Variation in the toxicity of sediment-associated substituted phenylamine antioxidants to an epibenthic (Hyalella azteca) and endobenthic (Tubifex tubifex) invertebrate

Chemosphere 181 (2017) 250e258

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Variation in the toxicity of sediment-associated substituted phenylamine antioxidants to an epibenthic (Hyalella azteca) and endobenthic (Tubifex tubifex) invertebrate R.S. Prosser a, *, 1, A.J. Bartlett a, D. Milani b, E.A.M. Holman a, H. Ikert a, D. Schissler a, J. Toito a, J.L. Parrott a, P.L. Gillis a, V.K. Balakrishnan a a b

Environment and Climate Change Canada, Aquatic Contaminants Research Division, Burlington, Ontario, Canada Environment and Climate Change Canada, Watershed Hydrology and Ecology Research Division, Burlington, Ontario, Canada

h i g h l i g h t s
28-d LC50s H. azteca exposed to sediment-associated SPAs ranged 22 e > 403 mg/g dw.
28-d EC50s T. tubifex exposed to sediment-associated SPAs ranged 3.6e15 mg/g dw.
Variation in toxicity between SPAs corresponded with Koc.
Variation in toxicity between species due to variation in pathway of exposure.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 6 January 2017 Received in revised form 13 April 2017 Accepted 15 April 2017 Available online 17 April 2017

Substituted phenylamine antioxidants (SPAs) are produced in relatively high volumes and used in a range of applications (e.g., rubber, polyurethane); however, little is known about their toxicity to aquatic biota. Therefore, current study examined the effects of chronic exposure (28 d) to four sedimentassociated SPAs on epibenthic (Hyalella azteca) and endobenthic (Tubifex tubifex) organisms. In addition, acute (96-h), water-only exposures were conducted with H. azteca. Mortality, growth and biomass production were assessed in juvenile H. azteca exposed to diphenylamine (DPA), N-phenyl-1napthylamine (PNA), N-(1,3-dimethylbutyl)-N’-phenyl-1,4-phenylenediamine (DPPDA), or 4,4’-methylene-bis[N-sec-butylaniline] (MBA). Mortality of adult T. tubifex and reproduction were assessed following exposure to the four SPAs. The 96-h LC50s for juvenile H. azteca were 1443, 109, 250, and >22 mg/L and 28-d LC50s were 22, 99, 135, and >403 mg/g dry weight (dw) for DPA, PNA, DPPDA, and MBA, respectively. Reproductive endpoints for T. tubifex (EC50s for production of juveniles > 500 mm: 15, 9, 4, 3.6 mg/g dw, for DPA, PNA, DPPDA, and MBA, respectively) were an order of magnitude more sensitive than endpoints for juvenile H. azteca and mortality of adult worms. The variation in toxicity across the four SPAs was likely related to the bioavailability of the sediment-associated chemicals, which was determined by the chemical properties of the SPAs (e.g., solubility in water, Koc). The variation in the sensitivity between the two species was likely due to differences in the magnitude of exposure, which is a function of the life histories of the epibenthic amphipod and the endobenthic worm. The data generated from this study will support effect characterization for ecological risk assessment. Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.

Handling Editor: Jim Lazorchak Keywords: Amphipod Oligochaete Organic carbon-water partition coefficient Bioavailability

1. Introduction

* Corresponding author. University of Guelph, School of Environmental Sciences, Guelph, Ontario, Canada. E-mail addresses: [email protected], [email protected] (R.S. Prosser). 1 Current address: University of Guelph, School of Environmental Sciences, Guelph, Ontario, Canada. http://dx.doi.org/10.1016/j.chemosphere.2017.04.066 0045-6535/Crown Copyright © 2017 Published by Elsevier Ltd. All rights reserved.

There are a large number of chemicals used in commerce for which there is little data on their potential fate and effects in the environment. For this reason, several initiatives have been adopted to prioritize these chemicals for hazard assessment. For example, in 2006, the Government of Canada introduced the Chemicals

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

Management Plan, to prioritize and evaluate the hazard of approximately 23,000 chemicals, used in commerce in the previous two decades, to human health and the environment by 2020 Government of Canada (2014). Similar programs to assess chemicals in commerce have been initiated in the United States and Europe (USEPA, 2015; ECHA, 2016). It is important to note that these lists of chemicals do not include pharmaceuticals and pesticides, which are extensively assessed under different federal programs (e.g., Health Canada (2009), USFDA (2015), European Commission (2016)). A class of chemicals that is present on these lists for prioritization and hazard assessment is substituted phenylamine antioxidants (SPAs) (Table 1). SPAs are used as a primary antioxidant in a variety of commercial products, e.g., rubber, polyurethane, and engine lubricants, for the purpose of extending the lifespan of the products (WHO, 1998; OECD, 2004; ECHA, 2008). SPAs prevent the formation of free radicals produced by heat, stress, exposure to oxygen and/or ozone, or UV radiation, which contribute to the degradation of polymers (Maier and Calafut, 1998). There is a dearth of data on the potential effects of SPAs on aquatic ecosystems. The majority of research has focused on the acute toxicity of SPAs due to exposure via water (Sikka et al., 1981; Tonogai et al., 1982; Geiger et al., 1990; Drzyzga et al., 1995; Murin et al., 1997). However, SPAs will likely partition into the sediment when they enter an aquatic system due to their relatively high organic carbon-water partition coefficient (Koc) and low solubility in water (Table 1). Hence, research is needed to investigate the effect of sediment-associated SPAs on benthic organisms. Therefore, the objective of our study was to examine the effect of sediment-associated SPAs on two species of benthic invertebrates that would be exposed via different pathways; Tubifex tubifex (oligochaete) are endobenthic (burrow into sediment) while Hyalella azteca (amphipod) are epibenthic (live at sediment-water interface). The magnitude and pathway of exposure to sediment-associated contaminants will vary between these two species, as T. tubifex live exclusively in sediment and are primarily exposed to contaminants bound to the sediment and within the pore water, while H. azteca live at the sediment-water interface and are primarily exposed to contaminants in the overlying water. These two species are also representative of the different pathways of exposure by which most aquatic invertebrates would be exposed to sediment-associated SPAs. As the primary route of exposure for H. azteca is via water, acute, water-only exposures with SPAs were

251

also conducted with this species. Four SPAs were chosen for this study: diphenylamine (DPA), N-phenyl-1-napthylamine (PNA), N(1,3-dimethylbutyl)-N’-phenyl-1,4-phenylenediamine (DPPDA), and 4,4’-methylene-bis[N-sec-butylaniline] (MBA) (Table 1). The four SPAs chosen are representative of the range of physicochemical properties (e.g., log KOW, solubility in water) that are found in this class of chemicals. The data generated from this study will provide information to assist in assessing the risk that SPAs may pose to aquatic ecosystems. 2. Methods 2.1. Sediment The sediment used in this study was a mixture of two reference sediments used by Environment and Climate Change Canada (ECCC) for invertebrate culturing and toxicity testing. The sediments were collected from two sites in Lake Erie in Ontario, Canada (Long Point marsh, 42.583683N, 80.443726W and Long Point Bay, 42.58472N 80.21806W). The physicochemical properties of the sediments are presented in the Supplementary Information (SI) (Table S1). The sediments were mixed to produce sediment with an organic carbon content of ~2%, which is representative of a large range of sediments across Canada. Long Point marsh and bay sediments were sieved through a 250-mm sieve before being mixed in a 2:3 ratio by volume. Tables S2 and S3 contain the physicochemical properties of the mixed sediment and the results of monitoring for potential contaminants. Physicochemical properties of sediment were determined by standard methods developed by ECCC's National Laboratory for Environmental Testing (NLET). Sediment was stored at ~4  C before being used in the study (stored 2e10 months). Due to their relatively low solubility in water, solid SPAs (DPA: purity > 99%, Sigma Aldrich; PNA: purity > 98.0%, TCI; DPPDA: purity > 98.0%, TCI; MBA: purity  99%, CHEMOS Gmbh) were dissolved in acetone to produce stock solutions of different concentrations for each individual SPA for spiking of sediment. Sediment was spiked with the same volume of stock solution to produce the different treatments used in each test and the volume of stock solution used for spiking represented <1% of the volume of sediment. However, the use of a solvent required the addition of a solvent control in each test. Solvent control treatments were

Table 1 Physicochemical properties of substitute phenylamine antioxidants diphenylamine (DPA), N-Phenyl-1-naphthylamine (PNA), N-(1,3-dimethylbutyl)-N’-phenyl-1,4phenylenediamine (DPPDA), and 4,40 -Methylene-bis(N-sec-butylaniline) (MBA). Physiochemical properties were modeled using USEPA EPI Suite. log KOW

log KOC

Solubility in water (25  C) (mg/L)

Henry's law constant (Pa$$ m3/mol)

122-39-4

3.3

2.9

63

0.106

N-phenyl-1-napthylamine (PNA)

90-30-2

4.5

4.5

9.0

0.010

N-(1,3-dimethylbutyl)-N’-phenyl-1,4-phenylenediamine (DPPDA)

793-24-8

4.7

4.4

2.8

0.003

4,4’-methylene-bis[N-sec-butylaniline] (MBA)

5285-60-9

6.1

4.9

0.07

0.0002

Name

CAS #

Diphenylamine (DPA)

Chemical structure

252

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

prepared by adding the same volume of acetone to the sediment that was used in the SPA treatments, and the control treatments received no SPA or solvent additions. Spiked sediments were mixed in 1-L amber jars for 24 h on rollers (Wheaton Industries Inc., Millville, NJ, USA) at 22 ± 2  C. After mixing, the solvent was allowed to evaporate from the spiked sediment by placing the 1-L jars with caps removed in a fume hood for 48 h. Following solvent evaporation, sediment was allowed to equilibrate at 4 ± 2  C for approximately three weeks. Sediment was then placed in test vessels with overlying culture water and aerated for 7 d before organisms were added. Culture water was City of Burlington (ON, Canada) tap water that had been dechlorinated. Tables S4 and S5 contain the physicochemical properties of the culture water and concentrations of selected contaminants, respectively. Water quality parameters (dissolved oxygen, conductivity, temperature, ammonia, and pH) were monitored in all treatments at the initiation and conclusion of each test and are summarized in Table S6 e S8. 2.2. Hyalella azteca 2.2.1. Acute water-only tests H. azteca used in this study were cultured at ECCC's Centre for Inland Waters in Burlington, Ontario, following procedures described in Borgmann et al. (1989). Juvenile amphipods aged 7e11 d were exposed to each SPA in water for 96 h. Exposure solutions with varying concentrations of SPA were prepared by spiking culture water with a concentrated solution of SPA in acetone. Culture water had been aerated for 24 h before spiking. Test vessels (250-mL glass beakers) contained 200 mL of exposure solution that was gently aerated throughout the test and a 3-cm2 piece of cotton gauze. Vessels were placed in an environmental chamber at 23 ± 2  C with a photoperiod of 16 h light: 8 h dark regime (~200 lux). Three test vessels were prepared for each treatment. A control containing only culture water and solvent control containing culture water spiked with the same volume of acetone used to prepare the exposure solution (<0.1% by volume) were included in each test. Nominal concentrations of SPA in exposure solutions (four treatments) used in each test are presented in Table S9. Ten amphipods were added to each test vessel for experiments with DPA, PNA, and MBA and fifteen amphipods were added to each test vessel for experiments with DPPDA, since a greater number of juvenile amphipods were available when the DPPDA test was conducted. Each test vessel of amphipods was fed 2.5 mg of ground TetraMin® fish food flakes at the initiation of the test. Samples of the exposure solutions were taken at the initiation and conclusion of the test for analyses of SPAs and stored at 80  C before being analyzed. After 96 h of exposure, the number of surviving amphipods was counted. 2.2.2. Chronic sediment tests The sediment tests with H. azteca and T. tubifex described below are based on ASTM (2010). As in the acute tests, juvenile amphipods aged 7e11 d were used in testing. Test vessels (600-mL glass beakers) contained 50 mL of sediment spiked with varying concentrations of SPAs and 350 mL of overlying water that were gently aerated. Overlying water was not changed over the course of the test. Test vessels were maintained in an environmental chamber at 23 ± 2  C with a photoperiod of 16 h light: 8 h dark (~200 lux) for the 7-d vessel aeration and subsequent 28-d test. The nominal concentration of SPA in sediment of each treatment (five treatments) for each test is presented in Table S10. Seven test vessels were prepared for each treatment. A control treatment consisting of sediment without the addition of SPA and a solvent control treatment consisting of sediment spiked with a volume of acetone equal

to the volume of stock solutions used to spike sediments were included in each test. Six test vessels in each treatment received 15 juvenile amphipods after the vessels had been aerated for 7 d. One test vessel from each treatment was sacrificed in order to sample water and sediment for chemical analysis at the start of the test before the addition of amphipods. Water and sediment were also sampled from replicate test vessels for each treatment at the conclusion of each test. Water and sediment samples were stored at 80  C before being analyzed. Each test vessel received 2.5 mg of ground TetraMin® fish food flakes twice a week in the first two weeks of the test, 2.5 mg of food three times in the third week, and 5 mg of food three times in the final week. After 28 d, surviving juvenile amphipods from six replicate vessels per treatment were counted and placed in a pre-weighed aluminum dish for drying. After being dried to constant weight at 60  C, dishes were weighed to determine growth and production of biomass for each replicate. Growth of amphipods for each replicate was determined by dividing the total dry mass of surviving amphipods by the total number of survival amphipods (i.e., g dw/amphipod). Production of biomass, which is a combined effect of survival and growth, was calculated for each replicate by dividing the total dry mass of surviving amphipods by the total number of amphipods in each replicate at the initiation of the experiment (i.e., 15). 2.3. Tubifex tubifex T. tubifex used in this study were taken from a permanent culture that is maintained at ECCC's Centre for Inland Waters. The worms were cultured in Long Point marsh sediment following procedures described in Milani et al. (2003). Test vessels (1-L glass jar) contained 100 mL of mixed sediment with varying concentration of SPA and 750 mL of overlying water, which was not changed over the course of the test. Nominal concentrations of SPA in the sediment of each treatment (five treatments) for each test are provided in Table S11. Test vessels were gently aerated and stored in a dark growth chamber at 23 ± 2  C for the 7-d vessel preparation and subsequent 28-d test. Five replicate vessels were prepared for each treatment. A control and solvent control treatment were also included in each test, as described above for the sediment tests with H. azteca. Water and sediment were sampled from one replicate at the initiation of the test to confirm the concentration of SPA and one replicate at the conclusion of the test. Prior to analyses, water and sediment samples were stored at 80  C. Mature worms with visible gonads were isolated from the culture for use in testing and four worms were added to each of three replicate test vessels for each treatment. Worms were not fed throughout the test as the sediment contained a sufficient amount of organic carbon to support survival of the worms and reproduction. Following the 28d exposure, the sediment in each test vessel was passed through a 500-mm and 250-mm sieve sequentially in order to remove adult worms, juvenile worms, and cocoons. Worms and cocoons were transferred from each sieve to separate Petri dishes and observed under a dissecting microscope. Adult worms were counted, and observations were made on visibility of gonads and overall health of the worms. Juvenile worms >500 mm and <500 mm in size and full and empty cocoons were also counted (Fig. S1). 2.4. Analysis of SPAs The analysis of SPAs in water and sediment is described in greater detail in Balakrishnan et al. (2016) and Prosser et al. (2017). Water samples were diluted with methanol and spiked with internal standard (13C-labeled caffeine). Sediment samples were freeze-dried and sub-samples were extracted using acetone and ultrasound-assisted extraction. Extracts were filtered and

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

reconstituted in methanol for analysis using liquid chromatography and tandem mass spectrometry (LC-MS/MS). Samples were analyzed using a XEVO-TQS tandem LC triple quadrupole mass spectrometer (Waters Corp., Milford, MA, USA) equipped with a Z-Spray electrospray ionization (ESI) source and operated in the positive-ion mode. The MS was operated in multiple reaction monitoring (MRM) and the MS apparatus was attached to an UPLC system (Waters Corp.) with a Kinetex C18 column (2.1 mm  100 mm, 2.6-mm pore; Phenomenex, Torrance, CA, USA). Water (0.1% formic acid; pH 3) and methanol (0.1% formic acid) were used as mobile phase solvents for gradient elution. All MS responses presented throughout this report were normalized against the labeled internal standard (internal standard quantification). Method detection limits (MDL) for DPA, PNA, DPPDA, and MBA in overlying water were 0.4, 0.07, 0.01, and 0.1 mg/ L and 0.4, 0.08, 0.01, and 0.1 mg/L, for H. azteca and T. tubifex tests, respectively. MDLs for DPA, PNA, DPPDA, and MBA in sediment were 0.3, 0.05, 0.2, and 0.3 ng/g, respectively. All spiked recoveries were between 80 and 120% and precision was within 15%. 2.5. Statistical analysis Non-linear regression using measured concentrations at the initiation and conclusion of the tests in overlying water and/or sediment was conducted to determine concentrations of SPAs causing 10, 25, and 50% mortality (e.g., LC10, LC25, and LC50) or a 10, 25, and 50% reduction in an endpoint relative to the control treatment (e.g., EC10, EC25, and EC50). LCs and ECs and associated standard errors and 95% confidence intervals were determined by fitting data to a 4-parameter log-logistic model (LL.4) with the drc package in R (Ritz and Streibig, 2005; R Development Core Team, 2016). Quantal data (e.g., mortality) were fit to the LL.4 model with the lower and upper limits set to 0 and 1, respectively, and based on a binomial distribution as opposed to the normal distribution used for continuous data (e.g., production of biomass, number of cocoons). The R coding used to determine LCs and ECs is presented in the SI. LCs and ECs were estimated using measured concentrations at the initiation and conclusion of the tests due to the decline in concentration of SPAs over the course of the tests. The true effect concentration will lie within the range of LCs and ECs reported. LCs and ECs estimated from measured concentrations at the initiation of the test are used in discussion and the LCs and ECs estimated from measured concentration at the conclusion of the test are presented in the SI. One-way analysis of variance (ANOVA) was used to identify significant differences among treatments (a ¼ 0.05). If significant differences were identified, a post-hoc Dunnett's test was conducted to determine which treatments were significantly different from the control treatment. Brown-Forsythe tests and Shapiro-Wilk tests were conducted to confirm that variance was equal between treatments and data were normally distributed, respectively. ANOVA was conducted using SigmaStat (Systat Software Inc., San Jose, CA, USA). 3. Results and discussion 3.1. Hyalella azteca 3.1.1. Acute water-only tests Measured concentrations in water were lower than nominal concentrations. The mean percent difference between nominal concentrations in water and measured concentrations for DPA, PNA, DPPDA, and MBA were 13, 35, 22, and 67%, respectively (Table S12). SPAs were not detected in water-only controls or solvent controls. The concentration of MBA was below the detection

253

limit in the three lowest treatments. This is likely related to the relatively low solubility of MBA in water, which made it difficult to ensure that all of the solid MBA was dissolved. The concentration of all SPAs in water declined over the course of experiment; DPPDA declined the most, with a mean percent difference of 84% (Table S12). Mortality was not significantly different between the control and solvent control treatments for any SPA (p > 0.01) (Table S9). Survival was 80% in all replicates of control and solvent control treatments (Table S9). DPA (LC50 ¼ 1440 mg/L) was the least toxic to juvenile amphipods relative to PNA and DPPDA (LC50 ¼ 109e250 mg/L) (Table 2). The greatest concentration of MBA, i.e., 22 mg/L, tested resulted in <10% mortality, therefore, the LC10, 25, and 50 were greater than 22 mg/L. A number of studies have investigated the acute toxicity of individual SPAs but few studies have compared the toxicity of different SPAs to a single organism. The LC50s for H. azteca exposed to DPA correspond with effects observed in acute studies with different species. Murin et al. (1997) investigated the effect of DPA on growth of the green algae Pseudokirchneriella subcapitata and immobilization of the freshwater crustacean Daphnia magna. P. subcapitata and D. magna were exposed to DPA for 48 and 72 h, respectively. The NOEC and EC50 values for growth inhibition of P. subcapitata were 370 and 2170 mg/L, respectively, and the EC50 value for immobilization of D. magna was 2000 mg/L. Acute LC50s for the fish species Pimephales promelas and Oryzias latipes exposed to DPA ranged from 2200 to 4000 mg/L (Tonogai et al., 1982; Geiger et al., 1990). H. azteca was more sensitive to PNA and DPPA than previously studied species. LC50s for 48-h and 96-h exposures to PNA for D. magna and Oncorhynchus mykiss, respectively, ranged from 300 to 740 mg/L (Sikka et al., 1981). Acute LC50s for D. magna exposed to DPPDA ranged from 400 to 1000 mg/L (Monsanto, 1978, 1984). The acute toxicity of MBA to aquatic organisms has not previously been investigated. In the current study, the greatest concentration that could be dissolved in water, i.e., 22 mg/L, did not cause significant mortality (Table 2 & Table S9). The modeled solubility of MBA in water is 70 mg/L (Table 1), but that concentration could not be achieved without solid MBA being visible in the solution. It would appear that an acute LC10 for H. azteca could not be reached without exceeding the solubility of MBA in water (Table S9). The glochidia (larvae) of the freshwater mussel species Lampsilis siliquoidea were exposed to DPA, PNA, DPPA, and MBA, which resulted in 48-h EC50s of 5951, 606, 439, and 258 mg/L, respectively (Prosser et al., 2017). The 48-h EC50 of 258 mg/L for MBA is greater than the modeled solubility in water of 70 mg/L (Table 1) and greater than the highest concentration achieved in the water-only exposures with H. azteca described above. Prosser et al. (2017) was able to achieve a concentration of DPPDA in water greater than the modeled solubility by introducing MBA to water in an acetone solution and mixing the water for 24 h on a stir plate. Similar 48-h EC50s were observed for another freshwater mussel species Lampsilis fasciola (Prosser et al., 2017). The trend in acute toxicity among the SPAs was similar for glochidia as was observed in H. azteca in the current study, but the glochidia were in general less sensitive to SPAs than H. azteca. This is contrary to previous studies that have shown that glochidia of freshwater mussels are more sensitive to some toxicants (e.g., copper, ammonia) relative to H. azteca (Wang et al., 2007), again underlining species specific sensitivities. 3.1.2. Chronic sediment tests Concentrations of all SPAs in overlying water declined considerably over the 28 d of the tests. For example, concentrations of DPA in overlying water of all treatments were below the MDL at the conclusion of the test (Table S13). Concentrations in overlying

254

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

Table 2 Measured concentrations of substituted phenylamine antioxidants (SPA) in water at the initiation of the test causing 10, 25, or 50% mortality in Hyalella azteca in 96-h wateronly tests and associated 95% confidence intervals. The number in brackets is the standard error for each LC. SPA

LC10 (mg/L)

95% CI

LC25 (mg/L)

95% CI

LC50 (mg/L)

95% CI

DPA PNA DPPDA MBA

665 (98) 81 (7) 104 (15) >22

473e856 83e105 73e134 e

979 (103) 94 (6) 161 (17) >22

778e1181 83e105 128e194 e

1440 (130) 109 (5) 250 (20) >22

1190e1700 100e118 210e290 e

water were greatest in DPA treatments (Table S13), which corresponds with DPA having the greatest solubility in water among the four SPAs (Table 1). MBA had the lowest concentrations in overlying water, which corresponds with it having the least solubility in water (Table 1). Concentrations of SPAs in the sediment of control and solvent control replicates were below MDLs (Tables S13 and 14). Measured concentrations of SPAs in sediment declined considerably (40e99%) from nominal concentrations at spiking over the course of equilibration until the addition of biota 28 d after spiking (Tables S13 and S14). The decline of SPAs in sediment following spiking was also observed by Prosser et al. (2017), who followed the same protocol for sediment spiking described in the current study. The concentration continued to decline over the course of the 28d exposure of biota. The percent difference between concentrations measured at day 0 compared to day 28 ranged from 20 to 93% (Tables S13 and S14). The data presented in the current study and by Prosser et al. (2017) would suggest that SPAs in sediment and overlying water are susceptible to biotic and/or abiotic degradation, which may result in a significant decline over a 28- to 56-d period. There is also the possibility that a fraction of the SPAs partitioned into organic matter within the sediment became unextractable using the method described in this study (Northcott and Jones, 2000; Rice et al., 2004). The current study and Prosser et al. (2017) are the first to document the fate of SPAs in sediment and overlying water. There were no significant differences between effect endpoints in the control versus the solvent control treatments (p < 0.05) (Table S10). Survival of amphipods was 87% in all replicates of control and solvent control treatments, except for one control replicate in the MBA test, which had a survival of 73% (Table S10). The 28-d LC50s based on the concentration in overlying water for DPA and PNA were an order of magnitude lower than 4-d LC50s described above (Tables 2 and 3). For PNA, the mean concentration in overlying water in the highest treatment was 10.7 mg/L and 97% mortality occurred in this treatment (Tables S10 and S13). However, in the second highest treatment, the concentration of PNA in overlying water was below the MDL (<0.07 mg/L) and less than 10% mortality occurred (Tables S10 and S13). Therefore, the 28-d LC50 for PNA based on the concentration in overlying water was reported as > 0.07 mg/L and <10.7 mg/L (Table 3). The acute-to-chronic ratio for mortality (96-h LC50/28-d LC10) (ACR) of DPA, PNA, and DPPDA to H. azteca, using measured concentrations in overlying water, were 80, >10, and 42, respectively (Tables 2 and 3). The ACR for PNA was >10 due to the LC50 calculated from the concentration in overlying water from the sediment test being <10.7 mg/L (Table 3). As described above, MBA was only detected in the overlying water of highest treatment (e.g., mean 10.7 mg/L) but 50% mortality occurred below the highest treatment. The ACR for MBA could not be determined, as 50% mortality did not occur in the highest treatments of the acute and chronic tests (Table 3). The ACRs for DPA, PNA, and DPPDA are in the higher percentiles of the range of ACRs for aquatic organisms exposed to organic chemicals reported in the literature (Lange et al., 1998; Ahlers et al., 2006). Ahlers et al. (2006) identified that primary

aniline derivatives could be a structural alert for an increased probability of a relatively high ACR. DPA, PNA, and DPPDA are secondary aromatic amines (Table 1), that when degraded could theoretically produce primary aniline derivatives. The concentration of SPAs declined considerably over the course of the tests (Table S12 e S14), a portion of this decline could be explained by degradation, which could have resulted in primary aniline production. Further investigation on the degradation products of SPAs under test conditions would need to be conduct to confirm this hypothesis. Relative toxicity of SPAs differed depending on how exposure was assessed. From the perspective of the concentration in overlying water, DPA was the least toxic of the four SPAs, but from the perspective of the concentration in sediment, DPA exhibited the greatest toxicity to H. azteca (Table 3). This observed opposite relationship in toxicity is a function of variation in the solubility of SPAs in water and the route of exposure to H. azteca (Table 1). The main route of exposure of sediment-associated SPAs to H. azteca will be via the overlying water and particles in the overlying water as these amphipods are epibenthic (Wang et al., 2004; Borgmann et al., 2005). The amphipods interacted with the surface of the sediment to some extent but they tended not to burrow into the sediment and primarily foraged in the overlying water of the system. DPA has the greatest solubility in water among the four SPAs, therefore, it has the greatest ability to partition into the overlying water from sediment, resulting in relatively higher concentrations in overlying water, and lower concentrations in sediment, than the other three SPAs (Table 1). This explains why from the perspective of the concentration in sediment that DPA exhibits the greatest toxicity relative to the other three SPAs (Table 3). Sedimentassociated DPA is more bioavailable than the other three SPAs. However, comparing the toxicity of SPAs based on measured concentration in overlying water is more relevant in the case of H. azteca, as the main route of exposure for H. azteca is through the overlying water. The greater toxicity of PNA and DPPDA relative to DPA from the perspective of the concentration in overlying water is likely a function of the greater hydrophobicity of PNA and DPPDA (Tables 1 and 3). The concentration of SPAs in overlying water that results in a significant effect (e.g., EC50) would suggest that the mechanism of action of this class of chemicals is not receptormediated but likely a result of narcosis. The greater octanol-water partition coefficient (Kow) of PNA and DPPDA relative to DPA suggests that they have a greater ability to partition into and disrupt cellular membranes, therefore, resulting in greater nonspecific toxicity (Veith et al., 1983). Solubility in water also explains why the treatment with the highest concentration of MBA did not cause 50% mortality in juvenile H. azteca (Table S10). The potential bioavailability of MBA was considerably less than the other three SPAs due to its relatively low solubility in water and greater Koc (Table 1). The mean mortality of the treatment with the greatest concentration of MBA was 15% (Table S10), therefore, the 28-d LC50s for MBA based on concentrations in overlying water and sediment were >1.62 mg/L and >403 mg/g dw, respectively (Tables 3 and S13). A similar relationship in toxicity from the perspective of

Table 3 Measured concentrations of substituted phenylamine antioxidants (SPA) in overlying water and sediment at the initiation of the test causing 10, 25, and 50% mortality or reduction in growth or production of biomass and associated 95% confidence intervals for Hyalella azteca in 28-day sediment tests. The number in brackets is the standard error for each LC/EC. Matrix

Endpoint

LC/EC10

95% CI

LC/EC25

95% CI

LC/EC50

95% CI

DPA

Overlying water (mg/L)

Mortality Growtha

18 (3) 171 (1.7  105)

12e24 3.5  105 e 3.5  105

30 (3) 179 (1.9  105)

24e37 3.8  105 e 3.8  105

50 (4) 188 (2.1  105)

Production of biomass Mortality Growtha Production of biomass Mortality Growth Production of biomass Mortality Growtha Production of biomass Mortality Growth Production of biomassa Mortality Growtha Production of biomass Mortality Growth Production of biomass Mortality Growth Production of biomass

33 (6) 9 (1) 46 (29) 17 (2) e e e 11 (3) 0.3 (4) 49 (21) 6 (1) 129 (22) 14 (10) 68 (6) 232 (197) 152 (214) e e e e e e

21e45 7e11 13e105 13e22 e e e 9e17 8e8 7e91 5e7 85e1733 e 56e80 165e630 280e584 e e e e e e

45 (5) 14 (1) 49 (30) 21 (2) e e e 33 (6) 4 (31) 62 (NA) 9 (1) 141 (14) 15 (5) 96 (6) 242 (192) 164 (116) e e e e e e

34e56 12e15 11e109 18e25 e e e 22e45 59e67 NA 8e10 112e170 4e26 84e109 147e631 71e398 e e e e e e

60 (6) 22 (1) 52 (31) 27 (2) <10.7 <10.7 <10.7 99 (15) 52 (118) 77 (NA) 13 (1) 155 (10) 16 (1) 135 (7) 253 (192) 176 (5) >1.62 >1.62 >1.62 >403 >403 >403

43e57 4.2  105 e 4.2  105 48e72 19e24 10e115 24e30 e e e 69e129 188e291 NA 11e14 135e175 15e17 122e149 136e641 165e187 e e e e

Sediment (mg/g dw)

PNA

Overlying water (mg/L)b

Sediment (mg/g dw)

DPPDA

Overlying water (mg/L)

Sediment (mg/g dw)

MBA

Overlying water (mg/L)c

Sediment (mg/g dw)c

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

SPA

NA: drc package not able to estimate standard error or 95% confidence interval due to relatively high variability in the data. a 95% confidence interval included zero. b Unable to calculate LC/ECs because of concentrations of PNA in overlying water were < MDL in all treatments except the highest concentration tested and the effect at the highest concentration tested was >50%. c LC/ECs could not be calculated and LC/EC50s were greater than the concentration of MBA in the overlying water and sediment in the treatment with the highest concentration of MBA due to 50% mortality or 50% reduction in production of biomass not being observed in this treatment.

255

256

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

Table 4 Measured concentrations of substituted phenylamine antioxidants (SPA) in sediment at the initiation of the test causing 10, 25, and 50% mortality in adult Tubifex tubifex and 10, 25, and 50% reduction in production of juvenile worms >500 mm, juvenile worms >500 mm, full cocoons, empty cocoons, total cocoons or total juveniles and associated 95% confidence interval for 28-day sediment tests. The number in brackets is the standard error for each EC. SPA

Endpoint

LC/EC10 (mg/g dw)

95% CI

LC/EC25 (mg/g dw)

95% CI

LC/EC50 (mg/g dw)

95% CI

DPA

Adult mortality Juvenile worms >500 mm Juvenile worms <500 mma Total juvenilesa Full cocoonsa Empty cocoons Total cocoons Adult mortalitya Juvenile worms >500 mma Juvenile worms <500 mm Total juveniles Full cocoonsa Empty cocoons Total cocoons Adult mortality Juvenile worms >500 mma Juvenile worms <500 mma Total juvenilesa Full cocoonsa Empty cocoonsa Total cocoons Adult mortality Juvenile worms >500 mma Juvenile worms <500 mm Total juvenilesa Full cocoonsa Empty cocoonsa Total cocoonsa

31 (12) 8 (3) 8 (4) 8 (4) 43 (39) 7 (3) 31 (7) 85 (763) 3 (2) 6 (2) 4 (2) 90 (107) 19 (7) 34 (9) 63 (2) 3 (231) 2 (20) 3 (206) 42 (27) 1 (1) 14 (2) 223 (92) 0.2 (0.6) 14 (6) 11 (7) 14 (15) 29 (14) 13 (7)

7e54 1e14 1e17 0e15 41e127 1e14 16e46 1410e1580 2e8 0.7e11 - 0.1e9 135e315 5e33 15e53 58e69 482e488 39e44 431e436 14e99 2e4 10e19 43e403 1.0e1.4 1e27 3e25 17e45 0.4e58 2e28

32 (3) 11 (3) 13 (4) 12 (4) 47 (42) 13 (4) 34 (3) 89 (798) 5 (3) 10 (3) 8 (3) 98 (113) 28 (6) 46 (6) 65 (2) 3 (290) 3 (27) 3 (255) 47 (23) 2 (2) 20 (2) 386 (103) 0.9 (1.6) 19 (4) 16 (6) 26 (14) 60 (19) 21 (7)

25e39 4e17 4e22 4e19 41e135 6e21 28e40 1480e1650 2e11 4e16 2e14 140e336 14e41 34e58 60e70 607e613 54e60 532e539 1e95 2e7 16e23 183e588 2.3e4.2 10e28 5e28 3e56 19e100 7e36

34 (6) 15 (3) 19 (4) 18 (3) 51 (45) 23 (4) 36 (13) 93 (840) 9 (4) 18 (3) 15 (3) 106 (122) 40 (6) 62 (6) 67 (7) 4 (364) 4 (37) 4 (314) 52 (18) 5 (3) 27 (2) 668 (168) 3.6 (3.9) 25 (6) 23 (4) 50 (22) 122 (25) 36 (7)

22e46 9e20 11e27 11e24 45e146 15e31 8e64 1550e1740 0.4e18 12e25 8e22 151e363 27e53 50e74 53e80 761e769 74e81 655e663 14e90 2e11 23e31 339e997 4.5e11.7 13e38 14e33 3e96 69e175 21e51

PNA

DPPDA

MBA

a

95% confidence interval(s) included zero.

concentration in sediment and water was observed in juvenile freshwater mussels chronically exposed to sediment-associated SPAs (Prosser et al., 2017). Based on concentrations of SPAs in water or sediment, the toxicity of SPAs was similar for juvenile freshwater mussels relative to H. azteca, except for MBA. It appears from the perspective of the concentration in sediment that H. azteca (28-d LC50: >430 mg/g dw) were much less sensitive to MBA compared to juvenile mussels (28-d LC50: 109 mg/g dw), but this is misleading, as the concentration in overlying water in the amphipod test did not exceed 1.62 mg/L (Table 3). The concentration in overlying water for the juvenile mussel test exceeded 10 mg/L and resulted in a 28-d LC50 (water) of 5 mg/L (Prosser et al., 2017). This illustrates the importance of measuring the concentration of a chemical in the sediment and overlying water when conducting sediment toxicity tests, particularly for epibenthic biota, in order to characterize all potential major pathways of exposure. Growth (total dry weight/surviving amphipods) of H. azteca was a less sensitive endpoint compared to mortality; 28-d EC50s for growth were 4e10 times greater than the 28-d LC50s (Table 3). This was a result of surviving amphipods at higher concentrations growing larger than surviving amphipods at lower concentrations (Table S10). The surviving individuals at higher concentrations where some mortality occurred likely grew larger due to reduced competition for food relative to lower concentrations where very little mortality occurred. Production of biomass (total dry weight/ total amphipods at initiation of test) was of similar sensitivity as mortality, with little difference between the 28-d ECs compared to LCs (Table 3). The relatively low ratio between 28-d LCs and 28d ECs (biomass production) indicates that the mode of action of SPAs is likely non-specific narcosis (Marinkovic et al., 2011). The endpoints growth and biomass production were considerably more variable than mortality, which results in wider confidence intervals for these endpoints. Mortality appeared to be a more reliable

endpoint for risk assessment relative to growth and biomass production. 3.2. Tubifex tubifex There were considerable differences between nominal concentration of SPAs spiked into sediment and measured concentration in sediment when worms were added to test vessels (Table S14). Concentrations of SPAs declined from 15 to 97% among the treatments during equilibration and aeration of test vessels, similar to the decline observed in test vessels used in H. azteca experiments (Table S13). Concentrations of SPAs in sediment and overlying water declined over the course of the 28-d experiment as was observed for H. azteca and L. siliquoidea (Prosser et al., 2017) (Table S14). SPAs were not detected in the sediment or overlying water of control and solvent control test vessels. Tubifex tubifex were more sensitive to sediment-associated SPAs than H. azteca and juvenile mussels (Tables 3 and 4) (Prosser et al., 2017). Lethal and effect concentrations were only estimated based on the concentrations of SPAs in the sediment, as this species is endobenthic. T. tubifex remains buried in the sediment in order to feed on organic matter. Thus, the major pathway of exposure for T. tubifex will be through sediment-bound SPAs and SPAs present in pore water that is consumed by the worm. Survival of adult worms in control and solvent control replicates was 100%, except for 50% survival in one solvent control replicate of the MBA test (Table S11). Effect endpoints were not significantly different between control and solvent control treatments, except for adult mortality in the MBA test, which was due to the mortality in one replicate of the solvent control (Table S11). The most sensitive measured effect endpoint for T. tubifex was the production of juvenile worms >500 mm in size, with 28-d EC50 for production of juveniles >500 mm was 15, 9, 4, and 3.6 mg/g dw for DPA, PNA, DPPDA, and

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

MBA, respectively (Table 4). Production of juveniles >500 mm was the most sensitive endpoint but it had considerably larger confidence intervals relative to production of juveniles <500 mm (Table 4). In general, EC50s for production of juveniles <500 mm may be a more reliable endpoint than the other endpoints of juvenile production. Total cocoon production was not as sensitive an endpoint as production of juveniles <500 mm but it exhibited smaller confidence intervals relative to the other juvenile and cocoon endpoints (Table 4). Adult mortality was the least sensitive endpoint for T. tubifex in this study, highlighting the importance of considering reproductive endpoints (Table 4). A particular level of exposure may not result in adult mortality but may inhibit reproduction, which would result in decline of the population. Variation in the toxicity of SPAs to adult worms was similar to the trend that was observed for H. azteca and juvenile mussels based on the concentration in sediment (Prosser et al., 2017). As discussed above, differences in chemical properties (e.g., solubility in water, Koc) of the four SPAs (Table 1) cause variation in their bioavailability, which translates into variation in toxicity. For example, DPA was the most toxic based on the concentration in sediment because it has the greatest solubility in water and least affinity for organic carbon in sediment (Table 1); therefore, DPA was the most bioavailable for interaction with adult worms relative to other SPAs. The measured concentration of DPA and other SPAs in overlying water of T. tubifex tests (Table S14) illustrates the positive relationship between bioavailability and Kow (Table 1). DPA exhibited the greatest toxicity to adult worms followed by DPPDA, PNA, and MBA (Table 4). However, this trend did not hold when reproductive endpoints were considered. The 28-d EC50s for total cocoon and juvenile production were similar across the SPAs, with the exception of total juveniles for DPPDA, which was a more sensitive endpoint relative to the other SPAs (Table 4). A similar threshold for reduction in reproductive output could be due to the SPAs causing a general reduction in worm consumption of organic matter at a similar concentration in sediment, which results in a reduction of resources for reproduction (Leppanen and Kukkonen, 1998). 4. Conclusions The concentration of all four SPAs in sediment and water declined over the course of equilibration, aeration of test vessels, and exposure of biota. It was not clear whether this was due to abiotic or biotic degradation and/or irreversible sorption to organic matter in the sediment. The relatively rapid dissipation of SPAs in these laboratory studies does call into question their bioavailability and/or persistence in freshwater ecosystems, further research should be conducted to investigate their persistence and bioavailability in the environment. Tubifex tubifex were more sensitive to sediment-associated SPAs compared to H. azteca in the current study and L. siliquoidea in a previous study (Prosser et al., 2017). The endobenthic life history of T. tubifex is the likely reason for this variation in toxicity compared to epibenthic biota. T. tubifex spend their life buried within the sediment, consuming particles to acquire nutrition from organic matter. Therefore, the magnitude of exposure to T. tubifex would be greater than epibenthic species for which the main route of exposure is via overlying water. The life history of these two species governs their sensitivity to sediment-associated SPAs. For the most part, the toxicity of the different sediment-associated SPAs to H. azteca and T. tubifex was driven by bioavailability, which was a function of solubility and Koc. A positive relationship between Kow and toxicity from the perspective of concentration of SPAs in water was observed but the opposite relationship was observed from the perspective of concentration in sediment. A conclusion on whether sediment-associated SPAs represent a hazard to benthic

257

invertebrates cannot be made at this time, as there are no measurements of environmental concentration of SPAs in water and/or sediment. Further research is necessary to characterize the magnitude and frequency of exposure of aquatic ecosystems to SPAs. The data generated in the current study will aid in completing a characterization of the hazards of SPAs to aquatic biota, which in combination with future data on the magnitude and frequency of potential exposures could be used to conduct an ecological risk assessment of this class of chemicals. Acknowledgements Authors would like to thank Lisa Brown and Amanda Hedges for providing H. azteca and Jennifer Unsworth for providing T. tubifex for testing. The Government of Canada's Chemicals Management Plan provided the funding for this study. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.chemosphere.2017.04.066. References Ahlers, J., Riedhammer, C., Vogliano, M., Ebert, R., Kuhne, R., Schuurmann, G., 2006. Acute to chronic ratios in aquatic toxicity - variation across trophic levels and relationships with chemical structure. Environ. Toxicol. Chem. 25, 2937e2945. ASTM, 2010. Standard Test Method for Measuring the Toxicity of Sedimentassociated Contaminants with Freshwater Invertebrates E1706-05. ASTM International, West Conshohocken, PA. Balakrishnan, V., Parrott, J., Bartlett, A., Milani, D., Gilroy, E., de Solla, S., Prosser, R., Gillis, P., 2016. Chemicals Management Plan Progress Report - the Environmental Fate, Distribution and Effects of Substituted Phenylamine Antioxidants (SPAs). Developing Analytical Methods, Investigating Toxicity and Evaluating Bioaccumulation. Final Report to Science and Risk Assessment Directorate of Environment and Climate Change Canada. Environment and Climate Change Canada, Ottawa, ON, Canada. Borgmann, U., Grapentine, L., Norwood, W., Bird, G., Dixon, D., Lindeman, D., 2005. Sediment toxicity testing with the freshwater amphipod Hyalella azteca: relevance and application. Chemosphere 61, 1740e1743. Borgmann, U., Ralph, K., Norwood, W., 1989. Toxicity test procedures for Hyalella azteca, and chronic toxicity of cadmium and pentachlorophenol to H. azteca, Gammarus fasciatus, and Daphnia magna. Arch. Environ. Contam. Toxicol. 18, 756e764. Drzyzga, O., Jannsen, S., Blotevogel, K., 1995. Toxicity of diphenylamine and some of its nitrated and aminated derivatives to the luminescent bacterium Vibrio fischeri. Ecotoxicol. Environ. Saf. 31, 149e152. ECHA, 2008. European Union Risk Assessment Report for Diphenylamine. European Chemical Agency, Helsinki, Finland. ECHA, 2016. REACH. http://echa.europa.eu/regulations/reach. April 4, 2016. European Commission, 2016. Legislation on Plant Protection Products (PPPs). http:// ec.europa.eu/food/plant/pesticides/legislation/index_en.htm. April 4, 2016. Geiger, D., Brooke, L., Call, D., 1990. Acute Toxicities of Organic Chemicals to Fathead Minnows (Pimephales promelas). Graduate thesis. Centre for Lake Superior Environmental Studies, University of Wisconsin. Government of Canada, 2014. Overview of the Chemicals Management Plan. http:// www.chemicalsubstanceschimiques.gc.ca/index-eng.php. April 4, 2016. Health Canada, 2009. Pest Management Regulatory Agency. http://www.hc-sc.gc. ca/ahc-asc/branch-dirgen/pmra-arla/index-eng.php. April 4, 2016. Lange, R., Hutchinson, T., Scholz, N., Solbe, J., 1998. Analysis of the ECETOC Aquatic Toxicity (EAT) Databased II - comparison of acute to chronic ratios for various aquatic organisms and chemical substances. Chemosphere 36, 115e127. Leppanen, M., Kukkonen, J., 1998. Factors affecting feeding rate, reproduction and growth an oligochaete Lumbriculus variegatus. Hydrobiologia 377, 183e194. Maier, C., Calafut, T., 1998. Additives. Polyproylene: the Definitive User's Guide and Databook. Plastics Design Library, Norwich, NY, USA. Marinkovic, M., Verweij, R., Nummerdor, G., Jonker, M., Kraak, M., Admiraal, W., 2011. Life cycle responses of the midge Chironomus riparius to compounds with different modes of action. Environ. Sci. Technol. 45, 1645e1651. Milani, D., Reynoldson, T., Borgmann, U., Kolasa, J., 2003. The relative sensitivity of four benthic invertebrates to metals in spiked-sediment exposures and application to contaminated field sediment. Environ. Toxicol. Chem. 22, 845e854. Monsanto, 1978. Aquatic Toxicity Studies with Santoflex 13. Monsanto, Unpublished study AB-78e121A. Monsanto, 1984. Santoflex 13 Degradation Toxicity Test with Daphnia Magna. Monsanto, Unpublished study MO-92e9050 ES-80-SS-11. Murin, M., Gavora, J., Drastichova, I., Duskova, E., Madsen, T., Torslov, J., Damborg, A.,

258

R.S. Prosser et al. / Chemosphere 181 (2017) 250e258

Tyle, H., Pedersen, F., 1997. Aquatic hazard and risk assessment of two selected substances produced in high volumes in the Slovak Republic. Chemosphere 34, 179e190. Northcott, G., Jones, K., 2000. Experimental approaches and analytical techniques for determining organic compound bound residues in soil and sediment. Environ. Pollut. 108, 19e43. OECD, 2004. SIDS Initial Assessment Report - N-(1,3-dimethylbutyl)-N'-phenyl-1,4phenylenediamine. Organization for Economic Co-oporation and Development, Paris, France. Prosser, R., Gillis, P., Holman, E., Ikert, H., Schissler, D., Toito, J., Gilroy, E., Campbell, S., Bartlett, A., Milani, D., Parrott, J., Balakrishnan, V., 2017. Effect of sediment-associated substituted phenylamine antioxidants on three life stages of the freshwater mussel Lampsilis siliquoidea. Aquat. Toxicol. (in press). R Development Core Team, 2016. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Rice, P., Anderson, T., Coats, J., 2004. Effect of sediment on the fate of metolachlor and atrazine in surface waters. Environ. Toxicol. Chem. 23, 1145e1155. Ritz, C., Streibig, J., 2005. Bioassay analysis using R. J. Stat. Softw. 12, 1e22. Sikka, H., Pack, E., Sugatt, R., Banerjee, S., Rosenberg, A., Simpson, B., 1981. Environmental Fate and Effects of N-phenyl-1-naphthylamine and its Disposition and Metabolism in the Rate Report No. AFOSR-TR-81-0703. Syracuse Research

Company, Syracuse, NY, USA. Tonogai, Y., Ogawa, S., Ito, Y., Iwaida, M., 1982. Actual survey of TLm values of environmental pollutants, especially on amines, nitriles, aromatic N compounds and artificial dyes. J. Toxicol. Sci. 7, 193e203. USEPA, 2015. TSCA Chemical Substance Inventory. https://www.epa.gov/tscainventory. April 4, 2016. USFDA, 2015. Pharmaceutical Guidance, Compliance & Regulatory Information. http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/. April 4, 2016. Veith, G., Call, D., Brooke, L., 1983. Structure-toxicity relationship for the fathead minnow, Pimephales promelas: narcotic industrial chemicals. Can. J. Fish. Aquatic Sci. 40, 743e748. Wang, F., Goulet, R., Chapman, P., 2004. Testing sediment biological effects with the freshwater amphipod Hyalella azteca: the gap between laboratory and nature. Chemosphere 57, 1713e1724. Wang, N., Ingersoll, C., Hardesty, D., Ivey, C., Kunz, J., May, T., Dwyer, F., Roberts, A., Augspurger, T., Kane, C., Neves, R., Barnhart, M., 2007. Acute toxicity of copper, ammonia, and chlorine to glochidia and juveniles of freshwater mussels (Unionidae). Environ. Toxicol. Chem. 26, 2036e2047. WHO, 1998. Concise International Chemical Assessment Document - N-phenyl-1Naphthylamine. World Health Organization, Geneva, Switzerland.