Article pubs.acs.org/est
Rapid in Vitro Metabolism of the Flame Retardant Triphenyl Phosphate and Effects on Cytotoxicity and mRNA Expression in Chicken Embryonic Hepatocytes Guanyong Su,†,‡ Doug Crump,*,† Robert J. Letcher,*,†,‡ and Sean W. Kennedy†,§ †
Environment Canada, National Wildlife Research Centre, Carleton University, Ottawa, Ontario K1A 0H3, Canada Department of Chemistry, Carleton University, Ottawa, Ontario K1S 5B6, Canada § Centre for Advanced Research in Environmental Genomics, Department of Biology, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada ‡
S Supporting Information *
ABSTRACT: The organophosphate flame retardant, triphenyl phosphate (TPHP), has been detected with increasing frequency in environmental samples and its primary metabolite is considered to be diphenyl phosphate (DPHP). Information on the adverse effects of these compounds in avian species is limited. Here, we investigate the effects of TPHP and DPHP on cytotoxicity and mRNA expression, as well as in vitro metabolism of TPHP, by use of a chicken embryonic hepatocyte (CEH) screening assay. After 36 h of exposure, CEH cytotoxicity was observed following exposure to >10 μM TPHP (LC50 = 47 ± 8 μM), whereas no significant cytotoxic effects were observed for DPHP concentrations up to 1000 μM. Using a custom chicken ToxChip polymerase chain reaction (PCR) array, the number of genes altered by 10 μM DPHP (9 out of 27) was greater than that by 10 μM TPHP (4 out of 27). Importantly, 4 of 6 genes associated with lipid/cholesterol metabolism were significantly dysregulated by DPHP, suggesting a potential pathway of importance for DPHP toxicity. Rapid degradation of TPHP was observed in CEH exposed to 10 μM, but the resulting concentration of DPHP accounted for only 17% of the initial TPHP dosing concentration. Monohydroxylated-TPHP (OH-TPHP) and two (OH)2-TPHP isomers were identified in TPHP-exposed CEH, and concentrations of these metabolites increased over 0 to 36 h. Overall, this is the first reported evidence that across 27 toxicologically relevant genes, DPHP altered more transcripts than its precursor, and that TPHP is also metabolized via a hydroxylation pathway in CEH.
■
INTRODUCTION
Several studies have addressed adverse effects of TPHP, including contact dermatitis,8 neurotoxicity,9,10 immunotoxicity,11 and cardiotoxicity.12 Recent in vitro studies13,14 showed that TPHP could bind to and activate peroxisome proliferator activated receptor γ (PPARγ) ligands in humans. TPHP was shown to be the most acutely toxic triaryl phosphate to fish, shrimp, and daphnia with toxicity indices (96 h; median lethal concentration [LC50]) for fish ranging from 0.36 mg/L (rainbow trout) to 290 mg/L (bluegills).15 Acute toxicity data for TPHP are also available for mammals (rat median lethal dose [LD50], 3500−10 800 mg/kg) and algae (median effective concentration [EC50], 0.26−2.0 mg/L).16 For chronic toxicity, estimated no-observed effect concentrations (NOECs) regarding survival and growth were 0.1 and 0.0014 mg/L for daphnia and fish, respectively.17 The high TPHP levels (up to 227 pg/m3) reported in the Great Lakes atmosphere22 means it is ubiquitous across the Great Lakes and thus avian species
As one of the potential substitutes for bromine-containing flame retardant (FR) formulations used in textile back-coatings, triphenyl phosphate (TPHP) is also used in hydraulic fluid, polyvinyl chloride, electronic equipment, casting resins, glues, engineering thermoplastics, phenylene-, and phenolic-based resins.1 TPHP is an additive FR that is not chemically bonded to products, and relative to reactive FRs, is more likely to be released into the environment.1 Stapleton et al. detected TPHP at concentrations ranging from 20 000 at m/z 322.0481 when the ESI interface was operated in the positive mode and the capillary voltage was 5500 V. The fragmentor and skimmer voltages were 180 and 80 V, respectively. Nitrogen was used as the drying and nebulizing gas, and helium was used as the collision gas when the system was operated in MS-MS mode. The gas temperature was 350 °C, dry gas 10 L/min, and nebulizer 25 psi. Full-scan data acquisition was performed by scanning from m/z 50 to 1700. Analysis of PHP, OH-TPHP and (OH)2-TPHP Metabolites in CEH by UPLC-ESI-TQ-S/MS. Determination of PHP and identified hydroxylated TPHP metabolites (OH-TPHP and (OH)2-TPHP) was conducted using a Waters ACQUITY UPLC I-Class system (UPLC) coupled to Waters Xevo TQ-S mass spectrometer (TQ-S/MS) (Milford, MA, U.S.A.) in the multiple reaction monitoring (MRM) mode. For PHP, the ESI was operated in negative mode. LC separation was carried out on a Cortecs UPLC C18 (2.1 × 50 mm2, 1.6 μm particle size) (Waters, Mississauga, ON, Canada). The LC mobile phases were water (A) and methanol (B), and both contained 2 mM of ammonium acetate. The mobile phase flow rate was 0.7 mL/ min and the gradient was as follows: 0 min, 1% B; hold for 2 min, 2−2.1 min, 95% B (linear); hold for 0.9 min; 3−3.1 min, 1% B (linear) and hold for 2.9 min. A volume of 10 μL of sample from the in vitro assay was injected into the LC system. Nitrogen was used as the drying and nebulizing gas, and highpurity argon was used as the collision gas. The capillary voltage was 1.0 kV. The source and desolvation temperatures were 150 and 500 °C, respectively. The desolvation and cone gas flow rates were 700 and 150 L/h, respectively. For OH-TPHP and (OH)2-TPHP analysis, the LC column and mobile phases were exactly same as those for PHP analysis. The mobile phase flow rate was 0.5 mL/min and the gradient was as follows: 0 min, 5% B; 0−5 min, 95% B (linear); hold for 1 min; 6−6.1 min, 5% B (linear) and hold for 4.9 min. ESI was operated in positive mode. The capillary voltage was 0.5 kV. The source and desolvation temperatures were 150 and 600 °C, respectively. The desolvation and cone gas flow rates were 800 and 150 L/h, respectively. The operation parameters for PHP, OH-TPHP and (OH)2TPHP are provided in SI Table S3. To our knowledge, this is the first UPLC-TQ-S/MS protocol for PHP, OH-TPHP, or (OH)2-TPHP determination. Accurate quantification of OHTPHP or (OH)2-TPHP was not possible due to lack of analytical standards. Here, relative comparisons of timedependent (0, 12, and 36 h) chromatographic peak areas were indicative of quantitative changes in the amount of metabolite formed in TPHP-exposed CEH. Data Analysis. For cytotoxic effects assessment, the luciferase intensity of each treatment was corrected for background (positive control, 300 μM of TDCIPP), and then normalized to a percent response value expressed relative to the response elicited by the DMSO control. Cytotoxicity data for TPHP and DPHP were fit to a nonlinear regression curve (log(agonist) vs response) using GraphPad software (version 5, San Diego, CA). PCR array data analysis was conducted using Eppendorf Mastercycler ep realplex 2.2 software (Eppendorf) and the cycle threshold (Ct) was set to 100. The fold change of
the protocol of the QuantiTect Reverse Transcription kit (Qiagen). A total of 36.4 μL of RNase-free water was added to each reaction and mixed by pipetting up and down. For the PCR array, 34 μL of the cDNA solution were added directly to the RT2 SYBR Green Mastermix (Qiagen). An aliquot of 25 μL cDNA/Mastermix was quickly added to each well containing a set of primers at preoptimized concentrations. The arrays were run using the Mastercycler ep realplex (Eppendorf, Hamburg, Germany) with the following thermal profile: 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min and ending with a dissociation curve segment of 95 °C for 15 s, 60 °C for 15 s, melting curve for 20 min, and 95 °C for 15 s. No amplification was observed in the genomic DNA contamination control, which ensured the robustness of the observed gene expression profiles from the PCR array. The positive PCR control and RT-control met the appropriate quality control criteria and the mRNA levels of the two internal control genes were invariable across all treatments. In Vitro Metabolism of TPHP. CEH were treated with a nominal concentration of 10 μM TPHP, which did not decrease hepatocyte viability significantly (see Cell Viability section). In parallel, the same amount of TPHP was added to wells that contained medium only in order to characterize any potential nonenzymatic hydrolysis of TPHP without the presence of CEH. DMSO-treated wells were included as experimental processing blanks. Samples (medium and cell layer) were collected at 0, 12, and 36 h. At each time point, the medium was aspirated gently from each well, and transferred into a 1.5 mL brown glass vial. The remaining cell layer was washed out twice with 100 μL of ethanol and transferred into a 1.5 mL brown glass vial. Before injection into the instrument for quantification, a 40 μL aliquot of medium or ethanol/cell mixture was diluted into 960 μL of fresh methanol spiked with 20 ng of two deuterated standards (d10-DPHP and d15-TPHP), and filtered through a centrifugal filter (0.2 μm Nylon membrane, 500 μL; VWR, Mississauga, ON, Canada). Quantification of TPHP and DPHP in CEH by LCESI(+)-MS/MS. Quantification of TPHP and DPHP was performed on a Waters 2695 high performance liquid chromatography (LC) system coupled to a Waters Quattro Ultima tandem quadrupole mass spectrometer (MS/MS) (Waters, Milford, MA, U.S.A.). The electrospray ionization (ESI) was operated in positive mode. TPHP was separated using a Waters Symmetry C18 column (100 mm L × 2.1 mm i.d., 3.5 μm particle size), and DPHP was separated using a Luna C18 column (100 mm L × 2.0 mm i.d., 3.0 μm particle size). TPHP was monitored by direct injection into the LCESI(+)-MS/MS, and DPHP was analyzed by use of dicationic ion technology. Detailed instrumental parameters and method validation for the analysis of TPHP and DPHP can be found in our previous publications,19,25 and are also provided in SI Table S2. Identification of OH-TPHP and (OH)2-TPHP Metabolites by LC-ESI(+)-Q-ToF-MS. An Agilent 1200 LC system, consisting of a degasser, binary high-pressure gradient pump, and autosampler and coupled to an Agilent 6520A Q-ToF-MS system (Agilent Technologies, Mississauga, ON, Canada), was used to identify TPHP metabolites. LC separation was carried out on a Luna C18(2) (2.0 × 50 mm2, 3 μm particle size) (Phenomenex, Torrance, CA, U.S.A.). The LC mobile phases were water (A) and methanol (B), and both contained 2 mM of ammonium acetate. The mobile phase flow rate was 0.4 mL/ 13513
dx.doi.org/10.1021/es5039547 | Environ. Sci. Technol. 2014, 48, 13511−13519
Environmental Science & Technology
Article
next subsection, the observed differences in CEH uptake of TPHP and DPHP might explain the observed cytotoxic effects to CEH, since TPHP was accumulated to a greater extent than DPHP. Cell death is a useful initial end point to rank/prioritize test chemicals,21 and the cytotoxic effects of TPHP in CEH can be used to compare to other organic flame retardant (OFR) chemicals. For example, a previous study examined 16 OFRs administered to CEH in a high-throughput screening study, and 6 of those, including tris(2-butoxyethyl) phosphate (TBOEP), bis(2-ethylhexyl) phosphate (HDEHP), allyl 2,4,6-tribromophenyl ether (ATE), α- and β-1,2-dibromo-4-(1,2-dibromoethyl)-cyclohexane (DBE-DBCH), and TDCIPP, significantly decreased CEH viability with LC50 values ranging from 40.6 (HDEHP) to 356 μM (β-DBE-DBCH).21 The LC50 of TPHP fell within the lower range of the other OFRs and in fact, HDEHP (LC50 40.6 ± 13 μM) was the only compound that was more cytotoxic to CEH than TPHP. Given the rapid metabolism of TPHP in CEH observed in the present study, the cytotoxic effect must either be extremely rapid or the result of exposure to other breakdown products. However, the lack of a significant cytotoxic response for one of the primary TPHP metabolites, DPHP, indicates that DPHP exposure did not elicit the large reduction in CEH viability. Alteration of mRNA Expression in CEH Exposed to TPHP/DPHP. The mRNA expression levels of 27 genes associated with phase I and II metabolism, immune function, glucose and fatty acid metabolism, oxidative stress, lipid/ cholesterol metabolism, thyroid hormone pathway, farnesoid X receptor (FXR) and liver X receptor (LXR), cell death, steatosis, and steroid metabolism were examined in CEH following a 36 h exposure to 10 μM of TPHP or DPHP (SI Table S1). Three of the 27 genes on the PCR array, THRSP, IGF1, and PDK4, were significantly down-regulated, and one gene, MT4, was up-regulated by TPHP in CEH (Figure 2; p < 0.05; fold-change ≥1.5). Two genes, THRSP and IGF1, are associated with the thyroid hormone pathway, which is critically important to normal central nervous system development, growth, and metabolism in avian species (Figure 2 and SI Table S4).28 As a transcription factor associated with the regulation of adipogenic enzymes, THRSP plays a role in thyroid hormone stimulation of lipogenesis.29 IGF1 encodes insulin-like growth factor 1, which is an important mediator of muscle growth and central to the modulation of adult mammalian muscles in mammals.30 TPHP was identified as a major contributor of Firemaster 550 in terms of binding to and activating human PPARγ,13 suggesting that disturbance of thyroid hormone signaling is a common adverse outcome for both avian species and humans following exposure to TPHP. Thyroid hormonerelated genes in CEH were also dysregulated following exposure to other flame retardants, such as hexabromocyclododecane (HBCDD), DE-71, α-DBE-DBCH, and tris(chloropropyl) phosphate (TCIPP),21,31 suggesting that disruption of the thyroid hormone pathway in avian species as a result of environmental chemical exposure represents an important adverse outcome and requires continued study and development of new assays. Pyruvate dehydrogenase kinase, isozyme 4 (PDK4) is associated with the glucose and fatty acid metabolism pathway, which is known to maintain the balance between carbohydrate and lipid metabolism.32 PDK4 was also dysregulated in chicken embryos following exposure to TDCIPP.33 Finally, MT4 is associated with oxidative stress, and induced by a variety of
target gene mRNA abundance relative to the vehicle control was calculated using the 2−ΔΔCt method,26 and significant differences in fold change compared to the DMSO vehicle control were determined using Student’s t-test (SABiosciences RT2 Profiler PCR Array Data Analysis Template v4.0, Microsoft Office Excel 2010). For visualization, nonsignificant fold changes (p > 0.05) and those less than 1.5 were set to 0 to minimize noise, and the gene expression profiling was performed on R 3.0.2 version using “gplots” package. Quantification of TPHP, DPHP, and PHP in CEH was conducted using MassLynx V4.1 software (Milford, MA, U.S.A.). The method limit of detection (MLOD) and method limit of quantification (MLOQ) were defined as the concentration of target compounds producing a peak in the chromatogram with a S/N ratio of 3 and 10, respectively. Concentrations lower than MLODs are reported as “not detected (ND)”. Concentrations lower than MLOQs but higher than MLODs are reported as “10 μM, and based on a fitted curve, the LC50 was 47 ± 8 μM (Figure 1). This LC50 value was similar to the calculated half maximal inhibitory concentration (IC50) value (37 μM) reported for Chinese Hamster Ovary (CHO) cells exposed to TPHP.14 However, no significant (p > 0.05) differences were observed between the DPHP-treated groups (up to a concentration of 1000 μM) and the DMSO control based on a one-way analysis of variance followed by Dunnett’s multiple comparisons test (SI Figure S2). As discussed in the 13514
dx.doi.org/10.1021/es5039547 | Environ. Sci. Technol. 2014, 48, 13511−13519
Environmental Science & Technology
Article
to the mammalian xenocensors: constitutive androstane receptor (CAR) and pregnane X receptor (PXR).40,41 In an OFRs in vitro screening study using CEH and the avian ToxChip PCR array, 8 of the 16 OFRs up-regulated CYP3A37 by a factor of 2.5- to 19.5-fold at the highest noncytotoxic dose administered.21 CYP3A37 induction was also observed in liver tissue of chicken embryos exposed via egg injection to OFRs including, HBCDD, TBOEP, triethyl phosphate, DBE-DBCH, tris(methylphenyl) phosphate, TDCIPP, and TCIPP demonstrating good concordance between in vitro and in ovo approaches for this end point.42−45 Only one OFR, 2,2bis(bromomethyl)-1,3-propanediol, down-regulated CYP3A37 in CEH (2.2-fold),21 a finding similar to that for DPHP. However, down-regulation of this phase I enzyme is not well understood in terms of toxicological importance. Overall, neither TPHP nor DPHP elicited a response typical of several other OFRs with respect to CYP3A37 mRNA expression. In Vitro Metabolism of TPHP, Part I: Mass Balance between TPHP and DPHP. Potential nonenzyme catalyzed hydrolysis of TPHP in the medium was assessed by spiking TPHP into fresh CEH-free medium. The concentration of TPHP in the medium was 2.27 ± 0.20, 2.38 ± 0.10, and 2.45 ± 0.05 nmol/well at 0, 12, and 36 h, respectively, and no significant differences were observed among TPHP concentrations (n = 3) measured at the three time points (One-way ANOVA, p > 0.05). Concomitantly, the amount of DPHP in the same wells was found to be extremely low; 0.03 ± 0.003 and 0.11 ± 0.01 nmol/well at 12 and 36 h, respectively. This accounted for 1 and 5% of the TPHP measured in the medium of the same wells at 12 and 36 h. These results indirectly demonstrate that enzyme-catalyzed metabolism is the major mechanism behind degradation as opposed to nonenzymatic hydrolysis. Concentrations of TPHP in both medium and cells at three time points, 0, 12, and 36 h, were determined to be 3.52 ± 0.22, 0.69 ± 0.17, and 0.06 ± 0.01 nmol/well, respectively, demonstrating that TPHP undergoes progressive degradation in CEH (Figure 3). Correspondingly, progressive DPHP
Figure 2. Transcriptional profiles of 27 target genes on the Avian ToxChip PCR array following exposure to 10 μM triphenyl phosphate (TPHP) or its metabolite diphenyl phosphate (DPHP). Hierarchical clustering was conducted based on mRNA expression fold-changes derived from a mean of three replicates. Genes with fold changes lower than 1.5 and p > 0.05 were set to 0 to minimize noise. Green-black-red represents down-, no-, and up-regulation. The detailed mRNA foldchange data are available in SI Table S4 for TPHP and DPHP.
metals, mytokines, stress hormones, oxyradicals, and xenbiotics in healthy cells.34,35 MT overexpression protects against hepatotoxicity whereas MT-knockout aggravates hepatotoxicity and metal-induced lethality in mice.36 In contrast to the mRNA expression results for TPHP, the metabolite, DPHP, altered more genes in CEH (9 of 27); THRSP, PDK4, CYP7B1, SLCO1A2, CYP3A37, CD36, DIO1, ACSL5, and HMGCR were all significantly down-regulated (p < 0.05, fold change >1.5) highlighting the variability in response among the two chemicals (Figure 2 and SI Table S4). In addition to the thyroid hormone pathway (DIO1, THRSP) and the glucose and fatty acid metabolism pathway (PDK4), genes associated with three other pathways including lipid/cholesterol metabolism (ACL5, HMGCR, SLCO1A2, CD36), FXR and LXR (CYP7B1), and Phase I and II metabolism (CYP3A37) were altered following a 36 h exposure to DPHP. Four of the 6 genes on the PCR array associated with lipid/cholesterol metabolism were down-regulated by DPHP, suggesting a possible adverse outcome pathway elicited by DPHP. Lipids and sterols play important roles in diverse biological processes in eukaryotes, such as membrane biosynthesis, intra- and extracellular signaling, and energy storage.37 In previous publications, TDCIPP was transcriptionally active with respect to lipid/cholesterol metabolism in CEH21 and liver tissue of chicken embryos exposed via egg injection.38 CYP7B1 encodes a member of the cytochrome P450 superfamily of enzymes, and its physiological roles include bile salt synthesis, steroid hormone metabolism, metabolism of estrogen receptor ligands and immunoglobulin production.39 CYP3A37 is associated with Phase I and II metabolism, and its expression is regulated by the chicken xenobiotic receptor (CXR) that is a nuclear orphan receptor equivalent in function
Figure 3. Mass balance between the degradation of triphenyl phosphate (TPHP) and formation of diphenyl phosphate (DPHP) following exposure of chicken embryonic hepatocytes to TPHP for 0, 12, or 36 h. The data are mean values and error bars (standard deviation) are based on three wells per treatment group.
formation was observed at the three time points with concentrations increasing from 0.03 ± 0.03 to 0.19 ± 0.05 to 0.60 ± 0.07 nmol/well (Figure 3). At 36 h, the remaining TPHP concentration only accounted for 0.2% of the original administered dose. The concentration of DPHP at 36 h accounted for approximately 17% of the initial TPHP dosing concentration (Figure 3). Interestingly, in our very recent 13515
dx.doi.org/10.1021/es5039547 | Environ. Sci. Technol. 2014, 48, 13511−13519
Environmental Science & Technology
Article
be further metabolized by CEH (e.g., Reemtsma et al. detected several OP monoesters in human urine samples);7 and/or (2) TPHP was directly metabolized to other chemicals in addition to DPHP. To test these two hypotheses, the monoester form of TPHP, PHP, was screened in TPHP-exposed CEH and the medium samples collected at 36 h were also scanned by LCESI-ToF/MS to identify potential undefined metabolites. However, PHP was consistently lower than the MLOD (
