Scholarly article on topic 'Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive endpoints in small-bodied fish exposed under different salinities'

Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive endpoints in small-bodied fish exposed under different salinities Academic research paper on "Biological sciences"

CC BY
0
0
Share paper
Academic journal
Chemosphere
OECD Field of science
Keywords
{"Endocrine disrupting chemicals" / "Fish reproductive tests" / Salinity / "Semi-quantitative review" / Sensitivity / "Small-bodied fish"}

Abstract of research paper on Biological sciences, author of scientific article — Thijs Bosker, Giacomo Santoro, Steven D. Melvin

Abstract The influence of salinity on toxicity outcomes has been demonstrated for various contaminants, but has received limited attention for endocrine disrupting chemicals (EDCs). Short-term laboratory tests using small-bodied fish are an important tool for evaluating impacts of EDCs on reproduction. Tests have been developed for both freshwater and estuarine/marine species, providing an opportunity to assess whether concentrations at which small-bodied fish respond to EDCs may be influenced by salinity. We conducted a semi-quantitative review of short-term laboratory tests with small-bodied fish exposed to EDCs, including 59 studies under freshwater conditions (7 species) and 23 studies conducted under saline conditions (5 species). We focused on two model estrogens [17α-ethinylestradiol and 17β-estradiol (E2)], and three androgens (17β-trenbolone, 5α-dihydrotestosterone and 17α-methyltestosterone). The lowest observed adverse effect concentration (LOAECLOW) for key reproductive endpoints was recorded, including sex-steroid and vitellogenin (VTG) levels, fecundity and fertilization. In 65.2% of cases, responses occurred at lower doses under freshwater compared to saline conditions, compared to only 4.3% of cases where fish responded to lower doses under saline conditions. The potential influence of salinity was more pronounced when estrogenic compounds were considered separately, with fish responding to lower doses under fresh compared to saline conditions in 90.5% of cases. Fecundity and E2 level were identified as the most sensitive endpoints for evaluating EDCs regardless of salinity. Interestingly, female VTG levels were a sensitive endpoint under freshwater but not saline conditions. Overall, our results suggest that salinity may be an important factor influencing how small-bodied fish respond to environmental EDCs.

Academic research paper on topic "Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive endpoints in small-bodied fish exposed under different salinities"

Accepted Manuscript

Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive endpoints in small-bodied fish exposed under different salinities

Thijs Bosker, Giacomo Santoro, Steven D. Melvin

PII: S0045-6535(17)30767-1

DOI: 10.1016/j.chemosphere.2017.05.063

Reference: CHEM 19278

To appear in: ECSN

Received Date: 25 January 2017 Revised Date: 20 April 2017 Accepted Date: 11 May 2017

Please cite this article as: Bosker, T., Santoro, G., Melvin, S.D., Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive endpoints in small-bodied fish exposed under different salinities, Chemosphere (2017), doi: 10.1016/j.chemosphere.2017.05.063.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 Salinity and sensitivity to endocrine disrupting chemicals: A comparison of reproductive

2 endpoints in small-bodied fish exposed under different salinities

3 Thijs Bosker+i , Giacomo Santorot, and Steven D. Melvin§

5 + Leiden University College, Leiden University, P.O. Box 13228, 2501 EE, The Hague, the

6 Netherlands

7 i Institute of Environmental Sciences, Leiden University, P.O. Box 9518, 2300 RA Leiden,

8 the Netherlands

9 § Australian Rivers Institute, Griffith University, Building G51, Edmund Rice Drive,

10 Southport, QLD 4215, Australia

12 Corresponding author: Thijs Bosker: t.bosker@luc.leidenuniv.nl

13 Abstract:

14 The influence of salinity on toxicity outcomes has been demonstrated for various

15 contaminants, but has received limited attention for endocrine disrupting chemicals (EDCs).

16 Short-term laboratory tests using small-bodied fish are an important tool for evaluating

17 impacts of EDCs on reproduction. Tests have been developed for both freshwater and

18 estuarine/marine species, providing an opportunity to assess whether concentrations at which

19 small-bodied fish respond to EDCs may be influenced by salinity. We conducted a semi-

20 quantitative review of short-term laboratory tests with small-bodied fish exposed to EDCs,

21 including 59 studies under freshwater conditions (7 species) and 23 studies conducted under

22 saline conditions (5 species). We focused on two model estrogens [17a-ethinylestradiol and

23 17P-estradiol (E2)], and three androgens (17P-trenbolone, 5a-dihydrotestosterone and 17a-

24 methyltestosterone). The lowest observed adverse effect concentration (LOAECLOW) for key

25 reproductive endpoints was recorded, including sex-steroid and vitellogenin (VTG) levels,

26 fecundity and fertilization. In 65.2% of cases, responses occurred at lower doses under

27 freshwater compared to saline conditions, compared to only 4.3% of cases where fish

28 responded to lower doses under saline conditions. The potential influence of salinity was

29 more pronounced when estrogenic compounds were considered separately, with fish

30 responding to lower doses under fresh compared to saline conditions in 90.5% of cases.

31 Fecundity and E2 level were identified as the most sensitive endpoints for evaluating EDCs

32 regardless of salinity. Interestingly, female VTG levels were a sensitive endpoint under

33 freshwater but not saline conditions. Overall, our results suggest that salinity may be an

34 important factor influencing how small-bodied fish respond to environmental EDCs.

35 Keywords: Endocrine disrupting chemicals; fish reproductive tests; salinity; semi-

36 quantitative review; sensitivity; small-bodied fish

1. Introduction:

Endocrine disrupting compounds (EDCs) upset hormone pathways in a variety of organisms, potentially negatively influencing reproductive performance (Martin and Voulvoulis, 2009). For this reason, EDCs have attracted significant scientific attention and are a major source of public concern. They are ubiquitous in aquatic systems and are found in freshwater, estuarine and marine environments. EDCs enter the aquatic environment through a variety of sources, including agricultural runoff (Gall et al., 2011; Bergman et al., 2013), sewage effluent (Fent et al., 2006; Coleman et al., 2008) and industrial effluents (Parks et al., 2001; Hewitt et al., 2008). Impacts of EDCs have been well documented under both laboratory and field conditions for various aquatic species, with the majority of the research pertaining to fish. Observed reproductive impacts in fish exposed to EDCs include changes in biochemical biomarkers such as vitellogenin (VTG) (Jobling et al., 1998), increased rates of intersex (Jobling et al., 1998; Kidd et al., 2007) and changes in sex ratios (Larsson et al., 2000). Importantly, such lower-level effects can have major ecological significance for fish, since they have the potential to scale up and can ultimately cause population failure (Kidd et al., 2007). It is therefore extremely important to identify factors that may influence the potency of EDCs to fish, so that at-risk populations can be better identified and protected.

Short-term reproductive bioassays using small-bodied fish provide a powerful tool to assess the impacts of EDCs on reproductive endpoints (Ankley and Johnson, 2004). Small-bodied fish species are advantageous because they are often readily available from commercial sources and are easily maintained under laboratory conditions. Reproductive bioassays therefore represent an important testing niche, and widely used protocols have been developed for a variety of species, including fathead minnow (Pimephales promelas), Japanese medaka (Oryzias latipes) and zebrafish (Danio rerio). Standardised bioassays using these particular small-bodied species are commonly applied by organizations such as the US

62 EPA (EPA, 2011) and the OECD (OECD) to study impacts of EDCs on fish reproduction.

63 Similar protocols are also frequently applied to improve the ecological relevance of the test

64 species (i.e. tests adapted to local species), for example to investigate effects under specific

65 environmental conditions (e.g., brackish or marine species). Protocols adapted for local small-

66 bodied freshwater species include tests with Chinese rare minnow (Gobiocypris rarus; Zha et

67 al., 2008), brook stickleback (Culaea inconstans; Muldoon and Hogan, 2016), Rio de la Plata

68 onesided livebearer (Jenynsia multidentata; Roggio et al., 2014), and the Australian crimson-

69 spotted rainbowfish (Melanotaeniafluviatilis; Pollino et al., 2007). Protocols for small-bodied

70 brackish and marine species include mummichog (Fundulus heteroclitus; Peters et al., 2007;

71 Bosker et al., 2010a), sheepshead minnow (Cyprinodon variegatus; Folmar et al., 2000),

72 three-spined stickleback (Gasterosteus aculeatus; Allen et al., 2008), sand goby

73 (Pomatoschistus minutus; Saaristo et al., 2009) and the brackish medaka (Oryzias melastigma;

74 Lee et al., 2014). Regardless the species, the standard approach for such tests involves

75 exposing fish for a relative short-period, ranging from 14-28 d depending on the protocol, to

76 either model EDCs (see Dang et al., 2011a) for a summary of studies) or environmental

77 samples, (e.g., municipal, agricultural or industrial effluents). Various reproductive endpoints

78 are subsequently assessed which span different levels of biological organization, most

79 commonly documenting changes in sex-steroid levels, relative gonad size, morphology and

80 broad indicators of fecundity (e.g., egg production and fertilization success).

81 As indicated, EDCs are ubiquitous globally and occur in a range of aquatic

82 environments. The characteristics of the receiving environment are therefore important to

83 consider for their potential influence on toxicity. Differences in salinity represent an obvious

84 environmental factor that may alter the potency of EDCs to fish, but this has received limited

85 research attention. On a physiological level, salinity is an important variable to consider, since

86 fish living under different salinities have adapted the way in which they osmoregulate (Evans

and Claiborne, 1997). Freshwater species are hyperosmotic to their environment and tend to drink very little water, with osmoregulation occurring predominantly through the gills (Evans and Claiborne, 1997). Contrarily, if species are hypoosmotic to their environment they tend to actively drink seawater to maintain their osmotic balance (Evans and Claiborne, 1997). Differences in osmoregulation can therefore result in contaminants entering an organism via different routes, and this in turn can potentially result in toxic effects being realised at different environmental concentrations. The influence of salinity on toxicity outcomes has been documented for various contaminants. For example, a variety of metals (Hall and Anderson, 1995; Wood et al., 2004; Blanchard and Grosell, 2005) and polycyclic aromatic hydrocarbons (PAHs) (Ramachandran et al., 2006; Shukla et al., 2007) have been shown to exhibit differential toxicity in fish exposed under freshwater compared to saline conditions. Considering the global threat that EDCs pose to fish populations, there is a need for research exploring whether salinity might be a factor mediating their toxicity.

A limited number of studies have been conducted directly comparing the impact of EDCs on reproductive parameters in small-bodied fish at different salinities, but the evidence seems to suggest that salinity may be an important factor. For example, Glinka et al. (2015), exposed mummichog to a potent androgen (DHT) under high and low salinity, and found a significant difference in response between freshwater and saline conditions. A direct comparison of the effects of pulp mill effluent, a known source of EDCs, on euryhaline mummichog and freshwater fathead minnow found limited differences between both species (Melvin et al., 2009). However, a comparison of the impacts of the sythetic estrogen EE2 on reproduction showed that in general freshwater fish respond to lower levels of EE2 compared to saline species for a select set of endpoints, including fecundity and VTG levels (Bosker et al., 2016). Freshwater species such as Chinese rare minnow and zebrafish exposed to EE2 exhibited reduced egg production at concentrations as low as 0.2 ng EE2/L (Chinese rare

minnow; Zha et al., 2008) and 1 ng EE2/L (zebrafish; Lin and Janz, 2006). In contrast, a study on mummichog under estuarine conditions found reductions in egg production only at exposure concentrations of 100 ng EE2/L (Peters et al., 2007) or no response at all (Bosker et al., 2016). A similar trend of differential sensitivity is apparent for androgens. For example, reduced egg production was observed in sheepshead minnow exposed to 17P-trenbolone (TB; a synthetic androgen used as a growth promoter in the cattle industry) at 5 [j,g TB/L (Hemmer et al., 2008), whereas fathead minnow responded to the same compound at concentrations 100-fold lower (Ankley et al., 2003). Finally, a review of short-term reproductive tests using three small-bodied freshwater species identified fecundity and gonad histology as two of the most sensitive endpoints to EDCs (Dang et al., 201 la). However, recent studies using the brackish mummichog found no effect of 5a-dihydrotestosterone (DHT) on male and female gonad morphology (Glinka et al., 2015) and no effect of EE2 on fecundity (Bosker et al., 2016). Limited experimental work directly addressing the influence of salinity on EDC potency precludes using purely quantitative techniques (e.g. meta-analysis) to investigate this question. Qualitatively, the existing literature seems to indicate that endpoint sensitivity could differ across species and salinities in fish exposed to EDCs, but given the disparities amongst studies there is a clear need for some form of systematic synthesise of the existing data.

The present study describes a semi-quantitative review of short-term reproductive laboratory bioassays with small-bodied fish. A novel approach to systematically compare endpoint sensitivity was applied to assess whether i) concentrations at which small-bodied fish respond to EDCs differ amongst studies performed under freshwater compared to saline conditions, and ii) whether sensitivity of specific endpoints differs amongst salinities.

2. Materials and Methods 2.1 Data collection

We performed a systematic review to collect data from short-term reproductive bioassays exposing fish to EDCs. Data was grouped based on the isosmotic point for fish, which is around 30-40% of full saltwater concentration (or approximately 9-13 ppt salinity) (Evans and Claiborne, 1997; Evans, 2008). For example, the isosmotic point for mummichog is estimated to be around 9 ppt (Marshall et al., 1999; Wood and Grosell, 2009). We defined a freshwater exposure as occurring under conditions in which fish were exposed at salinities below the isosmotic point, and saline conditions when the exposure concentration was near or above the isosmotic point.

Only studies in which adult, sexually mature, small-bodied (<150 mm; Environment Canada, 2012) fish were exposed to one of five model EDCs for a timeframe of 14-28 d were included in our analyses. Two EDCs were selected to represent an estrogenic mode of action: 17a-ethinylestradiol (EE2) and 17P-estradiol (E2). Two non-aromatizable androgenic compounds were selected: 17P-trenbolone (TB) and 17a-dihydrotestosterone (DHT), as well as one aromatizable androgen: 17a-methyltestosterone (MT). Studies were identified by searching the Thomson Reuters Web of Science™ database and the OECD website database for short-term reproductive tests. The cut-off date for inclusion in the review was 01 July 2016. Only laboratory experiments in which fish were exposed to at least two concentrations (excluding controls) were included in the analyses. In some cases, multi-generational tests or life-cycle tests were included, but only provided data for the F0 generation was reported for an exposure duration between 14-28 d.

Data was collected for a variety of commonly measured reproductive endpoints spanning different levels of biological organization, ranging from biochemical to functional endpoints. The endpoints selected were sex steroid levels [11-ketotestosterone (11KT) and testosterone (T) in males, and T and E2 in females], VTG levels, changes in secondary sex characteristics (SSC), gonadosomatic index (GSI), gonad histology, fecundity, fertilization

success and percent hatchability of eggs. Data were collected for both male and female fish whenever available. Measurements of hormone and VTG levels have been conducted using different methodologies in the literature, for example in blood plasma, in vitro (only for hormone levels) or measurements from specific tissues. In addition, there is considerable variation in the histological assessment of gonadal tissue. It is thus important to recognise that differences in protocols and amongst laboratories can influence study outcomes (Hutchinson et al., 2006). However, since various methods are applied under both saline and freshwater conditions, we assume limited impact of these differences on the overall outcomes of our analysis. Differences that are observable despite the inherent variability in methodologies amongst studies could instead add confidence in the conclusions. Nevertheless, for transparency the method of measuring hormone and VTG levels was reported.

The lowest observed adverse effect concentration (LOAEC) was recorded for each of the endpoints listed above, when available. If no effect was observed, the highest tested concentration was used since this would be expected to yield a conservative outcome and thus not contribute to erroneous conclusions. The following additional information was recorded for each experiment: data source, test species employed, concentrations at which the fish were exposed, length of the exposure, number of functional replicates, number of fish in each replicate, number of males and females, and the salinity of the water during exposure.

2.2 Effect concentrations under different salinities

Results were organized into summary tables presenting the LOAEC values in order to facilitate the identification of possible trends. Data was summarized for both freshwater and saline conditions for each endpoint by providing the lowest LOAEC (LOAECLOW). We defined LOAECLOW as the absolute lowest concentration at which an effect was observed for an endpoint, across all experiments at either the freshwater or saline conditions. If no effect

was observed in any of the experiments we reported the maximum concentration within the concentration range of all experiments as LOAEClow.

2.3 Comparative endpoint sensitivity under different salinities

We assessed whether endpoint sensitivity differed between studies carried out under freshwater compared to saline conditions, for both estrogenic and androgenic EDCs. The approach was adapted from a method recently develop by Dang et al. (Dang et al., 2011a; Dang et al., 2011b). For this approach, endpoints from each individual experiment were divided into three categories of effect:

1) If the LOAEC for a specific endpoint was the lowest of all other endpoints measured within that specific experiment, it was grouped in the first category.

2) If a significant effect was observed for an endpoint, but this occurred above the LOAEC for another endpoint in that study, it was grouped in the second category, and;

3) If no effect was observed for any endpoint at the maximum tested concentration the endpoint was grouped in the third category.

The number of observations for each of the three groups (LOAEC, > LOAEC but < no effect, and no observed effect) was summarized separately for estrogenic (E2 and EE2) and androgenic compounds (TB, DHT and MT). To allow for direct comparisons, the relative contribution of each category was calculated as the ratio between the numbers of observations in each category divided by the total number of observations in all three categories.

3. Results

Our search of the literature identified 43 publications containing 82 individual experiments that satisfied our criteria for inclusion in the study (Table S1, S2 and S3). Of

these, 59 were conducted under freshwater conditions, or at salinities below the isosmotic point of the specific test species. The remaining 23 experiments were conducted under saline conditions, at or above the isosmotic point of the test species, at salinities ranging from 15 to 35 ppt. In all papers fish were labelled either "sexually-mature" or "adult" We noted whether mature oocytes and/or spermatids were presents, either based on histological assessment, visual inspection or the ability to produce eggs (indication of mature oocyte) and the ability to fertilize eggs (indication of mature male spermatids) (Table S1). Table 1 (estrogens) and Table 2 (androgens) summarize the studies included in our analysis, including the specific endpoints measured within each individual study. Additional information for each study, such as the number of replicate tanks, the number of fish per sex per replicate, and measured concentration of the focal EDCs are presented in Table S1.

The euryhaline mummichog was the only species exposed under a range of salinities. When exposed at salinity below 9 ppt (isosmotic point), it was grouped among the freshwater studies, while if the exposure was conducted above 9 ppt the results were included in the saline studies. Importantly, Japanese medaka, a euryhaline species, was always exposed below the isosmotic point and results were thus interpreted as freshwater, while the eurahyline three-spined stickleback, brackish medaka and sheepshead minnow were always exposed above 13 ppt and thus were included as saline studies.

3.1 Difference in observed lowest, median and highest LOEC

The concentration ranges tested under freshwater and saline conditions were comparable for all chemicals, facilitating direct comparison between freshwater and saline conditions (Table 1-4). However, limited experimental data was available for both DHT and MT, and these results therefore need to be interpreted with caution. It was possible to directly compare freshwater against saline conditions for different endpoints in 47 cases (13 times for

EE2, 8 times for E2, 10 times for TB, 11 times for DHT and 4 times for MT).

In 30 out of 46 cases (65.2%) LOAEClow was less under freshwater compared to saline conditions (Table 3 and 4). In contrast, LOAECLOW was less under saline conditions in only 2 out of 46 (4.3%) cases (Table 3 and 4). For estrogenic compounds the influence of salinity was most evident, with 19 out of 21 (90.5%) cases reporting the lowest LOAECLOW under freshwater conditions. Responses for estrogenic EDCs never occurred at lower doses under saline compared to freshwater conditions (Table 3). For androgenic compounds this pattern was not as clear, with LOAECLOW observed under freshwater conditions in 11 out of 25 (44.0%) of the cases, compared to 2 out of 25 (8.0%) cases for saline conditions (Table 4).

Exposure concentrations at which the LOAECLOW was observed was considerably less under freshwater conditions compared to saline conditions. On average, for estrogenic compounds, LOAECLOW was >70-fold lower under freshwater conditions compared to saline conditions. For example, for estrogenic exposures, the lowest observed LOAEC for male VTG induction under freshwater exposures was 0.5 ng/L for EE2 and 10 ng/L for E2. In contrast, this was 50 ng EE2/L and 100 ng E2/L under saline conditions. For female GSI the lowest observed LOAEC was 0.2 ng EE2/L and 10 ng E2/L under freshwater condition (Table 3), whereas no effect was observed at exposure levels up to 100 ng EE2/L or at 500 ng E2/L under saline conditions. On average LOAECLOW for androgens was >17-fold lower under freshwater conditions compared to saline conditions.

3.2 Difference in Endpoint Sensitivity

Endpoint sensitivity for estrogenic and androgenic EDCs under freshwater and saltwater conditions is reported in Fig. 1 and 2, respectively. For both freshwater and saltwater conditions, endpoints presenting less than 2 observations were excluded. The most sensitive endpoints for estrogenic exposure were male VTG induction, female E2 levels and

fecundity, all exhibiting responsiveness in >65% of studies (Fig. 1). The same trend was identified when considering experiments conducted exclusively under freshwater conditions, but the prevalence of responsiveness increased to 80% of studies (Fig. 1 and 2). VTG levels for females were a sensitive endpoint to detect impacts of estrogenic exposure under freshwater conditions, with a significant effect measured in >80% of studies (Fig. 1). Contrarily, when females were exposed to estrogenic EDCs under saline conditions not a single significant difference in VTG-levels was reported (Fig. 1). The least sensitive endpoints for estrogenic exposure included male and female GSI, male and female testosterone levels, as well as histological assessment of gonadal tissue, with <33% of studies reporting significant effects of estrogenic exposure on these endpoints under both freshwater and saline conditions (Fig. 1).

When examining androgenic effects, endpoints measured in males were generally less sensitive compared to endpoints measured in females, with the exception of 11KT levels (Fig. 2). Male testosterone levels, VTG induction, histological assessment of gonad alteration and GSI showed effects in <40% of experiments, regardless of salinity (Fig. 2). The most sensitive endpoints for assessing androgenic effects, again regardless of salinity, were female E2 and T levels, female VTG levels and fecundity (Fig. 2). GSI in females was not a sensitive endpoint to assess androgenic effects. Histological alteration of female gonads was a sensitive endpoint under freshwater conditions (effects observed in nearly 70% of experiments), but not under saline conditions (effects only observed in 20% of experiments; Fig. 2).

Overall, E2 and fecundity were the only two endpoints identified as being sensitive for detecting both estrogenic and androgenic effects (Fig. 1 and 2). In >75% of experiments a significant change in fecundity was observed, regardless of salinity and estrogenic or androgenic mode of action. A significant response in E2 levels was observed in >65% of experiments.

4. Discussion

The outcomes of this semi-quantitative review suggest that the salinity at which standard reproductive bioassays (with small-bodied fish) are performed may be an important factor influencing effective concentration to EDCs. In general, the concentration at which LOAEClow was observed was more frequently lower when fish were exposed under freshwater conditions compared to saline conditions. This is especially true for the model estrogenic EDCs considered in our analysis (EE2 and E2), but also for model androgens (TB, DHT and MT), although the response pattern was less apparent. The influence of salinity on the observed LOAEC has been previously described for other contaminants, such as various metals and PAHs (Hall and Anderson, 1995; Wood et al., 2004; Blanchard and Grosell, 2005; Ramachandran et al., 2006; Shukla et al., 2007). However, to our knowledge this is the first study to confirm this phenomenon using an innovative approach for semi-quantitative review, based on available response data for common steroidal EDCs. Short-term reproductive tests using small-bodied freshwater species are commonly applied by regulatory agencies such as the US EPA (EPA, 2011) and the OECD (OECD) to investigate the potential impacts of EDCs on the environment. Our results are therefore important, since they highlight the need to consider both freshwater species, and species that normally inhabit estuarine or marine environments (e.g. three-spine sticklebacks, sheepshead minnow and mummichog) to accurately predict and assess the impacts of EDCs on aquatic biota.

The mechanism underlying differences in responsiveness to EDCs at varying salinities are poorly understood. Recent work exploring uptake of EE2 by mummichog under a range of salinities (0, 16 and 32 ppt) found a significant increase in EE2 uptake at brackish (16 ppt) compared to freshwater (0 ppt) and seawater (32 ppt) conditions (Blewett et al., 2013). This difference might be due to differences in gill morphology under different salinities (Blewett et al., 2013). One obvious explanation is that differences are associated with differential species

sensitivity, and that the influence of salinity on responsiveness may be more coincidence than cause. However, another study found no significant difference in EE2 uptake by the brackish mummichog exposed under freshwater conditions compared to several freshwater species (Blewett et al., 2014). This supports our findings because it suggests that differences in uptake, and potentially in responsiveness to EDCs may be more related to the salinity of the exposure medium, as opposed to basic differences in species sensitivity. Interestingly, tissue-specific accumulation differed across species in that study, with increased accumulation in the liver and gallbladder in mummichog, as well as Japanese medaka (Oryzias latipes), compared to fathead minnow, goldfish (Carassius auratus), zebrafish and rainbow trout (Blewett et al., 2014). As such, further research exploring differences in uptake, elimination, and bioaccumulation of EDCs are needed to better understand the influence of salinity.

A previous semi-quantitative review applied a similar approach to compare endpoint sensitivity in fathead minnow, zebrafish and Japanese medaka exposed to EDCs (Dang et al., 2011a). The present study expanded this evaluation to include a total of 12 different species to facilitate comparison of responsiveness in studies performed under freshwater versus saline conditions. The number of chemicals was also reduced for the present analysis, to include only those estrogenic and androgenic EDCs that have been studied under both freshwater and saline conditions with small-bodied reproductive fish bioassays. By focusing the analysis in this manner, our study identified several differences in comparative endpoint sensitivity between exposure under saline and freshwater conditions. Most notably, changes in VTG levels in female fish were identified as a sensitive endpoint to assess estrogenic EDCs under freshwater conditions, but this was not the case for studies carried out under saline conditions. Similarly, 11KT levels in males was found to be highly sensitive under freshwater conditions, but much less so under saline conditions. As discussed, a limited number of studies have explored the influence of salinity on responsiveness of fish to EDCs, but several studies have

explored the influence of salinity on sexual maturation, including vitellogenesis and steroidogenesis. For example, female striped mullet (Mugil cephalus) exhibited greater vitellogenesis in saline compared to freshwater conditions (Tamaru et al., 1994), and plasma steroid levels were unaffected by salinity in female black bream (Acanthopagrus butcheri) whereas males of this species exhibited increased 11KT in saline conditions (Haddy and Pankhurst, 2000). These examples support our hypothesis that salinity is an important factor that can influence sensitivity of fish to EDCs, and also corroborates the differences in responsiveness of VTG and 11KT identified between sexes.

Our results suggest that the most sensitive endpoints in fish exposed to both estrogenic and androgenic EDCs are E2 levels and altered fecundity in females. This is consistent with a previous review on short-term reproductive tests that similarly identified E2 as a highly sensitive endpoint. Importantly, that study also found E2 to exhibit the best correlation with changes in fecundity (Bosker et al., 2010b), highlighting the importance in assessing both of these endpoints when evaluating the effects of EDCs on fish reproduction. However, our results differ somewhat from the study by Dang et al., (2011a) who reported fecundity, VTG and gonad histology to be the most sensitive endpoints. Specifically, our results indicate that female VTG is not a sensitive endpoint for assessing estrogenic EDCs under saline conditions, and that male VTG levels are not sensitive to androgenic compounds under freshwater or saline conditions. Finally, histological assessment of the gonads showed only a moderate chance of finding significant effects for androgens under freshwater conditions, but not for any other scenario. The difference in outcomes may reflect the difference in approach. Specifically, the present study included a greater number of species but focussed on fewer chemicals compared to the study performed by Dang et al. (2011a).

To conclude, this is the first study to our knowledge to systematically assess the potential influence of salinity on reproductive effects in fish exposed to common

environmental EDCs. We found that fish generally respond to lower levels of both estrogenic and androgenic contaminants when exposed under freshwater conditions. In addition, our analysis revealed minor differences in endpoint sensitivity, which represents useful information for ensuring that the most sensitive endpoints are targeted for reproductive bioassays with small-bodied fish. The most sensitive endpoints in the literature, regardless of estrogenic or androgenic mode of action, or differences in salinity, were identified as E2 levels in female fish and fecundity. Overall, these findings support the hypothesis that salinity may be an important factor that can influence the effects of EDCs on fish reproduction, stressing the importance of taking this variable into account to achieve comprehensive environmental risk assessment. Considering the potential importance for influencing study outcomes, future experimental research is warranted to explicitly explore differential sensitivity in common model small-bodied fish species exposed under different salinities.

Supplementary information

Table S1: Background information on experimental design, dosing method, actual concentrations and reproductive stage of males and females.

Table S2: Summary table of reproductive data in which adult fish were exposed to either 17a-ethinylestradiol (EE2) or 17P-estradiol (E2) for a duration between 14-28 d

Table S3: Summary table of reproductive data in which adult fish were exposed to either 17P-trenbolone (TB), 5a-dihydrotestosterone (DHT) or methyltestosterone (MT) for a duration between 14-28 d. NOTE: nominal concentrations for TB in ng/L, for DHT and MT in |ig/L

Competing interests

388 The authors declare that they have no competing interests.

390 Funding

391 This research did not receive any specific grant from funding agencies in the public,

392 commercial, or not-for-profit sectors.

394 Acknowledgements

395 We thank Dr. Brid Walsh for providing feedback on this manuscript.

References

Allen, Y.T., Katsiadaki, I., Pottinger, T.G., Jolly, C., Matthiessen, P., Mayer, I., Smith, A.,

Scott, A.P., Eccles, P., Sanders, M.B., Pulman, K.G.T., Feist, S., 2008. Intercalibration exercise using a stickleback endocrine disrupter screening assay. Environ. Toxicol. Chem. 27, 404-412.

Ankley, G.T., Jensen, K.M., Makynen, E.A., Kahl, M.D., Korte, J.J., Hornung, M.W., Henry, T.R., Denny, J.S., Leino, R.L., Wilson, V.S., Cardon, M.C., Hartig, P.C., Gray, L.E., 2003. Effects of the androgenic growth promoter 17-beta-trenbolone on fecundity and reproductive endocrinology of the fathead minnow. Environ. Toxicol. Chem. 22, 1350-1360.

Ankley, G.T., Johnson, R.D., 2004. Small fish models for identifying and assessing the effects of endocrine-disrupting chemicals. ILAR J. 45, 469-483.

Armstrong, B.M., Lazorchak, J.M., Jensen, K.M., Haring, H.J., Smith, M.E., Flick, R.W., Bencic, D.C., Biales, A.D., 2016. Reproductive effects in fathead minnows (Pimphales promelas) following a 21 d exposure to 17 alpha-ethinylestradiol. Chemosphere 144, 366-373.

Bergman, A., Heindel, J.J., Jobling, S., Kidd, K.A., Zoeller, R.T., Jobling, S.K., 2013. State Of The Science Of Endocrine Disrupting Chemicals 2012: An Assessment Of The State Of The Science Of Endocrine Disruptors Prepared By A Group Of Experts For The United Nations Environment Programme And World Health Organization. World Health Organization.

Blanchard, J., Grosell, M., 2005. Effects of salinity on copper accumulation in the common killifish (Fundulus heteroclitus). Environ. Toxicol. Chem. 24, 1403-1413.

Blewett, T., MacLatchy, D.L., Wood, C.M., 2013. The effects of temperature and salinity on 17-alpha-ethynylestradiol uptake and its relationship to oxygen consumption in the model euryhaline teleost (Fundulus heteroclitus). Aquat. Toxicol. 127, 61-71.

Blewett, T.A., Chow, T.L., MacLatchy, D.L., Wood, C.M., 2014. A species comparison of 17-alpha-ethynylestradiol uptake and tissue-specific distribution in six teleost fish. Comp. Biochem. Phys. C 161, 33-40.

Bosker, T., Hewitt, L.M., Munkittrick, K.R., MacLatchy, D.L., 2010a. Validation of a refined short-term adult fish reproductive test with improved power for mummichog (Fundulus heteroclitus) to test complex effluents. Ecotoxicol. Environ. Saf. 73, 15961601.

Bosker, T., Munkittrick, K.R., Lister, A., MacLatchy, D.L., 2016. Mummichog (Fundulus heteroclitus) continue to successfully produce eggs after exposure to high levels of 17a-ethinylestradiol. Environ. Toxicol. Chem. 35, 1107-1112

Bosker, T., Munkittrick, K.R., MacLatchy, D.L., 2010b. Challenges and opportunities with the use of biomarkers to predict reproductive impairment in fishes exposed to endocrine disrupting substances. Aquat. Toxicol. 100, 9-16.

Brion, F., Tyler, C.R., Palazzi, X., Laillet, B., Porcher, J.M., Garric, J., Flammarion, P., 2004. Impacts of 17 beta-estradiol, including environmentally relevant concentrations, on reproduction after exposure during embryo-larval-, juvenile- and adult-life stages in zebrafish (Danio rerio). Aquat. Toxicol. 68, 193-217.

Environment Canada, 2012. Metal Mining Technical Guidance For Environmental Effects Monitoring. Environment Canada, Ottawa, Canada.

Caspillo, N.R., Volkova, K., Hallgren, S., Olson, P.E., Porsch-Hallstrom, I., 2014. Short-term

treatment of adult male zebrafish (Danio rerio) with 17 alpha-ethinyl estradiol affects the transcription of genes involved in development and male sex differentiation. Comp. Biochem. Phys. C 164, 35-42.

Coe, T.S., Hamilton, P.B., Hodgson, D., Paull, G.C., Stevens, J.R., Sumner, K., Tyler, C.R., 2008. An environmental estrogen alters reproductive hierarchies, disrupting sexual selection in group-spawning fish. Environ. Sci. Technol. 42, 5020-5025.

Coleman, H., Khan, S., Watkins, G., Stuetz, R., 2008. Fate and analysis of endocrine disrupting chemicals in some sewage treatment plants in Australia. Water Sci. Technol. 58, 2187-2194.

Cripe, G.M., Hemmer, B.L., Goodman, L.R., Fournie, J.W., Raimondo, S., Vennari, J.C., Danner, R.L., Smith, K., Manfredonia, B.R., Kulaw, D.H., Hemmer, M.J., 2009. Multigenerational exposure of the estuarine sheepshead minnow (Cyprinodon variegatus) to 17 beta-estradiol. i. organism-level effects over three generations. Environ. Toxicol. Chem. 28, 2397-2408.

Cripe, G.M., Hemmer, B.L., Raimondo, S., Goodman, L.R., Kulaw, D.H., 2010. Exposure of three generations of the estuarine sheepshead minnow (Cyprinodon variegatus) to the androgen, 17 beta-trenbolone: effects on survival, development, and reproduction. Environ. Toxicol. Chem. 29, 2079-2087.

Dammann, A.A., Shappell, N.W., Bartell, S.E., Schoenfuss, H.L., 2011. Comparing

biological effects and potencies of estrone and 17 beta-estradiol in mature fathead minnows, Pimephalespromelas. Aquat. Toxicol. 105, 559-568.

Dang, Z., Li, K., Yin, H., Hakkert, B., Vermeire, T., 2011a. Endpoint sensitivity in fish endocrine disruption assays: regulatory implications. Toxicol. Lett. 202, 36-46.

Dang, Z., Traas, T., Vermeire, T., 2011b. Evaluation of the fish short term reproduction assay for detecting endocrine disrupters. Chemosphere 85, 1592-1603.

EPA, U., 2011. Fish Short-Term Reproduction Assay. Washington, DC, USA.

Evans, D.H., 2008. Teleost fish osmoregulation: what have we learned since August Krogh, Homer Smith, and Ancel Keys. Am. J. Physiol. Regul. Integr. Comp. Physiol. 295, R704-R713.

Evans, D.H., Claiborne, J.B., 1997. The Physiology Of Fishes, Second Edition. Taylor & Francis, UK.

Fent, K., Weston, A.A., Caminada, D., 2006. Ecotoxicology of human pharmaceuticals. Aquat. Toxicol. 76, 122-159.

Feswick, A., Ings, J.S., Doyle, M.A., Bosker, T., Munkittrick, K.R., Martyniuk, C.J., 2014. Transcriptomics profiling and steroid production in mummichog (Fundulus heteroclitus) testes after treatment with 5 alpha-dihydrotestosterone. Gen. Comp. Endocrinol. 203, 106-119.

Folmar, L.C., Hemmer, M., Hemmer, R., Bowman, C., Kroll, K., Denslow, N.D., 2000.

Comparative estrogenicity of estradiol, ethynyl estradiol and diethylstilbestrol in an in vivo, male sheepshead minnow (Cyprinodon variegatus), vitellogenin bioassay. Aquat. Toxicol. 49, 77-88.

Forsgren, K.L., Qu, S., Lavado, R., Cwiertny, D., Schlenk, D., 2014. Trenbolone acetate metabolites promote ovarian growth and development in adult Japanese medaka (Oryzias latipes). Gen. Comp. Endocrinol. 202, 1-7.

Gall, H.E., Sassman, S.A., Lee, L.S., Jafvert, C.T., 2011. Hormone discharges from a

Midwest tile-drained agroecosystem receiving animal wastes. Environ. Sci. Technol. 45, 8755-8764.

Glinka, C.O., Frasca, S., Provatas, A.A., Lama, T., DeGuise, S., Bosker, T., 2015. The effects of model androgen 5 alpha-dihydrotestosterone on mummichog (Fundulus heteroclitus) reproduction under different salinities. Aquat. Toxicol. 165, 266-276.

Haddy, J.A., Pankhurst, N.W., 2000. The effects of salinity on reproductive development, plasma steroid levels, fertilisation and egg survival in black bream Acanthopagrus butcheri. Aquaculture 188, 115-131.

Hall, L.W., Anderson, R.D., 1995. The influence of salinity on the toxicity of various classes of chemicals to aquatic biota. Crit. Rev. Toxicol. 25, 281-346.

Hemmer, M.J., Cripe, G.M., Hemmer, B.L., Goodman, L.R., Salinas, K.A., Fournie, J.W., Walker, C.C., 2008. Comparison of estrogen-responsive plasma protein biomarkers and reproductive endpoints in sheepshead minnows exposed to 17 beta-trenbolone. Aquat. Toxicol. 88, 128-136.

Hewitt, L.M., Kovacs, T.G., Dube, M.G., MacLatchy, D.L., Martel, P H., McMaster, M.E., Paice, M.G., Parrott, J.L., Van den Heuvel, M.R., Van der Kraak, G.J., 2008. Altered reproduction in fish exposed to pulp and paper mill effluents: Roles of individual compounds and mill operating conditions. Environ. Toxicol. Chem. 27, 682-697.

Hogan, N.S., Currie, S., LeBlanc, S., Hewitt, L.M., MacLatchy, D.L., 2010. Modulation of steroidogenesis and estrogen signalling in the estuarine killifish (Fundulus heteroclitus) exposed to ethinylestradiol. Aquat. Toxicol. 98, 148-156.

Hutchinson, T.H., Ankley, G.T., Segner, H., Tyler, C.R., 2006. Screening and testing for endocrine disruption in fish-biomarkers as" signposts," not" traffic lights," in risk

assessment. Environ. Health Perspect. 114, 106.

Jobling, S., Nolan, M., Tyler, C.R., Brighty, G., Sumpter, J.P., 1998. Widespread sexual disruption in wild fish. Environ. Sci. Technol. 32, 2498-2506.

Jukosky, J.A., Watzin, M.C., Leiter, J.C., 2008. The effects of environmentally relevant mixtures of estrogens on Japanese medaka (Oryzias latipes) reproduction. Aquat. Toxicol. 86, 323-331.

Kang, I.J., Yokota, H., Oshima, Y., Tsuruda, Y., Shimasaki, Y., Honjo, T., 2008. The effects of methyltestosterone on the sexual development and reproduction of adult medaka (Oryzias latipes). Aquat. Toxicol. 87, 37-46.

Kang, I.J., Yokota, H., Oshima, Y., Tsuruda, Y., Yamaguchi, T., Maeda, M., Imada, N., Tadokoro, H., Honjo, T., 2002. Effect of 17 beta-estradiol on the reproduction of Japanese medaka (Oryzias latipes). Chemosphere 47, 71-80.

Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., Lazorchak, J.M., Flick, R.W., 2007. Collapse of a fish population after exposure to a synthetic estrogen. PNAS 104, 8897-8901.

Larsson, D.G.J., Hallman, H., Forlin, L., 2000. More male fish embryos near a pulp mill. Environ. Toxicol. Chem. 19, 2911-2917.

Lee, P.Y., Lin, C.Y., Chen, T.H., 2014. Environmentally relevant exposure of 17alpha-

ethinylestradiol impairs spawning and reproductive behavior in the brackish medaka Oryzias melastigma. Mar. Pollut. Bull. 85, 338-343.

Lin, L.L., Janz, D.M., 2006. Effects of binary mixtures of xenoestrogens on gonadal development and reproduction in zebrafish. Aquat. Toxicol. 80, 382-395.

Marshall, W., Emberley, T., Singer, T., Bryson, S., McCormick, S., 1999. Time course of

salinity adaptation in a strongly euryhaline estuarine teleost, Fundulus heteroclitus: a multivariable approach. J. Exp. Biol. 202, 1535-1544.

Martin, O., Voulvoulis, N., 2009. Sustainable risk management of emerging contaminants in municipal wastewaters. Phil. Trans. R. Soc. A 367, 3895-3922.

Meina, E.G., Lister, A., Bosker, T., Servos, M., Munkittrick, K., MacLatchy, D., 2013. Effects of 17 alpha-ethinylestradiol (EE2) on reproductive endocrine status in mummichog (Fundulus heteroclitus) under differing salinity and temperature conditions. Aquat. Toxicol. 134, 92-103.

Melvin, S.D., Munkittrick, K.R., Bosker, T., MacLatchy, D.L., 2009. Detectable effect size and bioassay power of mummichog (Fundulus heteroclitus) and fathead minnow (Pimephalespromelas) adult reproductive tests. Environ. Toxicol. Chem. 28, 24162425.

Miller, H.D., Clark, B.W., Hinton, D.E., Whitehead, A., Martin, S., Kwok, K.W., Kullman, S.W., 2012. Anchoring ethinylestradiol induced gene expression changes with testicular morphology and reproductive function in the medaka. Plos One 7.

Muldoon, B.M., Hogan, N.S., 2016. Biomarker responses to estrogen and androgen exposure in the brook stickleback (Culaea inconstans): A new bioindicator species for endocrine disrupting compounds. Comp. Biochem. Phys. C 180, 1-10.

OECD, 2006. Report Of The Initial Work Towards The Validation Of The 21-Day Fish

Screening Assay For The Detection Of Endocrine Active Substances. Series on testing and assessment, Number 60.

OECD, Test No. 229: Fish Short Term Reproduction Assay. OECD Publishing.

Panter, G.H., Hutchinson, T.H., Hurd, K.S., Sherren, A., Stanley, R.D., Tyler, C.R., 2004. Successful detection of (anti-) androgenic and aromatase inhibitors in pre-spawning adult fathead minnows (Pimephales promelas) using easily measured endpoints of sexual development. Aquat. Toxicol. 70, 11-21.

Parks, L.G., Lambright, C.S., Orlando, E.F., Guillette, L.J., Jr., Ankley, G.T., Gray, L.E., Jr., 2001. Masculinization of female mosquitofish in Kraft mill effluent-contaminated Fenholloway River water is associated with androgen receptor agonist activity. Toxicol. Sci. 62, 257-267.

Pawlowski, S., Sauer, A., Shears, J.A., Tyler, C.R., Braunbeck, T., 2004. Androgenic and estrogenic effects of the synthetic androgen 17 alpha-methyltestosterone on sexual development and reproductive performance in the fathead minnow (Pimephales promelas) determined using the gonadal recrudescence assay. Aquat. Toxicol. 68, 277291.

Peters, R E M., Courtenay, S.C., Cagampan, S., Hewitt, M.L., MacLatchy, D.L., 2007. Effects on reproductive potential and endocrine status in the mummichog (Fundulus heteroclitus) after exposure to 17 alpha-ethynylestradiol in a short-term reproductive bioassay. Aquat. Toxicol. 85, 154-166.

Pollino, C.A., Georgiades, E., Holdway, D.A., 2007. Use of the Australian crimson-spotted rainbowfish (Melanotaenia fluviatilis) as a model test species for investigating the effects of endocrine disruptors. Environ. Toxicol. Chem. 26, 2171-2178.

Ramachandran, S.D., Sweezey, M.J., Hodson, P.V., Boudreau, M., Courtenay, S.C., Lee, K., King, T., Dixon, J.A., 2006. Influence of salinity and fish species on PAH uptake from dispersed crude oil. Mar. Pollut. Bull. 52, 1182-1189.

Roggio, M., Guyon, N., Hued, A., Ame, M.V., Valdes, M.E., Giojalas, L., Wunderlin, D., Bistoni, M., 2014. Effects of the synthetic estrogen 17 alpha-ethinylestradiol on aromatase expression, reproductive behavior and sperm quality in the fish Jenynsia multidentata. Bull. Environ. Contam. Toxicol. 92, 579-584.

Runnalls, T.J., Beresford, N., Kugathas, S., Margiotta-Casaluci, L., Scholze, M., Scott, A.P., Sumpter, J.P., 2015. From single chemicals to mixtures - Reproductive effects of levonorgestrel and ethinylestradiol on the fathead minnow. Aquat. Toxicol. 169, 152167.

Rutherford, R., Lister, A., Hewitt, L.M., MacLatchy, D., 2015. Effects of model aromatizable (17 alpha-methyltestosterone) and non-aromatizable (5 alpha-dihydrotestosterone) androgens on the adult mummichog (Fundulus heteroclitus) in a short-term reproductive endocrine bioassay. Comp. Biochem. Phys. C 170, 8-18.

Saaristo, M., Craft, J.A., Lehtonen, K.K., Bjork, H., Lindstrom, K., 2009. Disruption of sexual selection in sand gobies (Pomatoschistus minutus) by 17 alpha-ethinyl estradiol, an endocrine disruptor. Horm. Behav. 55, 530-537.

Salierno, J.D., Kane, A.S., 2009. 17alpha-ethinylestradiol alters reproductive behaviors,

circulating hormones, and sexual morphology in male fathead minnows (Pimephales promelas). Environ. Toxicol. Chem. 28, 953-961.

Seki, M., Fujishima, S., Nozaka, T., Maeda, M., Kobayashi, K., 2006. Comparison of

response to 17 beta-estradiol and 17 beta-trenbolone among three small fish species. Environ. Toxicol. Chem. 25, 2742-2752.

Seki, M., Yokota, H., Matsubara, H., Tsuruda, Y., Maeda, N., Tadokoro, H., Kobayashi, K., 2002. Effect of ethinylestradiol on the reproduction and induction of vitellogenin and

testis-ova in medaka (Oryzias latipes). Environ. Toxicol. Chem. 21, 1692-1698.

Shappell, N.W., Hyndman, K.M., Bartell, S.E., Schoenfuss, H.L., 2010. Comparative

biological effects and potency of 17alpha- and 17beta-estradiol in fathead minnows. Aquat. Toxicol. 100, 1-8.

Sharpe, R.L., MacLatchy, D.L., Courtenay, S.C., Van Der Kraak, G.J., 2004. Effects of a model androgen (methyl testosterone) and a model anti-androgen (cyproterone acetate) on reproductive endocrine endpoints in a short-term adult mummichog (Fundulus heteroclitus) bioassay. Aquat. Toxicol. 67, 203-215.

Shukla, P., Gopalani, M., Ramteke, D., Wate, S., 2007. Influence of salinity on PAH uptake from water soluble fraction of crude oil in Tilapia mossambica. Bull. Environ. Contam. Toxicol. 79, 601-605.

Soffker, M., Stevens, J.R., Tyler, C.R., 2012. Comparative breeding and behavioral responses to ethinylestradiol exposure in wild and laboratory maintained zebrafish (Danio rerio) populations. Environ. Sci. Technol. 46, 11377-11383.

Sun, L.W., Zha, J.M., Wang, Z.J., 2009. Interactions between estrogenic chemicals in binary mixtures investigated using vitellogenin induction and factorial analysis. Chemosphere 75, 410-415.

Tamaru, C.S., Lee, C.-S., Kelley, C.D., Miyamoto, G., Moriwake, A., 1994. Oocyte growth in the striped mullet Mugil cephalus l. maturing at different salinities. J. World. Aquac. Soc. 25, 109-115.

Tilton, S.C., Foran, C.M., Benson, W.H., 2005. Relationship between ethinylestradiol-

mediated changes in endocrine function and reproductive impairment in Japanese medaka (Oryzias latipes). Environ. Toxicol. Chem. 24, 352-359.

Van den Belt, K., Verheyen, R., Witters, H., 2001. Reproductive effects of ethynylestradiol and 4t-octylphenol on the zebrafish (Danio rerio). Arch. Environ. Con. Toxicol. 41, 458-467.

Van den Belt, K., Verheyen, R., Witters, H., 2003. Comparison of vitellogenin responses in zebrafish and rainbow trout following exposure to environmental estrogens. Ecotoxicol. Environ. Saf. 56, 271-281.

Van den Belt, K., Wester, P.W., Van der Ven, L.T.M., Verheyen, R., Witters, H., 2002.

Effects of ethynylestradiol on the reproductive physiology in zebrafish (Danio rerio): Time dependency and reversibility. Environ. Toxicol. Chem. 21, 767-775.

Wood, C.M., Grosell, M., 2009. TEP on the tide in killifish (Fundulus heteroclitus): effects of progressively changing salinity and prior acclimation to intermediate or cycling salinity. J. Comp. Physiol. B 179, 459-467.

Wood, C.M., McDonald, M.D., Walker, P., Grosell, M., Barimo, J.F., Playle, R.C., Walsh, P.J., 2004. Bioavailability of silver and its relationship to ionoregulation and silver speciation across a range of salinities in the gulf toadfish (Opsanus beta). Aquat. Toxicol. 70, 137-157.

Xu, N., Chen, P.Y., Liu, L., Zeng, Y.Q., Zhou, H.X., Li, S., 2014. Effects of combined exposure to 17 alpha-ethynylestradiol and dibutyl phthalate on the growth and reproduction of adult male zebrafish (Danio rerio). Ecotoxicol. Environ. Saf. 107, 6170.

Zha, J.M., Sun, L.W., Spear, P.A., Wang, Z.J., 2008. Comparison of ethinylestradiol and nonylphenol effects on reproduction of Chinese rare minnows (Gobiocypris rarus). Ecotoxicol. Environ. Saf. 71, 390-399.

648 List of figures

650 Figure 1: Endpoint sensitivity for estrogens under fresh (a, c) and saline (b, d) exposure conditions.

651 The number of experiments reported with lowest observed effect concentration (LO) is shown in

652 orange. The number of studies in which observed effects were reported above the LOEC within the

653 same study is shown in yellow. The number of studies in which no observed effects was reported at

654 concentrations higher than the maximum tested concentration is shown in white

656 Figure 2: Endpoint sensitivity for androgens under fresh (a, c) and saline (b, d) exposure conditions.

657 The number of experiments reported with lowest observed effect concentration (LO) is shown in

658 orange. The number of studies in which observed effects were reported above the LOEC within the

659 same study is shown in yellow. The number of studies in which no observed effects was reported at

660 concentrations higher than the maximum tested concentration is shown in white.

Table 1: Lowest observed adverse effect concentration (LOAEC) for reproductive endpoints in adult fish exposed to either 17a-ethinylestradiol (EE2) or 17P-estradiol (E2) for a duration between 14-28 d

Condition

Salinity Species

Source

Nominal [ ] ng/L

Endpoint LOAECa

Males Females Other

11KTb Tb VTGc HIST GSI E2b Tb VTGc HIST GSI SSC FEC FERT HATCH

100 >100 1

1 5 1 5 25 5

1 3 10 1 100 1 100 10

10 10 10 20 10 10

25 0.5 25 0.5 25 >25 25 5

>4.5 1.5 >4.5 1.5 >4.5 4.5 >4.5 0.5 >4.5

62.5 62.5 500 500 >500

>500 500 500 5 >500 500 0.2 0.2 500 500

1000 1000 1000

>250 >250 >250 >250 >250

5 10 10 10

10 10 10 25 25 10

2 >10

5 >5 5

FW FW FW FW FW FW FW 1.64% FW FW

FW FW FW FW FW FW

C. inconstans G. rarus P. promelas

O. lapites

F. heteroclitus J.

multidentata D. rerio

Muldoon and Hogan 2015 Zha et al. 2008 Pawlowski et al. 2004 Salierno and Kane 2009 Runnalls et al. 2015 Armstrong et al. 2016

Seki et al. 2002 Tilton et al. 2005 Miller et al. 2012 Meina et al. 2013 Roggio et al. 2014

Van den Belt et al. 2001 Van den Belt et al. 2002 Coe et al 2008 Soffker et al. 2012 Caspillo et al. 2014 Xu et al. 2014

1-10-100 5/1/2025 0.1-1-3-10-100 10-20-40 0.5-5-25 0.5-1.5-4.5 31.3-62.5-125-250-500 0.2-5-500-2000 1-10 50-250

10-75-150

5-10-25-50 10-25 2-10 2-5 5-25 5-20

Saline

32% 20% 16% 16% 32% 16% 32% 1821%

O. melstigma F. heteroclitus

P. minutus C. variegatus

Lee et al. 2014 Peters et al. 2007 Hogan et al. 2010 Meina et al. 2013 Meina et al. 2013 Bosker et al. 2016 Saaristo et al. 2009

Folmar et al. 2000

1-10-50-100 0.1-1-10-100 100-500 50-250 50-250 3-30-300-3000 25-50

20-100-200-500-1000

>100 >500 >250 >250

100 100

100 500 >250 >250 3000 >50

250 >250

>100 >500 >250 >250

>100 >500 >250 >250 >3000

50 100

>3000 >3000 >3000

Condition

Salinity Species Source Nominal [ ] Endpoint LOAECa

ng/L Males Females Other

11KTb Tb VTGc HIST GSI E2b Tb VTGc HIST GSI SSC FEC FERT HATCH

FW M. fluviatilis Pollino et al. 2008 30-100-300-1000 >1000 >1000 >1000 1000 >1000 >1000 300 >1000

FW P. promelas OECD 2006 LAB1 10-32-100 100 100 >100 32 >100 >100

FW OECD 2006 LAB2 10-32-100 10 >100 100 100 100

FW OECD 2006 LAB4 10-32-100 10 >100 100 >100 >100

FW Shappell et al. 2010 9-18-44* 18 44 >44 >44 >44 >44 44

FW Dammann et al. 2011 5-25-50 >50 50 >50 >50 >50

FW Dammann et al. 2011 5-25-50 >50 >50 >50 >50 >50

FW Seki et al. 2006 10-32-100 100 >100 32 >100 100

FW O. lapites Kang et al. 2002 31.3-62.5-125-250-500 62.5 500 500 500 >500 500 500 500

FW Seki et al. 2006 10-32-100 10 >100 100 >100 >100

FW OECD 2006 LAB1 10-32-100 32 >100 100 >100 >100

FW OECD 2006 LAB2 10-32-100 10 >100 32 >100 >100

FW OECD 2006 LAB3 10-32-100 32 >100 100 >100

FW Jukosky et al. 2008 76-379-3793 3793 76 >3793 3793 >3793 3793

FW Sun et al. 2009 5-25-125-625-3125 125

FW D. rerio Van den Belt et al. 2003 20-100 20

FW Brion et al. 2004 5-25-100 25 >100 >100 25 >100 100

FW Seki et al. 2006 10-32-100 100 >100 100 >100

FW OECD 2006 LAB1 10-32-100 >100 >100 10 >100 >100 >100

FW OECD 2006 LAB2 10-32-100 32 >100 >100 32 >100 >100

FW OECD 2006 LAB3 10-32-100 >100 >100 10 100 >100 10

FW OECD 2006 LAB4 10-32-100 >100 32 >100 >100 >100 >100

21% C. variegatus Folmar et al. 2000 20-200-500-1000-2000 200

20% Cripe et al. 2009 10-30-80-200-500 >500 200 >500 >500 300 500

BW/SW G. aculeatus Allen et al. 2008 10-32-100 100 >100 >100 >100 >100 >100

BW/SW Allen et al. 2008 10-32-100 100 >100 >100 >100 >100 >100

BW/SW Allen et al. 2008 10-32-100 100 >100 >100 >100

a 11KT: 11-ketotestosterone; T: testosterone; E2: 17ß-estradiol; VTG: vitellogenin; SSC: secondary sex characteristics; HIST: gonad histology; GSI: gonadosomatic index; FEC: fecundity; SPAWN: number of spawning events; FERT: fertility; HATCH: hatchability. b Bold and italic: in vitro measurements, Bold and underscored: tissue measurement otherwise plasma measurement c Bold and italic: mRNA measurements, Bold and underscored: tissue measurement, otherwise plasma measurement

Table 2: Lowest observed effect concentration (LOAEC) for reproductive endpoints in adult fish exposed to 17P-trenbolone (TB), 5a-dihydrotestosterone (DHT) or methyltestosterone (MT) for a duration between 14-28 d. NOTE: nominal concentrations for TB in ng/L, for DHT and MT in ^/L

Condition Salinity Species Source Nominal [ ] Endpoint LOAECa

ng/L Males Females Other

11KTb Tb VTGc HIST GSI E2b Tb VTGc HIST GSI SSC FEC FERT HATCH

TB Fresh FW P. promelas Ankley et al. 2003 5-50-500-5000-50000 50000 >50000 >50000 >50000 500 500 50 50 >50000 50 50 500 500

FW Seki et al. 2006 50-500-5000 >5000 >5000 5000 >5000 >5000

FW OECD 2006 LAB1 50-500-5000 >5000 >5000 >5000 50 500 >5000 >5000

FW OECD 2006 LAB2 50-500-5000 >5000 >5000 >5000 500 >5000 >5000 >5000

FW OECD 2006 LAB4 50-500-5000 500 50 >5000 5000 500 >5000

FW O. lapites Seki et al. 2006 50-500-5000 >5000 >5000 50 >5000 500

FW OECD 2006 LAB1 50-500-5000 >5000 >5000 50 500 >5000 500

FW OECD 2006 LAB2 50-500-5000 >5000 >5000 >5000 500 >5000 >5000 500

FW OECD 2006 LAB3 50-500-5000 >5000 >5000 >5000 . 50 500 500

FW Forsgren et al. 2014 10-100-1000 >1000 >1000 10

FW D. rerio Seki et al. 2006 50-500-5000 >5000 >5000 500 5000

FW OECD 2006 LAB1 50-500-5000 >5000 >5000 >5000 50 >5000

FW OECD 2006 LAB2 50-500-5000 >5000 >5000 >5000 50 >5000 >5000

FW OECD 2006 LAB3 50-500-5000 >5000 >5000 >5000 50 500 5000

FW OECD 2006 LAB4 50-500-5000 >5000 >5000 >5000 >5000 500 >5000

Saline 19,1% C. variegatus Hemmer et al. 2008 5-50-5000 >5000 >5000 >5000 5000 >5000 >5000

20% Cripe et al. 2010 10-40-200-1000-5000 >5000 1000 1000 1000 200 1000 1000 5000

BW/SW G. aculeatus Allen et al. 2008 50-500-5000 >5000 >5000 >5000 50 >5000 >5000

BW/SW Allen et al. 2008 50-500-5000 >5000 >5000 >5000 >5000 >5000 >5000

BW/SW Allen et al. 2008 50-500-5000 >5000 >5000 5000 >5000

Condition

Salinity Species

Source

Nominal [ ] Hg/L

Endpoint LOAECa

Females

VTGc HIST

Tb VTGc HIST GSI

FEC FERT HATCH

DHT Fresh FW p. promelas Panter et al. 2004 2 °%° F. heteroclitus Glinka et al. 2015

10-32-100 0.05-0.5-5

>100 >5

0.5 0.5

>100 >5

Saline 16%o f. heteroclitus Feswick et al. 2014

16%0 Rutherford et al. 2015

16% Glinka et al. 2015

5-50 10-100 0.05-0.5-5

>50 >100 >5

10 0.5

10 0.5

>100 >5

MT Fresh FW c. inconstans Muldoon and Hogan 2015 FW p. promelas Pawlowski et al. 2004 _FW O. lapites_Kang et al. 2008_

0.001-0.01-0.1 0.1-1-5-50 0.025-0.05-0.1-0.2-0.4

>50 0.4

>0.1 >50 >0.4

50 0.2

0.1 0.025

>0.1 50 0.05

0.025 0.05

Saline 15% f. heteroclitus Sharpe et al. 2004

16% Rutherford et al. 2015

0.001-0.01-0.1 0.1-1

0.1 >1

0.01 >1

>0.1 >1

>0.1 >1

0.01 0.01 1 >1

0.1 >1

>0.1 >1

a 11KT: 11-ketotestosterone; T: testosterone; E2: 17ß-estradiol; VTG: vitellogenin; SSC: secondary sex characteristics; HIST: gonad histology; GSI: gonadosomatic index; FEC: fecundity; SPAWN: number of spawning events; FERT: fertility; HATCH: hatchability.

b Bold and italic: in vitro measurements, Bold and underscored: tissue measurement otherwise plasma measurement

c Bold and italic: mRNA measurements, Bold and underscored: tissue measurement, otherwise plasma measurement

Table 3. The lowest reported LOEC across studies for individual endpoints for fish exposed to 17a-ethinylestradiol (EE2) or 170-estradiol under either freshwater or saline conditions. Red indicates under which salinity the lowest LOAEClow was observed.

Contaminant Condition _Endpoint LOAECa_

Males Females Other

_11KT_T_VTG_HIST_GSI_E2_T_VTG_HIST_GSI_SSC_FEC_FERT HATCH

EE2 Concentration Fresh 0.5-25 0.2-500 0.1-2025 0.1-2025 0.1-2025 0.2-2000 0.5-500 0.1-2025 1-2025 0.1-2025 0.1-100 0.1-2000 0.1-2000 0.2-2000

range (ng/L) Saline 0.1-100 0.1-500 0.1-1000 0.1-100 0.1-3000 0.1-250 0.1-500 0.1-100 0.1-100 0.1-3000 25-50 0.1-3000 0.1-3000 3-3000

Number of Fresh 3 4 13 6 9 4 3 7 3 9 3 8 6 1

experiments Saline 1 4 4 1 6 3 4 1 0 5 1 3 2 1

LOAECLow Fresh 2 10 0.5 3 1 0.5 >4.5 1 25 0.2 1 0.2 5 500

Saline >100 >100 50 >100 100 10 >100 >100 - >100 >50 50 100 >3000

Saline:Fresh >50 >10 100 >33 100 20 - >100 - >500 >50 250 20 >6

Concentration range (ng/L)

Number of experiments

Lowest

Fresh Saline

Fresh Saline

Fresh Saline

Saline:Fresh

30-3793

5-3125 10-2000

100 10

10-1000 10-500

5-3973 10-500

>100 >10

30-1000

30-3793

5-500 10-500

>100 >4

10-500

>44 300

5-3793 10-500

500 50

5-10000

30-3793 10-500

31.3-500 10-500

>500 >1

30-1000

11KT: 11-ketotestosterone; T: testosterone; E2: 17|3-estradiol; VTG: vitellogenin; SSC: secondary sex characteristics; HIST: gonad histology; GSI: gonadosomatic index; FEC: fecundity; SPAWN: number of spawning events; FERT: fertility; HATCH: hatchability

Table 4. The lowest reported LOEC across studies for individual endpoints for fish exposed to 170-trenbolone (TB), 5a-dihydrotestosterone (DHT) or methyltestosterone (MT) under either freshwater or saline conditions. Red indicates under which salinity the lowest LOAEClow was observed.

Contaminant Condition _Endpoint LOAECa_

Males Females Other

_11KT_T_VTG_HIST_GSI_E2_T_VTG_HIST_GSI_SSC_FEC_FERT HATCH

TB Concentration Fresh 5-50000 5-50000 5-50000 50-5000 5-50000 5-50000 5-50000 5-50000 5-50000 5-50000 5-50000 5-50000 5-50000 5-50000

range (ng/L) Saline 50-500 - 5-5000 50-5000 50-5000 - - 5-5000 10-5000 5-5000 10-5000 5-5000 5-5000 5-5000

Number of Fresh 1 1 13 9 13 2 2 13 10 13 8 1 1 1

experiments Saline 0 0 5 2 3 0 0 5 3 5 1 2 2 2

Lowest Fresh 50000 >50000 500 50 >5000 500 500 50 10 500 50 50 500 500

Saline - - >5000 >5000 >5000 - - 50 1000 1000 200 1000 1000 5000

Saline:Fresh - - >10 >100 - - - 1 100 2 4 20 2 10

Concentration range (ng/L)

Fresh Saline

0.05-5 0.05-100

0.05-5 0.05-100

10-100 10-100

0.05-5 0.05-50

0.05-100 0.05-100

0.05-5 0.05-100

0.05-5 0.05-100

10-100 10-100

0.05-5 0.05-5

0.05-100 0.05-100

10-100

0.05-5 0.05-5

Number of experiments

Fresh Saline

Lowest

Fresh Saline

Saline:Fresh

0.5 0.5

0.5 0.5

100 10

>5 0.05

Concentration range (ng/L)

Fresh Saline

0.001-1

0.001-1

0.001-50 0.001-1

0.025-50

0.001-50 0.001-1

0.001-1

0.001-1

0.025-50 0.001-1

0.025-50

0.001-50 0.001-1

0.025-50 0.025-50 0.025-50 0.025-50

Number of experiments

Fresh Saline

Lowest

Fresh Saline

Saline:Fresh

>0.1 >0.1

0.2 0.1

>0.1 >2

11KT: 11-ketotestosterone; T: testosterone; E2: 17ß-estradiol; VTG: vitellogenin; SSC: secondary sex characteristics; HIST: gonad histology; GSI: gonadosomatic index; FEC: fecundity; SPAWN: number of spawning events; FERT: fertility; HATCH: hatchability

Z. 20 -

Estrogens, Freshwater

Estrogens, Freshwater

11KT T VTG HIST GSI E2 T VTG HIST GSI SSC FEC FERT HATCH 11KT T VTG HIST GSI E2 T VTG HIST GSI

SSC FEC FERT HATCH

Estrogens, Saline

Estrogens, Saline

Androgens, Freshwater

Androgens, Freshwater

Androgens, Saline

Androgens, Saline

Highlights:

• The influence of salinity was determined in small-bodied fish exposed to EDCs

• Responses occurred at lower levels under freshwater conditions compared to saline

• This effect if most pronounced when fish were exposed to estrogenic EDCs

• Fecundity and female E2 levels were most sensitive to detect impacts of EDCs