Environ. Sci. Technol. 2007, 41, 8506–8511
Identification of Estrogenic Compounds Emitted from the Combustion of Computer Printed Circuit Boards in Electronic Waste C L Y D E V . O W E N S , J R . , * ,† CHRISTY LAMBRIGHT,‡ KATHY BOBSEINE,‡ BRYCE RYAN,‡ L. EARL GRAY, JR.,‡ BRIAN K. GULLETT,† AND VICKIE S. WILSON‡ National Risk Management Research Laboratory, and National Health and Environmental Effects Research Laboratory, Reproductive Toxicology Division, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711
Received June 13, 2007. Accepted September 18, 2007.. Revised manuscript received September 11, 2007
Rapid changes in technology have brought about a surge in demand for electronic equipment. Many of these products contain brominated flame-retardants (BFRs) as additives to decrease the rate of combustion, raising concerns about their toxicological risk. In our study, emissions from the combustion of computerprinted circuit boards were evaluated in the T47D-KBluc estrogen-responsive cell line at a series of concentrations. There was significant activity from the emission extract when compared to the positive control, 0.1 nM estradiol. After HPLC fractionation, GC/MS identified ten chemicals which included bisphenol A; the brominated derivates mono-, di-, and tribisphenol, triphenyl phosphate, triphenyl phosphine oxide, 4′bromo-[1,1′-biphenyl]-4-ol, 3,5-dibromo-4-hydroxybiphenyl, 3,5dibromo-2-hydroxybiphenyl, and the oxygenated polyaromatic hydrocarbon benzanthrone. Commercially available samples of these ten compounds were tested. The compound 4′-bromo[1,1′-biphenyl]-4-ol resulted in dose-dependent significant increases for luciferase activity at concentrations ranging from 0.1 to 10 µM in the T47D-KBluc assay. The chemical also demonstrated an affinity for binding to the estrogen receptor (ER) with an IC50 of 2 × 10-7 M. To determine the uterotrophic activity, three doses (50, 100, and 200 mg/kg/day) of 4′-bromo[1,1′-biphenyl]-4-ol were administered to adult ovariectomized Long–Evans rats for 3 days. Treatment of the animals with 200 mg/ kg/day showed an increase in uterine weight. Hence one new chemical, released by burning of electrical wastes, was identified which displays estrogenic activity both in vitro and in vivo. However, it was about 1000-fold less potent than ethynyl estradiol.
Introduction In recent years, there has been an enormous increase in the demand for consumer electronics equipment such as com* Corresponding author fax: (919) 541-0554; e-mail:
[email protected]. † National Risk Management Research Laboratory. ‡ National Health and Environmental Effects Research Laboratory, Reproductive Toxicology Division.
puters, cell phones, stereo equipment, and televisions. Environmental concern rises from the incorporation of brominated organic compounds, which are used as flame retardants (BFRs), to slow down the initial phase of a developing fire (1). Because these consumer products are being produced and subsequently discarded at such a high rate, safe disposal has become a concern (2). In the year 2000, the U.S. Environmental Protection Agency (EPA) reported that 4.6 million tons of electronic waste (e-waste) entered U.S. landfills (3). In China, an undetermined amount of e-waste is imported where it undergoes rudimentary recycling of metal and, in some cases, open burning of the nonrecycled residuals (4–7). This low-tech combustion of e-waste can produce toxic halogenated organic pollutants (4). For example, the combustion of e-waste containing brominated flame materials may constitute a major source of polybrominated dibenzodioxins and dibenzofurans (PBDD/ Fs) to the overall global environment (8). The most-common BFRs are tetrabromobisphenol A (TBBPA), hexabromocyclododecane (HBCDD), and decabromodiphenyl ether (DeBDE), which together represent about 150 000 t or ∼25% of the total annual volume of flame-retardant chemicals produced worldwide (9–11). Research has shown that during pyrolysis or thermal decomposition of polyvinyl chloride (PVC) material contained in e-waste plastics, copper may operate as a catalyst for the formation of PBDD/Fs, resulting in human and environmental exposure (12–14). Epoxy resins made of bisphenol A are widely used in production of circuit boards, combustion of which can lead to the production of polybrominated organic pollutants (15). While bisphenol A and some individual PBDEs have been shown to be estrogenic in vitro, it is unknown whether these or other potentially estrogenic compounds are present in sufficient quantities elicit an estrogenic response (16, 17). Thus, the initial focus of this study was to determine if e-waste exhaust contained estrogenic activity. If so, bioassay-directed chemical analysis would be used to identify compounds contributing to estrogenicity from open-fire electronic combustion.
Materials and Methods Chemicals. 17-β estradiol (E2, 99%), 17R-ethynylestradiol (EE, 98%), and corn oil were purchased from Sigma Chemical Co. (St. Louis, MO). Solvent-grade ethanol and HPLC-grade acetonitrile were provided by LabCore Chemical Supply (Durham, NC). Benzanthrone (99%), 3,5-dibromo-biphenyl2-ol (99%), 3,5-dibromo-biphenyl-4-ol (99%), 4′-bromo-[1,1′biphenyl]-4-ol (99%), triphenyl phosphate (99%), and triphenyl phosphine oxide (99%) were purchased from Aldrich Chemical Co. (St. Louis, MO). Printed Circuit Board Burning. Combustion tests were conducted in a controlled open-burning facility at the EPA in Research Triangle Park, NC. The building includes an enclosed burn chamber adjacent to a covered sampling area where equipment is protected from direct sun or rain. The structure is composed of sheet metal fitted to a frame of structural galvanized square tubing. The interior was finished in sheetrock which forms the backing to which aluminum foil was attached. Electronic waste (printed circuit boards with the chips removed to simulate rudimentary recycling operations) material was placed on a sand-filled, loss-inweight platform. To aid in ignition, kerosene was poured over the 5–10 kg sample of electronics waste stacked in pile. Because the fire was not self-sustaining, charcoal was used as a support to ignite and maintain the fire. Charcoal burns fairly clean without contributing large quantities of halogens
8506 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007 10.1021/es071425p CCC: $37.00 Not subject to U.S. Copyright. Publ. 2007 Am. Chem. Soc.
Published on Web 11/16/2007
FIGURE 1. Schematic diagram overview of the in vitro bioassay-directed fractionation and characterization of the printed circuit board combustion samples. or other elements of interest. Combustion blanks for the charcoal and kerosene were tested showing primarily light and medium aliphatic compounds (18). In a pilot study to screen for estrogenic activity, two exactly duplicated quartz filters laden with particulate matter from burning printed circuit boards were donated from Gullett et al. (18). Soxhlet extraction of individual filters was performed for 24 h in 500 mL of ethanol (EtOH) solvent. The polar solvent EtOH was chosen for extraction for two reasons: (1) to avoid toxicity in the cell-based assays from nonpolar solvents,for example, hexane or benzene, and (2) the high solubility of most steroidlike compounds in ethanol. For the first filter, the EtOH solvent was removed by vaporization to approximately 1 mL, resuspended to 2 mL of EtOH, and placed inside a glass screwtop vial. The 500 mL solvent extract of the second filter was vaporized to dryness then reconstituted in 6 mL of DMSO and stored in a glass screw-top vial until needed for quantification work. The extract was stored in DMSO to ensure chemical stability over time. A brief description of the overall approach used in this study is summarized in Figure 1. Chemical Screening. Estrogen-responsive T47D-KBluc cells were maintained in standard growth media as described by Wilson et al. (19). Growth media was RPMI (Gibco) supplemented with 2.5 g/L glucose, 10 mM HEPES, 1 mM sodium pyruvate, 1.5 g/L NaHCO2, 0.2 u/mL insulin, 10% FBS (fetal bovine serum), 100 mg/mL penicillin, 100 U/mL streptomycin, and 0.25 mg/mL amphotericin B. Cells were placed in growth media modified by replacement of 10% FBS with 10% dextran–charcoal-treated FBS without antibiotic supplement one week prior to the assay to remove trace estrogens. In the media used for dosing, the dextran–charcoal-treated FBS concentration was further reduced to 5%. Cells were seeded at 104 cells per well in 96-well plates and allowed to attach overnight. Under these conditions, a typical dose response curve for Estradiol (E2) in this assay ranges from about 0.1 pM to just over 0.1 nM. Both extracts and test chemical solutions were serially diluted in EtOH such that the solvent concentration was consistent throughout each assay. Dosing solutions from the electronic waste extract sample were prepared by aliquoting 1 µL of the undiluted or diluted extract into 1 mL of medium. There were at least two replicate wells of media blank without any treatment and four wells of corresponding solvent control on each plate. E2 positive-control wells, in quadruplicate,
each contained 0.1 nM in 100 µL/well; a concentration which produces a consistent, near-maximal response in this assay. Extracts, HPLC fractions, and test compounds were evaluated in quadruplicate wells on each plate with three replicate plates per experiment. After addition of dosing medium, the cells were incubated overnight at 37 °C. After 24 h, the medium was removed by aspiration; the cells were washed with 25 µL of phosphate-buffered saline and lysed with 25 µL of lysis buffer (Promega) at room temperature. Luciferase activity was determined by measurement of substrate-induced luminescence or relative lights units using a microtiter -plate luminometer (Dynex, Chantilly, VA). Animals and Housing. Adult Long–Evans female rats were provided with Purina Rat Chow 5001 and watered ad libitum. Environmental conditions were 21–24 °C, 40–55% humidity, and a 14L:10D cycle (lights on at 2100 h). Rats (2–3 per group) were ovariectomized and allowed a three week recovery period prior to starting experiments. All animals and animal surgery were conducted according to the guidelines for good animal laboratory practices, and the study was conducted under a protocol that had been approved by the National Health and Environmental Effects Research Laboratory Institutional Animal Care and Use Committee (20). Uterotrophic Assay. The test chemical, 4′-bromo-[1,1′biphenyl]-4-ol was dissolved in corn oil and administered daily for three consecutive days by subcutaneous injection at doses of 50, 100, and 200 mg/kg prior to 10:00 a.m. EST. As a positive control, 65 µg/kg 17R-ethynylestradiol (EE, 98%) was administered to one group of females. The body weights of individual animals were recorded. Approximately 3 h after the final treatment, the animals were anesthetized with CO2, decapitated, and the uterus was removed with fluid and weighed. All rats were checked to confirm a complete lack of ovarian tissue. The rats were also checked to ensure that the subcutaneous injections of the 4′-bromo-[1,1′-biphenyl]4-ol solution were fully absorbed into the body. Pan Vera (hERr-binding) Estrogen Receptor Binding. To evaluate the ability of 4′-bromo-[1,1′-biphenyl]-4-ol to competitively bind to the estrogen receptor (ER), final doses of 0.1–50 µM were evaluated using the Pan Vera fluorescent polarization competitive binding assay kit (Invitrogen, Carlsbad, CA, #P2698) per manufacturer protocol except that concentrations of ER and labeled ligand used were 5nM ERR and 0.125nM Fluomone. Stocks of 4′-bromo-[1,1′-biphenyl]4-ol were prepared in 100% ethanol and then diluted in the VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8507
supplied assay buffer. Assays were incubated two hours at room temperature in 6 × 50 mm glass cuvettes (Kimble), and then read in a Beacon 2000 fluorescent polarization machine. HPLC Fractionation. For HPLC (Beckman Coulter, Palo Alto, CA) separation, 50 µL of the sample (printed circuit board extract) was injected onto a 125 × 4 mm Chromosphere 100 RP-18 column (5 mm, Merck) at 25 °C. Gradient elution was employed using a water and acetonitrile mixture. The standard gradient started with 95% water/5% acetonitrile for the first 5 min following the injection onto the column, after which the polarity was gradually increased to 95% acetonitrile over a linear period of 50 min The solvent was maintained at 95% acetonitrile for a further 10 min, after which it was returned to 95% water to re-equilibrate the column. As the sample eluted from the column, using a flow rate of 1 mL/min and ultraviolet detection monitoring at 254 nm, discrete fractions were collected at 5-min intervals: a total of 10 fractions were collected into glass vials. After a sequence of twenty 50 µL injections (totaling 1000 µL or 1 mL), all the fractions produced were vaporized to dryness andresuspended in 200 µL of ethanol, and 1 µL aliquots were assayed for estrogenic activity. Gas Chromatography/Mass Spectroscopy. Chemical analyses were performed by using a Hewlett-Packard 5890 Series II gas chromatograph and a Hewlett-Packard 5971 mass selective detector. Helium was used as the carrier gas with constant flow (1.00 mL/min) maintained using electronic pressure programming. The column was a diphenyl (5%)–dimethylsiloxane (95%) copolymer stationary phase (30 m length, 0.25 mm i.d., and 0.25 mm film thickness; oven temperature programming (40 °C for 10 min to 300 °C for 50 min at 10 °C/min). The MS detector was scanned from 50 to 500 m/z (3 scans/s). For quantification, a response factor was obtained by analysis ofa known quantity of the analyte and measurement of the area of the relevant peak from the GC/MS. An internal standard was added for determination of the ratio of the analyte signal to the internal standard signal as a function of the analyte concentration of the standard. The original sample, spiked with a known amount of internal standard and the analyte of interest, can be quantified using the response from the known relevant peak area divided by the concentration of the analyte injected. Statistical Analysis. Data were collected from several independent experiments (minimum of 3) with 3–5 replicate wells per plate per experiment. All T47D-KBluc-assay data were electronically imported into a spreadsheet (Microsoft Excel, Microsoft, Seattle, WA). Data were analyzed by twoway ANOVA using Statistical Analysis System (SAS), version 6.09 (SAS Institute, Cary, NC.), on an IBM mainframe computer. Relative fold-induction versus vehicle control data were analyzed in a general linear means model (PROC GLM) which included the concentrations and replicates. When necessary, data were log-transformed to reduce variance heterogeneity. Statistically significant effects (p < 0.05) were further examined using the least-squares means procedure. IC50 values were computed using a nonlinear fit of the data with Origin 6.1 graphing software (OriginLab, Northampton, MA) with maximum and minimum values of the curve constrainedto100%andzerobound,respectively.Uterotrophic assay data were analyzed using PROC GLM after log10 transformation to correct for heterogeneity of variance, and individual groups were compared to control by least squared means posthoc analyses.
Results Initial screening of the EtOH Soxhlet extract in the T47DKBluc assay indicated significant induction of estrogenic activity in a concentration dependent manner shown in Figure 2A. Estrogenic activity compared to vehicle control 8508
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007
FIGURE 2. (A) Dose-dependent estrogenic activity in the circuit board extract. Dilutions up to 1:200 elicited significant estrogenic activity that was completely blocked by the addition of 1 µM of the antiestrogen ICI 182,780. Estrogenic activity was comparable to vehicle control at a dilution of 1:400: a indicates a statistically significant increase in the response from the extract dilution compared to vehicle control, and b indicates a significant decline in estrogenic activity compared to the next most-concentrated sample. (B) High-performance liquid chromatogram (HPLC) fractionated printed circuit boards extracts collected at 5 min intervals were evaluated in the T47D-KBluc-assay. The greatest estrogenic activity eluted from the column in the 25, 30, 35, and 40 min fractions. Data represent the mean ( standard error of three replicate experiments each conducted in quadruplicate: **, p e 0.001; ***, p e 0.0001. was significantly increased at dilutions up to 1:200. Estrogenic activity began to decline at dilutions of 1:100 and greater, reaching a statistically significant decline at 1:200 until all activity was lost at dilutions of 1:400. Coadministration of the ER-specific inhibitor, ICI 182,780, blocked the estrogenic response elicited by the extract, giving added evidence that the effect observed was mediated through ER binding. Upon determination of a positive response in estrogenic activity, the sample extract was fractionated using HPLC methods. Ten fractions were collected at 5 min intervals, and each was evaluated using the T47D-KBluc cell line, as shown in in Figure 2B. The positive control, 0.1 nM E2 produced a 13-fold induction of luciferase activity over media. The strongest estrogenic activity was detected in the 25, 30, 35, and 40 min fractions. Other fractions only exhibited weak activity or activity that was not statistically significant versus vehicle control. The 25 min fraction produced a 16-fold induction over media which was about 18% higher than that for the E2 control. The 30, 35, and 40-min fractions produced a 10-, 8-, or 5-fold induction over vehicle, respectively.
FIGURE 4. Dose–response curve for 4′-bromo-[1,1′-biphenyl]-4-ol in theT47D-KBluc-assay. The compound showed significantly increased luciferase activity at concentrations of 0.1 µM and greater, compared to that of the vehicle control. Data are presented as mean ( standard error of three independent experiments, each conducted in quadruplicate: ***, p e 0.001. FIGURE 3. Chemical structures of the ten compounds identified by gas chromatogram/mass spectroscopy from the 25, 30, 35, and 40 min HPLC fractions Although the activities from these three fractions were lower than the that for the 25-min fraction, the significant luciferase response warranted further investigation. In Figure 3, the chemical structures of ten compounds identified by the GC/MS from the HPLC fractions (25, 30, 35, and 40 min fractions) demonstrating the highest estrogenic response are shown. The ten compounds were identified by retention time and mass fragmentation patterns matching those of an authentic primary standard and a National Institute of Standards and Technology (NIST) library mass spectrum. In the 25 min fraction, only a subset of the compounds could be identified with confidence; the principle component was bisphenol A, a known estrogenic compound, with a retention time of 29.05 min (16, 17). In the 30 min fraction, the compounds triphenyl phosphate and triphenyl phosphine oxide were noted with retention times of 32.09 and 34.55 min, respectively. The mono and dibromo derivatives of bisphenol, bromo-bisphenol and dibromo-bisphenol, were also identified from the identification library with retention times of 31.45 and 32.65 min, respectively. The oxygenated polyaromatic hydrocarbon benzanthrone and tribromo-bisphenol showed retention times of 35.60 and 36.89 min, respectively, in the 35 min fraction. In the 40 min fraction, GC/MS identified the compounds 4′-bromo-[1,1′biphenyl]-4-ol, 3,5-dibromo-4-hydroxybiphenyl, and 3,5dibromo-2-hydroxybiphenyl. Of the ten chemicals identified from the estrogenic fractions, six (triphenyl phosphine oxide, triphenyl phosphate, benzanthrone, 3,5-dibromo-4-hydroxybiphenyl, 3,5dibromo-2-hydroxybiphenyl, and 4′-bromo-[1,1′-biphenyl]4-ol) were tested for estrogenicity in vitro in the T47D-KBluc assay. The compounds bisphenol A, and the mono-, di-, and tribromo derivatives of bisphenol were not evaluated because of previously confirmed estrogenicity ranging between 2 and 100 µM by Meerts et al. (16). Of the tested compounds, only 4′-bromo-[1,1′-biphenyl]-4-ol produced significant dosedependent increases in luciferase activity at concentrations ranging from 0.1 to 20 µM, as shown in Figure 4, with a maximal response attained at 10 µM. This is believed to be the first report of estrogenic activity for 4′-bromo-[1,1′biphenyl]-4-ol. To confirm interaction of 4′-bromo-[1,1′-biphenyl]-4-ol with the ER and rule out a nonspecific effect on luciferase expression, the chemical was assessed with regard to its
FIGURE 5. Concentration–response curve of 4′-bromo-[1,1′biphenyl]-4-ol binding to the estrogen receptor in the ER competitive binding assay. An estradiol dose response curve is shown from relative comparison. Data represent the mean ( standard error of three replicate experiments each conducted in duplicate. competitive binding to ER at concentrations from 0.01 nM to 50 µM in Figure 5. The dose range from 0.5 to 50 µM 4′-bromo-[1,1′-biphenyl]-4-ol strongly inhibited the binding of the [3H] E2 to the ER with an IC50 value of 2 × 10-7 M. For comparison, the IC50 of E2 in the same assay was 3 × 10-9 M. To confirm whether 4′-bromo-[1,1′-biphenyl]-4-ol was estrogenic in vivo, a uterotrophic assay was conducted using adult ovariectomized female Long–Evans rats. There were no signs of overt toxicity at any of the doses tested. Although there was no significant difference in body weight at the end of the study, there was a significant body-weight loss (∼7%) in the 200 mg/kg/day dose group. Uterine weights in the controls were typical of that expected for long-term ovariectomized female rats. The compound, 4′-bromo-[1,1′biphenyl]-4-ol increased uterine weights in a dose-dependent manner, which was significant at 200 mg/kg/day. Mean uterine wet weight in this group increased 1.85-fold, compared to the vehicle-treated control group, as shown in Table 1. As expected, the positive control (EE) at doses of 65 µg produced a significant increase (2.92-fold higher) in mean uterine wet weight when compared to the vehicle control. At necropsy, ovariectomy was confirmed; no females had ovarian tissue present. In addition, the solutions appeared to be fully absorbed at the infection site in all animals. VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8509
TABLE 1.
Effect of Subcutaneous Administration of 4′-Bromo-[1,1′-biphenyl]-4-ol on Uterine Weights in Long–Evans
Rats treatment vehicle 17R-ethynylestradiol 4′-bromo-[1,1′-biphenyl]-4-ol
a
dose (mg/kg/day)
no. of rats
starting body weight ( SE (g)
terminal body weight ( SE (g)
uterus weight ( SE (mg)
0 0.065 50 100 200
2 3 3 3 3
462 ( 58 446 ( 24 424 ( 33 448 ( 20 455 ( 16
465 ( 55 440 ( 23 424 ( 32 448 ( 22 422 ( 12
128.9 ( 5.9 377.1 ( 60a 134.6 ( 18 140.9 ( 4 238.4 ( 60a
Significant effects (p < 0.05)
Once estrogenic activity of 4′-bromo-[1,1′-biphenyl]-4-ol was confirmed, the concentration of the 4′-bromo-[1,1′biphenyl]-4-ol was determined in archived DMSO extract from the duplicate circuit board burn filter. Using chemical analysis and quantification by means of the relative response factor from the GC-MS, a total of about 1.98 µg (i.e.0.33 µg/ ml in the 6 mL extract) of 4′-bromo-[1,1′-biphenyl]-4-ol particulate matter had been extracted from the filter.
Discussion Using bioassay-directed chemical fractionation analysis, we were able to show the compound 4′-bromo-[1,1′-biphenyl]4-ol exhibited significant estrogenic activity in vitro by inducing estrogen-dependent gene expression and binding to the ER. Furthermore, 4′-bromo-[1,1′-biphenyl]-4-ol did induce an estrogenic response in vivo, stimulating growth of the uterus of adult ovariectomized female rats. In vivo, this chemical was about 1000-fold less potent than the potent synthetic estrogenic drug ethinyl estradiol, whereas in vitro it was about 100-fold less potent than the natural steroidal 17-β estradiol. GC/MS identification confirmed the structures of ten compounds which included triphenylphosphate, triphenyl phosphine oxide, benzanthrone, bisphenol A, the bromonated bisphenol derivaties, (mono-, di-, and tribromo bisphenol), and the brominated hydroxybiphenyl derivatives (4′-bromo-[1,1′-biphenyl]-4-ol, 3,5-dibromo-4-hydroxybiphenyl, and 3,5-dibromo-2-hydroxybiphenyl). The organophosphorus compound triphenyl phosphate is used as a flame retardant in phenolic and phenylene-oxide-based resins for the manufacture of electrical and automobile components and as a nonflammable plasticizer in cellulose acetate for photographic films. The oxygenated compounds triphenyl phosphine oxide and benzanthrone are hypothesized to form in the oxidation process of triphenyl phosphate and benzanthracene. Bisphenol A is used primarily to make polycarbonate plastic and epoxy resins providing toughness, high heat, and excellent electrical resistance. Estrogenicity has been confirmed for bisphenol A at 100 µM in the ERCALUX assay (16). The three brominated bisphenol compounds (mono-, di-, and tribromo bisphenol) identified from the GC/MS profile, which were possibly formed through the thermal decomposition of tetrabromobisphenol, were previously determined to have estrogenic potencies with concentrations ranging from 2 µM for the tri derivative and 10 µM for the mono and di derivatives in the ER-CALUX assay (16). Of the ten additional compounds identified in this study from the HPLC fractionated samples, an estrogenic response was seen only with 4′-bromo-[1,1′-biphenyl]-4-ol. Because of the apparent similarity in the chemical structure of 4′bromo-[1,1′-biphenyl]-4-ol, we suspected similar estrogenic responses from the two additional halogenated hydroxylbiphenyl derivatives (3,5-dibromo-4-hydroxybiphenyl and 3,5-dibromo-2-hydroxybiphenyl). However, our results for estrogenicity were comparable to that determined for similar 8510
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007
structural isomers of halogenated compounds (21, 22). Estrogenicity for 3,5-dibromo-4-hydroxybiphenyl and 3,5dibromo-2-hydroxybiphenyl derivatives was inactive when the –OH group was in the ortho or para position and the halogenated group was in an adjacent position. The position of the hydroxyl on the phenolic ring affects estrogenicity, possibly, because this affects the alignment of the hydroxybiphenyl in the receptor. The ortho- and para-directing influence of the hydroxyl group on the phenol ring may not significantly produce induction of luciferase activity because of steric hindrance from the substituted bromines. The competitive binding binding of 4′-bromo-[1,1′biphenyl]-4-ol for the ER determined an IC50 value of 2 × 10-7 M. Our results were similar to Korach et al. reporting that some hydroxyl-PCBs competitively bind to the estrogen receptor (23). The most estrogenic halogenated substituted compounds, 2′,4′,6′-trichloro-4-biphenylol and 2′,3′,4′,5′tetrachloro-4-biphenylol, contained a single p-hydroxyl group on one phenyl ring and chlorine substitution on the second ring, which were similar to the structure of 4-bromo-4hydroxybiphenyl. Although in vitro assays can demonstrate mechanistically that a compound can interact with the ER and elicit a response, in vitro assays cannot replicate the myriad of pharmacokinetic and toxicodynamic interactions that may influence the alleged estrogenic activity of a compound. Therefore, the uterotrophic assay, a standard short-term in vivo screening assay for estrogenic compounds was used to determine whether 4′-bromo-[1,1′-biphenyl]-4-ol was also estrogenic in vivo (24, 25). In our uterotrophic assay, subcutaneous treatment with 4′-bromo-[1,1′-biphenyl]-4-ol in ovariectomized adult Long–Evans rats at 200 mg/kg/day for three consecutive days induced a significant increase in uterine wet weight. Fielden et al. evaluated in vivo estrogenic activities of hydroxylated polychlorinated biphenyls with similar structure to 4′-bromo-[1,1′-biphenyl]-4-ol in ovariectomized CD-1 mice (26). At a dose of 202 mg/kg, 2,4,6,2′,6′pentachlorobiphenyl increased uterine wet weight and induced vaginal epithelial cell cornification in a dosedependent manner. Chemical analysis and quantification by GC/MS revealed that computer-printed circuit boards contained levels of 4′bromo-[1,1′-biphenyl]-4-ol totaling about 1.9 µg (0.33 µg/ ml). In contrast, the identified nonestrogenic compounds triphenylphosphate, triphenyl phosphine oxide, benzanthrone, and the brominated hydroxybiphenyl derivative 3,5dibromo-2-hydroxybiphenyl was detected in amounts ranging from 0.72 to 1.62 µg/mL. While 4′-bromo-[1,1′-biphenyl]4-ol contributes to the significant estrogenic response produced from combustion of printed circuit boards, bisphenol A, the brominated bisphenol derivatives, (mono-, di-, and tribromo bisphenol), and possibly other as yet unidentified compounds also contribute to the estrogenic activity. The prominent amount of 4′-bromo-[1,1′-biphenyl]-4-ol detected from the circuit board extract confirmed the anticipated conversion of brominated flame retardants to
these types of chemically combusted byproducts. This supports nomination of the combustion of flame retardantcontaining waste as a potential source of environmental contamination of estrogenic compounds. These findings suggest that open burning of e-waste could be a major significant health and environmental hazard. The major significance in the identification of 4′-bromo[1,1′-biphenyl]-4-ol’s estrogenic potential may be its route of human exposure. Compounds such as 4′-bromo-[1,1′biphenyl]-4-ol or other pollutants may be inhaled directly into the lungs. In addition, the particles may subsequently be swallowed and absorbed via the gastrointestinal tract. Inhalation of compounds to which brominated flame retardant additives are associated would thus seem to be a very important route of exposure. Given that there has been little or no research reported on the compound 4′-bromo[1,1′-biphenyl]-4-ol, the biological relevance and health risks of this compound are currently unknown. The estrogenicity of 4′-bromo-[1,1′-biphenyl]-4-ol described herein, combined with the confirmed presence of both known and possibly yet-to-be-identified estrogenic compounds emitted from e-wastecombustion,wouldseemtowarrantfurtherinvestigation.
Literature Cited (1) Sjodin, A.; Carlsson, H.; Thursesson, K.; Sjolin, S.; Bergman, A.; Ostman, C. Flame retardants in indoor air at an electronics recycling plant and at other work environments. Environ. Sci. Technol. 2001, 35 (3), 448–454. (2) Electronic Waste Management, http://www.noewaste.com. (3) Mayfield, K. E-waste: Dark side of digital age, http://www. wired.com/news/technology/0,1282,57151,00.html. (4) Li, H.; Yu, L.; Sheng, G.; Fu, J.; Peng, P. Severe PCDD/F and PBDD/F pollution in air around an electronic waste dismantling area in China. Environ. Sci. Technol. 2007, 41 (16), 5641–5646. (5) Li, J.; Lu, H.; Guo, J.; Xu, Z.; Zhou, Y. Recycle technology for recovering resources and products from waste printed circuit boards. Environ. Sci. Technol. 2007, 41 (6), 1995–2000. (6) Leung, A.; Wei Cai, Z.; Wong, M. H. Environmental contamination from electronic waste recycling at Guiyu, southeast China. J. Mater. Cycles Waste Manage. 2006, 8 (1), 21–33. (7) Deng, W.; Louie, P.; Liu, W.; Bi, X.; Fu, J.; Wong, M. Atmospheric levels and cytotoxicity of PAHs and heavy metals in TSP and PM2.5 at an electronic waste recycling site in southeast China. Atmos. Environ. 2006, 40 (36), 6945–6955. (8) Hardy, M. The toxicology of the commercial polybrominated diphenyl oxide flame retardants: DBDPO, OBDPO, PeBDPO. Organohalogen Compd. 2000, 47, 41–44. (9) De Wit, C. Brominated Flame Retardants.; Report 5065; Swedish Environmental Protection Agency: Stockholm, Sweden 2000; p 94. (10) Hardy, M. Distribution of decabromodiphenyl oxide in the environment. Organohalogen Compd. 2000, 47, 237–240. (11) International Programme on Chemical Safety. In Brominated Diphenyl Ethers.; Environmental Health Criteria; World Health Organization: Geneva, 1994.
(12) Buser, H. Polybrominated dibenzofurans and dibenzo-p-dioxins: Thermal reaction products of polybrominated diphenyl ether flame retardants. Environ. Sci. Technol. 1986, 20, 404–408. (13) Lorenz, W.; Bahadir, M. Recycling of flame retardants containing printed circuits: A study of the possible formation of polyhalogenated dibenzodioxins/furans. Chemosphere 1993, 26, 2221– 2229. (14) Luijk, R.; Wever, H.; Olie, K.; Covers, H.; Boon, J. The influence of the polymer matrix on the formation of polybrominated dibenzo-p-dioxins (PBDDs) and polybrominated dibenzofurans (PBDFs). Chemosphere 1991, 23, 1173–1183. (15) Williams, P.; Hall, W. Separation and recovery of materials from scrap printed circuit boards. Resour., Conserv. Recycl. 2007, 51 (3), 691–709. (16) Meerts, I.; Letcher, R.; Hoving, S.; Marsh, G.; Bergman, A.; Lemmen, J.; Van der Burg, B.; Brouwer, A. In vitro estrogenicity of polybrominated diphenyl ethers, hydroxylated PBDEs, and polybrominated bisphenol A compounds. Environ. Health Perspect. 2001, 109 (4), 399–407. (17) Wetherill, Y.; Akingbemi, B.; Kanno, J.; McLachlan, J.; Nadal, A.; Sonnenschein, C.; Watson, C.; Zoeller, T.; Belcher, S. In vitro molecular mechanisms of bisphenol A action. Reprod. Toxicol. Published online 29 May 2007. (18) Gullett, B.; Linak, W.; Touati, A.; Gatica, S.; King, C.; Wasson, S. Emissions and ash characterization from open burning of electronic wastes during rudimentary recycling operations. J. Mater. Cycles Waste Manage. 2007, 9 (1), 69–79. (19) Wilson, V.; Bobseine, K.; Gray, E., Jr. Development and characterization of a cell that stably expresses an estrogen responsive luciferase reporter for the detection of estrogen receptor agonist and antagonists. Toxicol. Sci. 2004, 81 (1), 69– 77. (20) Guide for the Care and Use of Laboratory Animals; National Academy Press: Washington, DC, 1996. (21) Beresford, N.; Routledge, E.; Harris, C.; Sumpter, J. Issues arising when interpreting results from an in vitro assay for estrogenic activity. Toxicol. Appl. Pharmacol. 2000, 162, 22–33. (22) Routledge, E.; Sumpter, J. Structural features of alkylphenolic chemicals associated with estrogenic activity. J. Biol. Chem. 1997, 272 (6), 649–652. (23) Korach, K.; Sarver, P.; Chae, K.; McLachlan, J.; McKinney, J. Estrogen receptor-binding activity of polychlorinated hydroxybiphenyls: Conformationally restricted structural probes. Mol. Pharmacol. 1988, 33, 120–126. (24) Hisaw, F. L. Comparative effectiveness of estrogens on fluid imbibition and growth of the rat’s uterus. Endocrinology 1959, 54 (85), 276–289. (25) Jones, R. C.; Edgren, R. A. The effects of various steroids on the vaginal histology in the rat. Fertil. Steril. 1973, 24, 284–291. (26) Fielden, M.; Chen, I.; Chittim, B.; Safe, S.; Zacharewski, T. Examination of the estrogenicity of 2,4,6,2′,6′-pentachlorobiphenyl (PCB 104), its hydroxylated metabolite, 2,4,6,2′,6′pentachloro-4-biphenylol (HO-PCB 104), and a further chlorinated derivative, 2,4,6,2′,4′,6′-hexachlorobiphenyl (PCB 155). Environ. Health Perspect. 1997, 105, 1238–1248.
ES071425P
VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8511
