- Email: [email protected]
Determination of estrogenic potential in waste water without sample extraction
Journal of Hazardous Materials 260 (2013) 527–533
Contents lists available at SciVerse ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Determination of estrogenic potential in waste water without sample extraction Miha Avberˇsek a,b , Bojana Zˇ egura c , Metka Filipiˇc c , Nataˇsa Uranjek-Zˇ evart d , Ester Heath a,b,∗ a
“Joˇzef Stefan” Institute, Jamova 39, 1000 Ljubljana, Slovenia International Postgraduate School Joˇzef Stefan, Jamova 39, 1000 Ljubljana, Slovenia c National Institute of Biology, Veˇcna pot 111, 1000 Ljubljana, Slovenia d Komunalno podjetje Velenje d.o.o., Koroˇska cesta 37b, 3320 Velenje, Slovenia b
h i g h l i g h t s • • • •
A modified ER-Calux® (NE-(ER-Calux® )) does not need pre-extraction of raw water samples. NE-(ER-Calux® ) enables determination of estrogenic potential of raw water samples. The sensitivities of NE-(ER-Calux® ) and conventional ER-Calux® assay are comparable. NE-(ER-Calux® ) assay is recommended as a screening assay in multi sample studies.
a r t i c l e
i n f o
Article history: Received 11 March 2013 Received in revised form 29 May 2013 Accepted 3 June 2013 Available online 10 June 2013 Keywords: ER-Calux® No extraction GC–MSD Waste water Estrogenic potential
a b s t r a c t This study describes the modification of the ER-Calux® assay for testing water samples without sample extraction (NE-(ER-Calux® ) assay). The results are compared to those obtained with ER-Calux® assay and a theoretical estrogenic potential obtained by GC–MSD. For spiked tap and waste water samples there was no statistical difference between estrogenic potentials obtained by the three methods. Application of NE(ER-Calux® ) to “real” influent and effluents from municipal waste water treatment plants and receiving surface waters found that the NE-(ER-Calux® ) assay gave higher values compared to ER-Calux® assay and GC–MSD. This is explained by the presence of water soluble endocrine agonists that are usually removed during extraction. Intraday dynamics of the estrogenic potential of a WWTP influent and effluent revealed an increase in the estrogenic potential of the influent from 12.9 ng(EEQ)/L in the morning to a peak value of 40.0 ng(EEQ)/L in the afternoon. The estrogenic potential of the effluent was
1. Introduction Naturally excreted steroid estrogens are ubiquitously present in waste water and other environmental samples [1–3]. They have the highest estrogenic activity among endocrine disrupting compounds and account for the majority of estrogenic potential in municipal waste water [4–6]. Concentrations of steroid estrogens in WWTP effluent are typically in the low ng/L range [1], but this is still sufficient to affect the endocrine system of living organisms [7,8].
∗ Corresponding author at: “Joˇzef Stefan” Institute, Jamova cesta 39, 1000 Ljubljana, Slovenia. Tel.: +386 1 4773 584; fax: +386 1 2519 385. E-mail address: [email protected] (E. Heath). 0304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2013.06.009
To detect such low concentrations of estrogens sensitive methods are necessary. Gas or liquid chromatography coupled with mass spectrometry are the techniques of choice for quantitative chemical analyses [1,3,9] while total estrogenic potential of the sample is obtained by biological in vivo or in vitro assays [2,7]. Several cellbased in vitro bioassays like MELN [10–13], MVLN [14], E-Screen [13,15,16], ER-Calux® assay [13,17–19] and MMV-Luc [20] have been developed for environmental samples. Alternatively, recombinant yeast based assays are available [2]. These are easier to use, but they lack complex estrogenic interactions [21,21], are less sensitive than mammalian cell-based assays and are unable to detect anti-estrogenic compounds [19]. Low concentrations of estrogens also mean that extraction from complex environmental matrices and pre-concentration of analytes is essential for both detection and quantification with the bio-assays. Liquid–liquid or solid phase extraction (SPE) with
528
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
additional clean-up is generally applied for chemical analysis and bioassays [2,22] and only a few published studies describe the determination of estrogenic potential without extensive sample preparation and extraction. Of these the majority employ a recombinant yeast assay [6,22,23] and just one applies a mammalian cell-based assay (MELN cells) without presenting any results [10]. The aim of this study was to modify the ER-Calux® assay for the determination of estrogenic potential of water samples without extensive sample handling and extraction of analytes. The modified method that we named NE-(ER-Calux® ) assay, was tested with tap and waste water samples spiked with steroid estrogens, and compared to conventional ER-Calux® assay and chemical analysis with gas chromatography–mass selective detection (GC–MSD). The modified method was then applied for investigating estrogenic potential of “real” environmental samples and for studying intraday dynamics of estrogenic potential in influent and effluent samples of WWTP. 2. Materials and methods 2.1. Standards, chemicals, growth media Standards estrone (E1; min 99%), 17-estradiol (E2; min 98%), 17␣-ethinylestradiol (EE2; min 98% (HPLC)), estriol (E3; min 99%), a deuterated internal standard (bisphenol A)-d16 (98 atom% D) were purchased from Sigma (Steinheim, Germany). Standards were prepared freshly in ethyl acetate, for ER-Calux® assay calibration curve and in methanol for spiking water samples used for analyses. Methanol, ethyl acetate “Baker ultra resi-analysed® ” grade were purchased from J.T. Baker (Deventer, the Netherlands). Pyridine (max 0.01% H2 O) was purchased from Merck (Darmstadt, Germany). The derivatising agent N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA; derivatisation grade), was purchased from Sigma. For ER-Calux® assay, media Gibco® D-MEM/F-12 with GlutaMAXTM (with phenol red), Gibco® D-MEM/F-12 with lglutamine (without phenol red) and Stripped FBS (foetal bovine serum) were purchased from Invitrogen (Paisley, UK). EDTA, nonessential amino acids (MEM 100×), and penicillin/streptomycin were purchased from Sigma. FBS (foetal bovine serum) and PBS
(Phosphate Buffered Saline) were purchased from PAA (Pasching, Austria) while Difco trypsin was obtained from Becton Dickinsen (Heidelberg, Germany). 2.2. Sampling and sample handling Sampling was performed for three different purposes as follows. Spiked tap and waste water effluent samples were used for optimisation purposes, while waste water treatment plant (WWTP) influent and effluent samples as well as surface waters were used to study “real” samples. Hourly samples were collected at one WWTP to measure intraday variations of estrogenic potential in influent and effluent samples. 2.2.1. Spiked samples Tap water from our laboratory and grab waste water effluent sample from WWTP2 (Table 1) were spiked with estrone (E1), 17-estradiol (E2), 17␣-ethinylestradiol (EE2) and estriol (E3) at environmentally relevant concentrations (0–40 ng/L) at levels that were chosen randomly within this range (see Supplementary material: Table S1). After spiking, samples were homogenised by shaking at 300 rpm for 30 min. For the NE-(ER-Calux® ) assay, 10 mL of each sample was stored at −20 ◦ C, while 200 mL was immediately used for SPE (see Section 2.3). Extracts of the samples were analysed by ER-Calux® assay and GC–MSD, while un-extracted samples were analysed by NE-(ER-Calux® ) assay. 2.2.2. “Real” waste water and surface water samples Grab samples (250 mL, glass bottles) of WWTP influent and effluent samples and surface river water samples (upstream and downstream of the effluent site) were collected from seven different WWTPs (Table 1). Samples at each WWTP were collected on four consecutive weeks (two WWTPs per week), on Monday morning. In order to assure the same sample storage and preparation time, effluent and river samples were taken without considering hydraulic retention time. Similar to spiked samples, 10 mL of the sample was stored at −20 ◦ C for NE-(ER-Calux® ) assay and 200 mL was used for the extraction and analysis by ER-Calux® assay and GC–MSD. To avoid sample degradation, the extraction was performed within 4 h after the sampling.
Table 1 Main characteristics of waste water treatment plants involved in the study. Name
Treatment
WWTP1
Biofiltration with P and N removal
WWTP2
Capacity (PU)
Design
Flow/year (m3 )
Mean influent COD (BOD) (mg/L)
Mean effluent COD (BOD) (mg/L)
SRT (days)
Estimated surface water flow (m3 /s)
Actual NAa
2
18
15–32
7
43 (<10)
19
8
800 (300)
80 (15)
22
20
110
10,000,000
740 (400)
26 (6)
29
12
20
110,805
4,795,963
1013 (506)
128 (28)
22
4–10
85,000
8,486,259
429 (214)
17 (4)
22.5
20–24
50,000
45,000
6,161,222
400 (196)
Activated sludge, nitrification; no P removal
200,000
143,623
7,303,085
576 (66)
294 (15)
WWTP3
Activated sludge; nitrification; no P removal
360,000
420,000
29,928,900
590 (312)
WWTP4
Activated sludge; nitrification; no P removal
100,000
85,000
5,500,000
WWTP5
Activated sludge, anoxic zones, P removal
250,000
160,000
WWTP6
Activated sludge, no P removal
68,000
WWTP7
Activated sludge, nitrification, P removal
70,000
a
HRT (h)
39.7 (8.2)
Sludge age as determined for suspended biomass is not relevant in water treatment with biofiltration.
2.5
20
0.5 35
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
529
2.2.3. Intraday waste water samples from WWTP1 In this part of the study, samples were taken at WWTP1. Hourly samples and time proportional samples were collected at influent and effluent site. Grab influent samples were taken every hour from 7:00 to 19:00. Appropriate effluent samples were taken within a 2.5 h shift (9:30–21:30) according to the hydraulic retention time at WWTP1 (Table 1). Hourly waste water samples were analysed with NE-(ER-Calux® ) assay only and therefore, 10 mL of sample volume was sampled in polypropylene flasks. Samples were stored at −20 ◦ C and were analysed the next morning as described in Section 2.6. Time proportional samples for the whole sampling interval (sampling every 15 min) were collected from the influent (7:00–19:00) and effluent (9:30–21:30) and analysed as well. To check the stability of water treatment system the following parameters were monitored: temperature, pH, chemical oxygen demand – COD (sensor name: UVAS), dissolved oxygen – O2 (LDO), ammonium – NH4 (AMTAX), phosphates – PO4 (PHOSPHAX), nitrates – NO3 (NITRAX) were measured with HACH LANGE controller system (Düsseldorf, Germany).
luminometer (top reading plate reader) with a set integration time of 4000 ms and a gain of 200 (Tecan Genios, Maennedorf, Switzerland). Results expressed in relative luminescent units (RLU) were processed using a sigmoidal calibration curve constructed with a MS Excel template (provided by BDS) together with an add-in “Solver”. The E2 equivalents (EEQ) were calculated as described by Avberˇsek et al. [25]. LOD (0.136 ng(EEQ)/L (=0.5 pM)) and LOQ (0.409 ng(EEQ)/L (=1.5 pM)) of ER-Calux® assay was determined by manufacturer of the assay (BDS, b.v.) on the basis of their long term experience with the assay. However, LODs and LOQs were also determined from calibration curve each time the assay was performed. These values were only used to test the performance of the bioassay. Therefore, together with concentration factor of SPE–ER Calux® (0.2×), limits of detection (LOD) and quantification (LOQ) assay were 0.68 ng(EEQ)/L and 2.05 ng(EEQ)/L, respectively. Recovery (97 %), and repeatability (4%) were determined with tap water (n = 5), spiked with E2 (c = 4.09 ng/L (=15 pM)) [25].
2.3. Sample extraction
All spiked and “real” sample extracts (90%) that remain after ER-Calux® assay were derivatised with MSTFA and analysed with GC–MSD (HP 6890 Series, Hewlet-Packard, Waldbron, Germany) [24]. Separation was achieved on a HP5-MS column. SIM mode was used for quantification of E1, E2, EE2 and E3. Estrogenic potential was calculated from concentrations of each compound as described by Avberˇsek et al. [25]. Validation of the method was performed using tap and waste water as a matrix [24,25].
2.5.1. NE-(ER-Calux® ) assay The standard ER-Calux® assay has been adapted for testing the water samples without prior extraction and we named it NE-(ERCalux® ) assay. The frozen samples (10 mL) were thawed, sterilised by filtration (Whatman, ANOTOP25, pore size 0.2 m) and diluted 1:1, 1:3, 1:10, 1:30 and 1:100 with PBS. For the cell treatment 20 vol% of diluted samples were mixed with “Test medium”. For the control “Test medium” was mixed with 20 vol% PBS and for the calibration E2 was added (final concentrations 0.3, 0.6, 1, 3, 6, 10, and 30 pM/well). Further steps of the assay were the same as in the conventional ER-Calux® assay. In comparison to ER-Calux® , no significant difference was observed in LODs and LOQs determined from standard calibration curve performed on each test plate. For that reason, we keep the same LOD and LOQ for NE-(ER-Calux® ) as well. Together with concentration factor (0.2×), limit of detection (LOD) and quantification (LOQ) of NE-(ER-Calux® ) assay were 0.68 ng(EEQ)/L and 2.05 ng(EEQ)/L, respectively. Recovery (92%), and repeatability (5%) were determined with tap water (n = 5), spiked with E2 (c = 4.09 ng/L (=15 pM)).
2.5. Determination of estrogenic potential with ER-Calux® assay
2.6. Toxicity testing
The ER-Calux® assay was performed as described by Avberˇsek et al. [25]. Briefly, T47D-EREtata-Luc cells were used (provided and licenced by BioDetection Systems (BDS) b.v. Amsterdam, the Netherlands) and maintained at 37 ◦ C; 5% CO2 in “Growth medium” (D-MEM/F12; GlutaMAXTM ; phenol red as pH indicator; 7.5% FBS, 1% nonessential amino acids (MEM) and a 0.2% penicillin/streptomycin solution). For the assay purposes the “Growth medium” was replaced with the “Test medium” (D-MEM/F12; lglutamine; without phenol red; 5% stripped FBS, 1% nonessential amino acids (MEM) and 0.2% penicillin/streptomycin solution). The cells in the “Test medium” were seeded (10,000 cells/well) into 96-well white microtiter plates (Nunc, Roskilde, Denmark), incubated for 24 h (37 ◦ C), then the medium was replaced with fresh “Test medium” and incubated for additional 24 h. The medium was then replaced with 100 L of “Test medium”, supplemented with 0.1% sample extracts corresponding to 1 L of sample extract or its dilution (1:1, 1:3, 1:10, 1:30 and 1:100) in 1 mL of “Test medium”. E2 (0.6–30 pM/well) was used the calibration standard and 0.1 vol% ethyl acetate as the solvent control. All the samples were tested in triplicate wells. After 24 h incubation, the luminescence kit “Steadylite plusTM ” (Perkin Elmer, Shelton, USA) was added (100 L) to the medium and gently shaken for 15 min at room temperature. Luminescence was measured using a TECAN
MTT assay was used for toxicity testing of all waste water samples. MTT assay was performed on ER-Calux® cells according to Mosmann [26] with minor modifications [27]. ER-Calux® cells were exposed to the samples in the same way as in ER-Calux® or NE-(ER-Calux® ) assays. However, instead of the luminescence kit, MTT reagent was added after 24 h exposure. After 3 h incubation, the difference in optical density at 570 and 690 nm wavelength was measured (Tecan Genios, Maennedorf, Switzerland). The results were compared to control sample (20% PBS).
Sample extraction for ER-Calux® and GC–MSD was performed as described by Avberˇsek et al. [24,25]. Samples (200 mL) were filtered and afterwards extracted with SPE (Waters, Oasis HLB) and eluted with ethyl acetate. Extracts were cleaned-up on silica-gel cartridges (Biotage, Isolute SI), dried, reconstituted in ethyl acetate and used (10%) for ER-Calux® assay. The remainder of the sample extract (90%) was derivatised and used for GC–MSD analysis. 2.4. GC–MSD
3. Results and discussion 3.1. LOD and LOQ LOD and LOQ of methods with and without extraction are the same, since the same bioassay, same concentration factor and assay dilutions are used. We have shown that in case of testing the waste waters, both assays are appropriate to use. However, in surface and tap water expected concentrations of steroid estrogens are much lower than in waste water influent and effluent (< 0.5 ng/L) and therefore the improvement of LOD is necessary. Despite time consuming, SPE concentration factor can be raised to 1000× or 10,000×
530
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
Fig. 1. Comparison of ER-Calux® with NE-(ER-Calux® ) on spiked tap (a) and waste water (c) and comparison of GC–MSD with NE-(ER-Calux® ) on spiked tap (b) and waste water (d). The results of NE-(ER-Calux® ) and ER-Calux assay are presented as means ± standard deviation of three parallels, while the results of GC–MSD are represented as mean ± relative standard deviation of measurement by GC–MSD.
(instead of 200× used in this study) and therefore lower LODs can be reached. On the other hand, in NE-(ER-Calux® ) assay, where un-concentrated water samples are tested and no concentration step of the sample is included, we are limited with certain LOD (0.68 ng(EEQ)/L). In this respect the challenge for future research is to additionally improve the method to reach LOD values <0.1 ng/L without sample extraction. 3.2. Spiked samples A comparison of the methods: NE-(ER-Calux® ), ER-Calux® and GC–MSD (Fig. 1) reveals good correlation and optimal slope of the regression line for both, tap and waste water spiked samples (for detailed results see Supplementary Table S1). A paired t-test was performed to determine if the NE-(ER-Calux® ) values are statistically the same as obtained by ER-Calux® assay and theoretical estrogenicity calculated from the results of GC–MSD. At the 0.05 level, the difference in the population means was not significantly different with the test difference (0) in any tested combination (NE(ER-Calux® ):ER-Calux® ; NE-(ER-Calux® ):GC–MSD; tap or waste water).
Several authors prove that complexity of the matrix might affect biological assay [6,19,28]. Therefore, higher differences were expected (but not observed) in case of spiked waste water samples (Fig. 1c and d) due to the presence of impurities, and water soluble estrogenic agonists and antagonists. It is possible, that in our spiked effluent sample, no other agonists or antagonists of estrogenic receptor were present. In general, the results proved that at the tested concentration range NE-(ER-Calux® ) assay was comparable to other two methods and can therefore be used for analyzing environmental samples. Anyway, comparison of methods was performed in concentration range common to waste waters [1], and the results are in agreement with Leusch et al. [13] where spiked waste waters were tested. However, for reliable testing of surface and tap water with NE-(ER-Calux® ), comparison should be made in concentrations just above LOD of the method, where higher differences could be observed, especially while comparing the results to chemical analysis. As mentioned in Section 3.1, the possibility of SPE to raise concentration factor for 50 times can make this sample preparation method more appropriate to test samples with low concentrations of estrogenic potential (<0.5 ng/L). Moreover, recent advances
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
531
in analytical instrumentation, particularly in mass spectrometry, allow even lower detection limits. 3.3. “Real” waste water and surface water samples Since different waste water samples have different complex compositions that might influence the results, the performance of the all three methods was evaluated by testing influents, effluents from seven different WWTPs as well as corresponding river water samples upstream and downstream the WWTPs. As complex environmental samples may contain cytotoxic components that may affect the results obtained with the bioassays, cytotoxicity of all the samples was tested with the MTT assay to exclude cytotoxic potential. The experiments were performed under the same exposure conditions as used in NE-(ER-Calux® ) and ER-Calux® assays for determination of estrogenic potential. None of the samples was toxic for the T47D cells (data not shown). The data obtained with GC–MSD analysis showed that the concentrations of E1 in the influents were from 6.2 to 119 ng/L, E2 from 1.3 to 10.1 ng/L, E3 from 37 to 119 ng/L, while EE2 was not detected. In the effluents the concentrations of E1 ranged from 2.2 to 51.1 ng/L, E2 from LOD to 9.0 nf/L and E3 from LOD to 45.7 ng/L. In the river water samples downstream the WWTPs the concentrations of E1 ranged from LOD to 7.4 ng/L, E2 from LOD to 3.1 ng/L and E3 from LOD to 79.8 ng/L. Of the river water samples upstream the WWTPs only one sample contained 2 ng/L E1 and 1.6 ng/L E2 (for detailed results see Supplementary material: Table S2). The calculated estrogenic potential (cEEQ) derived from GC-MDS analysis varied from 13 to 48.3 ng(cEEQ)/L in the influents, from 0.9 to 35.8 ng(cEEQ)/L in the influents, from LOD to 17 ng(cEEQ)/L in river water samples downstream the WWTPs and 2.4 ng(cEEQ)/L in the one river water sample upstream the WWTP (Fig. 2). The results of chemical analysis are in agreement with published results [1]. The values determined by ER-Calux® assay were 9.8–50.5 ng(EEQ)/L,
Fig. 2. Estrogenic potential of “real” samples from seven WWTPs, tested with NE(ER-Calux® ), ER-Calux® and GC–MSD. Error bars represent standard deviation of three parallels in NE-(ER-Calux® ) and ER-Calux® and relative standard deviation of measurement with GC–MSD.
by ER-Calux® . Substantially higher EEQ values obtained with NE(ER-Calux® ) compared to ER-Calux® were observed also in the effluent samples of WWTP2 and WWTP3 (about 1.3-fold) and river water samples obtained downstream the WWTP2, WWTP3 and WWTP6 (about 1.7-fold). The opposite was observed in WWTP4 effluent sample and river water sample upstream of WWTP1 where EEQ values obtained with NE-(ER-Calux® ) were more than twofold lower from those determined after sample extraction. The river water samples upstream WWTPs exhibited no measureable or lower estrogenic potential (WWTP1) compared to river water samples downstream WWTPs, indicating that the estrogenic
532
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
potential was most likely a result of components present in waste water effluent. This was especially observed in the cases where effluent is discharged to smaller surface waters (WWTP 1, 2 and 6). The estrogenic potential in these river water downstream samples exceed 50% of the estrogenic potential detected in the corresponding effluent sample. Similar values of estimated estrogenic activity by GC–MSD (cEEQ) and the results of NE-(ER-Calux® ) and ER-Calux® assays suggest that the estrogenic potential in our samples can be explained by the four most ubiquitous and naturally present estrogenic compounds (E1, E2, EE2 and E3). This is in agreement with other studies [4,6,12]. However, in case of WWTP6 effluent sample, theoretical estrogenic potential was higher than measured in both bioassays, possibly due to the presence of antagonists in the sample. On the other hand, in WWTP3 and WWTP5 influent and WWTP2 effluent samples the GC–MSD result was lower than the EEQ values of NE-(ER-Calux® ) and ER-Calux® assays. This might be due to the possible presence of agonists in waste water sample that were not detected with the GC–MSD. Also the observed differences between the EEQ values obtained with extracted and non-extracted samples of WWTP5 and WWTP6 influents, WWTP2 and WWTP4 effluents, river water sample downstream WWTP2 and river water sample upstream suggest that in addition to steroid estrogens unidentified water soluble agonists as well as antagonist may be present in samples. Such compounds cannot be efficiently extracted with SPE, together with steroid estrogens, and are therefore undetectable with ER-Calux® assay.
Removal was calculated with proposition, that estrogenic potential in
3.4. Intraday variability of estrogenic potential of waste water samples
An important advantage that distinguishes NE-(ER-Calux® ) assay from other two methods is sample preparation. Since there is no sample extraction, water soluble agonists or antagonist can be detected. Thus, NE-(ER-Calux® ) assay gives us more accurate information about estrogenic potential to which environmental organisms are exposed. Since low sample volumes are necessary, sample storage is easier and freezing and thawing faster. The very important aspect of NE-(ER-Calux® ) assay is the time needed for sample preparation as there is no extraction phase; compared to ER-Calux® assay the time is reduced for approximately 95%. Only several minutes are necessary for the sample to be ready for testing. Moreover, uncertainties, analytes losses (and contaminations) derived from sample handling procedures (extraction, evaporation, etc.) and sample storage can be avoided. Advantages of NE-(ER-Calux® ) assay have been exploited in a study where intraday (hourly) dynamics of estrogenic potential in WWTP were investigated. With NE-(ER-Calux® ) assay all the samples were prepared within 1 h which represents approximately 5% of time needed if the samples would be extracted for ER-Calux® or GC–MSD. However, there are also several disadvantages of NE-(ERCalux® ) method. As mentioned before, it lacks possibility for further sample concentration and the use for testing surface, tap and ground waters is limited due to low concentrations of estrogens in such waters. Without extraction, cells used in the assay are exposed to physiological factors in samples like pH, salinity, etc. that may affect the viability of cells and they might respond differently as if combination of SPE and ER-Calux® was performed. Consequently this can lead to false negative or positive results. Moreover, in case of testing without extraction the metabolism and cell survival can be affected by toxic compounds present in water samples, especially waste water samples which can lead to false negative results. For that reason we have tested in parallel to NE-(ER-Calux® ) also the cytotoxic potency of all samples with MTT assay to provide the evidence that the samples were not cytotoxic at conditions tested.
The NE-(ER-Calux® ) assay was applied for studying variation of estrogenic activity of hospital wastewater during the day. As it is evident from Fig. 3, the estrogenic potential of influent samples was increasing from 12.9 ng(EEQ)/L at 7:00 reaching the maximum value of 40.0 ng(EEQ)/L at 16:00. The calculated average estrogenic potential of the hourly samples (27.1 ng(EEQ)/L) were similar to the estrogenic potential of the time proportional sample (25.2 ng(EEQ)/L). The estrogenic potential of the effluent (with 2.5 h delay in sampling) was constantly below the LOQ (or LOD) in all cases except at 17:00. The calculated average estrogenic activity of all effluent hourly samples and time proportional sample were
Fig. 3. Intraday variations of estrogenicity in WWTP influent and effluent.
3.5. Performance of NE-(ER-Calux® ) assay
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
4. Conclusions One of the difficulties in environmental monitoring for the presence of endocrine disruptors are expensive and time consuming sample preparation methods required for chemical analysis as well as for biological testing. In this study we modified the ER-Calux® assay to enable direct determination of estrogenic potential of raw environmental samples. We confirmed that NE-(ER-Calux® ) assay gives results comparable to those obtained with conventional ERCalux® assay and theoretical estrogenic potential (cEEQ) obtained by chemical analysis with GC–MSD. As NE-(ER-Calux® ) assay is simple, fast and gives reliable results it can be recommended as screening assay in multi sample studies and for environmental monitoring, especially for testing waste waters. Acknowledgements The authors would like to acknowledge the financial support given by Slovenian Research Agency (Program groups P1-0143, P10245, Projects J1-0005 and L1 5457 as well as young researcher grant to Miha Avberˇsek). We would also like to thank all the representatives from WWTPs for their support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat. 2013.06.009. References [1] C. Miège, J.M. Choubert, L. Ribeiro, M. Eusèbe, M. Coquery, Fate of pharmaceuticals and personal care products in wastewater treatment plants – conception of a database and first results, Environ. Pollut. 157 (2009) 1721–1726. [2] G. Streck, Chemical and biological analysis of estrogenic, progestagenic and androgenic steroids in the environment, TrAC – Trends Anal. Chem. 28 (2009) 635–652. [3] G. Sándor, Advances in the analysis of steroid hormone drugs in pharmaceuticals and environmental samples (2004–2010), J. Pharm. Biomed. Anal. 55 (2011) 728–743. [4] H.-R. Aerni, B. Kobler, B. Rutishauser, F. Wettstein, R. Fischer, W. Giger, A. Hungerbühler, M.D. Marazuela, A. Peter, R. Schönenberger, A.C. Vögeli, M.F. Suter, R.L. Eggen, Combined biological and chemical assessment of estrogenic activities in wastewater treatment plant effluents, Anal. Bioanal. Chem. 378 (2004) 688–696. [5] T. Furuichi, K. Kannan, J.P. Giesy, S. Masunaga, Contribution of known endocrine disrupting substances to the estrogenic activity in Tama River water samples from Japan using instrumental analysis and in vitro reporter gene assay, Water Res. 38 (2004) 4491–4501. [6] L. Salste, P. Leskinen, M. Virta, L. Kronberg, Determination of estrogens and estrogenic activity in wastewater effluent by chemical analysis and the bioluminescent yeast assay, Sci. Total Environ. 378 (2007) 343–351. [7] C.G. Campbell, S.E. Borglin, F.B. Green, A. Grayson, E. Wozei, W.T. Stringfellow, Biologically directed environmental monitoring, fate, and transport of estrogenic endocrine disrupting compounds in water: a review, Chemosphere 65 (2006) 1265–1280. [8] E.J. Routledge, D. Sheahan, C. Desbrow, G.C. Brighty, M. Waldock, J.P. Sumpter, Identification of estrogenic chemicals in STW effluent. 2. In vivo responses in trout and roach, Environ. Sci. Technol. 32 (1998) 1559–1565.
533
[9] V. Gabet, C. Miège, P. Bados, M. Coquery, Analysis of estrogens in environmental matrices, TrAC – Trends Anal. Chem. 26 (2007) 1113–1131. [10] M. Cargouët, D. Perdiz, A. Mouatassim-Souali, S. Tamisier-Karolak, Y. Levi, Assessment of river contamination by estrogenic compounds in Paris area (France), Sci. Total Environ. 324 (2004) 55–66. [11] M.L. Jugan, L. Oziol, M. Bimbot, V. Huteau, S. Tamisier-Karolak, J.P. Blondeau, Y. Lévi, In vitro assessment of thyroid and estrogenic endocrine disruptors in wastewater treatment plants, rivers and drinking water supplies in the greater Paris area (France), Sci. Total Environ. 407 (2009) 3579–3587. [12] C. Miège, V. Gabet, M. Coquery, S. Karolak, M.L. Jugan, L. Oziol, Y. Levi, M. Chevreuil, Evaluation of estrogenic disrupting potency in aquatic environments and urban wastewaters by combining chemical and biological analysis, TrAC – Trends Anal. Chem. 28 (2009) 186–195. [13] F.D.L. Leusch, C. de Jager, Y. Levi, R. Lim, L. Puijker, F. Sacher, L.A. Tremblay, V.S. Wilson, H.F. Chapman, Comparison of five in vitro bioassays to measure estrogenic activity in environmental waters, Environ. Sci. Technol. 44 (2010) 3853–3860. [14] J.H. Shen, B. Gutendorf, H.H. Vahl, L. Shen, J. Westendorf, Toxicological profile of pollutants in surface water from an area in Taihu Lake, Yangtze Delta, Toxicology 166 (2001) 71–78. [15] C. Bicchi, T. Schilirò, C. Pignata, E. Fea, C. Cordero, F. Canale, G. Gilli, Analysis of environmental endocrine disrupting chemicals using the E-screen method and stir bar sorptive extraction in wastewater treatment plant effluents, Sci. Total Environ. 407 (2009) 1842–1851. [16] J.B. Gadd, L.A. Tremblay, G.L. Northcott, Steroid estrogens, conjugated estrogens and estrogenic activity in farm dairy shed effluents, Environ. Pollut. 158 (2010) 730–736. [17] C.J. Houtman, P.E.G. Leonards, W. Kapiteijn, J.F. Bakker, A. Brouwer, M.H. Lamoree, J. Legler, H.J.C. Klamer, Sample preparation method for the ER-CALUX bioassay screening of (xeno-)estrogenic activity in sediment extracts, Sci. Total Environ. 386 (2007) 134–144. [18] J. Legler, M. Dennekamp, A.D. Vethaak, A. Brouwer, J.H. Koeman, B. van der Burg, A.J. Murk, Detection of estrogenic activity in sediment-associated compounds using in vitro reporter gene assays, Sci. Total Environ. 293 (2002) 69–83. [19] A.J. Murk, J. Legler, M.M.H.v. Lipzig, J.H.N. Meerman, A.C. Belfroid, A. Spenkelink, B.v.d. Burg, G.B.J. Rijs, D. Vethaak, Detection of estrogenic potency in wastewater and surface water with three in vitro bioassays, Environ. Toxicol. Chem. 21 (2002) 16–23. [20] K. Cai, C.T. Elliott, D.H. Phillips, M.L. Scippo, M. Muller, L. Connolly, Treatment of estrogens and androgens in dairy wastewater by a constructed wetland system, Water Res. 46 (2012) 2333–2343. [21] E.J. Routledge, J.P. Sumpter, Estrogenic activity of surfactants and some of their degradation products assessed using a recombinant yeast screen, Environ. Toxicol. Chem. 15 (1996) 241–248. [22] J.C. Colosi, A.D. Kney, A yeast estrogen screen without extraction provides fast, reliable measures of estrogenic activity, Environ. Toxicol. Chem. 30 (2011) 2261–2269. [23] H.A. Balsiger, R. de la Torre, W.-Y. Lee, M.B. Cox, A four-hour yeast bioassay for the direct measure of estrogenic activity in wastewater without sample extraction, concentration, or sterilization, Sci. Total Environ. 408 (2010) 1422–1429. ˇ [24] M. Avberˇsek, J. Somen, E. Heath, Dynamics of steroid estrogen daily concentrations in hospital effluent and connected waste water treatment plant, J. Environ. Monit. 13 (2011) 2221–2226. [25] M. Avberˇsek, B. Zˇ egura, M. Filipiˇc, E. Heath, Integration of GC–MSD and ERCalux® assay into a single protocol for determining steroid estrogens in environmental samples, Sci. Total Environ. 409 (2011) 5069–5075. [26] T. Mosmann, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods 65 (1983) 55–63. ˇ [27] B. Zˇ egura, E. Heath, A. Cernoˇ sa, M. Filipiˇc, Combination of in vitro bioassays for the determination of cytotoxic and genotoxic potential of wastewater, surface water and drinking water samples, Chemosphere 75 (2009) 1453–1460. [28] L. Viganò, E. Benfenati, A.v. Cauwenberge, J.K. Eidem, C. Erratico, A. Goksøyr, W. Kloas, S. Maggioni, A. Mandich, R. Urbatzka, Estrogenicity profile and estrogenic compounds determined in river sediments by chemical analysis, ELISA and yeast assays, Chemosphere 73 (2008) 1078–1089. [29] S. Teske, R. Arnold, Removal of natural and xeno-estrogens during conventional wastewater treatment, Rev. Environ. Sci. Biotechnol. 7 (2008) 107–124.
Contents lists available at SciVerse ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Determination of estrogenic potential in waste water without sample extraction Miha Avberˇsek a,b , Bojana Zˇ egura c , Metka Filipiˇc c , Nataˇsa Uranjek-Zˇ evart d , Ester Heath a,b,∗ a
“Joˇzef Stefan” Institute, Jamova 39, 1000 Ljubljana, Slovenia International Postgraduate School Joˇzef Stefan, Jamova 39, 1000 Ljubljana, Slovenia c National Institute of Biology, Veˇcna pot 111, 1000 Ljubljana, Slovenia d Komunalno podjetje Velenje d.o.o., Koroˇska cesta 37b, 3320 Velenje, Slovenia b
h i g h l i g h t s • • • •
A modified ER-Calux® (NE-(ER-Calux® )) does not need pre-extraction of raw water samples. NE-(ER-Calux® ) enables determination of estrogenic potential of raw water samples. The sensitivities of NE-(ER-Calux® ) and conventional ER-Calux® assay are comparable. NE-(ER-Calux® ) assay is recommended as a screening assay in multi sample studies.
a r t i c l e
i n f o
Article history: Received 11 March 2013 Received in revised form 29 May 2013 Accepted 3 June 2013 Available online 10 June 2013 Keywords: ER-Calux® No extraction GC–MSD Waste water Estrogenic potential
a b s t r a c t This study describes the modification of the ER-Calux® assay for testing water samples without sample extraction (NE-(ER-Calux® ) assay). The results are compared to those obtained with ER-Calux® assay and a theoretical estrogenic potential obtained by GC–MSD. For spiked tap and waste water samples there was no statistical difference between estrogenic potentials obtained by the three methods. Application of NE(ER-Calux® ) to “real” influent and effluents from municipal waste water treatment plants and receiving surface waters found that the NE-(ER-Calux® ) assay gave higher values compared to ER-Calux® assay and GC–MSD. This is explained by the presence of water soluble endocrine agonists that are usually removed during extraction. Intraday dynamics of the estrogenic potential of a WWTP influent and effluent revealed an increase in the estrogenic potential of the influent from 12.9 ng(EEQ)/L in the morning to a peak value of 40.0 ng(EEQ)/L in the afternoon. The estrogenic potential of the effluent was
1. Introduction Naturally excreted steroid estrogens are ubiquitously present in waste water and other environmental samples [1–3]. They have the highest estrogenic activity among endocrine disrupting compounds and account for the majority of estrogenic potential in municipal waste water [4–6]. Concentrations of steroid estrogens in WWTP effluent are typically in the low ng/L range [1], but this is still sufficient to affect the endocrine system of living organisms [7,8].
∗ Corresponding author at: “Joˇzef Stefan” Institute, Jamova cesta 39, 1000 Ljubljana, Slovenia. Tel.: +386 1 4773 584; fax: +386 1 2519 385. E-mail address: [email protected] (E. Heath). 0304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2013.06.009
To detect such low concentrations of estrogens sensitive methods are necessary. Gas or liquid chromatography coupled with mass spectrometry are the techniques of choice for quantitative chemical analyses [1,3,9] while total estrogenic potential of the sample is obtained by biological in vivo or in vitro assays [2,7]. Several cellbased in vitro bioassays like MELN [10–13], MVLN [14], E-Screen [13,15,16], ER-Calux® assay [13,17–19] and MMV-Luc [20] have been developed for environmental samples. Alternatively, recombinant yeast based assays are available [2]. These are easier to use, but they lack complex estrogenic interactions [21,21], are less sensitive than mammalian cell-based assays and are unable to detect anti-estrogenic compounds [19]. Low concentrations of estrogens also mean that extraction from complex environmental matrices and pre-concentration of analytes is essential for both detection and quantification with the bio-assays. Liquid–liquid or solid phase extraction (SPE) with
528
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
additional clean-up is generally applied for chemical analysis and bioassays [2,22] and only a few published studies describe the determination of estrogenic potential without extensive sample preparation and extraction. Of these the majority employ a recombinant yeast assay [6,22,23] and just one applies a mammalian cell-based assay (MELN cells) without presenting any results [10]. The aim of this study was to modify the ER-Calux® assay for the determination of estrogenic potential of water samples without extensive sample handling and extraction of analytes. The modified method that we named NE-(ER-Calux® ) assay, was tested with tap and waste water samples spiked with steroid estrogens, and compared to conventional ER-Calux® assay and chemical analysis with gas chromatography–mass selective detection (GC–MSD). The modified method was then applied for investigating estrogenic potential of “real” environmental samples and for studying intraday dynamics of estrogenic potential in influent and effluent samples of WWTP. 2. Materials and methods 2.1. Standards, chemicals, growth media Standards estrone (E1; min 99%), 17-estradiol (E2; min 98%), 17␣-ethinylestradiol (EE2; min 98% (HPLC)), estriol (E3; min 99%), a deuterated internal standard (bisphenol A)-d16 (98 atom% D) were purchased from Sigma (Steinheim, Germany). Standards were prepared freshly in ethyl acetate, for ER-Calux® assay calibration curve and in methanol for spiking water samples used for analyses. Methanol, ethyl acetate “Baker ultra resi-analysed® ” grade were purchased from J.T. Baker (Deventer, the Netherlands). Pyridine (max 0.01% H2 O) was purchased from Merck (Darmstadt, Germany). The derivatising agent N-methyl-N-(trimethylsilyl)trifluoroacetamide (MSTFA; derivatisation grade), was purchased from Sigma. For ER-Calux® assay, media Gibco® D-MEM/F-12 with GlutaMAXTM (with phenol red), Gibco® D-MEM/F-12 with lglutamine (without phenol red) and Stripped FBS (foetal bovine serum) were purchased from Invitrogen (Paisley, UK). EDTA, nonessential amino acids (MEM 100×), and penicillin/streptomycin were purchased from Sigma. FBS (foetal bovine serum) and PBS
(Phosphate Buffered Saline) were purchased from PAA (Pasching, Austria) while Difco trypsin was obtained from Becton Dickinsen (Heidelberg, Germany). 2.2. Sampling and sample handling Sampling was performed for three different purposes as follows. Spiked tap and waste water effluent samples were used for optimisation purposes, while waste water treatment plant (WWTP) influent and effluent samples as well as surface waters were used to study “real” samples. Hourly samples were collected at one WWTP to measure intraday variations of estrogenic potential in influent and effluent samples. 2.2.1. Spiked samples Tap water from our laboratory and grab waste water effluent sample from WWTP2 (Table 1) were spiked with estrone (E1), 17-estradiol (E2), 17␣-ethinylestradiol (EE2) and estriol (E3) at environmentally relevant concentrations (0–40 ng/L) at levels that were chosen randomly within this range (see Supplementary material: Table S1). After spiking, samples were homogenised by shaking at 300 rpm for 30 min. For the NE-(ER-Calux® ) assay, 10 mL of each sample was stored at −20 ◦ C, while 200 mL was immediately used for SPE (see Section 2.3). Extracts of the samples were analysed by ER-Calux® assay and GC–MSD, while un-extracted samples were analysed by NE-(ER-Calux® ) assay. 2.2.2. “Real” waste water and surface water samples Grab samples (250 mL, glass bottles) of WWTP influent and effluent samples and surface river water samples (upstream and downstream of the effluent site) were collected from seven different WWTPs (Table 1). Samples at each WWTP were collected on four consecutive weeks (two WWTPs per week), on Monday morning. In order to assure the same sample storage and preparation time, effluent and river samples were taken without considering hydraulic retention time. Similar to spiked samples, 10 mL of the sample was stored at −20 ◦ C for NE-(ER-Calux® ) assay and 200 mL was used for the extraction and analysis by ER-Calux® assay and GC–MSD. To avoid sample degradation, the extraction was performed within 4 h after the sampling.
Table 1 Main characteristics of waste water treatment plants involved in the study. Name
Treatment
WWTP1
Biofiltration with P and N removal
WWTP2
Capacity (PU)
Design
Flow/year (m3 )
Mean influent COD (BOD) (mg/L)
Mean effluent COD (BOD) (mg/L)
SRT (days)
Estimated surface water flow (m3 /s)
Actual NAa
2
18
15–32
7
43 (<10)
19
8
800 (300)
80 (15)
22
20
110
10,000,000
740 (400)
26 (6)
29
12
20
110,805
4,795,963
1013 (506)
128 (28)
22
4–10
85,000
8,486,259
429 (214)
17 (4)
22.5
20–24
50,000
45,000
6,161,222
400 (196)
Activated sludge, nitrification; no P removal
200,000
143,623
7,303,085
576 (66)
294 (15)
WWTP3
Activated sludge; nitrification; no P removal
360,000
420,000
29,928,900
590 (312)
WWTP4
Activated sludge; nitrification; no P removal
100,000
85,000
5,500,000
WWTP5
Activated sludge, anoxic zones, P removal
250,000
160,000
WWTP6
Activated sludge, no P removal
68,000
WWTP7
Activated sludge, nitrification, P removal
70,000
a
HRT (h)
39.7 (8.2)
Sludge age as determined for suspended biomass is not relevant in water treatment with biofiltration.
2.5
20
0.5 35
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
529
2.2.3. Intraday waste water samples from WWTP1 In this part of the study, samples were taken at WWTP1. Hourly samples and time proportional samples were collected at influent and effluent site. Grab influent samples were taken every hour from 7:00 to 19:00. Appropriate effluent samples were taken within a 2.5 h shift (9:30–21:30) according to the hydraulic retention time at WWTP1 (Table 1). Hourly waste water samples were analysed with NE-(ER-Calux® ) assay only and therefore, 10 mL of sample volume was sampled in polypropylene flasks. Samples were stored at −20 ◦ C and were analysed the next morning as described in Section 2.6. Time proportional samples for the whole sampling interval (sampling every 15 min) were collected from the influent (7:00–19:00) and effluent (9:30–21:30) and analysed as well. To check the stability of water treatment system the following parameters were monitored: temperature, pH, chemical oxygen demand – COD (sensor name: UVAS), dissolved oxygen – O2 (LDO), ammonium – NH4 (AMTAX), phosphates – PO4 (PHOSPHAX), nitrates – NO3 (NITRAX) were measured with HACH LANGE controller system (Düsseldorf, Germany).
luminometer (top reading plate reader) with a set integration time of 4000 ms and a gain of 200 (Tecan Genios, Maennedorf, Switzerland). Results expressed in relative luminescent units (RLU) were processed using a sigmoidal calibration curve constructed with a MS Excel template (provided by BDS) together with an add-in “Solver”. The E2 equivalents (EEQ) were calculated as described by Avberˇsek et al. [25]. LOD (0.136 ng(EEQ)/L (=0.5 pM)) and LOQ (0.409 ng(EEQ)/L (=1.5 pM)) of ER-Calux® assay was determined by manufacturer of the assay (BDS, b.v.) on the basis of their long term experience with the assay. However, LODs and LOQs were also determined from calibration curve each time the assay was performed. These values were only used to test the performance of the bioassay. Therefore, together with concentration factor of SPE–ER Calux® (0.2×), limits of detection (LOD) and quantification (LOQ) assay were 0.68 ng(EEQ)/L and 2.05 ng(EEQ)/L, respectively. Recovery (97 %), and repeatability (4%) were determined with tap water (n = 5), spiked with E2 (c = 4.09 ng/L (=15 pM)) [25].
2.3. Sample extraction
All spiked and “real” sample extracts (90%) that remain after ER-Calux® assay were derivatised with MSTFA and analysed with GC–MSD (HP 6890 Series, Hewlet-Packard, Waldbron, Germany) [24]. Separation was achieved on a HP5-MS column. SIM mode was used for quantification of E1, E2, EE2 and E3. Estrogenic potential was calculated from concentrations of each compound as described by Avberˇsek et al. [25]. Validation of the method was performed using tap and waste water as a matrix [24,25].
2.5.1. NE-(ER-Calux® ) assay The standard ER-Calux® assay has been adapted for testing the water samples without prior extraction and we named it NE-(ERCalux® ) assay. The frozen samples (10 mL) were thawed, sterilised by filtration (Whatman, ANOTOP25, pore size 0.2 m) and diluted 1:1, 1:3, 1:10, 1:30 and 1:100 with PBS. For the cell treatment 20 vol% of diluted samples were mixed with “Test medium”. For the control “Test medium” was mixed with 20 vol% PBS and for the calibration E2 was added (final concentrations 0.3, 0.6, 1, 3, 6, 10, and 30 pM/well). Further steps of the assay were the same as in the conventional ER-Calux® assay. In comparison to ER-Calux® , no significant difference was observed in LODs and LOQs determined from standard calibration curve performed on each test plate. For that reason, we keep the same LOD and LOQ for NE-(ER-Calux® ) as well. Together with concentration factor (0.2×), limit of detection (LOD) and quantification (LOQ) of NE-(ER-Calux® ) assay were 0.68 ng(EEQ)/L and 2.05 ng(EEQ)/L, respectively. Recovery (92%), and repeatability (5%) were determined with tap water (n = 5), spiked with E2 (c = 4.09 ng/L (=15 pM)).
2.5. Determination of estrogenic potential with ER-Calux® assay
2.6. Toxicity testing
The ER-Calux® assay was performed as described by Avberˇsek et al. [25]. Briefly, T47D-EREtata-Luc cells were used (provided and licenced by BioDetection Systems (BDS) b.v. Amsterdam, the Netherlands) and maintained at 37 ◦ C; 5% CO2 in “Growth medium” (D-MEM/F12; GlutaMAXTM ; phenol red as pH indicator; 7.5% FBS, 1% nonessential amino acids (MEM) and a 0.2% penicillin/streptomycin solution). For the assay purposes the “Growth medium” was replaced with the “Test medium” (D-MEM/F12; lglutamine; without phenol red; 5% stripped FBS, 1% nonessential amino acids (MEM) and 0.2% penicillin/streptomycin solution). The cells in the “Test medium” were seeded (10,000 cells/well) into 96-well white microtiter plates (Nunc, Roskilde, Denmark), incubated for 24 h (37 ◦ C), then the medium was replaced with fresh “Test medium” and incubated for additional 24 h. The medium was then replaced with 100 L of “Test medium”, supplemented with 0.1% sample extracts corresponding to 1 L of sample extract or its dilution (1:1, 1:3, 1:10, 1:30 and 1:100) in 1 mL of “Test medium”. E2 (0.6–30 pM/well) was used the calibration standard and 0.1 vol% ethyl acetate as the solvent control. All the samples were tested in triplicate wells. After 24 h incubation, the luminescence kit “Steadylite plusTM ” (Perkin Elmer, Shelton, USA) was added (100 L) to the medium and gently shaken for 15 min at room temperature. Luminescence was measured using a TECAN
MTT assay was used for toxicity testing of all waste water samples. MTT assay was performed on ER-Calux® cells according to Mosmann [26] with minor modifications [27]. ER-Calux® cells were exposed to the samples in the same way as in ER-Calux® or NE-(ER-Calux® ) assays. However, instead of the luminescence kit, MTT reagent was added after 24 h exposure. After 3 h incubation, the difference in optical density at 570 and 690 nm wavelength was measured (Tecan Genios, Maennedorf, Switzerland). The results were compared to control sample (20% PBS).
Sample extraction for ER-Calux® and GC–MSD was performed as described by Avberˇsek et al. [24,25]. Samples (200 mL) were filtered and afterwards extracted with SPE (Waters, Oasis HLB) and eluted with ethyl acetate. Extracts were cleaned-up on silica-gel cartridges (Biotage, Isolute SI), dried, reconstituted in ethyl acetate and used (10%) for ER-Calux® assay. The remainder of the sample extract (90%) was derivatised and used for GC–MSD analysis. 2.4. GC–MSD
3. Results and discussion 3.1. LOD and LOQ LOD and LOQ of methods with and without extraction are the same, since the same bioassay, same concentration factor and assay dilutions are used. We have shown that in case of testing the waste waters, both assays are appropriate to use. However, in surface and tap water expected concentrations of steroid estrogens are much lower than in waste water influent and effluent (< 0.5 ng/L) and therefore the improvement of LOD is necessary. Despite time consuming, SPE concentration factor can be raised to 1000× or 10,000×
530
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
Fig. 1. Comparison of ER-Calux® with NE-(ER-Calux® ) on spiked tap (a) and waste water (c) and comparison of GC–MSD with NE-(ER-Calux® ) on spiked tap (b) and waste water (d). The results of NE-(ER-Calux® ) and ER-Calux assay are presented as means ± standard deviation of three parallels, while the results of GC–MSD are represented as mean ± relative standard deviation of measurement by GC–MSD.
(instead of 200× used in this study) and therefore lower LODs can be reached. On the other hand, in NE-(ER-Calux® ) assay, where un-concentrated water samples are tested and no concentration step of the sample is included, we are limited with certain LOD (0.68 ng(EEQ)/L). In this respect the challenge for future research is to additionally improve the method to reach LOD values <0.1 ng/L without sample extraction. 3.2. Spiked samples A comparison of the methods: NE-(ER-Calux® ), ER-Calux® and GC–MSD (Fig. 1) reveals good correlation and optimal slope of the regression line for both, tap and waste water spiked samples (for detailed results see Supplementary Table S1). A paired t-test was performed to determine if the NE-(ER-Calux® ) values are statistically the same as obtained by ER-Calux® assay and theoretical estrogenicity calculated from the results of GC–MSD. At the 0.05 level, the difference in the population means was not significantly different with the test difference (0) in any tested combination (NE(ER-Calux® ):ER-Calux® ; NE-(ER-Calux® ):GC–MSD; tap or waste water).
Several authors prove that complexity of the matrix might affect biological assay [6,19,28]. Therefore, higher differences were expected (but not observed) in case of spiked waste water samples (Fig. 1c and d) due to the presence of impurities, and water soluble estrogenic agonists and antagonists. It is possible, that in our spiked effluent sample, no other agonists or antagonists of estrogenic receptor were present. In general, the results proved that at the tested concentration range NE-(ER-Calux® ) assay was comparable to other two methods and can therefore be used for analyzing environmental samples. Anyway, comparison of methods was performed in concentration range common to waste waters [1], and the results are in agreement with Leusch et al. [13] where spiked waste waters were tested. However, for reliable testing of surface and tap water with NE-(ER-Calux® ), comparison should be made in concentrations just above LOD of the method, where higher differences could be observed, especially while comparing the results to chemical analysis. As mentioned in Section 3.1, the possibility of SPE to raise concentration factor for 50 times can make this sample preparation method more appropriate to test samples with low concentrations of estrogenic potential (<0.5 ng/L). Moreover, recent advances
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
531
in analytical instrumentation, particularly in mass spectrometry, allow even lower detection limits. 3.3. “Real” waste water and surface water samples Since different waste water samples have different complex compositions that might influence the results, the performance of the all three methods was evaluated by testing influents, effluents from seven different WWTPs as well as corresponding river water samples upstream and downstream the WWTPs. As complex environmental samples may contain cytotoxic components that may affect the results obtained with the bioassays, cytotoxicity of all the samples was tested with the MTT assay to exclude cytotoxic potential. The experiments were performed under the same exposure conditions as used in NE-(ER-Calux® ) and ER-Calux® assays for determination of estrogenic potential. None of the samples was toxic for the T47D cells (data not shown). The data obtained with GC–MSD analysis showed that the concentrations of E1 in the influents were from 6.2 to 119 ng/L, E2 from 1.3 to 10.1 ng/L, E3 from 37 to 119 ng/L, while EE2 was not detected. In the effluents the concentrations of E1 ranged from 2.2 to 51.1 ng/L, E2 from LOD to 9.0 nf/L and E3 from LOD to 45.7 ng/L. In the river water samples downstream the WWTPs the concentrations of E1 ranged from LOD to 7.4 ng/L, E2 from LOD to 3.1 ng/L and E3 from LOD to 79.8 ng/L. Of the river water samples upstream the WWTPs only one sample contained 2 ng/L E1 and 1.6 ng/L E2 (for detailed results see Supplementary material: Table S2). The calculated estrogenic potential (cEEQ) derived from GC-MDS analysis varied from 13 to 48.3 ng(cEEQ)/L in the influents, from 0.9 to 35.8 ng(cEEQ)/L in the influents, from LOD to 17 ng(cEEQ)/L in river water samples downstream the WWTPs and 2.4 ng(cEEQ)/L in the one river water sample upstream the WWTP (Fig. 2). The results of chemical analysis are in agreement with published results [1]. The values determined by ER-Calux® assay were 9.8–50.5 ng(EEQ)/L,
Fig. 2. Estrogenic potential of “real” samples from seven WWTPs, tested with NE(ER-Calux® ), ER-Calux® and GC–MSD. Error bars represent standard deviation of three parallels in NE-(ER-Calux® ) and ER-Calux® and relative standard deviation of measurement with GC–MSD.
by ER-Calux® . Substantially higher EEQ values obtained with NE(ER-Calux® ) compared to ER-Calux® were observed also in the effluent samples of WWTP2 and WWTP3 (about 1.3-fold) and river water samples obtained downstream the WWTP2, WWTP3 and WWTP6 (about 1.7-fold). The opposite was observed in WWTP4 effluent sample and river water sample upstream of WWTP1 where EEQ values obtained with NE-(ER-Calux® ) were more than twofold lower from those determined after sample extraction. The river water samples upstream WWTPs exhibited no measureable or lower estrogenic potential (WWTP1) compared to river water samples downstream WWTPs, indicating that the estrogenic
532
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
potential was most likely a result of components present in waste water effluent. This was especially observed in the cases where effluent is discharged to smaller surface waters (WWTP 1, 2 and 6). The estrogenic potential in these river water downstream samples exceed 50% of the estrogenic potential detected in the corresponding effluent sample. Similar values of estimated estrogenic activity by GC–MSD (cEEQ) and the results of NE-(ER-Calux® ) and ER-Calux® assays suggest that the estrogenic potential in our samples can be explained by the four most ubiquitous and naturally present estrogenic compounds (E1, E2, EE2 and E3). This is in agreement with other studies [4,6,12]. However, in case of WWTP6 effluent sample, theoretical estrogenic potential was higher than measured in both bioassays, possibly due to the presence of antagonists in the sample. On the other hand, in WWTP3 and WWTP5 influent and WWTP2 effluent samples the GC–MSD result was lower than the EEQ values of NE-(ER-Calux® ) and ER-Calux® assays. This might be due to the possible presence of agonists in waste water sample that were not detected with the GC–MSD. Also the observed differences between the EEQ values obtained with extracted and non-extracted samples of WWTP5 and WWTP6 influents, WWTP2 and WWTP4 effluents, river water sample downstream WWTP2 and river water sample upstream suggest that in addition to steroid estrogens unidentified water soluble agonists as well as antagonist may be present in samples. Such compounds cannot be efficiently extracted with SPE, together with steroid estrogens, and are therefore undetectable with ER-Calux® assay.
Removal was calculated with proposition, that estrogenic potential in
3.4. Intraday variability of estrogenic potential of waste water samples
An important advantage that distinguishes NE-(ER-Calux® ) assay from other two methods is sample preparation. Since there is no sample extraction, water soluble agonists or antagonist can be detected. Thus, NE-(ER-Calux® ) assay gives us more accurate information about estrogenic potential to which environmental organisms are exposed. Since low sample volumes are necessary, sample storage is easier and freezing and thawing faster. The very important aspect of NE-(ER-Calux® ) assay is the time needed for sample preparation as there is no extraction phase; compared to ER-Calux® assay the time is reduced for approximately 95%. Only several minutes are necessary for the sample to be ready for testing. Moreover, uncertainties, analytes losses (and contaminations) derived from sample handling procedures (extraction, evaporation, etc.) and sample storage can be avoided. Advantages of NE-(ER-Calux® ) assay have been exploited in a study where intraday (hourly) dynamics of estrogenic potential in WWTP were investigated. With NE-(ER-Calux® ) assay all the samples were prepared within 1 h which represents approximately 5% of time needed if the samples would be extracted for ER-Calux® or GC–MSD. However, there are also several disadvantages of NE-(ERCalux® ) method. As mentioned before, it lacks possibility for further sample concentration and the use for testing surface, tap and ground waters is limited due to low concentrations of estrogens in such waters. Without extraction, cells used in the assay are exposed to physiological factors in samples like pH, salinity, etc. that may affect the viability of cells and they might respond differently as if combination of SPE and ER-Calux® was performed. Consequently this can lead to false negative or positive results. Moreover, in case of testing without extraction the metabolism and cell survival can be affected by toxic compounds present in water samples, especially waste water samples which can lead to false negative results. For that reason we have tested in parallel to NE-(ER-Calux® ) also the cytotoxic potency of all samples with MTT assay to provide the evidence that the samples were not cytotoxic at conditions tested.
The NE-(ER-Calux® ) assay was applied for studying variation of estrogenic activity of hospital wastewater during the day. As it is evident from Fig. 3, the estrogenic potential of influent samples was increasing from 12.9 ng(EEQ)/L at 7:00 reaching the maximum value of 40.0 ng(EEQ)/L at 16:00. The calculated average estrogenic potential of the hourly samples (27.1 ng(EEQ)/L) were similar to the estrogenic potential of the time proportional sample (25.2 ng(EEQ)/L). The estrogenic potential of the effluent (with 2.5 h delay in sampling) was constantly below the LOQ (or LOD) in all cases except at 17:00. The calculated average estrogenic activity of all effluent hourly samples and time proportional sample were
Fig. 3. Intraday variations of estrogenicity in WWTP influent and effluent.
3.5. Performance of NE-(ER-Calux® ) assay
M. Avberˇsek et al. / Journal of Hazardous Materials 260 (2013) 527–533
4. Conclusions One of the difficulties in environmental monitoring for the presence of endocrine disruptors are expensive and time consuming sample preparation methods required for chemical analysis as well as for biological testing. In this study we modified the ER-Calux® assay to enable direct determination of estrogenic potential of raw environmental samples. We confirmed that NE-(ER-Calux® ) assay gives results comparable to those obtained with conventional ERCalux® assay and theoretical estrogenic potential (cEEQ) obtained by chemical analysis with GC–MSD. As NE-(ER-Calux® ) assay is simple, fast and gives reliable results it can be recommended as screening assay in multi sample studies and for environmental monitoring, especially for testing waste waters. Acknowledgements The authors would like to acknowledge the financial support given by Slovenian Research Agency (Program groups P1-0143, P10245, Projects J1-0005 and L1 5457 as well as young researcher grant to Miha Avberˇsek). We would also like to thank all the representatives from WWTPs for their support. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhazmat. 2013.06.009. References [1] C. Miège, J.M. Choubert, L. Ribeiro, M. Eusèbe, M. Coquery, Fate of pharmaceuticals and personal care products in wastewater treatment plants – conception of a database and first results, Environ. Pollut. 157 (2009) 1721–1726. [2] G. Streck, Chemical and biological analysis of estrogenic, progestagenic and androgenic steroids in the environment, TrAC – Trends Anal. Chem. 28 (2009) 635–652. [3] G. Sándor, Advances in the analysis of steroid hormone drugs in pharmaceuticals and environmental samples (2004–2010), J. Pharm. Biomed. Anal. 55 (2011) 728–743. [4] H.-R. Aerni, B. Kobler, B. Rutishauser, F. Wettstein, R. Fischer, W. Giger, A. Hungerbühler, M.D. Marazuela, A. Peter, R. Schönenberger, A.C. Vögeli, M.F. Suter, R.L. Eggen, Combined biological and chemical assessment of estrogenic activities in wastewater treatment plant effluents, Anal. Bioanal. Chem. 378 (2004) 688–696. [5] T. Furuichi, K. Kannan, J.P. Giesy, S. Masunaga, Contribution of known endocrine disrupting substances to the estrogenic activity in Tama River water samples from Japan using instrumental analysis and in vitro reporter gene assay, Water Res. 38 (2004) 4491–4501. [6] L. Salste, P. Leskinen, M. Virta, L. Kronberg, Determination of estrogens and estrogenic activity in wastewater effluent by chemical analysis and the bioluminescent yeast assay, Sci. Total Environ. 378 (2007) 343–351. [7] C.G. Campbell, S.E. Borglin, F.B. Green, A. Grayson, E. Wozei, W.T. Stringfellow, Biologically directed environmental monitoring, fate, and transport of estrogenic endocrine disrupting compounds in water: a review, Chemosphere 65 (2006) 1265–1280. [8] E.J. Routledge, D. Sheahan, C. Desbrow, G.C. Brighty, M. Waldock, J.P. Sumpter, Identification of estrogenic chemicals in STW effluent. 2. In vivo responses in trout and roach, Environ. Sci. Technol. 32 (1998) 1559–1565.
533
[9] V. Gabet, C. Miège, P. Bados, M. Coquery, Analysis of estrogens in environmental matrices, TrAC – Trends Anal. Chem. 26 (2007) 1113–1131. [10] M. Cargouët, D. Perdiz, A. Mouatassim-Souali, S. Tamisier-Karolak, Y. Levi, Assessment of river contamination by estrogenic compounds in Paris area (France), Sci. Total Environ. 324 (2004) 55–66. [11] M.L. Jugan, L. Oziol, M. Bimbot, V. Huteau, S. Tamisier-Karolak, J.P. Blondeau, Y. Lévi, In vitro assessment of thyroid and estrogenic endocrine disruptors in wastewater treatment plants, rivers and drinking water supplies in the greater Paris area (France), Sci. Total Environ. 407 (2009) 3579–3587. [12] C. Miège, V. Gabet, M. Coquery, S. Karolak, M.L. Jugan, L. Oziol, Y. Levi, M. Chevreuil, Evaluation of estrogenic disrupting potency in aquatic environments and urban wastewaters by combining chemical and biological analysis, TrAC – Trends Anal. Chem. 28 (2009) 186–195. [13] F.D.L. Leusch, C. de Jager, Y. Levi, R. Lim, L. Puijker, F. Sacher, L.A. Tremblay, V.S. Wilson, H.F. Chapman, Comparison of five in vitro bioassays to measure estrogenic activity in environmental waters, Environ. Sci. Technol. 44 (2010) 3853–3860. [14] J.H. Shen, B. Gutendorf, H.H. Vahl, L. Shen, J. Westendorf, Toxicological profile of pollutants in surface water from an area in Taihu Lake, Yangtze Delta, Toxicology 166 (2001) 71–78. [15] C. Bicchi, T. Schilirò, C. Pignata, E. Fea, C. Cordero, F. Canale, G. Gilli, Analysis of environmental endocrine disrupting chemicals using the E-screen method and stir bar sorptive extraction in wastewater treatment plant effluents, Sci. Total Environ. 407 (2009) 1842–1851. [16] J.B. Gadd, L.A. Tremblay, G.L. Northcott, Steroid estrogens, conjugated estrogens and estrogenic activity in farm dairy shed effluents, Environ. Pollut. 158 (2010) 730–736. [17] C.J. Houtman, P.E.G. Leonards, W. Kapiteijn, J.F. Bakker, A. Brouwer, M.H. Lamoree, J. Legler, H.J.C. Klamer, Sample preparation method for the ER-CALUX bioassay screening of (xeno-)estrogenic activity in sediment extracts, Sci. Total Environ. 386 (2007) 134–144. [18] J. Legler, M. Dennekamp, A.D. Vethaak, A. Brouwer, J.H. Koeman, B. van der Burg, A.J. Murk, Detection of estrogenic activity in sediment-associated compounds using in vitro reporter gene assays, Sci. Total Environ. 293 (2002) 69–83. [19] A.J. Murk, J. Legler, M.M.H.v. Lipzig, J.H.N. Meerman, A.C. Belfroid, A. Spenkelink, B.v.d. Burg, G.B.J. Rijs, D. Vethaak, Detection of estrogenic potency in wastewater and surface water with three in vitro bioassays, Environ. Toxicol. Chem. 21 (2002) 16–23. [20] K. Cai, C.T. Elliott, D.H. Phillips, M.L. Scippo, M. Muller, L. Connolly, Treatment of estrogens and androgens in dairy wastewater by a constructed wetland system, Water Res. 46 (2012) 2333–2343. [21] E.J. Routledge, J.P. Sumpter, Estrogenic activity of surfactants and some of their degradation products assessed using a recombinant yeast screen, Environ. Toxicol. Chem. 15 (1996) 241–248. [22] J.C. Colosi, A.D. Kney, A yeast estrogen screen without extraction provides fast, reliable measures of estrogenic activity, Environ. Toxicol. Chem. 30 (2011) 2261–2269. [23] H.A. Balsiger, R. de la Torre, W.-Y. Lee, M.B. Cox, A four-hour yeast bioassay for the direct measure of estrogenic activity in wastewater without sample extraction, concentration, or sterilization, Sci. Total Environ. 408 (2010) 1422–1429. ˇ [24] M. Avberˇsek, J. Somen, E. Heath, Dynamics of steroid estrogen daily concentrations in hospital effluent and connected waste water treatment plant, J. Environ. Monit. 13 (2011) 2221–2226. [25] M. Avberˇsek, B. Zˇ egura, M. Filipiˇc, E. Heath, Integration of GC–MSD and ERCalux® assay into a single protocol for determining steroid estrogens in environmental samples, Sci. Total Environ. 409 (2011) 5069–5075. [26] T. Mosmann, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods 65 (1983) 55–63. ˇ [27] B. Zˇ egura, E. Heath, A. Cernoˇ sa, M. Filipiˇc, Combination of in vitro bioassays for the determination of cytotoxic and genotoxic potential of wastewater, surface water and drinking water samples, Chemosphere 75 (2009) 1453–1460. [28] L. Viganò, E. Benfenati, A.v. Cauwenberge, J.K. Eidem, C. Erratico, A. Goksøyr, W. Kloas, S. Maggioni, A. Mandich, R. Urbatzka, Estrogenicity profile and estrogenic compounds determined in river sediments by chemical analysis, ELISA and yeast assays, Chemosphere 73 (2008) 1078–1089. [29] S. Teske, R. Arnold, Removal of natural and xeno-estrogens during conventional wastewater treatment, Rev. Environ. Sci. Biotechnol. 7 (2008) 107–124.