Science of the Total Environment 430 (2012) 119–125 Contents lists available at SciVerse ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv The occurrence of steroidal estrogens in south-eastern Ontario wastewater treatment plants Susanna K. Atkinson ⁎, Vicki L. Marlatt, Lynda E. Kimpe, David R.S. Lean, Vance L. Trudeau, Jules M. Blais Centre for Advanced Research in Environmental Genomics, Department of Biology, University of Ottawa, 30 Marie Curie, Ottawa, ON, Canada K1N 6N5 a r t i c l e i n f o Article history: Received 10 October 2011 Received in revised form 24 April 2012 Accepted 28 April 2012 Available online 25 May 2012 Keywords: Steroidal estrogen Biodegradation Wastewater treatment plant Removal efficiency Estrogenic potency a b s t r a c t We measured steroidal estrogens in wastewater in Ottawa and Cornwall (Ontario, Canada) to determine removal efficiency of these steroids during the treatment process, and whether removal varies during a seasonal cycle. Estrone (E1), 17β-estradiol (E2) and 17α-ethinylestradiol (EE2) were found at maximum concentrations in raw sewage (RS), at 104, 66.9 and 5.7 ng L− 1, respectively. For the Ottawa wastewater treatment plant (WWTP), there was sufficient data to show that E1 concentrations in RS correlated with both ambient air temperature and mean daily flow of the WWTP (R2 =0.792, p =0.003 and R2 =0.757, p =0.005). E1 removal was correlated with the percent difference in cBOD from RS to FE (final effluent) (R2 =0.435, p = 0.075). However estrogenic potency, as determined by a sensitive in vitro reporter gene assay, did not decrease during the water treatment process, suggesting that many estrogenic chemicals are conserved in FE. E1 and EE2 were found in river water, both upstream and downstream of the WWTPs, and at much lower concentrations than in FE. Our study demonstrates the persistence of steroidal estrogens and estrogenic potency in Ontario WWTP effluents and surface waters, and has uncovered temporal patterns of release that may be used to help predict risks to aquatic organisms in these environments. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The excretion of steroidal hormones from human populations is a major contributor to the presence of endocrine-disrupting chemicals (EDCs) in the aquatic environment. 17β-estradiol (E2), the major estrogen in vertebrates, estrone (E1), the precursor as well as a breakdown product of E2, and 17α-ethinylestradiol (EE2), the synthetic estrogen used in the birth control pill, are released in conjugated forms in urine (Ternes et al., 1999b). Their principal entry into the environment is via wastewater treatment plant (WWTP) effluents. Chemical fractionation and in vitro biological screening with yeast have identified E1, E2 and EE2 as the greatest sources of estrogenic activity in domestic sewage effluents (Desbrow et al., 1998). Moreover, significant progress has been made in quantifying the presence of steroidal estrogens in waste and surface waters, which have been measured in the ng L− 1 to μg L− 1 range (Williams et al., 2003; Ternes et al., 1999a; Carbella et al., 2004; Barel-Cohen et al., 2006; Fernandez et al., 2007; Lishman et al., 2006). The next challenge is to quantify the ⁎ Corresponding author at: Pest Management Regulatory Agency, Health Canada, 2720 Riverside Drive, Ottawa, ON, Canada K1S 0M2. Tel.: + 1 613 736 3898; fax: + 1 613 736 3540. E-mail addresses: susannakatkinson@gmail.com (S.K. Atkinson), vicki.marlatt@ufv.ca (V.L. Marlatt), lkimpe@uottawa.ca (L.E. Kimpe), drslean@gmail.com (D.R.S. Lean), trudeauv@uottawa.ca (V.L. Trudeau), jules.blais@uottawa.ca (J.M. Blais). 0048-9697/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2012.04.069 persistence and patterns of degradation of estrogens in these environments, enabling the estimation and reduction of risks to aquatic organisms. The degradation of steroidal estrogens is thought to be microbially mediated, with deconjugation of sulfates and glucuronides and oxidation of E2 and EE2 into E1 as the principal transformations (Carbella et al., 2004). Microbial activity in a WWTP may therefore be a proxy to predict elimination of steroids from raw sewage to final effluent (Johnson and Sumpter, 2001). This activity is often measured as the carbonaceous biochemical oxygen demand (cBOD), which varies greatly with the type of treatment (activated sludge, biological trickling filter, ultraviolet treatment) (Kirk et al., 2001; Carbella et al., 2004; Schlüsener and Bester, 2008). Increased hydraulic and sludge retention times (HRTs and SRTs), which allow time for the biological reactions to occur, and increased temperatures, which promote microbial degradation, can increase steroid elimination rates (Lee and Liu, 2002; Lishman et al., 2006; Shiwei et al., 2008; Ternes et al., 1999a). Despite these degradation reactions in WWTPs, steroidal estrogens continue to be detected in the natural aquatic environment. Barel-Cohen et al. (2006) determined that although natural and synthetic estrogen concentrations were reduced by half within 25 km from a sewage effluent source, they could still be detected up to 100 km along the river. The most potent but often least abundant steroidal estrogen EE2 has even been found in German drinking water at 0.5 ng L− 1 (Aherne and Briggs, 1989). The purpose of this study was (1) to quantify steroidal estrogens in the waste and receiving surface waters surrounding two southeastern Ontario WWTPs, in Ottawa and Cornwall, Ontario Canada; and (2) to 120 S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 Table 1 Latitude, longitude and altitude of sampling locations. Sampling location WWTP Ottawa River Cornwall Ottawa St. Lawrence Raw sewage Final effluent N/A Upstream Downstream Site 1 Site 2 Latitude (N) Longitude (W) Altitude (masl) 45° 45° 45° 45° 45° 45° 45° 75° 75° 74° 75° 75° 74° 74° 54 52 65 40 42 47 46 27′ 27′ 01′ 27′ 28′ 01′ 01′ 5.69″ 50.64″ 46.21″ 55.97″ 45.35″ 06.57″ 29.82″ 35′ 35′ 40′ 35′ 33′ 41′ 40′ 29.29″ 33.32″ 42.39″ 46.57″ 37.52″ 45.74″ 56.87″ Coordinates and altitudes obtained from Google Earth (2007). masl = meters above sea level. determine whether the efficiency of steroid removal by WWTPs relates to treatment type, HRT and temperature. We also incorporated precipitation in our model, as it may affect HRTs, and, like temperature, will vary seasonally. We predicted that E1, E2 and EE2 loadings would be greater to the Ottawa WWTP than the Cornwall WWTP, because of the larger population being served. Furthermore, we hypothesized that steroidal estrogen removal would be more efficient in the Ottawa WWTP than the Cornwall WWTP because Ottawa uses secondary treatment whereas Cornwall uses only primary treatment. Finally, we predicted that E1 would be the least persistent of the three estrogens, and that EE2 would be the most persistent. 2. Methods 2.1. Study sites The City of Ottawa WWTP is a secondary treatment facility with year round activated sludge and phosphorus removal, and seasonal chlorine disinfection occurring from May 16th to November 15th (Sierra Legal, 2004). Total daily flow is 4.22 × 10 8 L (Ontario Centre for Municipal Best Practices, 2006). The City of Cornwall WWTP is a primary treatment facility with a total daily flow of 1.63 × 107 L (United States Environmental Protection Agency, 2008; Department of Infrastructure & Municipal Works, City of Cornwall, personal communications, 2008). The locations of the study sites are given in Table 1. and the Laboratory Services of the Environmental Programs and Technical Support (personal communications, 2006; Appendix A). Details on the sampling and analysis of the hydrology and water chemistry data at the Ottawa WWTP can be found in its Annual Operations Report (City of Ottawa, 2005). 2.3. Ottawa meteorological data The daily mean ambient air temperature and total precipitation for the City of Ottawa were collected for the five days prior to each Ottawa wastewater sampling date (National Climate Data and Information Archive, 2011). For each sampling event, a mean temperature was calculated and a sum for the total daily precipitation was tabulated (Appendix B). The five days chosen to calculate the mean temperature and total precipitation for each sampling event was an estimate of the average time it takes for City of Ottawa RS to arrive at the WWTP. 2.4. Prediction of estrogen concentrations entering WWTPs Estimates of the population being served by the Ottawa and Cornwall WWTPs were obtained from 2006 Community Profiles (2011) (Appendices C and D). While 96.8% of the City of Ottawa population is served by the Ottawa wastewater treatment plant (Ontario Centre for Municipal Best Practices, 2006), the population being served by the Cornwall wastewater treatment plant was assumed to be the total population of the City of Cornwall. Population data were separated into groups of males and females, with the latter divided into pre-menstruating, menstruating, menopausal and pregnant women (Tables 2 and 3). There was a further subdivision of women taking the contraceptive pill, and women on hormone replacement therapy. Menstruating women were defined as females between the ages of 15 and 59 years, and the proportions of the female population that are pregnant (1/75), use the contraceptive pill (13%) and are on hormone replacement therapy (3%) were defined as reported by Johnson et al. (2000). As well, it was assumed that premenstruating females release the same amounts of E1 and E2 as males. This information, along with the steroidal estrogen excretion rates described by Johnson et al. (2000) and Christiansen et al. (2002), was used to calculate the total daily loadings of E1, E2 and EE2 in raw sewages in the Ottawa and Cornwall WWTPs. The daily loadings of estrogens were then divided by the average daily flow in the WWTPs to predict concentrations of E1, E2 and EE2 entering the plants (Tables 2 and 3). 2.2. Hydrology and water chemistry at the Ottawa WWTP 2.5. Sampling, extraction and analysis The daily WWTP flow (Q), hydraulic retention time (HRT), FE temperature (FET), and RS and FE carbonaceous biochemical oxygen demand (cBOD) for the wastewater sampling dates were obtained from monitoring records provided by the Ottawa WWTP technicians 2.5.1. Sampling and isotope dilution Triplicate grab wastewater composite samples were taken between June 2005 and August 2006 in 1-L amber glass bottles. An isotope Table 2 Predicted concentrations of estrone (E1), 17β-estradiol (E2) and 17α-ethinylestradiol (EE2) in Ottawa raw sewage. Study Group Population E1 excreted daily (μg) E2 excreted daily (μg) EE2 excreted daily (μg) Johnson et al. (2000) Males Pre-menstruating females Menstruating females Menopausal women Pregnant women Women on pill Women on hormone therapy Total Predicted concentration (ng L− 1) 380,418 67,880 257,050 75,362 5409 34,120 7874 1.48 × 106 2.65 × 105 2.06 × 106 3.01 × 105 1.40 × 106 n/a n/a 5.51 × 106 13.1 6.09 × 105 1.09 × 105 1.23 × 106 1.73 × 105 3.25 × 106 n/a n/a 5.37 × 106 12.7 n/a n/a n/a n/a n/a 3.10 × 105 5.63 × 104 3.67 × 105 0.9 Christiansen et al. (2002) Total Predicted concentration (ng L− 1) 786,130 1.01 × 107 23.9 5.25 × 106 12.4 4.70 × 105 1.1 Data based on Johnson et al. (2000), Christiansen et al. (2002), 2006 Community Profiles (2011) and Ontario Centre for Municipal Best Practices (2006); an average daily flow of 4.22 × 108 L was used. n/a = not applicable. S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 121 Table 3 Predicted concentrations of estrone (E1), 17β-estradiol (E2) and 17α-ethinylestradiol (EE2) in Cornwall raw sewage. Study Group Population E1 excreted daily (μg) E2 excreted daily (μg) EE2 excreted daily (μg) Johnson et al. (2000) Males Pre-menstruating females Menstruating females Menopausal women Pregnant women Women on pill Women on hormone therapy Total Predicted concentration (ng L− 1) 21,795 3810 13,683 6345 322 1821 420 8.50 × 104 1.49 × 104 1.09 × 105 2.54 × 104 8.35 × 104 n/a n/a 3.18 × 105 7.3 3.49 × 104 6.10 × 103 6.57 × 104 1.46 × 104 1.93 × 105 n/a n/a 3.15 × 105 7.2 n/a n/a n/a n/a n/a 1.66 × 104 3.00 × 103 1.96 × 104 0.4 Christiansen et al. (2002) Total Predicted concentration (ng L− 1) 45,965 5.89 × 105 13.5 3.07 × 105 7.0 2.75 × 104 0.6 Data based on Johnson et al. (2000), Christiansen et al. (2002), 2006 Community Profiles (2011) and personal communications with the Department of Infrastructure & Municipal Works, City of Cornwall (2008); an average daily flow of 1.63 × 107 L was used. n/a = not applicable. dilution was performed immediately after sampling by adding 80 μL of a 100 μg mL− 1 internal standard of 13C-labeled E1, E2 and EE2 in acetonitrile (E1: 3, 4-13C, 90% purity; E2: 3, 4-13C, 99% purity; EE2: 20, 21-13C, 99% purity; Cambridge Isotope Laboratories, Andover, MA). Adding the 13 C-labeled standards prior to filtration means the analysis provides the total amount in all phases, providing the basic assumption, that the isotopes added are distributed in the same way as the native compounds, is fulfilled (de Hoffman and Stroobant, 2002). 2.5.2. Filtration and solid phase extraction Filtration and extraction of samples were based on Lee and Peart's method (1998). Samples were filtered using glass fiber filters (Whatman GF/F 47 mm) and celite (Aldrich Grade 545), when needed, to prevent the filter from plugging. Samples were vacuum-loaded onto 1 g ENVI-18 tube solid phase extraction (SPE) tubes (Sigma-Aldrich, Oakville, ON) pre-conditioned with 5 mL acetone, 5 mL methanol and 10 mL high-performance liquid chromatography (HPLC) water (HPLC, GC pesticide grade; Burdick & Jackson, Muskegon, MI) using Visiprep large-volume sampler 1/8 in. Teflon tubes and an ENVI-Disk Holder manifold for 6 × 1 L samples (both from Sigma-Aldrich, Oakville, ON). The SPE tubes were eluted with 5 mL acetone and 5 mL methanol. An average flow rate of 10 mL/min was maintained for the pre-conditioned elution, the water sample, and the elution. Extracts were evaporated to dryness under a stream of nitrogen gas using a mini-evaporator. 2.5.3. Derivatization and clean-up Residues were redissolved in 2 mL HPLC water and a derivatization method was done by adding 10 μL 10% pentafluorobenzoyl chloride (2, 3, 4, 5, 6 — PFBCl) (Sigma-Aldrich, Oakville, ON) in toluene (Omnisolve, 99.98% purity; EMD Chemicals, Gibbstown, NJ) and 50 μL 2 M potassium hydroxide (KOH) buffer in water (Sigma-Aldrich, Oakville, ON). The mixture was then extracted twice with 2.5 mL hexane (HPLC, GC pesticide grade; Burdrick & Jackson, Muskegon, MI) and reduced to 0.5 mL with the mini-evaporator. Subsequently, the organic phase was passed through a glass pipette column plugged with glass wool (presonicated in dichloromethane, DCM) and filled with anhydrous sodium sulfate (ACS Grade, EMD granular) to remove impurities and any traces of water from the HPLC water and KOH buffer in water used in the derivatization reaction. The sample was dried with the mini-evaporator and resuspended in 200 μL isooctane (2,2,4-trimethylpentane, optima grade, 99% purity; Fisher Scientific, Fair Lawn, NJ). 2.5.4. Separation and detection by GC/MS The pentafluorobenzoyl-derivatives of E1, E2 and EE2 were separated and detected by gas chromatography–mass spectrometry (GC−MS). Samples of 4 μL were injected in splitless mode into a gas chromatograph with a Zebron ZB-5 ms column (25 m×250.00 μm×0.25 μm), a 10 cm guard, and a helium carrier gas at 65 cm s− 1 (6890 series; Agilent Technologies, Palo Alto, CA). The injector was set at 250 °C and the oven temperatures were programmed as follows: 80 °C min− 1 for 1 min, 24 °C min− 1 until 200 °C and held for 0.5 min, 73 °C min− 1 until 245 °C and held for 5 min, and then 1 °C min− 1 to 260 °C. The mass spec transfer line temperature was set at 300 °C. A Hewlett Packard 5973 quadrupole mass spectrometer and chemical ion source (Palo Alto, CA), with the quad temperature set at 100 °C, the source at 200 °C, and the methane reaction gas at 40% (26 cm s− 1), were used for quantification of estrogens in negative chemical ionization (NCI) mode. The following quantification ions were used in selected ion monitoring (SIM) mode: m/z 464 (C-12 E1), m/z 466 (C-13 E1, C-12 E2), m/z 468 (C-13 E2), m/z 490 (C-12 EE2) and m/z 492 (C-13 EE2). Since the 13 C-labeled standards were added to each sample prior to workup, any problems with recovery that may have occurred during workup and analysis should be compensated by isotopic dilution. Supplementary identifier masses and characteristic fragments were used in selected samples for analyte confirmation of E1 and EE2 following methods in Kuch and Ballschmiter (2001). Chromatographic data were collected and analyzed using Agilent MSD Chemstation data analysis software (Palo Alto, CA). 2.5.5. Detection limits Instrument detection limits (IDLs) and method detection limits (MDLs) were calculated using the methods outlined by the United States Environmental Protection Agency (1993). IDLs were obtained from seven injections of one water sample extract (prepared from E1, E2 and EE2 standards of 40, 80 and 5 ng L − 1, respectively, in 1 L double distilled water) where the standard deviation for the mean concentration was multiplied by three. IDLs for E1, E2 and EE2 were 1.2, 14.2 and 0.4 ng L − 1, respectively. MDLs were obtained by the injection of seven extracts from seven water samples (prepared from E1, E2 and EE2 standards of 5, 5 and 1 ng L − 1, respectively, in 1 L double distilled water) extracted identically, where the standard deviation for the mean concentration was multiplied by the one-tailed t statistic at a 95% confidence level. MDLs for E1, E2 and EE2 were 1.8, 24.3 and 0.5 ng L− 1, respectively. The detection limits for the extraction of E2 from water was elevated due to difficulty in separating E1 and E2. In addition, the linear responses of steroidal estrogens were confirmed from sample extracts of 0, 12.5, 25, 37.5 ng L− 1 of E1 and E2 standards and from sample extracts of 0, 1, 2, 3 and 4 ng L − 1 of EE2 standards. Plotting the responses versus the concentration gave R 2 values of 0.96, 0.97 and 0.99 for E1, E2 and EE2, respectively. Blank samples of double distilled water were spiked with the 13C standards 122 S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 β-glucuronidase, as well as the oxidation of E2 to E1 (D'Ascenzo et al., 2003; Ternes et al., 1999b; Lee and Liu, 2002). Other estrogens (E2 and EE2) were less frequently detected in this study, as has been seen elsewhere in the literature (Carbella et al., 2004; Fernandez et al., 2007). E2 was detected consistently in RS and FE samples from both WWTPs, but concentrations were much lower than for E1, and only quantifiable in three samples. E2 was found at a maximum concentration of 66.9 ng L − 1 (SE ± 15.7 ng L − 1) in Ottawa RS on September 14, 2005 (Table 4). EE2 was detected in every Cornwall WWTP sample, at a maximum concentration of 9.8 ng L− 1 (SE± 8.4 ng L− 1) in FE from June 29, 2005 (Table 4). In Ottawa, EE2 was only detected in two samples, one RS and one FE, from two different sampling dates. The sporadic occurrences of EE2 are similar to those reported by Williams et al. (2003), and may be because it is not released at such an elevated and constant rate from a population, compared to the other estrogens. Aside from their degradation, estrogens may not be detected in wastewater because they have sorbed to sewage sludge, which may be difficult to gather in a composite sample. EE2 in particular, which has the highest Kow (1.41 × 104 compared to 2.69 × 103 for E1 and 8.71× 103 for E2), shows this tendency (Suidan et al., 2005; Lai et al., 2002). Difficulties in analyzing sludge-bound estrogens may prevent a complete mass-balance analysis of E1, E2 and EE2 around each treatment unit in a plant. That concentrations in a sewage sludge are not well studied is a concern, because sludge is frequently applied as fertilizer to agricultural lands. (Cambridge Isotope Laboratories, Andover, MA) and taken through the extraction process alongside each series of sample analysis. 2.6. In vitro estrogenicity assay Four Ottawa wastewater samples (2 RS and 2 FE) collected on October 11, 2005 were analyzed for estrogen concentrations, as described in Section 2.5. A second set of wastewater samples (again, 2 RS and 2 FE), along with two positive E1 controls (500 ng L− 1), were only filtered and passed through an SPE tube, as described in Section 2.5.2, omitting the isotope dilution step. The dried residues of the four samples and two controls were resuspended in 100 μL of dimethyl sulfoxide (DMSO). Each extract in DMSO was assessed in a fish-specific, estrogen receptor-dependent reporter gene assay, as previously reported (Ackermann et al., 2002) and validated in our laboratory (Atkinson et al., 2011). Briefly, in this assay the rainbow trout (Oncorhynchus mykiss) gonad cell line RTG-2 is transiently transfected with an expression vector containing the rainbow trout estrogen receptor alpha complimentary DNA, an estrogen-dependent firefly luciferase reporter plasmid (pERETK-Luc) and a renilla luciferase control vector (PRL-TK). The cells are subsequently exposed to cell culture test medium (phenol-red free) containing the DMSO solvent control or test substance dissolved in DMSO in triplicate wells; the DMSO concentration in the cell culture test medium per well for all treatments was 0.05%. The mean fold induction above the DMSO control of the positive E1 control, and the samples of RS and FE, are reported. 3.2. Comparison of predicted concentrations versus measured concentrations 3. Results and discussion The E1 concentrations measured in raw sewage are higher than what would be predicted with a model using the excretion values reported by Johnson et al. (2000) and Christiansen et al. (2002) by about a factor of 2 (Tables 2 and 3). Similarly, where the E2 concentrations are quantifiable, these are also higher than what would be predicted by the model. Discrepancies between the predicted concentrations and the measured concentrations may be due to the different sex ratios and age distribution of the Danish and Ottawa populations, along with the difficulties in estimating the number of pregnancies and number of people taking the oral contraceptive pill or undergoing hormonal therapy, and deviations from the average daily flow (Johnson et al., 2000). The higher daily loadings of E1 and E2 in Ottawa wastewater compared to Cornwall are linked to the larger human population being served by the Ottawa sewage treatment plant (approximately 17 times higher than Cornwall). This is also in keeping with 3.1. Concentrations of steroidal estrogens in wastewater The concentrations of steroidal estrogens measured in samples of Ottawa and Cornwall RS and FE (Table 4) compare well to the values obtained in other Canadian studies (Cicek et al., 2007; Fernandez et al., 2007; Servos et al., 2005). E1 was found consistently in all the RS and FE samples collected, from both the Ottawa and Cornwall WWTPs. E1 was measured at a maximum concentration of 370 ng L− 1 (SE±26 ng L− 1) on August 16, 2006, and at a minimum concentration of 11.2 ng L− 1 (SE± 0.9 ng L− 1) on November 30, 2005, both in Ottawa FE samples (Table 4). Generally, the E1 concentration increased during the treatment process. This is likely due to both the decoupling of estrogen sulfates and estrogen glucuronide conjugates by the fecal coliform bacteria (Escherichia coli) enzyme Table 4 Concentrations of estrone, 17β-estradiol, 17α-ethinylestradiol in Ottawa and Cornwall wastewater samples; means (standard error) of raw sewage (RS) and final effluent (FE) concentrations, and percent difference (% Δ) from raw sewage to final effluent are listed. Site Date Estrone Mean (ng L Ottawa Cornwall June 9/05 July 20/05 Aug. 4/05 Sept. 14/05 Oct. 11/05a Nov. 30/05 Jan. 19/06 May 10/06 June 6/06 Aug. 16/06 June 29/05 Aug. 11/05 17β-estradiol −1 %Δ ) RS FE 47.3 (4.1) 50.4 (1.5) 46.7 (1.4) 44.9 (9.3) 46.0 (3.7) 35.0 (1.7) 35.2 (0.9) 48.6 (3.0) not sampled 104 (5.0) 13.1 (1.5) 29.3 (0.1) 86.8 20.6 205 158 113 11.2 44.9 11.3 15.7 370 16.0 22.9 (7.3) (0.4) (2.0) (2.0) (70) (0.9) (1.3) (0.4) (2.1) (26) (0.2) (0.2) − 83.5 59.1b − 340b − 252b − 148 68.2b − 27.6b 76.8b n/a − 257b − 22.5 21.9b Mean (ng L 17α-ethinylestradiol −1 %Δ ) RS FE 24.7 (0.1) b MDL b MDL 66.9 (15.7) b MDL b MDL nd nd not sampled nd b MDL b MDL b MDL nd nd b MDL 26.7 nd b MDL nd b MDL nd b MDL b MDL MDL = method detection limit; n/a = not applicable; nd = not detected. a Samples analyzed for inductions of reporter gene activity above a DMSO solvent control. b Significant difference between raw sewage and final effluent concentration. n/a 100 100 n/a n/a 100 n/a n/a n/a n/a n/a n/a Mean (ng L− 1) %Δ RS FE nd nd nd nd nd nd 1.3 (1.3) nd not sampled nd 5.7 (4.1) 0.5 (0.1) nd nd nd 0.6 (0.4) nd nd nd nd nd nd 9.8 (8.4) 1.0 (0.1) n/a n/a n/a n/a n/a n/a 100 n/a n/a n/a − 73.1 − 119 S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 3.3. Effect of meteorological variables and WWTP parameters on steroidal estrogen concentrations Regression analyses were used to examine the relationships between meteorological variables/WWTP parameters and steroidal estrogen concentrations for E1, because it is the most consistently measured steroidal estrogen and it provides the most complete data set from which to examine patterns over time and space. Studentized residuals were used to justify leaving the August 2006 Ottawa E1 concentrations out of the regression analyses; concentrations were over two times greater than for all other values. A significant positive relationship was found between the concentrations of E1 in RS and the mean ambient air temperature (R2 = 0.792, p = 0.003; Fig. 1); this is perhaps due to increased microbial activity at higher temperatures. Although breakdown of E1 may occur, at this stage of wastewater's pathway into the environment, the oxidation of conjugates and the conversion of E2 into E1 are greater than the breakdown of E1. Shiwei et al. (2008) reported higher concentrations of E1 in Chinese raw sewage in summer months, also attributing this to a more rapid degradation of E2 under higher temperatures. A significant inverse relationship was also found between E1 in RS and mean daily flow (R2 = 0.757, p = 0.005; Fig. 2), which is likely a dilution effect. No significant relationships were found between E1 in FE and any meteorological variable or WWTP parameter. 55 Concentration (ng/L) Desforges et al. (2010), who showed that estrogenic potency, as measured by plasma vitellogenin from male teleost fish in rivers around the world, was correlated with upstream human population size. 123 50 45 R2 = 0.757 p = 0.005 40 35 30 5.8 6 6.2 6.4 6.6 6.8 LN [Q] (ML)] Fig. 2. Concentration of estrone in Ottawa raw sewage versus mean daily flow. The percent removal in cBOD from RS to FE explained some of the variations observed in percent removal of E1 from RS to FE (R 2 = 0.435, p= 0.075, Fig. 3). Advancements in upgrading sewage treatment processes are usually designed to limit biological oxygen demand (BOD), total suspended solids (TSS) and total phosphorus in effluent, and sometimes ammonia nitrogen, E. coli and total residual chlorine; the removal of natural hormones is generally not a criterion (Sierra Legal, 2006). Here, the use of secondary treatment for removing organic material may also promote reactions that produce E1 (namely deconjugation of estrogen sulfates and glucorinides and oxidation of E2), and release of E1 in FE, a trend seen in other secondary WWTPs (Carbella et al., 2004; Schlüsener and Bester, 2008). 3.4. Removal of steroidal estrogens through the WWTP 3.5. Estrogenicity of wastewater In the Ottawa WWTP, there is a reduction of both E1 and E2 concentrations from raw sewage to final effluent. For 7 of the 9 Ottawa sampling dates for which percent E1 removal was calculated, a two-tailed t-test showed a significant difference (p≤ 0.05) between RS and FE. In Ottawa, E1 concentrations increased in FE up to 340%, compared to Cornwall FE, where the increase was only 22%. As discussed, the increase in E1, also seen in studies by Fernandez et al. (2007) and Schlüsener and Bester (2008), is likely from microbial deconjugation of steroidal esters and the conversion of E2 into E1, which is more efficient in the secondary plant than in the primary plant. On several dates, removal of E2 to a non-detectable level is likely due to oxidation of E2 into E1 and its subsequent breakdown products, (Table 4). However, secondary treatment does not seem to remove E2 more efficiently than the primary treatment; in both plants, E2 can still be detected in the final effluent. Due to the sporadic occurrences of EE2 found in wastewater, no conclusion can be made about its removal in secondary versus primary treatments. The E1 concentration in Ottawa wastewater sampled on October 11, 2005, increased from RS (E1 = 46.0 ng L − 1 ± 3.7 ng L − 1) to FE (113 ng L − 1 ± 70 ng L − 1, E2 b MDL, EE2 = nd), but this increase was not significant (Table 4). Estradiol and EE2 concentrations did not change (RS and FE: E2b MDL, EE2 = nd). Both the RS and FE samples caused significant elevation (one-way ANOVA, Tukey's post hoc test; p b 0.05) of estrogen-induced luciferase activity in the reporter gene assay (4.2± 0.3 and 3.5 ± 0.4 fold, respectively; mean± standard error (SEM)), and the magnitude of these responses was lower but not significantly different from the 500 ng L− 1 E1 positive control (8.7 ± 1.4 fold; mean± SEM; Fig. 4; p > 0.05). In addition, the estrogenicity of the RS compared to the FE samples was not significantly different in the reporter gene assay (Fig. 4; p > 0.05), which coincides with the absence of significantly higher steroidal estrogens measured during the chemical analyses of the FE samples compared to the RS samples. Collectively, these results demonstrate that sewage treatment was not very effective 200 100 55 % Difference Concentration (ng/L) 60 50 45 40 35 30 -10 0 -100 -200 R2 = 0.435 p = 0.075 -300 R2 = 0.792 p = 0.003 -400 -5 0 5 10 15 20 25 30 Temperature (oC) Fig. 1. Concentration of estrone in Ottawa raw sewage versus mean environmental temperature five days prior to sampling. 90 92 94 96 98 100 % Difference cBOD Fig. 3. Percent difference of estrone versus percent difference of carbonaceous biological oxygen demand in Ottawa wastewater. 124 S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 Fold induction above DMSO control 12 10 8 6 4 2 0 E1 control Raw Sewage Final Effluent Fig. 4. Estrogen-dependent luciferase induction by positive estrone control (500 ng L− 1), and raw sewage and final effluent samples collected from the Ottawa WWTP in a rainbow trout estrogen receptor reporter gene assay. Mean fold induction (± SE) above the DMSO control of two independent experiments are presented. An ANOVA shows no significant difference between the E1 control, the raw sewage and the final effluent samples. in removing the estrogen bioactivity caused by compounds binding to the trout ERα in the in vitro reporter gene assay. Although a full dose response curve with E1 was not included in the present study, the relative potency of E1 relative to E2 ranges from 0.1 to 0.7 in various in vitro and in vivo bioassays (Cespedes et al., 2004; Van den Belt et al., 2004; Salste et al., 2007; Thorpe et al., 2006). These relative potency values reported in the literature are similar to the relative potency value estimated when comparing the single 500 ng L− 1 E1 dose tested to the E2 standard curve obtained in the present experiments. Without a complete chemical analysis of the wastewater samples tested including non-hormone based estrogenic compounds (i.e. alkylphenols, phthalates, and pesticides), we cannot completely account for all of the compounds that are contributing to the estrogenic or anti-estrogenic activity of the RS and FE samples. Our results suggest that the breakdown products in a WWTP may include E1, and most likely, additional uncharacterized compounds. This may include lumiestrone, a little known photochemical derivative of E1 that we have recently discovered to be biologically active in vivo and in vitro (Trudeau et al., 2011). It is possible that the levels of E1 present and additional uncharacterized compounds in these Ottawa wastewater samples have the potential to induce an estrogenic response in exposed aquatic wildlife. Future investigations aimed at a more complete chemical analysis of wastewater constituents and continued monitoring of the Ottawa WWTP for endocrine disrupting effects are warranted. 3.6. Concentrations of steroidal estrogens in rivers E1 and EE2 were detected in both rivers at maximum concentrations of 107 ng L− 1 (SE±35 ng L− 1) in the St. Lawrence River, and 2.2 ng L− 1 (SE ± 1.0 ng L − 1) in the Ottawa River, respectively (Table 5). Whereas the EE2 value is within the range of measurements found elsewhere, the E1 value is higher than measurements previously reported in surface waters (Labadie and Budzinski, 2005; Williams et al., 2003). E2 was not detected in either of the rivers. Because sampling of both WWTP and rivers on the same day was not often possible, river sample results are compared to final effluent samples taken one month before, and one month after river sampling dates. The detection of 20.6 ng L − 1 E1 in Ottawa FE on July 20, 2005, and its absence from the downstream sample, suggest that degradation, sorption and dilution of estrogens occur in rivers. These reactions are further studied by Jürgens et al. (2002) and Barel-Cohen et al. (2006). The high E1 concentration (107 ng L − 1 SE ± 35 ng L − 1) measured in the St. Lawrence upstream of the Cornwall wastewater treatment plant, is at least four times higher than any measured or predicted concentration of E1 in Cornwall wastewater (Tables 3, 4 and 5). As well, EE2 is found upstream of both wastewater treatment plants, even though it is never detected in raw sewage. These values suggest that a source of estrogens exists upstream from the wastewater treatment plants, which may be from storm sewer overflow or agricultural sources. In addition to release from domestic wastewater, E1, E2 and EE2 can also enter watersheds through runoff from land, where sources include cattle or pig excrement and fertilizers. This type of transport and degradation has not been well investigated in the field (Gagné et al., 2001; Johnson and Sumpter, 2001; Khanal et al., 2006; Johnson et al., 2006). The river water sampling results also demonstrate the resilience of estrogenic compounds in the natural environment, particularly ethinylestradiol. In Ottawa and Cornwall, E1 and EE2 were detected downstream of the wastewater treatment plant; whereas E1 was measured at much lower concentrations in downstream samples than in final effluents, the concentrations of EE2 did not decrease. Some have hypothesized that estrogenic compounds may originate from the Ottawa River after observing significantly higher VTG levels in mussels and fish in Rivière Des Prairies, to which the Ottawa River drains (Aravindakshan et al., 2004), and the present study supports this hypothesis. Whether these compounds originate from Ottawa effluent or via runoff, these compounds exhibit environmental persistence. 4. Conclusion We found patterns of occurrence, removal and persistence of E1, E2 and EE2 in Ottawa and Cornwall WWTPs and surface waters. Secondary treatment does not seem to improve removal efficiency of E2 and EE2, in comparison to primary treatment, and in addition, it also produces higher concentrations of E1 in final effluent. The production of E1 appears to be microbially mediated because of the relationship between E1 production and cBOD removal, and may be more efficient at higher temperatures, resulting in a seasonal pattern of estrogen release. The sampling results suggest that in addition to wastewater effluent, other sources of estrogens to the Ottawa River and St. Lawrence River Table 5 Concentrations of estrone, 17β-estradiol, 17α-ethinylestradiol in Ottawa and Cornwall final effluent (FE) samples and in Ottawa and St. Lawrence River upstream and downstream samples; means (standard error) of concentrations are listed. Site Date Estrone Mean (ng L 17β-estradiol −1 FE Ottawa Cornwall June 9/05 June 22/05 July 20/05 Aug. 4/05 June 15/05 June 29/05 86.8 – 20.6 86.8 – 16.0 (7.3) (0.4) (7.3) (0.2) MDL = method detection limit; nd = not detected. “–” indicates that for this date, no sample was taken. ) Mean (ng L −1 17α-ethinylestradiol Mean (ng L− 1) ) US DS FE US DS FE US DS – nd nd – 107 (35) – – bMDL nd – nd – b MDL – nd b MDL – b MDL – nd nd – nd – – nd nd – nd – nd – nd nd – 9.8 (8.4) – 0.8 (0.3) bMDL – 0.4 (0.2) – – 2.2 (1.0) b MDL – 0.7 (0.1) – S.K. Atkinson et al. / Science of the Total Environment 430 (2012) 119–125 exist and need to be better described (e.g., farming activity, storm sewers, coastal currents). Continued monitoring of these steroidal estrogens is necessary in order to improve wastewater treatment processes, and to ensure the safety of drinking water. It is important to examine low potency steroidal estrogens (i.e. estrone) because they are commonly detected in surface waters worldwide, and can significantly contribute to the overall estrogenic activity of a sample. Monitoring should also be expanded to include conjugates, sludge-bound estrogens, and other estrogen agonists and antagonists. Given the persistence of estrogenicity in effluent and rivers, we recommend surveillance studies for indicators of estrogenization of aquatic wildlife populations within these waters. Acknowledgments This research was funded by grants from the National Science and Engineering Research Council of Canada to JMB, DRSL and VLT. Thanks are given to Gary Robidoux (City of Ottawa), Serena Maharaj (University of Ottawa), Adrienne Fowlie and Jeff Ridal (St. Lawrence River Institute of Environmental Sciences) for help with sampling, and to Fida Ahmed (University of Ottawa) for her assistance in the laboratory. The donation of the ERE-TK-luc construct from Dr. Farzad Pakdel (Rennes) is acknowledged with gratitude. Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.scitotenv.2012.04.069. References 2006 Community Profiles [internet]. Ottawa (ON): Statistics Canada. c2011 — [cited 2011 Aug 10. Available from: http://www12.statcan.ca/english/census06/data/ profiles/community/Index.cfm?Lang=E. Ackermann GE, Brombacher E, Fent K. Development of a fish reporter gene system for the assessment of estrogenic compounds and sewage treatment plant effluents. Environ Toxicol Chem 2002;21:1864–75. Aherne GW, Briggs R. The relevance of the presence of certain synthetic steroids in the aquatic environment. J Pharm Pharmacol 1989;41:735–6. Aravindakshan J, Paquet V, Gregory M, Dufresne J, Fournier M, Marcogliese DJ, et al. Consequences of xenoestrogen exposure on male reproductive function in spottail shiner (Notropis hudsonius). Toxicol Sci 2004;78:156–65. Atkinson SK, Marlatt VL, Kimpe LE, Lean DRS, Trudeau VL, Blais JM. Environmental factors affecting ultraviolet photodegradation rates and estrogenicity of estrone and ethinylestradiol in natural waters. Arch Environ Contam Toxicol 2011;60:1–7. Barel-Cohen K, Shore LS, Shemesh M, Wenzel A, Mueller J, Kronfeld-Schor N. Monitoring of natural and synthetic hormones in a polluted river. J Environ Manage 2006;78:16–23. Carbella M, Omil F, Lema JM, Llompart M, García-Jares C, Rodríguez I, et al. Behaviour of pharmaceuticals, cosmetics and hormones in a sewage treatment plant. Water Res 2004;38:2918–26. Cespedes R, Petrovic M, Raldua D, Saura U, Pina B, Lacorte S, et al. Integrated procedure for determination of endocrine-disrupting activity in surface waters and sediments by use of biological technique recombinant yeast assay and chemical analysis by LC −ESI-MS. Anal Bioanal Chem 2004;378:697–708. Christiansen LB (University of Southern Denmark, Denmark), Winther-Nielsen M, Helweg C (DHI, Water & Environment, Denmark). Feminization of fish. Final report. Danish Environmental Protection Agency, Danish Ministry of the Environment; 2002. Environmental project no. 729. Cicek N, Londry K, Oleszkiewicz JA, Wong D, Lee Y. Removal of selected natural and synthetic estrogenic compounds in a Canadian full-scale municipal wastewater treatment plant. Water Environ Res 2007;79(7):795–800. City of Ottawa. Public Works and Services Department, Utility Services Branch, Wastewater and Drainage Services Division. Robert O. Pickard Environmental Centre annual operations report. Ottawa (ON): City of Ottawa; 2005.. 47 pp. D'Ascenzo G, Di Corcia A, Gentili A, Mancini R, Mastropasqua R, Nazzari M, et al. Fate of natural estrogen conjugates in municipal sewage transport and treatment facilities. Sci Total Environ 2003;302:199–209. de Hoffman E, Stroobant V. Mass spectrometry: principles and applications. John Wiley and Sons; 2002. 407 pp. Desbrow C, Routledge EJ, Brighty GC, Sumpter JP, Waldock M. Identification of estrogenic chemicals in STW effluent. 1. Chemical fractionation and in vitro biological screening. Environ Sci Technol 1998;32(11):1549–58. Desforges JP, Peachey BDL, Sanderson PM, White PA, Blais JM. Plasma vitellogenin in male teleost fish from 43 rivers worldwide is correlated with upstream human population size. Environ Pollut 2010;158:3279–84. 125 Fernandez MP, Ikonomou MG, Buchanan I. An assessment of estrogenic organic contaminants in Canadian waters. Sci Total Environ 2007;373:250–69. Gagné F, Marcogliese DJ, Blaise C, Gendron AD. Occurrence of compounds estrogenic to freshwater mussels in surface waters in an urban area. Environ Toxicol 2001;16:260–8. Google Earth (version 4.2.0198.2451) [software]. Mountain View, CA: Google Inc.; 2007. Johnson AC, Sumpter JP. Removal of endocrine-disrupting chemicals in activated sludge treatment works. Environ Sci Technol 2001;35(24):4697–703. Johnson AC, Belfroid A, Di Corcia A. Estimating steroid oestrogen inputs into activated sludge treatment works and observations on their removal from the effluent. Sci Total Environ 2000;256:163–73. Johnson AC, Williams RJ, Matthiessen P. The potential steroid hormone contribution of farm animals to freshwaters, the United Kingdom as a case study. Sci Total Environ 2006;362(1–3):166–78. Jürgens MD, Holthaus KIE, Johnson AC, Smith JJL, Hetheridge M, Williams RJ. The potential for estradiol and ethinylestradiol degradation in English rivers. Environ Toxicol Chem 2002;21(3):480–8. Khanal SK, Xie B, Thomspon M, Sung S, Ong S, VanLeeuwen J. Fate, transport, and biodegradation of natural estrogens in the environment and engineered systems. Environ Sci Technol 2006;40(21):6537–46. Kirk LA, Tyler CR, Lye CM, Sumpter JP. Changes in estrogenic and androgenic activities at different stages of treatment in wastewater treatment works. Environ Toxicol Chem 2001;21:972–9. Kuch HM, Ballschmiter K. Determination of endocrine-disrupting phenolic compounds and estrogens in surface and drinking water by HRGC−(NCI)−MS in the picogram per liter range. Environ Sci Technol 2001;35:3201–6. Labadie P, Budzinski H. Determination of steroidal hormone profiles along the Jalle d'Eysines River (near Bordeaux, France). Environ Sci Technol 2005;39(14):5113–20. Lai KM, Scrimshaw MD, Lester JN. Prediction of the bioaccumulation factors and body burden of natural and synthetic estrogens in aquatic organisms in the river systems. Sci Total Environ 2002;289:159–68. Lee HB, Liu D. Degradation of 17β-estradiol and its metabolites by sewage bacteria. Water Air Soil Pollut 2002;134:353–68. Lee HB, Peart TE. Determination of 17β-estradiol and its metabolites in sewage effluent by solid-phase extraction and gas chromatography/mass spectrometry. J AOAC Int 1998;81:1209–16. Lishman L, Smyth SA, Sarafin K, Kleywegt S, Toito J, Peart T, et al. Occurrence and reductions of pharmaceuticals and personal care products and estrogens by municipal wastewater treatment plants in Ontario, Canada. Sci Total Environ 2006;367(2–3):544–58. National Climate Data and Information Archive. [Internet]. Fredericton (NB): Environment Canada. [cited 2011 Aug 10]. Available from:http://www.climate.weatheroffice.ec.gc. ca/climateData/canada_e.html2011. Ontario Centre for Municipal Best Practices. Best practice summary report. Water and wastewater alternative sources of energy. Case study of Ottawa. Toronto (ON): Association of Municipalities of Ontario; 2006. 5 pp. Salste L, Leskinen P, Virta M, Kronberg L. Determination of estrogens and estrogenic activity in wastewater effluent by chemical analysis and the bioluminescent yeast assay. Sci Total Environ 2007;378:343–51. Schlüsener MP, Bester K. Behaviour of steroid hormones and conjugates during wastewater treatment — a comparison of three sewage treatment plants. Clean 2008;36(1):25–33. Servos MR, Bennie DT, Burnison BK, Jurkovic A, McInnis R, Neheli T, et al. Distribution of estrogens, 17β-estradiol and estrone, in Canadian municipal wastewater treatment plants. Sci Total Environ 2005;336(1–3):155–70. Shiwei J, Yang F, Liao T, Hui Y, Xu Y. Seasonal variation of estrogenic compounds and their estrogenicities in influent and effluent from a municipal sewage treatment plant in China. Environ Toxicol Chem 2008;27(1):146–53. Sierra Legal Defence Fund. The national sewage report card (number three): grading the sewage treatment of 22 Canadian cities. Toronto (ON): Sierra Legal Defence Fund; 2004. 68 pp. Sierra Legal Defence Fund. The Great Lakes sewage report card. Toronto (ON): Sierra Legal Defence Fund; 2006. 57 pp. Suidan MT, Esperanza M, McCauley P, Brenner RC, Venosa AD. Challenges in biodegradation of trace organic contaminants — gasoline oxygenates and sex hormones. Water Environ Res 2005;77(1):4-11. Ternes TA, Stumpt M, Mueller J, Harbere K, Wilken RD, Servos M. Behaviour and occurrence of estrogens in municipal sewage treatment plants — I. Investigations in Germany, Canada and Brazil. Sci Total Environ 1999a;225:81–90. Ternes TA, Kreckel P, Mueller J. Behaviour and occurrence of estrogens in municipal sewage treatment plants — II. Aerobic batch experiments with activated sludge. Sci Total Environ 1999b;225:91–9. Thorpe KL, Gross-Sorokin M, Johnson I, Brighty G, Tyler CR. An assessment of the model of concentration addition for predicting the estrogenic activity of chemical mixtures in wastewater treatment works effluents. Environ Health Perspect 2006;114:90–7. Trudeau VL, Heyne B, Blais JM, Temussi F, Atkinson SK, Pakdel F, et al. Lumiestrone is photochemically derived from estrone and may be released to the environment without detection. Front Exp Endocrinol 2011. http://dx.doi.org/10.3389/fendo.2011.00083/full. United States Environmental Protection Agency. U.S. Code of Federal Regulations, title 40, part 136, appendix B. Washington (VA): US EPA; 1993. United States Environmental Protection Agency. Lake Ontario lakewide management plan status 2008. Chicago (Il): US EPA; 2008. 130 pp. Van den Belt K, Berckmans P, Vangenechten C, Verheyen R, Witters H. Comparative study on the in vitro/in vivo estrogenic potencies of 17β-estradiol, estrone, 17αethynylestradiol and nonylphenol. Aquat Toxicol 2004;66:183–95. Williams RJ, Johnson AC, Smith JJL, Kanda R. Steroid estrogens profiles along river stretches arising from sewage treatment works discharges. Environ Sci Technol 2003;37:1744–50.