated temperature, had a greater effect on IGF2 regulation than either stressor alone.

It has been shown that temperature can exacerbate the feminizing effects of EE2 exposure and delay the development of zebrafish testes (Luzio et al., 2016). Previous studies have found that effects of estrogenic EDCs on molecular endpoints and development can be influenced by abiotic factors such as temperature and photoperiod (Brian et al., 2008; Jin et al., 2009). For example, Brian et al. (2008) found that the vitellogenin production increased at higher temperatures in fish exposed to a mixture of estrogenic EDCs. In the case of 17β-HSD, we found that increased temperature can also influence effects of bifenthrin exposure in the F2 generation. Bifenthrin has been shown to alter the expression of genes involved in hormone regulation producing higher order effects on reproduction (Brander et al., 2016b). It has been suggested that effects produced by concurrent exposure to EDCs and elevated temperature can negatively impact survival, development and reproduction, resulting in population declines over time (Brown et al., 2015).

Our results indicate that organisms exposed at early life stages are more susceptible to molecular effects of EDCs, as found with previous studies (Guillette Jr et al., 1995; Van Aerle et al., 2002), as we determined a greater effect of EDCs and temperature at the larval life stage than adult. In the case of FSHR expression, we found that the same pattern measured in the previous generation was also observed in their offspring. Parental exposure to environmental factors can result in phenotypic changes to offspring to allow for rapid response to environmental stressors (Burgess & Marshall, 2014). It has previously been shown that the effects of EDCs can be observed across multiple generations following exposure (Corrales et al., 2014). Transgenerational effects have been linked to lowered survival and reproduction in generations following parental exposure, including generations only exposed as germ cells (Schwindt et al., 2014; Bhandari, Vom Saal & Tillitt, 2015). However, similar effects between the F1 and F2 larvae were not seen in all genes with the expression of 17β-HSD and GPR30 differing between generations. For example, we saw a greater influence of bifenthrin on the expression of 17β-HSD and GRP30 in the F2 larvae, which were exposed indirectly as germ cells, than the previous generation. Our previous findings suggest that EDCs can have more deleterious effects on development in indirectly exposed larvae (DeCourten & Brander, 2017). Our results indicate that the full impact of some EDCs may not be realized when conducting exposures on a single generation. Multigenerational testing is important to determine the effects of parental exposure and potential epigenetic modifications caused by EDC exposure that may result in persistent effects to exposed populations (Brander, Biales & Connon, 2017).

Conclusions

We found that increases in temperature can increase the effects of EDC exposure, as we observed more alterations of gene expression in higher temperatures. Based off predictions of temperature increases over the next 100 years by the IPCC 2013 report (Stocker et al., 2013), we can expect to see rising temperatures influence how organisms respond to exposure to EDCs. Our findings confirm the importance for research incorporating multiple stressors to assess the risk of anthropogenic stressors to aquatic organisms. We found that the changes in larval gene expression resulting from exposure to a combination of EDCs and temperature can result in higher order effects that are measurable later in life after molecular endpoints are no longer affected. These findings imply that organisms exposed briefly as larvae can be affected by these stressors throughout their lives regardless of time allotted for recovery. Additionally, we found that effects of EDCs can be carried over to the next generation or in some cases produce different effects in subsequent generations. These findings suggest mechanisms dictating generational transfer of effects following parental exposure may not affect all molecular endpoints equally. Further research is needed to determine the effects of parental exposure on subsequent generations. We found that larval life stages were more sensitive to EDC exposure than adults, and that changes in gene expression following EDC exposure can be further influenced by temperature. Our findings will inform the long-term management of aquatic systems vulnerable to pollution under future climate change scenarios, many of which provide critical habitat for fishes during early life stages.

Supplemental Information