Effects of kraft pulp mill condensates on plasma testosterone levels in mummichog (Fundulus heteroclitus)

ARTICLE IN PRESS

Ecotoxicology and Environmental Safety 67 (2007) 140–148 www.elsevier.com/locate/ecoenv

Effects of kraft pulp mill condensates on plasma testosterone levels in mummichog (Fundulus heteroclitus)$ Kevin S. Shaughnessya, Andrew M. Belknapb,c, L. Mark Hewittc, Monique G. Dube´d, Deborah L. MacLatchya, a

Department of Biology and Canadian Rivers Institute, University of New Brunswick, Saint John, NB, Canada b Department of Environmental Biology, University of Guelph, Guelph, ON, Canada c National Water Research Institute, Environment Canada, Burlington, ON, Canada d Toxicology Centre, University of Saskatchewan, Saskatoon, SK, Canada Received 26 June 2005; received in revised form 4 April 2006; accepted 6 April 2006 Available online 5 June 2006

Abstract Past studies at a bleached kraft pulp mill located in Saint John, NB, Canada have shown that chemical recovery condensates significantly depress circulating and gonadal steroids in mummichog (Fundulus heteroclitus), an endemic fish species. In the present study, compounds were extracted from the condensates, and a toxicity identification evaluation (TIE) was conducted to characterize the hormonally active substances (HASs) in the condensates. Extracts were fractionated by high performance liquid chromatography (HPLC) and mummichog were exposed to the fractions in a 7-day bioassay. Plasma testosterone was measured for each sex following exposure. Responses in fish exposed to the whole extract at 1% v/v were not as pronounced as had been observed previously (female plasma testosterone was reduced by 16% in the current study compared to 75% previously). Dose–response experiments showed an exposure concentration of 4% v/v was required to elicit significant plasma steroid reductions. Despite these responses, individual condensate fractions actually increased steroid levels in mummichog, which suggests that multiple HASs may need to act synergistically or additively to elicit effects, and if separated, the compounds may have different hormonal activity. The HASs in question caused a reduction in male gonad size at 4% v/v, and have sex-dependent mechanisms of action (males were more responsive to exposure than females). r 2006 Elsevier Inc. All rights reserved. Keywords: Fundulus heteroclitus; Bleached kraft pulp mill; Condensates; Hormonally active substances; Testosterone

1. Introduction Effluents from some bleached kraft pulp mills have been shown to reduce gonad size and egg production, cause significant delays in maturation, and/or depress reproductive steroid hormones in different species of fish (Sandstro¨m et al., 1988; Munkittrick et al., 1991, 1994; Leblanc et al., 1997; Dube´ and MacLatchy, 2000; Sepu´lveda et al., 2004). In Canada, recent studies at a large number of pulp $ Ethical review: All fish exposures were approved by the University of New Brunswick (Saint John)’s Local Animal Care Committee according to the guidelines of the Canadian Council on Animal Care. Corresponding author. Fax: +1 506 648 5650. E-mail address: [email protected] (D.L. MacLatchy).

0147-6513/$ - see front matter r 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2006.04.003

mills have found a common pattern of increased condition, increased liver size and reduced gonad size (metabolic disruption) in wild fish collected from the receiving environment, relative to reference sites (Munkittrick et al., 2002). These effects are presumed to be caused by the presence of endocrine-active contaminants in the effluent (McMaster, 2001). However, despite a great deal of effort, it is unclear what components are responsible. Research on characterizing the causative compounds has been hindered by the complex nature of bleached kraft mill effluents and the variability in sensitivity of different fish species (Munkittrick, 2003). Cause and effect relationships between the components of the effluent and fish responses are difficult to determine given the range of chemicals present (e.g., wood-derived, production-derived,

ARTICLE IN PRESS K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

treatment-derived), and their potential combinations. As well, protocols in only a few fish species have been described that can link endocrine responses to chemical exposures (McMaster et al., 1996; Parrott and Wood, 2002; MacLatchy et al., 2003). Studies done at Irving Pulp and Paper Ltd. (IPP), a bleached kraft pulp mill in Saint John, New Brunswick, Canada, were some of the first to implement an investigation of cause approach (Hewitt et al., 2003a) for identifying specific hormonally active waste streams within the mill (Dube´ and MacLatchy, 2000, 2001; Hewitt et al., 2002). Those studies have focused attention on the chemical recovery condensate stream as a key contributor to reproductive responses in an estuarine killifish, the mummichog (Fundulus heteroclitus). Using laboratory exposures, Dube´ and MacLatchy (2001) showed that mummichog had significantly depressed circulating testosterone levels when exposed to an environmentally relevant concentration (1% v/v) of the condensates for 21 d. The source was later confirmed using a shorter 7-day bioassay (Hewitt et al., 2002). Installation of a reverse osmosis (RO) system for condensate treatment significantly raised the threshold for the steroid depression effect from the condensates and in final effluent (Dube´ and MacLatchy, 2000). The advantage of this waste stream identification approach was that it determined biologically active substances at their source of origin in the mill, thereby facilitating efforts to identify them. To further isolate compounds associated with hormone depressions, a solid phase extraction (SPE) technique was developed for the condensates at the Saint John mill (Hewitt et al., 2002). This procedure optimizes recovery of bioavailable compounds responsible for depressing circulating sex steroids in the 7-day bioassay in mummichog (Hewitt et al., 2002). GC/MS analyses indicated that bioactive fractions contain chemicals consistent with kraft lignin functionalities (Hewitt et al., 2002), indicating that the sources of the hormonally active substances (HASs) are naturally derived from the wood feedstock. Recently, we have developed an HPLC protocol to separate the bioactive condensate extract into specific fractions for use in toxicity identification evaluations (TIE) (Belknap et al., 2004). It is highly desirable to obtain information about which compounds are eliciting effects on the reproductive performance of fish in Canadian receiving environments. Current effluent treatment systems were implemented in the mid-1990s to meet regulatory requirements for the reduction of adsorbable organic halogens, total suspended solids, effluent acute toxicity and biological oxygen demand. The reductions in gonad size were first noted prior to these process changes at a number of mills (Munkittrick et al., 1992a, 1994) and have since been observed at a larger number of sites through three cycles of the Environmental Effects Monitoring (EEM) program. While the reductions in gonad size need to be addressed, it has proven to be a complex problem to investigate. At sites

141

where in-depth studies of wild fish responses have been conducted (Munkittrick et al., 1992a, 1994) alterations in sex steroid levels have coincided with gonad size depressions. The linkage of chemical recovery condensates to steroid depressions and the development of a short-term mummichog bioassay have now provided the opportunity to investigate factors causing steroid depressions and to determine if they are also related to reduced gonad size. The objective of the present study was to fractionate bioactive condensate extracts at the Saint John mill to isolate and identify condensate components associated with testosterone depressions. 2. Materials and methods 2.1. Condensate collection A total of three TIE experiments, and two bioassay refinement experiments, were conducted in this study. Depending on doses and number of replicates for each experiment, appropriate volumes of condensates (from 6 to 35 L) were collected during softwood production runs from the 5th effect evaporator (i.e., the feed to the RO apparatus) in solvent-rinsed 20-L stainless-steel canisters. Condensates were sent to Burlington, ON, Canada by overnight courier for extraction and fractionation.

2.2. Solid phase extraction Condensates were extracted using the optimized protocol developed by Hewitt et al. (2002). Condensate preparation and specific details regarding the extraction process for this study are described in Belknap et al. (2006). Following extraction (250 mL aliquots of condensates were processed at a time; number of aliquots processed depended on experiment), condensate extracts for each treatment were reconstituted into 4 mL of methanol in TIE experiments 1 and 2, and into 60 mL of methanol for TIE experiment 3. For exposures 1 and 2, 1 mL of the extract was equivalent to 1 or 1.5 L, respectively, of the original condensate. One-hundred and sixty microliters of each extract was added to 16 L of tank water for 1% v/v (exposure 1) and 1.5% v/v (exposure 2) final concentrations. The concentration– response experiment was conducted with the SPE-2 extract (0.5%, 1%, 2%, and 4% v/v final concentrations), and fish were dosed with 2 mL of a 0.32 L/mL equivalent extract for TIE exposure 3 (4% v/v final tank concentration in 16 L water). For the TIE experiments, part of the SPE-2 extract was kept for each exposure as a positive control in exposures 1 and 2 while whole (untreated) condensates were employed as the positive control for TIE exposure 3. Clean laboratory RO water was processed through exactly the same procedure as condensate samples to generate the negative control treatment (method blank SPE-2) for all experiments.

2.3. HPLC fractionation SPE-2 extracts were injected onto a HPLC system and all extracts were fractionated using the solvent program and conditions optimized by Belknap et al. (2004). Fraction cutoff times for the six fractions were adjusted between experiments by calibration with photo diode arraydetectable peaks in maxplot mode (maximum absorbance from 180 to 800 nm wavelengths detected) for SPE-2 extracts. Compound retention times were 0–5, 5–9, 9–17, 17–29, 29–37, and 437 min for fractions 1–6, respectively (Belknap et al., 2004). For exposure 3 in the TIE, fractions 1 and 2 were combined, and fractions 3 and 4 were combined, which reduced the number of treatments and increased the number of tank replicates (statistical power) in the study design.

ARTICLE IN PRESS 142

K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

Fractions, all of which contained different proportions of water from the HPLC mobile phase, were evaporated to just-dryness by either vortex nitrogen evaporation (approx. 10 L/min flow rate; 55 1C; model ZW6403C Turbovap Evaporator, Zymark Corp., Hopkinton, MA, USA) for exposures 1 and 2, or by evaporation at ambient temperature from broad glass pans (exposure 3). Fractions were made up to appropriate final condensate equivalent volumes (see above) in methanol, and were sent frozen to the University of New Brunswick (Saint John, New Brunswick, Canada) in screw-capped vials by overnight courier for mummichog exposures.

2.4. Bioassay Mummichog were collected from clean reference estuaries in the Northumberland Strait, New Brunswick (Horton’s Creek, and Black River) and transported to laboratory facilities at UNB Saint John, where they were acclimated in laboratory stock tanks (approx. 750 L) for at least 2 weeks prior to experimentation. Three adult mummichog (minimum 65 mm) of each sex were weighed (to 0.01 g) and randomly allocated to 20–24 glass aquaria, each containing 16 L of water. Fish were held in the experimental aquaria for 1 week prior to the commencement of treatments. Fish were kept at a 14-h L:10-h D photoperiod (late spring conditions) and fed commercial trout pellets (approx. 1% of total body weight) daily. Each tank was aerated to maintain dissolved oxygen levels above 70%, and water quality parameters (dissolved oxygen, temperature, salinity, conductivity) were recorded daily prior to feeding using a YSI meter (Yellow Springs Instruments, Yellow Springs, OH, USA). Temperature was approximately 16 1C for each exposure. A static system was used for the bioassay and water was completely renewed and extracts/ effluents were re-administered every 24 h for 7 days. Extracts were dissolved in methanol; reference tanks received equivalent doses of methanol [i.e., 0.0125% v/v (2 mL methanol in 16 L water) in TIE experiment 3; highest methanol dose used in the study] for control purposes. Previous trials have shown that methanol administered at or above the concentrations used in this study have no effect on the reproductive endpoints measured in the bioassay (D. MacLatchy, unpublished data). Fish were naturally cycling prior to TIE exposures 1 and 2, and underwent artificial recrudescence for 6 weeks prior to the RO feed potency evaluation, SPE-2 dose–response, and TIE exposure 3, using a protocol developed by MacLatchy et al. (2003). Artificial recrudescence was achieved by an increase in water temperature and photoperiod from approximately 4 1C and short day length (9:15 h light:dark) to 16 1C and long day length (14:10 h light:dark) over 6 weeks. A noticeable increase in male coloration was observed, indicating that fish were reproductively active prior to experimentation.

2.5. TIE exposures 1 and 2 Exposure 1 commenced on April 19, 2003, and exposure 2 commenced on September 8, 2003. The objective was to determine which fractions of the SPE-2 bioactive condensate extract cause significant depressions in plasma testosterone, at 1% v/v, and 1.5% v/v for exposures 1 and 2, respectively. The experimental design consisted of a positive control (total SPE-2), a reference (lab blank; distilled water processed through SPE and dissolved in methanol), and six treatment groups (one of six SPE-2 fractions mixed with methanol). All groups consisted of three replicate aquaria (16 L).

impact on responses observed. Therefore, for two treatments (1% batch and 4% batch), RO feed was collected once from the mill, and the same stock of effluent was used throughout the exposure. For the other two treatments (1% daily and 4% daily), RO feed was collected and administered daily. The reference group received water alone, and four replicate aquaria were used for each treatment.

2.7. SPE-2 dose–response exposure This exposure commenced on February 23, 2004. The primary objective was to determine if higher concentrations of SPE-2 would elicit responses in plasma testosterone, and the secondary objective was to determine if SPE-2 is an accurate representation of the RO feed. Treatments consisted of 0.5%, 1%, 2%, and 4% SPE-2, and the reference group received methanol. Four replicate aquaria were used for each treatment.

2.8. TIE exposure 3 Exposure 3 commenced on April 27, 2004. The objective was to determine the responses in plasma testosterone following exposure to specific fractions at 4% v/v. SPE-2 treatments consisted of fractions 1 and 2 combined, fractions 3 and 4 combined, fraction 5, and fraction 6 for a total of four treatment groups, each consisting of four aquaria. Reference tanks received the laboratory blank, and whole condensates (RO feed) were used as a positive control.

2.9. Sampling protocol Details of the sampling protocol are provided by MacLatchy et al. (2005). Briefly, at the termination of exposure, fish were anaesthetized with 0.05% tricaine methane sulfonate (Syndel International Inc., Vancouver, BC, Canada), weighed (0.01 g), and measured for standard length (mm). Blood was collected using 26 3/8 gauge needles (Beckton–Dickenson) inserted into the caudal vasculature. Blood was centrifuged (2400g) to isolate plasma, which was frozen at 20 1C and preserved for later ether extraction and radioimmunoassay (RIA). Fish were killed by spinal severance following blood collection and the gonads and liver were dissected and weighed (to 0.001 g) for determination of somatic indices. Condition factor was calculated as Cf ¼ (total weight (g)/standard length3)  100. Liversomatic and gonadosomatic indices were calculated as LSI or GSI ¼ (tissue weight (g)/total weight (g))  100).

2.10. Radioimmunoassay For determination of circulating testosterone levels, plasma was thawed and steroid hormones were isolated from the blood proteins by ether extraction (McMaster et al., 1992). Steroid hormones were re-suspended in 1 mL of phosgel buffer, frozen at 20 1C, and later thawed for RIA. The RIA protocols used in this study, modified from McMaster et al. (1992), are described in MacLatchy et al. (2005). The testosterone antibody used has less than 0.1% cross-reactivity with closely related steroids and was purchased from Medicorp, Montreal, QC, Canada. 3H-Radiolabelled testosterone was purchased from Amersham Pharmacia Biotech, Baie d’Urfe´, QC. Unlabelled testosterone was purchased from Sigma-Aldrich (Oakville, ON). All plasma testosterone samples were measured in duplicate. Intra-assay variability was 5.8% and interassay variability was 7.5%.

2.6. RO feed potency evaluation 2.2. Data analysis This exposure was designed to retest the potency of the condensates (RO feed or 5th effect condensates). The experiment commenced on February 13, 2004 and the primary objective was to determine the responses in plasma testosterone following exposure to 1% and 4% RO feed. A secondary objective was to determine if potential compound degradation throughout the duration of the 7-day exposure would have an

Circulating testosterone was measured for each sex in all exposures; however, sexes were separated for statistical analyses. Prior to statistical analyses for determination of treatment effects, Dixon tests were used for determination of the presence/absence of outliers (Kanji, 1993; MacLatchy et al., 2005), and outliers were removed from further analysis.

ARTICLE IN PRESS K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148 A nested analysis of variance (ANOVA; Pp0:05) was used to determine the presence/absence of tank effects (nested factor), and treatment effects (fixed-effect factor) (Zar, 1999). In the absence of tank effects, fish were used as the units of replication, and significant differences between SPE-2 fractions and the reference treatment in the TIE exposures were assessed using a Dunnet’s test. For the bioassay refinement exposures, a Tukey’s post hoc test was used for assessment of treatment differences. If assumptions of normality failed, data were log transformed, and re-tested. If assumptions were not met following transformation, the non-parametric Kruskal–Wallis test was used for the assessment of treatment differences. Analysis of co-variance (ANCOVA) was used for comparisons of liver weight and gonad weight with total fish weight as the co-variate of organ weight. To evaluate condition, an ANCOVA was conducted for total fish weight with standard length as the covariate. Statistical analyses were conducted using Sigmastat 3.0 (SPSS, Chicago, IL, USA) and Systat 9.0 (Systat Software Inc., Richmond, CA, USA).

3. Results 3.1. TIE exposures 1 and 2 (1% and 1.5% condensate equivalents) There were no significant treatment differences in fish length, weight, gonad weight, condition factor, or liver weight (data not shown) for males or females for either exposure. Fish were naturally recrudescing at the termination of exposure 1 and regressed at the termination of exposure 2. In exposure 1, mean plasma testosterone in female fish exposed to SPE-2 was significantly reduced in comparison to reference levels (P ¼ 0:049). Mean male plasma testosterone levels in the SPE-2 treatment were not significantly different than reference levels (P ¼ 0:07). Female fish showed a significant reduction in circulating testosterone in the FR6 treatment (P ¼ 0:045). Males showed no significant differences relative to reference values for any of the fractions (FR1 through FR6). In exposure 2, mean plasma testosterone did not significantly differ in any treatment relative to reference levels for either sex. Mean gonad weight was approximately 8-fold lower in females, and 2.5-fold lower in males than levels observed in exposure 1. 3.2. RO feed potency evaluation There were no significant differences in fish length, weight (Table 1), condition factor, or liver weight (data not shown) among treatments. Mean gonad weight was significantly higher in males exposed to 1% batch than males exposed to 4% batch (P ¼ 0:028), but no other gonad weight differences existed among treatments for either sex (Table 1). Fish were in a state of gonadal recrudescence at the end of the exposure. Female mummichog did not show any responses in circulating testosterone when exposed to 1% or 4% condensate treatments. Significant decreases in circulating testosterone were observed in male mummichog exposed to 4% RO feed whether collections were made daily or only once at the beginning of the exposure (Fig. 1).

143

3.3. SPE-2 dose–response exposure There were no differences in length, weight (Table 1), condition factor, or liver weight (data not shown) among treatments for either males or females. There was no significant difference in female gonad weight among treatments, but male gonad weight was significantly lower in the 4% SPE-2 treatment than the reference treatment (P ¼ 0:046) (Table 1). Fish were in a state of gonadal recrudescence at the termination of exposure. Female mummichog did not show any responses in circulating testosterone levels for any treatment. Male fish exposed to 4% SPE-2 showed the greatest depression in plasma testosterone relative to the other treatments when compared to control values. The only other treatment that caused a significant reduction in male plasma testosterone was 1% SPE-2. No responses were observed in male fish exposed to 0.5% or 2% SPE-2 (Fig. 2). 3.4. Exposure 3 (4% condensate equivalents) There were no significant treatment differences in fish length, weight (Table 1), condition factor, or liver weight (data not shown) for males or females. There was no significant difference in male gonad weight among treatments, however, mean female gonad weight was significantly lower in the FR1,2 treatment than in the FR 3,4 treatment (Table 1). Fish were pre-spawning at the termination of exposure. No significant responses in plasma testosterone levels were observed in females. Male fish showed significant increases in plasma testosterone for FR3,4 and FR6. Plasma testosterone was significantly reduced compared to reference fish in males exposed to 4% RO feed, which was the positive control treatment for this exposure (Fig. 3). Mean gonad weights were not different from those observed in exposure 1. 4. Discussion In the present study we exposed mummichog to the various condensate fractions generated by softwood production (comprised of spruce, pine and balsam fir feedstocks). The level of circulating testosterone (the precursor to 11-ketotestosterone in males and estradiol in females) was selected as the primary endpoint for assessment of bioactivity, as depressions in plasma testosterone are a common response of fish to pulp mill effluents (Munkittrick et al., 1992a, 1998; McMaster et al., 1996, Karels et al., 1998) and mummichog have consistently responded to androgen-active compounds and effluents in past studies (Dube´ and MacLatchy, 2000, 2001; Hewitt et al., 2002; Sharpe et al., 2004). Experiments indicated that the condensates are capable of depressing plasma testosterone levels in mummichog, but the potency has been reduced since previous studies (Dube´ and MacLatchy, 2000, 2001; Hewitt et al., 2002). Additionally, sexdependent differences in responses of plasma testosterone

ARTICLE IN PRESS K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

144

Table 1 Mean (71SE) length, weight and gonadosomatic indices (GSI) of male and female Fundulus heteroclitus February 2004: RO feed potency experiment Sex

Variable

Treatments REF

1% daily

1% batch

4% daily

4% batch

F

Length (mm) Weight (g) GSI (%)

72.071.57a 7.8170.81a 7.271.11a

73.271.60a 8.3570.55a 7.3471.01a

73.571.20a 7.6270.33a 6.7771.04a

72.371.65a 7.3670.64a 7.4571.61a

74.971.81a 8.0570.66a 7.7471.69a

M

Length (mm) Weight (g) GSI (%)

74.071.22a 7.6770.50a 2.0470.14ab

72.771.44a 7.1670.45a 2.0670.15ab

73.870.94a 7.2470.20a 2.3870.13a

72.671.82a 6.9870.53a 1.9470.17ab

73.671.00a 7.1070.39a 1.8170.13b

REF

0.50%

1%

2%

4%

April 2004a: SPE-2 dose response experiment Sex

Variable

Treatments

F

Length (mm) Weight (g) GSI (%)

76.573.13a 9.3971.14a 6.2570.90a

74.472.31a 9.2370.86a 9.6671.93a

76.372.51a 9.3271.05a 6.8871.28a

73.971.33a 8.3670.55a 8.6671.42a

73.872.93a 8.6570.98a 6.0771.25a

M

Length (mm) Weight (g) GSI (%)

70.072.51a 6.6370.83a 2.2170.12a

72.471.47a 6.5670.40a 2.1370.16a

70.571.28a 6.5270.38a 2.0270.14a

73.173.00a 7.2271.05a 2.3170.12a

72.972.23a 7.0870.58a 1.8670.13b

April 2004b: TIE experiment testing fractions Sex

Variable

Treatments REF

FR1,2

FR3,4

FR5

FR6

POS CON

F

Length (mm) Weight (g) GSI (%)

81.872.98a 8.9070.62a 15.371.55ab

79.372.72a 8.6770.68a 10.771.55b

77.972.37a 8.6470.50a 20.372.71a

76.272.70a 8.4470.45a 15.371.55ab

77.572.91a 8.2870.53a 13.471.77ab

77.872.88a 8.5670.49a 17.871.99ab

M

Length (mm) Weight (g) GSI (%)

71.471.30a 6.4270.39a 2.8670.18a

71.871.70a 6.5670.47a 2.7070.17a

71.471.43a 6.6270.48a 2.6070.15a

71.971.62a 6.7270.51a 2.6470.19a

73.172.56a 7.3670.97a 2.6270.16a

72.271.04a 6.7370.29a 2.5570.17a

Values with different letters are significantly different (Pp0:05) from each other within a variable within an experiment. February 2004: REF refers to the reference treatment (water alone); daily treatments refer to RO feed that was collected daily throughout the exposure; batch treatments refer to RO feed that was collected once at the beginning of exposure. April 2004a: REF refers to distilled water sent through SPE and dissolved in methanol; 0.5%, 1%, 2%, and 4% refer to the concentration of SPE-2 added to each tank. April 2004b: REF refers to distilled water sent through SPE and dissolved in 4% methanol; FR1,2 through FR6 refers to specific SPE-2 fractions separated by HPLC; POS CON refers RO feed.

levels were observed following exposure to both the condensate extract and the RO feed. In exposure 1 (April, 2003), the SPE-2 extract did not significantly depress steroids in males, and although reductions were significant in females, the magnitude of the depression was minimal compared to those seen previously (Hewitt et al., 2002). Specifically, previously we have observed significant 75% and 58% reductions in plasma testosterone levels in males and females exposed to 1% (v/v) SPE-2, respectively. In the current study, mean plasma testosterone levels in fish exposed to 1% (v/v) SPE2 were not significantly different than reference levels. The experiment was repeated in September of 2003, using a 1.5% v/v condensate equivalent concentration. Since plasma testosterone levels were not affected in any of the treatments, including the whole extract, we speculated that

either fish were not responsive to the HASs during this exposure, or the composition of condensates had changed since previous work. Fish in exposure 2 were regressed and plasma testosterone was at or near basal levels. It is possible that effects may be diminished when fish are regressed, however, we observed significant depressions in plasma testosterone in regressed male and female mummichog in November 1999 (MacLatchy et al., 2005) and December 2000 (Hewitt et al., 2002), which indicates that mummichog can be responsive to HASs when regressed. These observations led us to retest (February 2004) the potency of the RO feed (whole condensates before RO treatment) at 1% and 4% v/v. In order to reduce inherent variability we increased our sample size (number of tanks per treatment) from three to four for all exposures. Dube´ and MacLatchy (2001) observed a 58% reduction in

ARTICLE IN PRESS K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148 p = 0.015

0.6

Females Males

Plasma testosterone (ng /ml)

a

0.6

A

ab

0.4

b A

Females Males

p = 0.045 *

*

0.5

a

A

A

b A

0.2

Plasma testosterone (ng/ml)

0.8

145

0.4

0.3

0.2 * 0.1

0.0

0.0

REF

1% daily

1% batch

4% daily

4% batch

Treatment

p = 0.032

Plasma testosterone (ng /ml)

a

Females Males

0.8 ab

ab

0.6 b b 0.4

A A

A

A

A

0.2

0.0 REF

0.5%

1%

4% RO feed FR1,2

FR3,4

FR5

FR6

Treatment

Fig. 1. Mean (71SE) plasma testosterone levels (ng/mL) in female and male Fundulus heteroclitus exposed to water alone (REF), 1% and 4% RO feed collected daily (1% daily, 4% daily), and 4% RO feed collected once at beginning of exposure (1% batch, 4% batch) for 7 days in February 2004. For each sex, bars with different letters are significantly different (Pp0:05).

1.0

REF

2%

4%

Treatment (percent SPE-2)

Fig. 2. Mean (71SE) plasma testosterone levels (ng/mL) in female and male Fundulus heteroclitus exposed to methanol (REF), and four different concentrations of SPE-2 (0.5%, 1%, 2%, 4%) for 7 days in March 2004. For each sex, bars with different letters are significantly different (Pp0:05).

plasma testosterone in males and a 39% reduction in females when exposed to 1% RO feed for 7 days, while the current study observed no significant effects at the same concentration (Fig. 1). Exposure to 4% RO feed elicited responses more comparable to those observed by Dube´ and MacLatchy (2001) at 1% v/v. Consultation with mill personnel revealed no process changes had been made to their chemical recovery phase of operations between 2000 and 2004. It is possible that unknown/uncontrolled alterations in the RO feed may have occurred during this time, but these cannot be specifically identified by mill

Fig. 3. Mean (71SE) plasma testosterone levels (ng/mL) in female and male Fundulus heteroclitus exposed to methanol (REF), 4% RO feed, and one of four SPE-2 fractions at 4% v/v for 7 days in April 2004. For each sex, bars with asterisks are significantly different from the reference (REF) (Pp0:05).

personnel (John Leroy, environmental engineer, personal communication). A secondary question that was addressed in this exposure was the possibility of degradation of the HASs from the time of collection to the end of exposure. It was unknown whether the chemical constituents within the condensates are altered over the course of time from collection through SPE-2 processing and the 7-day exposure (approximately 3 weeks). The exposure was designed such that two treatments (1% and 4% v/v; daily) involved the collection of RO feed on a daily basis, and two treatments (1% and 4% v/v; batch) involved a single collection at the beginning of the exposure. Male plasma testosterone was significantly reduced in both 4% treatments (Fig. 1), indicating that potential compound degradation over the course of exposure does not seem to have an impact on the reproductive endpoints examined in this study. Mean plasma testosterone levels in fish exposed to 1% RO feed were not different than those in reference fish, indicating that the potency of the condensates had decreased since 2001. Although males demonstrated concentration-dependent responses in steroid levels, plasma testosterone levels were unaffected in females (Fig. 1). This finding indicates that that the HASs possess sex-dependent mechanisms of action in mummichog. In the SPE-2 dose–response exposure (Fig. 2), fish were exposed to one of four different concentrations of the condensate extract (0.5%, 1%, 2%, and 4% SPE-2). These concentrations were specifically chosen so that direct comparisons in effects could be made with the previous RO feed potency evaluation and was conducted immediately following the prior experiment. It should be noted that reference GSI values were significantly higher for both males and females in the later exposure (Table 1) despite a

ARTICLE IN PRESS 146

K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

difference in sampling of just 9 d. This indicated that fish were rapidly investing in gonadal growth during this period. Fish responses were highly comparable between the two bioassay refinement experiments (Figs. 1 and 2). Male plasma testosterone levels were approximately 2-fold lower than reference levels for both the extract (4% v/v) and the RO feed (4% v/v) treatments, indicating the SPE-2 extract is an accurate representation of RO feed. Male GSI in fish exposed to 4% SPE-2 was significantly lower than reference GSI levels, the only exposure in this study to show an effect of a treatment on gonad size. A 55% reduction in plasma testosterone was associated with a significant 16% reduction in GSI compared to reference levels. Early studies have documented a correlation between reduced GSI and depressed circulating steroid levels in wild fish downstream from a bleached kraft pulp mill (McMaster et al., 1991; Munkittrick et al., 1991, 1992a, b). This finding is important, as it indicates that the HASs in question are capable of affecting mummcihog reproductive capacity. Although 1% SPE-2 caused a significant reduction in plasma testosterone, this may not be a suitable concentration for further TIE experiments as the effect disappeared at 2% (Fig. 2). This phenomenon is consistent with findings from the preceding exposures in which 1% SPE-2 appeared to have a borderline effect in TIE exposure 1, and no effect in TIE exposure 2, while 1% RO feed treatments had no effect in the RO feed potency experiment. Males consistently showed testosterone reductions at 4% v/v for both the extract and the RO feed, providing further evidence to indicate that the potency of the RO feed has been reduced since previous studies. Further TIE experiments were, therefore, carried out at 4% v/v to ensure that the concentrations being used were capable of eliciting effects. The following changes were made in the design for TIE exposure 3 (April 2004) based on the conclusions made following the bioassay refinement exposures: (1) sample size was increased from three tanks to four tanks per treatment; (2) the concentration of the exposure was increased to 4% v/v; (3) RO feed was designated as the positive control group since it induced comparable effects to the extract at 4% v/v; and (4) the number of fractions was reduced from six (as in exposures 1 and 2) to four. This allowed for the recombining of some compounds into specific fractions that were separated in previous exposures and reduced the number of treatments required for the experiment. It is understood that compounds may elicit additive or synergistic effects within fractions. The intention was to make no direct comparisons of results to exposures 1 and 2 given the challenges in those experiments. If in the third TIE the specific fractions elicited significant changes in steroid levels, the compounds in these fractions would be the focus of follow-up studies (further TIE experiments). In exposure 3, two fractions caused significant increases in male plasma testosterone levels, while no responses were observed in females, which supports our previous conclu-

sion that the HASs possess sex-dependent (as well as concentration-dependent) mechanisms of action. Comparatively, we previously observed 4.5- and 4-fold increases in plasma testosterone in males exposed to 0.5% and 1% combined mill effluent, respectively, at IPP, while females showed no significant responses (Dube´ and MacLatchy, 2000). We have also observed depressions in plasma steroid levels in females exposed to high concentrations of ethinyl estradiol (4250 ng/mL), while low concentrations (o100 ng/mL) caused increases in plasma steroid levels (MacLatchy et al., 2003). In this study, male mummichog were more sensitive than females when exposed to whole condensates (RO feed) and the condensate extract (SPE-2). Robinson (1994) observed that in vitro production of plasma testosterone in fathead minnows (Pimephales promelas) responded to pulp mill effluent exposure in a concentration-dependent manner in males but not in females. Robinson (1994) also concluded that various factors including life stage, reproductive development, or changes in effluent quality over time may play significant roles in steroid responses following exposure. These conclusions are further supported by this study. The specific mechanisms of action of HASs in pulp mill condensates are currently unknown, however, it is understood that xenobiotics capable of altering sex steroids can have multiple hormonal activities (Van Der Kraak et al., 1992; Sohoni and Sumpter, 1998). Targets of HASs in pulp mill effluents may include the hypothalamo–pituitary– gonadal axis (Van Der Kraak et al., 1992), the steroidogenic pathway (Leusch and MacLatchy, 2003), plasma binding proteins (Hewitt et al., 2003b), receptor binding (Hewitt et al., 2000), and clearance (Hewitt et al., 1996). Further studies should focus on understanding mechanisms of action of the HASs in the condensates. The overall results from this study indicate that condensates from IPP are capable of depressing plasma testosterone levels in mummichog despite a reduction in potency since 2001. Individual condensate fractions did not reduce steroid levels in mummichog which suggests that multiple HASs may need to act synergistically or additively to elicit effects. Future studies should compare the hormonal activity of the SPE-2 fraction, the separated SPE-2 fractions and a recombined SPE-2 fraction to provide better insight into the effects of compound separation. We verified that the SPE-2 extract is an accurate representation of the whole condensates, indicating that the extraction process retains the HASs in question. Additionally, we determined that HASs within the condensates retain their potency from the time of collection and throughout experimentation, indicating that the potential for compound degradation is minimal within this time-frame. We also determined that each sex responds differently to condensate exposure. These sex-dependent effects of the condensates may provide clues to the mechanisms of action of the HASs. Lastly, we noted that the HASs in the condensates are capable of affecting gonad

ARTICLE IN PRESS K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

size in only 7 days of exposure. Compound identification at IPP will aid investigation of cause studies at other mills as putative HASs will be easier to trace once identified at this mill. Acknowledgments We thank J. Adams, C. Arens, M. Beyea, C. Blanar, S. Brasfield, K. Gormley, R. Peters, J. Ings, R. Kassie, R. Sharpe, G. Vallie´res, and L. Vallis for sampling and/or field collection assistance. We also thank D. Muir, P. Webber, and J. LeRoy from Irving Pulp and Paper for their cooperation and support. Funding was provided by a Natural Science and Engineering Research Council of Canada Collaborative Research and Development grant to DLM, LMH, and MGD; Irving Pulp and Paper Ltd.; and a Canadian Water Network grant to DLM (K. Munkittrick, PI). References Belknap, A., Shaughnessy, K.S., MacLatchy, D.L., Solomon, K., Hewitt, M.L., 2004. Method development for the identification of hormonally active components in bleached kraft chemical recovery condensates. In: Borton, D.L., Hall, T.J., Fisher, R.P., Thomas, J.F. (Eds.), Pulp and Paper Mill Effluent Environmental Fate and Effects. DEStech Publications, Inc., Lancaster, PA, USA, pp. 374–383. Belknap, A., Solomon, K.R., Dube´, M.G., MacLatchy, D.L., Hewitt, M.L., 2006. Identification of compounds associated with testosterone depressions in fish exposed to bleached kraft pulp and paper mill chemical recovery condensates. Environ. Toxicol. Chem. 25 (9), in press. Dube´, M.G., MacLatchy, D.L., 2000. Endocrine responses of Fundulus heteroclitus to effluent from a bleached-kraft pulp mill before and after installation of reverse osmosis treatment of a waste stream. Environ. Toxicol. Chem. 19, 2788–2796. Dube´, M.G., MacLatchy, D.L., 2001. Identification and treatment of a waste stream at a bleached-kraft pulp mill that depresses a sex steroid in the mummichog (Fundulus heteroclitus). Environ. Toxicol. Chem. 20, 985–995. Hewitt, M.L., Carey, J.H., Dixon, G.D., Munkittrick, K.R., 1996. Examination of bleached kraft mill effluent fractions for potential inducers of mixed function oxygenase activity in rainbow trout. In: Servos, M.R., Munkittrick, K.R., Carey, J.H., Van Der Kraak, G.J. (Eds.), Environmental Fate and Effects of Pulp and Paper Mill Effluents. St. Lucie Press, Delray Beach, FL, USA, pp. 79–94. Hewitt, M.L., Parrott, J.L., Wells, K.L., Calp, M.K., Biddiscombe, S., McMaster, M.E., Munkittrick, R., Van Der Kraak, G.J., 2000. Characteristics of ligands for the Ah receptor and sex steroid receptors in hepatic tissues of fish exposed to bleached kraft mill effluent. Environ. Sci. Technol. 34, 4327–4334. Hewitt, M.L., Smythe, S.A.M., Dube´, M.G., Gilman, C.I., MacLatchy, D.L., 2002. Isolation of compounds from bleached kraft mill recovery condensates associated with reduced levels of testosterone in mummichog (Fundulus heteroclitus). Environ. Toxicol. Chem. 21, 1359–1367. Hewitt, M.L., Dube´, M.G., Culp, J.M., MacLatchy, D.L., Munkittrick, K.R., 2003a. A proposed framework for investigation of cause for environmental effects monitoring. Hum. Ecol. Risk Assess. 19, 195–211. Hewitt, M.L., Pryce, A.C., Parrott, J.L., Marlatt, V., Wood, C., Oaks, K., Van Der Kraak, G.J., 2003b. Accumulation of ligands for aryl hydrocarbon and sex steroid receptors in fish exposed to treated effluent from a bleached sulfite/groundwood pulp and paper mill. Environ. Toxicol. Chem. 22, 2890–2897.

147

Kanji, G.K., 1993. 100 Statistical Tests. SAGE Publications, London, UK, 224pp. Karels, A.E., Soimasuo, M., Lappivaara, J., Leppa¨nen, H., Aaltonen, T., Mellanen, P., Oikari, A.O.J., 1998. Effects of ECF-bleached kraft mill effluent on reproductive steroids and liver MFO activity in populations of perch and roach. Ecotoxicology 7, 123–132. Leblanc, J., Coulliard, C.M., Breˆthes, J.-C.F., 1997. Modifications of the reproductive period in mummichog (Fundulus heteroclitus) living downstream from a bleached kraft pulp mill in the Miramichi Estuary, New Brunswick, Canada. Can. J. Fish. Aquat. Sci. 54, 2564–2573. Leusch, F.D.L., MacLatchy, D.L., 2003. In vivo implants of b-sitosterol cause reductions of reactive cholesterol pools in mitochondria isolated from gonads of male goldfish (Carassius auratus). Gen. Comp. Endocrinol. 134, 255–263. MacLatchy, D.L., Courtenay, S.C., Rice, C.D., Van Der Kraak, G.J., 2003. Development of a short-term reproductive endocrine bioassay using steroid hormone and vitellogenin endpoints in the estuarine mummichog (Fundulus heteroclitus). Environ. Toxicol. Chem. 22, 996–1008. MacLatchy, D.L., Gormley, K.L., Ibey, R.E.M., Sharpe, R.L., Shaughnessy, K.S., Courtenay, S.C., Dube´, M., Van der Kraak, G.J., 2005. A short-term mummichog (Fundulus heteroclitus) bioassay to assess endocrine response to hormone-active compounds and mixtures. In: Ostrander, G.K. (Ed.), Techniques in Aquatic Toxicology, vol. 2. CRC Press, New York, pp. 55–92. McMaster, M.E., 2001. A review of the evidence for endocrine disruption in Canadian aquatic ecosystems. Water Qual. Res. J. Can. 36, 215–231. McMaster, M.E., Van Der Kraak, G.J., Portt, C.B., Munkittrick, K.R., Sibley, P.K., Smith, I.R., Dixon, D.G., 1991. Changes in hepatic mixed function oxygenase (MFO) activity, plasma steroid levels and age at maturity of a white sucker (Catostomus commersoni) population exposed to bleached kraft pulp mill effluent. Aquat. Toxicol. 21, 199–218. McMaster, M.E., Munkittrick, K.R., Van Der Kraak, G.J., 1992. Protocol for measuring circulating levels of gonadal sex steroids in fish. Can. Tech. Rep. Fish. Aquat. Sci. 1836, 1–29. McMaster, M.E., Munkittrick, K.R., Van Der Kraak, G.J., Flett, P.A., Servos, M.R., 1996. Detection of steroid hormone disruptions associated with pulp mill effluent using artificial exposures of goldfish. In: Servos, M.R., Munkittrick, K.R., Carey, J.H., Van Der Kraak, G.J. (Eds.), Environmental Fate and Effects of Pulp and Paper Mill Effluents. St. Lucie Press, Delray Beach, FL, USA, pp. 425–437. Munkittrick, K.R., 2003. Environmental effects of effluents from pulp and paper mills. In: Northcote, T.G., Hartman, G.F. (Eds.), Fishes and Forests: Worldwide Watershed Interactions and Management. Blackwell Science, Oxford, pp. 336–361. Munkittrick, K.R., Portt, C.B., Van Der Kraak, G.J., Smith, I.R., Rokosh, D.A., 1991. Impact of bleached kraft mill effluent on population characteristics, liver MFO activity, and serum steroid levels of a Lake Superior white sucker (Catostomus commersoni) population. Can. J. Fish. Aquat. Sci. 48, 1371–1380. Munkittrick, K.R., McMaster, M.E., Portt, C.B., Van Der Kraak, G.J., Smith, I.R., Dixon, D.G., 1992a. Changes in maturity, plasma sex steroid levels, hepatic MFO activity and the presence of external lesions in lake whitefish exposed to bleached kraft mill effluent. Can. J. Fish. Aquat. Sci. 49, 1560–1569. Munkittrick, K.R., Van Der Kraak, G.J., McMaster, M.E., Portt, C.B., 1992b. Response of hepatic MFO activity and plasma sex steroids to secondary treatment of bleached kraft pulp mill effluent and mill shutdown. Eviron. Toxicol. Chem. 11, 1427–1439. Munkittrick, K.R., Van Der Kraak, G.J., McMaster, M.E., Portt, C.B., van den Heuvel, M.R., Servos, M.R., 1994. Survey of receiving water environmental impacts associated with discharges from pulp mills. II. Gonad size, liver size, hepatic EROD activity and plasma sex steroid levels in white sucker. Environ. Toxicol. Chem. 13, 1089–1101.

ARTICLE IN PRESS 148

K.S. Shaughnessy et al. / Ecotoxicology and Environmental Safety 67 (2007) 140–148

Munkittrick, K.R., McMaster, M.E., McCarthy, L.H., Servos, M.R., Van Der Kraak, G.J., 1998. An overview of recent studies on the potential of pulp-mill effluents to alter reproductive parameters in fish. J. Toxicol. Environ. Health B1, 347–371. Munkittrick, K.R., McGeachy, S.A., McMaster, M.E., Courtenay, S.C., 2002. Overview of freshwater fish studies from the pulp and paper environmental effects monitoring program. Water Qual. Res. J. Can. 37, 49–77. Parrott, J.L., Wood, C.S., 2002. Fathead minnow lifecycle tests for detection of endocrine-disrupting substances in effluents. Water Qual. Res. J. Can. 37, 651–667. Robinson, R.D., 1994. Evaluation and development of laboratory protocols for estimating reproductive impacts of pulp mill effluent on fish. Ph.D. Thesis, University of Guelph, Guelph, ON, Canada. Sandstro¨m, O., Neuman, E., Kara˚s, P., 1988. Effects of bleached pulp mill effluent on growth and gonad function in Baltic coastal fish. Water Sci. Technol. 20, 107–118.

Sepu´lveda, M.S., Gallagher, E.P., Weiser, C.M., Gross, T.S., 2004. Reproductive and biochemical biomarkers in largemouth bass sampled downstream of a pulp and paper mill in Florida. Ecotoxicol. Environ. Saf. 57, 431–440. 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. Sohoni, P., Sumpter, J.P., 1998. Several environmental estrogens are also anti-androgens. J. Endocrinol. 158, 327–339. Van Der Kraak, G.J., Munkittrick, K.R., McMaster, M.E., Portt, C.B., Chang, J.P., 1992. Exposure to bleached kraft pulp mill effluent disrupts the pituitary–gonadal axis of white sucker at multiple sites. Toxicol. Appl. Pharmacol. 115, 224–233. Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice-Hall, Upper Saddle River, NJ, pp. 303–311.