Environ. Sci. Technol. 2008, 42, 1773–1779
Microarray Analysis Reveals a Mechanism of Phenolic Polybrominated Diphenylether Toxicity in Zebrafish ANTONIUS L. VAN BOXTEL,† JORKE H. KAMSTRA,† PETER H. CENIJN,† BART PIETERSE,‡ MARIJKE J. WAGNER,§ MAARTJE ANTINK,| KLAAS KRAB,§ BART VAN DER BURG,‡ GÖRAN MARSH,⊥ ABRAHAM BROUWER,† AND J U L I E T T E L E G L E R * ,† Institute for Environmental Studies and Department of Molecular Cell Physiology, Faculty of Earth and Life Sciences, VU University, Amsterdam, The Netherlands, BioDetection Systems BV, Amsterdam, The Netherlands, Service XS BV, Leiden, The Netherlands, and Department of Environmental Chemistry, Stockholm University, Stockholm, Sweden
Received August 21, 2007. Revised manuscript received December 5, 2007. Accepted December 7, 2007.
Polybrominated diphenylethers (PBDEs) are ubiquitous in the environment, with the lower brominated congener 2,2′,4,4′tetrabromodiphenylether (BDE47) among the most prevalent. The phenolic PBDE, 6-hydroxy-BDE47 (6-OH-BDE47) is both an important metabolite formed by in vivo metabolism of BDE47 and a natural product produced by marine organisms such as algae. Although this compound has been detected in humans and wildlife, including fish, virtually nothing is known of its in vivo toxicity. Here we report that 6-OH-BDE47 is acutely toxic in developing and adult zebrafish at concentrations in the nanomolar (nM) range. To identify possible mechanisms of toxicity, we used microarray analysis as a diagnostic tool. Zebrafish embryonic fibroblast (PAC2) cells were exposed to 6-OH-BDE47, BDE47, and the methoxylated metabolite 6-MeO-BDE47. These experiments revealed that 6-OH-BDE47 alters the expression ofgenesinvolvedinprotontransportandcarbohydratemetabolism. These findings, combined with the acute toxicity, suggested that 6-OH-BDE47 causes disruption of oxidative phosphorylation (OXPHOS). Therefore, we further investigated the effect of 6-OHBDE47 on OXPHOS in zebrafish mitochondria. Results show unequivocally that this compound is a potent uncoupler of OXPHOS and is an inhibitor of complex II of the electron transport chain. This study provides the first evidence of the in vivo toxicity and an important potential mechanism of toxicity of an environmentally relevant phenolic PBDE of both anthropogenic and natural origin. The results of this study emphasize the need for further investigation on the presence and toxicity of this class of polybrominated compounds.
* Corresponding author telephone: (31) 20 598 9516; fax: (31) 20 598 9553; e-mail:
[email protected]. † Institute for Environmental Studies, VU University. ‡ BioDetection Systems BV. § Department of Molecular Cell Physiology, VU University. | Service XS BV. ⊥ Stockholm University. 10.1021/es0720863 CCC: $40.75
Published on Web 02/01/2008
2008 American Chemical Society
Introduction Polybrominateddiphenyl ethers (PBDEs) are industrial chemicals that are produced in large quantities and are used extensively as flame-retardants in a wide range of products. Deep concerns were raised after PBDEs were found to be rapidly accumulating in the environment over the past decades (1). A large number of PBDE congeners has been detected in a wide variety of wildlife species, including a number of fish species (2–4) and also in humans (5, 6). Despite the extensive amount of exposure data on these compounds, there are pronounced data gaps with respect to their toxicity. To date, PBDEs seem to be only mildly toxic at high concentrations in rodents, so it is unclear whether these compounds pose a threat to the environment or to human health (7, 8). However, there is a definite need for the investigation of the toxic effects of specific congeners, in particular their metabolites (8). Of all the PBDE congeners found in the environment, 2,2′,4,4′-tetrabromodiphenylether (BDE47) is one of the most abundant and persistent (9). BDE47 has been shown to undergo phase I metabolism in vertebrates, as studies in both rodents and fish have shown the formation of hydroxylated (OH-) BDE47 metabolites after administration of a single dose of BDE47 (10, 11). Importantly, the natural production of OH-PBDEs and methoxylated (MeO-) PBDEs has been shown as high concentrations of these compounds have been detected in algae and sponges (12–14). The structures of naturally occurring OH-and MeO-PBDEs seem to differ from the OH-PBDE metabolites of anthropogenic origin. For the naturally occurring compounds, the OH-/ MeO-group is predominantly attached to an ortho position, (3, 12) whereas the meta and para positions are preferentially substituted with an OH-group in the case of an anthropogenic PBDE metabolite (15). Some of the natural PBDEs have antibacterial properties, and it is thought that they are produced as a natural defense against unwanted bacterial infection (14). The presence of MeO-PBDEs in fish appears to be dependent on their natural occurrence (12). In the case of the OH-PBDEs, the precise origin in fish cannot be determined because there may be a contribution from PBDE metabolism as well as from natural formation. In this study, we focused on the hydroxylated BDE47 metabolite, 6-hydroxy-2,2′,4,4′- BDE47 (6-OH-BDE47), an important metabolite within the pool of hydroxylated PBDEs retained in biological systems, which has been detected in the blood of several wildlife species and in humans (2, 6, 11). In addition to being formed by metabolism of BDE47 (16), this phenolic PBDE has also been shown to be produced by algae and sponges (12). Although in vitro studies have demonstrated the potential endocrine disrupting and cytotoxic properties of 6-OH-BDE47 (16, 17) little is known of the in vivo toxicity of this both naturally occurring and anthropogenic PBDE. The main goal of this study was to determine the in vivo toxicity of 6-OH-BDE47 and to investigate possible mechanisms of toxicity using gene expression analysis. We used developing and adult zebrafish as our model and compared the toxicity of 6-OH-BDE47 with the structurally related methoxylated PBDE, 6-methoxy2,2′,4,4′-BDE47 (6-MeO-BDE47), and the parent compound BDE47.
Materials And Methods A full and detailed description of the experimental procedures and compounds used can be found in the Supporting Information. VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1773
Compounds. 2,2′,4,4′,-tetrabromodiphenylether (BDE47), 6-OH-BDE47. and 6-MeO-BDE47 were synthesized in-house (18, 19) and were dissolved in dimethyl sulfoxide (DMSO). Zebrafish Toxicity Experiments. Zebrafish were maintained under standard conditions (20). Eggs were exposed to the diluted compounds in 25 mL of embryo water. Solvent (DMSO, 0.01%) and negative controls were incorporated in all experiments. The developing zebrafish were scored daily for 72 h for mortality and developmental abnormalities based on sublethal and teratogenic end points previously described (20). To investigate adult toxicity, five 1.5 year old zebrafish were exposed to the compounds for 96 h in 4 L of acclimated tap water in glass aquaria at 27 °C. Measurement of 6-OH-BDE47 in Water and Zebrafish Liver. Water samples (10 mL) were taken at the beginning (t ) 0) and termination of the experiment (t ) 96 h) of the toxicity experiments with adult zebrafish. Following sacrifice, the livers were removed from exposed adult zebrafish, frozen in liquid nitrogen, and pooled (n g 5 per concentration). Water and tissue concentrations of 6-OH-BDE47 were analyzed according to a previously described method (21). Microarray Analysis of Embryonic Zebrafish Fibroblasts (PAC2) Cells. Zebrafish embryonic fibroblast (PAC2) cells were cultured as described elsewhere (22). After 16 h, compounds dissolved in DMSO were added to complete culture medium to a final concentration of 1 µM at a volume of 0.1% DMSO. DMSO served as the solvent control, and exposures were performed for 24 h. All exposures were done in triplicate, and RNA was isolated using a RNA purification kit (Macherey-Nagel, Germany). The isolated RNA samples were processed according to Affymetrix protocols and were used for the hybridization to Affymetrix GeneChip zebrafish genome arrays. All (pre-) processing and statistical procedures were performed in Matlab. Significance analysis was performed on LOESS normalized data by means of ANOVA and a Tukey HSD test. The data were analyzed by principal component analysis (PCA) for target selection. The set of genes that was found to be significantly regulated compared to control samples was used for further GO-ontology based analysis using the online Affymetrix Netaffx analysis center. Mitochondrial Experiments. To measure changes in mitochondrial respiration and inner mitochondrial membrane potential (∆Ψ), mitochondria from the gastrointestinal tract, heart, liver, and kidneys from 10-25 adult zebrafish of 1.5 years of age were isolated, transferred to cold homogenization medium, and kept on ice. The tissue was homogenized and centrifugated for 6 min at 750 g at 4 °C. The supernatant was transferred to a clean tube and was centrifuged for 10 min at 10 000g. The formed pellet was transferred to a 23% Percoll gradient, and the mixture was centrifuged for 30 min at 40 000g. Mitochondria were collected, washed twice in wash buffer, and resuspended in a small volume of wash buffer. For simultaneous measurement of respiration and ∆Ψ, the mitochondria were incubated at 25 °C in reaction medium in a closed and stirred perspex vessel equipped with an oxygen electrode and a tetraphenylphosphonium ion (TPP+)-sensitive electrode (23). The amount of free TPP+ in the reaction medium is a relative measurement of ∆Ψ. Respiration was measured after addition of mitochondria, succinate (10 mM), ATP (1µM), and ADP (0.1 mM) and was followed by the titration of compounds that were dissolved in DMSO or ethanol. The reduction level of ubiquinone (Q) was measured with the same batches of mitochondria that were used in the mitochondrial membrane potential experiments, using the reaction vessel described above with glassy carbon and platinum electrodes (23). In short, mitochondria were incubated in 1 mL reaction medium containing 2 µM Q. The ratio between reduced and oxidized Q at the start of the 1774
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008
FIGURE 1. Toxicity of 6-OH-BDE47 in zebrafish. (a) Zebrafish embryos were exposed in triplicate to 6-OH-BDE47 and were photographed through a stereomicroscope at 48 h post fertilization (hpf). Depicted are four representative images at different concentrations of 6-OH-BDE47. (b) Abnormal development (% relative to control) at 72 hpf for 25 embyos (N ) 3). (c) Mortality (% relative to control) in adult zebrafish after exposure to 6-OH-BDE47 (N ) 5). Concentrations of 400 and 500 nM were toxic within the first hour of exposure whereas 300 nM was toxic after 2 h. Concentrations shown are nominal. Data points depict averages and bars depict standard deviation. experiment was determined by HPLC analysis as previously described (24). Subsequently, the reduction was expressed as the percentage of reduced Q in the total pool of Q present in the mitochondria. All data were recorded digitally using a PowerLab/4SP system connected to an Apple Macintosh running Chart v. 3.6s software (ADInstruments Pty Ltd., United Kingdom). All experiments were performed in triplicate.
Results To gain more insight in the embryotoxic effects of BDE47, the phenolic PBDE 6-OH-BDE47 and the methylated derivative 6-MeO-BDE47, zebrafish embryos were exposed from 3 h post fertilization (hpf) for the first 72 h of development. For BDE47 and 6-MeO-BDE47, no toxic or teratogenic effects were observed at nominal concentrations up to 10 µM (data not shown). Exposure to 100 nM 6-OH-BDE47, however, caused acute toxicity with embryos undergoing developmental arrest around the 18 somite stage (Figure 1a). Lower concentrations (25–50 nM) caused a wide range of developmental defects such as pericardial edema, yolk sac deformations, reduced pigmentation, lowered heart rate, and developmental delay (Figure 1a). The developmental effects
TABLE 1. Measured 6-OH-BDE47 Concentrations in Water and Pooled Zebrafish Liver after 96 h Exposure concentration
water liver
nominal (nM) t ) 0 actual (nM) t ) 0 actual (nM) t ) 96 average concentration (nM) measured (ng gr-1 ww) measured concentration (nM) bioconcentration factord
6-OH-BDE47 0
