Environ. Sci. Technol. 2006, 40, 2594-2600
Assessment of the Reproductive-Endocrine Disrupting Potential of Chlorine Dioxide Oxidation Products of Plant Sterols M I C H A E L R . V A N D E N H E U V E L , * ,† FREDERIC D. L. LEUSCH,‡ SEAN TAYLOR,† NICHOLAS SHANNON,† AND A. BRUCE MCKAGUE§ Scion, Sala Street, Private Bag 3020, Rotorua, New Zealand, CRC for Water Quality and Treatment, Queensland Health Scientific Services, 39 Kessels Road, Coopers Plains, Brisbane, Qld 4108, Australia, and Pulp and Paper Centre, University of Toronto, 200 College Street, Toronto, Ontario, Canada, M5S 3E5
This study examined the hypothesis that chlorine dioxide bleaching used in pulp and paper production causes the formation of reproductive-endocrine disrupting compounds from plant sterols. This was tested by conducting a laboratory simulation of the chlorine dioxide oxidation of two plant sterols, β-sitosterol and stigmasterol. Oxidation products of the plant sterol β-sitosterol were purified and identified and found to be cholestan-24-ethyl-3-one, 4-cholestene-24-ethyl-3-one, and 4-cholestene-24-ethyl-3,6dione. The first two compounds were found in a number of pulp and paper effluents and biosolids. The sterols and their oxidation products were tested in vitro using bioassays for androgenicity and estrogenicity. A 28 d in vivo bioassay was employed to examine masculinization in female mosquitofish. In vitro bioassays revealed little estrogenic activity in the parent sterols or in mixtures of their oxidation products. Androgenic activity as measured by the androgen receptor binding bioassay was in the order of 19-96 µg/g testosterone equivalents but with no increase or decrease with chlorine dioxide oxidation. The mosquitofish bioassay did not show significant masculinization for any of the preparations tested. A number of androstane steroids were identified in the sterols tested, however, those compounds could only account for a small fraction of the androgenic activity in the sterols. It was clear that the parent sterols were not themselves acting as androgens in the bioassays used. This study indicated that chlorine dioxide oxidation of sterols produced predominantly oxidized sterols that were not likely to act through androgenic or estrogenic mechanisms.
* Corresponding author current address: Canada Research Chair in Watershed Ecological Integrity, Department of Biology, University of Prince Edward Island, 550 University Avenue, Charlottetown, Prince Edward Island C1A 4P3, Canada; phone: 1-902-566-6072; fax: 1-902566-0740; e-mail:
[email protected]. † Scion. ‡ CRC for Water Quality and Treatment. § University of Toronto. 2594
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 8, 2006
Introduction The observation that pulp and paper effluents cause subtle reproductive alterations to fishes has been documented throughout the world. Studies in Scandinavia have shown reduced gonad growth, reduced sex steroid hormones (1), and altered sex ratios (2). Studies in Canada have observed reduced gonad development and depressed sex steroid hormones (3, 4) with similar results observed in the United States (5, 6). Studies in New Zealand have demonstrated reduced gonad growth related to an endocrine mechanism as well as the exposure to both androgens and estrogens in effluent (7, 8). The mechanism and compound(s) responsible for reproductive dysfunction in fishes exposed to pulp and paper effluents remain ambiguous. Evidence for estrogen receptor agonists in effluent has been suggested due to the observation of the induction of vitellogenin (7, 9) or vitellogenin mRNA (10) in fishes exposed to pulp and paper effluent. Evidence for an androgen mediated response has been suggested by the masculinization of female mosquitofish and by altered sex ratios in eelpout (2, 8, 11-13). Given the wide range of effluents and reproductive endpoints observed, it is unlikely that all reproductive effects of pulp and paper mill effluents are mediated by a single mechanism caused by one compound, or by a single group of compounds (14). Plant sterols have long been suspected as being bioactive agents in pulp and paper effluents. However, closer examination reveals that the fish reproductive response patterns for the plant sterol, β-sitosterol, estradiol, and for pulp mill effluent differs significantly (15). There is evidence that plant sterols do not act through steroid hormone receptors but can act via interfering with cholesterol transport across the mitochondrial membrane, preventing side chain cleavage, the first reaction in steroid synthesis (16, 17). The possibility that steroidal compounds could lead to the observed effects in fishes has long been supported by the observation that bacteria are capable of producing steroids from sterols (18, 19). Nonsteroidal estrogen agonists have also been identified in pulp and paper effluents, including the plant flavone genistein (20). Recently, the androstane steroid androstenedione has been identified as a putative androgen in pulp and paper effluent (21), though the possibility that this compound is the primary androgen has subsequently been refuted (22). A quite different result was found in a study using in-mill waste streams where compounds that had the potential for reducing steroidogenesis were found in polar fractions (23). Compounds in these fractions were consistent with polyphenolics, potentially lignin breakdown products. While the hypothesis of microbial breakdown of sterols remains an active one, little effort has been directed toward the possibility that chemical reactions within pulping and bleaching processes may lead to the production of reproductively bioactive compounds. Chlorine dioxide is the oxidant that is most often used for the bleaching of kraft pulp. The purpose of this study was to determine if chlorine oxidation products of plant sterols can result in compounds that have the potential to influence reproduction in fishes through a steroid hormone receptor-mediated mechanism. This hypothesis was tested in vitro using rainbow trout estrogen and androgen receptor binding assays. The potential for androgenicity was validated in vivo using the masculization of female mosquitofish as an endpoint. 10.1021/es060089u CCC: $33.50
2006 American Chemical Society Published on Web 03/21/2006
TABLE 1. Compounds Screened, Retention Time (RT), and Quantification Ions for Using Single Ion Monitoring GC-MS formula name
trivial name
RT
SIM Ions (M/z)
5-androsten-3R-ol-17-one 5β-androstan-3β-ol-17-one 5R-androstan-3R-ol-17-one 5β-androstan-3β,17β-diol 5β-androstan-3,17-dione 5-androsten-3β-ol-17-one 5R-androstan-3,17-dione 5R-androstan-3β-ol-17-one 4-androsten-3β,17β-diol 5R-androstan-17β-ol-3-one 5-androsten-3β,17β-diol 4-androsten-17R-ol-3-one 1,3,5(10)-estratrien-3-ol-17-one 5-androsten-3,17-dione 4-androsten-3,17-dione 1,3,5(10)-estratrien-3,17β-diol 1,4-androstadien-3,17-dione 4-androsten-17β-ol-3-one 1,4-androstadien-17β-ol-3-one 1,3,5(10)-estratrien-17R-ethynyl-3,17β-diol 3-methyl ether 4-androsten-4-ol-3,17-dione 5-pregnen-3β-ol-20-one 1,3,5(10)-estratrien-17R-ethynyl-3,17β-diol 4-androsten-17β-ol-3,11-dione 4-pregnen-3,20-dione 1,3,5-estratrien-3,16R,17β-triol 5β-cholestan-3β-ol 5-cholesten-3β-ol 5-cholesten-24R-methyl-3β-ol 5,22-cholestadien-24β-ethyl-3β-ol 5-cholesten-24β-ethyl-3β-ol 5R-cholestan-24β-ethyl-3β-ol 5R-cholestan-24-ethyl-3-one 4-cholesten-24-ethyl-3-one 4-cholesten-24-ethyl-3,6-dione
dehydroandrosterone epietiocholanolone androsterone dihydroxyetiocholane etiocholane-3,17-dione dehydroepiandrosterone androstanedione epiandrosterone 4-androstenediol dihydrotestosterone 5-androstenediol epitestosterone estrone 5-androstenedione Androstenedione E2, estradiol androstadienedione testosterone dehydrotestosterone mestranol 4-hydroxyandrostenedione pregnenolone ethynylestradiol 11-ketotestosterone progesterone estriol coprostanol cholesterol campesterol stigmasterol β-sitosterol sitostanol
39.36 39.80 40.01 41.30 41.90 43.42 43.78 44.01 44.66 45.11 45.24 45.50 45.86 46.07 46.89 47.91 48.18 48.50 50.07 50.41 51.06 51.44 53.51 54.46 55.48 58.35 65.74 70.62 76.40 78.03 81.17 81.81 83.08 85.38 94.00
304, 360, 270 272, 257, 244 272, 347, 362 256, 241, 346 288, 244, 199 304, 360, 231 288, 217, 229 347, 272, 362 434, 344,239 272, 257, 362 239, 344, 215 360, 270, 226 342, 257, 218 358, 343, 208 286, 244,124 416, 285, 232 122, 159, 284 360, 270,147 122, 147, 195 367, 227, 174 359, 169, 331 298, 388, 241 425, 285, 440 446, 431, 208 314, 272, 229 504, 345, 311 370, 215, 257 129, 329, 368 129, 343, 382 129, 255, 394 129, 357, 396 215, 306, 488 143, 486, 457 484, 207, 281 207, 281, 317
Materials and Methods Preparation of Sterol Oxidation Products. β-sitosterol and stigmasterol were obtained from Steraloids, (Newport, RI). The β-sitosterol and stigmasterol were measured to be 86.2% and 98% pure, respectively. Oxidations were carried out by adding the sterols (2 mmol) to solutions of chlorine dioxide (6 mmol) in water (45 mL) at room temperature with stirring. The mixtures were heated to 70 °C and stirred for 30 min protected from light. After cooling, the products were extracted with ethyl acetate and the extracts were washed with water and dried. The extracts from each oxidation were split into two equal portions and evaporated separately. One portion of each oxidation product was stirred with 10% NaOH (20 mL) at 60 °C for 20 min. After cooling, the products were acidified with 10% HCl and extracted with ethyl acetate. The extracts were washed with water, dried, and evaporated. The chlorine dioxide products of β-sitosterol were purified by fractionation over silica gel. A total of 28 fractions were collected by eluting with hexane/ether at 9:1, 4:1, and 1:1 ratios followed by 100% ether and 100% methanol. Fractions were examined for oxidation products using GC with FID detection and TLCs developed with 2,4-DNP/H2SO4. Fractions containing the same components were combined and crystallized. Melting points and mass spectra were obtained for the unknown compounds and compounds were identified by interpretation of the mass spectra. The identified compounds were compared to standards obtained by the oxidation of stigmasterol and β-sitosterol with chromic acid according to Rowe (24). Compound identification of was further verified by proton NMR. Sterol Fractionation. Sterols were fractionated by passage through a 300 × 3.9 mm analytical C-18 column. The column was eluted with a linear gradient from 35 to 95% methanol/
water in 45 min and held until 60 min, at a flow rate of 1 mL/min. The method was calibrated using wide spectrum photodiode array (PDA) and mass-spectrometry (MS) detection for a number of steroids and sterols. Retention times of several steroids (in min) were as follows: testosterone 18.8, 17β-estradiol 16.7, androstenedione 16.5, and pregnenolone 24.8. The sterols and their retention times (in min) were as follows: stigmasterol 48.8, campesterol 48.9, β-sitosterol 49.4, and sitostanol 50.1. Fractions were collected at 11 min intervals. For each fraction the methanol was first evaporated under gentle nitrogen stream, the remaining water was frozen at -80 °C, and samples were brought to dryness by freezedrying. The fractions were reconstituted in 150 µL of ethanol, and estrogenic and androgenic activity were tested by in vitro binding bioassays. After bioassay analysis, the remainder of the fractions were exchanged into methylene chloride, derivitised by adding 50 µL of bis(trimethylsilyl) trifluoroacetamide + 1% trimethylchlorosilane silylation reagent (Alltech, Auckland), and dibromoanthracene (TCI, Tokyo, Japan) was added as the internal standard. Analysis was performed on an Agilent 6890N gas chromatograph/5973N mass selective detector in single ion monitoring mode on an Agilent Ultra-2 column (50 m × 200 µm i.d., film thickness 0.33 µm) using a screen for sterols and steroids shown in Table 1. Androgen and Estrogen Bioassays. Immature trout were injected intraperitoneally with estradiol (Sigma, 5 mg/kg, suspended in corn oil) weekly for three weeks, then sacrificed. The protocol to isolate estrogen receptors from rainbow trout liver was described in detail elsewhere (25). The resultant liver supernatant, containing the nuclear estrogen receptor, was stored at -80 °C until use. Approximately 70 adult male rainbow trout were anaesthetized with MS222 (0.1 g/L) and VOL. 40, NO. 8, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2595
brains were excised for androgen receptor purification. The protocol to isolate androgen receptors from rainbow trout brains was adapted from Sperry and Thomas (26). The final supernatant containing the nuclear androgen receptor was stored at -80 °C for later use. The number of binding sites (Bmax) for estradiol with the pooled rainbow trout hepatic estrogen receptors was 1.1 pmol/mg of protein, and the dissociation constant (Kd) was 1.7 nM. With rainbow trout brain androgen receptors, Bmax for T was 1.1 pmol/mg and the Kd was 0.41 nM. Both estrogen and androgen receptor competitive binding assays were adapted from published protocols in (27). Total bioassay volume was 250 µL with either 1 nM [2,4,6,7-3H] estradiol or 2.5 nM [1,2,6,7-3H] testosterone, 15 µL of the receptor preparation. After 18 h at 4 °C, 500 µL of dextrancoated charcoal was added to strip any unbound radioligand from the supernatant. After 5 min incubation at 4 °C, the tubes were centrifuged at 2000g for 15 min and the supernatant was decanted into a liquid scintillation vial. The β radiation was measured by liquid scintillation in a Packard Tri-Carb 2100-TR. Ovaries from seven mid-vitellogenic, three-year-old rainbow trout were pooled and homogenized for aromatase enzyme preparations. The aromatase isolation protocol was adapted from existing protocols (28) and the product was stored at -80 °C until used in the aromatase assay. The total assay volume was 300 µL including 25 µL of aromatase preparation and 5 µL of sample in ethanol. Concentrations of reagents were 0.1 M KH2PO4, 0.1 M KCl, 1 mM EDTA, 10 mM MgCl2, 1 mM NADPH, and 10 nM androst-4-ene-3, [1,2,6,7-3H]-17-dione (NEN, Boston, MA). The tubes were incubated for 90 min in a water bath at 22 °C. The reaction was terminated by addition of 1 mL of chloroform to each tube, and 1.5 mL of distilled water was added to each tube. The tubes were vortexed and centrifuged at 1000g for 10 min. The top 1 mL was gently pipetted into a new tube, and 1 mL of dextran-coated charcoal was added to each tube. Tubes were vortexed and centrifuged at 2000g for 15 min. The supernatant was decanted into scintillation counting tubes and counted. Mosquitofish Masculization Bioassay. Mosquitofish were collected from Lake Tarawera, New Zealand using a seine net and were maintained in the laboratory in 80 L glass aquaria with Lake Tarawera water supplemented with 0.2% NaCl to inhibit disease. All mosquitofish exposures were conducted under a 12:12 h photoperiod. Each tank was maintained at 25(1 °C and each aquarium received gentle aeration. Female mosquitofish were exposed in duplicate aquaria containing 8 L of water. To prevent mortality due to aggression, each female was contained within an 8 cm diameter stainless steel mesh (1.5 mm) cage allowing a total of 8 females per aquaria. Aquaria were dosed with sterols and sterol oxidation products at doses of 1, 10, and 100 µg/L using acetone as the carrier solvent (0.0006% v/v). Carrier solvent alone was used as the control treatment. Each aquarium received a 50% water replacement each day and was re-dosed. A positive control, 17-methyl testosterone (Sigma, 1000 ng/L) was run with all bioassays. The masculinization of the mosquitofish anal fin was quantified by taking digital photos of the anal fins using a stereomicroscope with an attached digital camera. The anal fin ray 4:6 ratio was determined for each fish in each treatment by measuring the lengths of rays 4 and 6 using UTHSCSA Image Tool 3.0. Statistics. Displacement curves for binding assays were obtained by plotting the proportion of radioligand bound to the receptor against the chemical concentration. The EC50, or effective concentration required to displace 50% of the radioligand from the receptor binding sites, was calculated 2596
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 8, 2006
FIGURE 1. Major products identified in β-sitosterol and stigmasterol oxidized using simulated chlorine dioxide bleaching. by fitting a Verhulst curve to the data by least-squares regression (29). The affinity of a test chemical for the receptor binding site was assessed by determining its relative binding affinity (RBA), a ratio between the EC50 of the standard and the EC50 of the test chemical. Analysis of anal fin ray 4:6 ratio was conducted using one-way analysis at variance (ANOVA) followed by Dunnett’s post-hoc test.
Results Three oxidized sterols were purified and identified as products of the oxidation of β-sitosterol (Figure 1). Other products in the mixture were identified as 5(6)-chlorocholestan-24-ethyl3β,6(5)-diol, and cholestan-24-ethyl-3β,5,6-triol. We subsequently identified two of the oxidation products in pulp and paper mill final effluents from New Zealand (integrated bleached kraft/thermomechanical mill with aerated lagoon treatment) and Canada (bleached kraft and thermomechanical mills with activated sludge treatment) with concentrations of 4-cholestene-24-ethyl-3-one ranging from nondetect to 8 µg/L. The compound cholestan-24-ethyl-3one was present at up to 1 µg/L. Cholestene-24-ethyl-3-one was also found in activated sludge treated sewage effluent from Rotorua, New Zealand (0.16 µg/L). In samples of biosolids from a different New Zealand pulp and paper mill treatment system (bleached kraft mill with aerated lagoon treatment), cholestan-24-ethyl-3-one was most predominant with concentrations ranging from 1 to 16 µg/g dw while the 4-cholestene variant was found from 0.5 to 12 µg/g dw. The relative concentration of the two oxidation products in the biosolids samples ranged from 10 to 65% of the total sterols. Rainbow trout androgen and estrogen receptor binding assays revealed that both the sterols and the oxidation products had measurable levels of androgen activity (Table
TABLE 2. Activity of Sterols and Sterol Oxidation Products Assessed Using in Vitro Bioassays for Androgenicity and Estrogenicity sample
androgen receptor binding (µg/g testosterone equivalent concentration)
estrogen receptor binding (µg/g estradiol equivalent concentration)
β-Sitosterol β-Sitosterol with ClO2 oxidation β-Sitosterol with ClO2 oxidation/NaOH Stigmasterol Stigmasterol with ClO2 oxidation Stigmasterol with ClO2 oxidation/NaOH
96 55 93 18 31 19
