Identification of Benzothiazole Derivatives and Polycyclic Aromatic Hydrocarbons as Aryl Hydrocarbon Receptor Agonists Present in Tire Extracts

Materials

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD, dioxin) was obtained from Steve Safe (Texas A&M University, TX, USA), [γ³²P]ATP (>5000 Ci/mmol) was purchased from Amersham (Piscataway, NJ, USA), dimethylsulfoxide (DMSO), hexane, toluene, ethyl acetate (EA), and methanol (MeOH) were from Fisher Scientific (PA, USA) and poly dI dC was from Roche Molecular Biochemicals (IN, USA). Cell culture reagents and media were purchased from Gibco/BRL (NY, USA), G418 was from Gemini Bio-Products (CA, USA), and luciferase lysis and assays reagents from Promega (WI, USA). Truck tire tread was collected from a highway roadside. Benzothiazole derivatives were purchased from several sources: N-cyclohexyl-2-benzothiazolylsulfenamide, 2-(4-chlorophenyl)-benzothiazole, 2-benzothiazolyl diethyldithiocarbamate, 2-(4-morpholinyldithio)-benzothiazole and (2-benzothiazolylthio)acetic acid from TCI (OR, USA); 2-morpholinobenzothiazole and 2-benzothiazolyl disulfide from Pfaltz and Bauer (CT, USA); N-t-butyl-2-benzothiazolesulfenamide and N,N-dicyclohexyl-benzothiazolesulfenamide from Wako (VA, USA); BT, 2-methylthiobenzothiazole (MTBT), 2-mercaptobenzothiazole (MBT), 2-hydroxybenzothiazole (OBT), 2-(2-benzothiazolylthio)-ethanol, 2-benzothiazolamine and 2-chloro-benzothiazole from Aldrich Chemical (MO, USA); 4-(2-benzothiazolyl)-2-methyl-benzenamine, N-(2-benzothiazolylthio)-N-cyclohexyl-2-benzothiazolesulfenamide and 2,2'-dithiobis (6-nitro-benzothiazole) from TimTec (DE, USA); 2-(benzylsulfinyl)-1,3-benzothiazole from ChemBridge (CA, USA).

Tire sample preparation, fractionation and chemical analyses

For analysis of total tire extract, 1 gram of washed inner tire tread was cut into small pieces and immersed into 2 ml of DMSO overnight in a Teflon^®-lined screw-cap tube, followed by centrifugation for 15 min at 1500 rpm. The resulting supernatant (referred to as the original tire extract) was transferred into a fresh tube and an aliquot was serially diluted (10–10,000-fold range) in DMSO and 1 µl of each concentration analyzed for AhR agonist activity. For fractionation studies, 16 g of tire tread was incubated in 100 ml of toluene overnight, the supernatant collected and the tire fragments extracted twice more with 100 ml of toluene. The combined toluene extract was evaporated to dryness, chromatographed through a hexane-washed silica gel column and the column sequentially eluted with 60 ml hexane, 60 ml hexane–toluene (1:1, v/v), 60 ml toluene and 60 ml MeOH to yield twenty-two primary fractions of approximately 10 ml each. Fractions were dried and resuspended in 1 ml of hexane and the AhR agonist activity in 1 µl of each fraction was determined using the mouse hepatoma H1L1.1c2 cells bioassay as described below, with fractions 4 to 22 showing agonist activity. This concentration of hexane had no adverse effect on the cells (data not shown). The most active fraction (number 20) was subjected to a secondary fractionation by silica gel column chromatography and eluted with 20 ml hexane–toluene (1:1, v/v), 20 ml toluene, 20 ml toluene–EA (1:1, v/v), 20 ml EA, and 20 ml MeOH to obtain twenty secondary fractions of approximately 5 ml each. Fractions were dried and resuspended in 1 ml of hexane (fractions 1–17) or methanol (fractions 18–20) and agonist activity in 1 µl of each fraction was determined, with positive AhR activity identified in fractions 8 to 12 and 18 to 20. Hexane solvent blanks were also run before sample fractionation to check for contaminants released from the chromatographic matrix. Analysis of these solvent blanks by gas chromatography-mass spectrometry (GC-MS) and gel retardation analysis revealed no chemical contamination or AhR activity. Active fractions were further analyzed by GC/MS in full scan mode using an electron impact ionization mode on a HP5973 mass spectral detector tuned with perfluorotributylamine using the system auto-tune parameters. The mass spectrometer was interfaced with a HP 6890 gas chromatograph equipped with a 30 m × 0.25 m, 0.25 µM DB-5ms phase column. Optimized oven program used inlet, transfer line and quadrapole temperatures of 250°C, 280°C, and 160°C, respectively. The oven program was an initial temperature of 50°C held for 1 min, ramped at 10°C/min to 320°C and held 10 min. Helium was used as the carrier gas at a constant flow of 0.8 ml/min. Analyses were executed in the splitless mode.

Cell culture and induction of luciferase activity

Analysis of chemicals, extracts and chromatographic fractions for their ability to stimulate AhR-dependent gene expression was carried out using recombinant mouse hepatoma (Hepa1c1c7) cell-based CALUX (H1L1.1c2 and H1L6.1c2) and CAFLUX (H1G1.1c3) clonal cell lines. The CALUX clonal cell lines are essentially identical and contain a stably integrated AhR-/dioxin-responsive element (DRE)-driven firefly luciferase plasmid (pGudLuc1.1 or pGudLuc6.1, respectively) whose time course of luciferase induction varies as a result of differences in intracellular localization and stability of the luciferase gene product (with H1L1.1c2 cells inducing rapidly and maximally by 4 to 6 h after agonist treatment and H1L6.1 cells inducing more slowly with little activity at 4 to 6 h but with maximal induction observed by 24 h after agonist treatment [20,21]). The CAFLUX clonal cell line contains a stably integrated AhR-/DRE-driven enhanced green fluorescent protein (EGFP) reporter gene plasmid whose activity can be repeated measured in intact cells for time course studies [21,24]. Transcriptional activation in the CALUX and CAFLUX cell lines occurs in a ligand-, dose- and AhR-dependent manner [20,21,24]. For CALUX analysis, cells grown in white clear-bottomed 96-well microplates were incubated with carrier solvent DMSO (10 µl/ml), TCDD (1 nM), the indicated concentration of total tire extract, chromatographic fraction or the indicated compound for 4 h (in H1L1.1c2 cells) or 24 h (in H1L6.1c2 cells) at 37°C and luciferase activity measured in a Berthold microplate luminometer as previously described [19–22]. For CAFLUX analysis, cells grown in black, clear-bottomed 96-well microplates were incubated with carrier solvent DMSO (10 µl/ml), TCDD (1 nM) or the indicated concentration of total tire extract at 33°C for the indicated time. Concentrations of DMSO up to 2% produced no visual signs of cytotoxicity in CALUX or CAFLUX cell lines. EGFP was measured in intact cells (without removal of media) using a Fluostar microtiter plate fluorometer (Phenix Research Products) with an excitation wavelength of 485 nm (25 nm bandwidth) and an emission wavelength of 515 nm (10 nm bandwidth) as previously described [24]. Normalization between experiments involved adjusting the instrument fluorescence gain setting so that the level of EGFP fluorescence induced by 1 nM TCDD resulted in 9,000 relative fluorescence units. Samples were run in triplicate and the fluorescent activity present in wells containing media only were subtracted from the fluorescence in all samples.

Preparation of cytosol and gel retardation analysis

Hepatic cytosol was prepared from male Hartley guinea pigs (250–300 g) in buffer containing 25 mM Hepes (pH 7.5), 1 mM ethylenediamine tetraacetic acid, 1 mM dithiothreitol, and 10% (v/v) glycerol as previously described and aliquots stored at −80°C [25]. Incubation of cytosol with TCDD or the indicated extracts or chemicals and subsequent gel retardation analysis was carried out as described in detail [25]. Briefly, a complementary pair of synthetic oligonucleotides containing the DRE3 binding site for the transformed AhR:Arnt complex (5'-GATCTGGCTCTTCTCAC-GCAACTCCG-3' and 5'-GATCCGGAGTTGCGTGAGAAGAGCCA-3') were synthesized, purified, annealed, and radiolabeled with [γ³²P]ATP. An aliquot (125 µl) of liver cytosol (8 mg of protein/ml) was incubated with DMSO (2.5 µl), TCDD in DMSO (20 nM final concentration), 1.25 µl of the original tire extract in DMSO or the indicated compound (200 µM each in DMSO) for 2 h at 20°C. A 10 µl aliquot of each incubation was analyzed by gel retardation analysis, and protein-DNA complexes were resolved by non-denaturing polyarylamide gel electrophoresis and autoradiography of the dried gel. The amount of ³²P-labeled DRE present in the induced protein-DNA complex was determined using a molecular dynamics phosphoimager.

Tire extract induces AhR-dependent gene expression

The presence of AhR agonists in leachates of tires have been suggested from several studies in aquatic organisms [2,6,12]. In order to systematically identify the responsible chemical(s), we first needed to confirm the presence of AhR agonists in tire extracts. The indicated concentration of tire extract was incubated for four hours with recombinant mouse hepatoma (H1L1.1c2) cells, which contain a stably integrated AhR-responsive luciferase reporter gene [22,21]. These results demonstrated that tire extract induced luciferase in a concentration-dependent manner in the H1L1.1c2 cells (Fig. 1A), producing a response between 40 to 170% of that maximally induced by TCDD. This induction response was back calculated to result from a range of extract that was equivalent to 0.5 µg to 0.5 mg of original tire. Superinduction of reporter gene activity by this extract was unexpected, but not surprising given our previous observation of the ability of other sample solvent extracts to produce a synergistically enhanced induction response [26]. Similar concentration-dependent induction/superinduction by this crude tire extract was also observed in recombinant AhR-responsive guinea pig, rat and human cell lines containing a stably transfected AhR-responsive reporter gene (Supplemental Data, Fig. S1) demonstrating that this induction response was not unique to the mouse AhR and that this extract contained AhR agonists. Maximal induction in these cell lines was approximately 210, 170, and 230% of that induced by 1 nM TCDD, respectively. While the exact mechanism(s) responsible for the observed superinduction response is not known, we have previously described several possible mechanisms that may contribute to synergistic increases in AhR-dependent gene expression [26].

To determine whether the induction response was by metabolically stable HAHs that might have been extracted from the tire and/or by other metabolically labile AhR ligands (such as PAHs and others), we examined the time course of induction of AhR-dependent gene expression in recombinant rat hepatoma (H1G1.1c3) cells that contain a stably transfected AhR-responsive EGFP reporter gene. A major advantage of these cells is that the same cells can be analyzed repeatedly in time course studies since measurement of EGFP activity is carried out using intact cells in culture media [21,24]. While EGFP was induced in a concentration-dependent manner, the magnitude of gene induction by the tire extract initially increased over 48 hours and then progressively decreased over time (Fig. 1B); TCDD continued to induce EGFP to a stable maximal activity. These results not only indicate that the responsible AhR agonists extracted from tire are metabolically labile and produce only a transient induction response, they also confirm that the responsible chemicals are not HAHs.

Tire extract stimulates AhR transformation and DNA binding in vitro

While the cell-based bioassays demonstrate induction of AhR-dependent reporter gene activity by the tire extract, they do not directly demonstrate the ability of chemicals in the tire extract to bind to and activate the AhR. To confirm the ability of the tire extract to directly activate the AhR, we examined its ability to stimulate AhR transformation and DNA binding in vitro. Incubation of guinea pig hepatic cytosol with crude tire extract and subsequent gel retardation analysis revealed that the tire extract did indeed induce AhR transformation and DNA binding (Fig. 1C). Phosphoimager quantitation revealed that the amount of AhR:ARNT:DRE complex formed by incubation with tire extract was comparable to that produced by a maximally stimulating concentration of TCDD (Fig. 1D). These results suggest that the ability of the tire extract to stimulate AhR-dependent gene expression is due to the ability of chemicals present in the extract to bind to and stimulate AhR transformation and DNA binding.

TIE identification of AhR-active compounds in tire extracts

In order to identify the specific chemical(s) responsible for activating the AhR and AhR signaling pathway, a toxicant identification evaluation approach was used. Tire extract was fractionated by silica gel column chromatography and progressively eluted with a series of solvents (hexane, hexane/toluene (1:1, v/v), toluene and then methanol) into 22 fractions (Fig. 2A). Aliquots of each fraction were exchanged into hexane and AhR agonist activity in each fraction determined using the recombinant mouse hepatoma (H1L1.1c2) cells. Interestingly, 19 of 22 fractions exhibited AhR agonist activity with a range in overall induction between 15 to 130% of that maximally induced by TCDD (Fig. 2B). The presence of AhR agonist activity in so many fractions suggests the presence of a large number of structurally different AhR active compounds. To further separate, analyze and identify responsible AhR agonists, the most active fraction (#20) from the first fractionation scheme was subjected to additional silica gel chromatography and progressively eluted with a different series of solvents (hexane/toluene (1:1, v/v), toluene, toluene/ethyl acetate (1:1, v/v), ethyl acetate and methanol) into 20 fractions (Fig. 2A). Aliquots of each fraction were exchanged into hexane (fractions 1–17) or methanol (fractions 18–20) and AhR agonist activity of each determined using the H1L1.1c2 cells. Positive AhR agonist activity (defined as >10% of that of that maximally induced by TCDD) was contained in fractions 8 to 12 and 18 to 20 (Fig. 2C). Instrumental analysis by GC/MS was carried out on each active fraction and comparison of mass spectra and retention times to authentic standards revealed that active fraction 18 was shown to contain several BT derivatives (specifically BT, MTBT, MBT, and OBT (Fig.3 and Supplemental Data, Fig. S2); a number of PAHs were also tentatively identified in fraction 9, although some of these PAHs were also present in other fractions (structures shown in Fig. 3). Many of these PAHs are known AhR agonists [15,27–29] or would be expected to have AhR agonist activity. The lack of chemical contaminants or AhR activity in the hexane solvent blanks run before extract fractionation confirmed that the AhR activity was derived from the tire extract.

MBT and OBT are AhR agonists

While instrumental analysis indicated the presence of several BT derivatives in active fraction 18, they do not confirm that these compounds are AhR active. Accordingly, we examined the ability of commercially obtained BT, MTBT, MBT, and OBT to stimulate AhR-dependent gene expression in the mouse H1L1.1c2 cell line (Fig. 4A). Although these four BTs have similar structures (Fig. 3), only MBT and OBT induced AhR-dependent gene expression and they were relatively weak inducers (as compared to TCDD); BT and MTBT were inactive. While OBT was relatively efficacious and demonstrated a full concentration-dependent induction of luciferase activity (with an EC50 of ~4 µM), MBT was only a partial AhR agonist, with a concentration of 100 µM MBT inducing luciferase to approximately 50% of that produced by a maximal inducing concentration of TCDD (Fig. 4A). To examine the metabolic stability of MBT and OBT as AhR agonists, we examined the ability of 10 µM of each compound to induce luciferase activity in two mouse hepatoma (Hepa1c1c7)-based CALUX clonal cell lines whose time course of maximal luciferase gene induction is relatively rapid (4–6 h for H1L1.1c2 cells) or slow (24 h for H1L6.1c3 cells) [21]. The lower relative magnitude of luciferase gene induction at 24 h as compared to 4 h, compare to that induced by 1 nM TCDD, indicates that these inducers are metabolically labile (Supplemental Data, Fig. S3), consistent with the transient induction response by the crude tire extract [Fig. 1B]. To demonstrate that the identified BTs can bind to and activate the AhR, we examined their ability to stimulate AhR transformation and DNA binding by gel retardation analysis. These analyses not only demonstrated that OBT and MBT could stimulate AhR transformation and DNA binding to between 40 to 50% of that of TCDD and confirmed them as AhR agonists, but interestingly revealed that BT and MTBT also stimulated AhR transformation and DNA binding, albeit only to 15 to 20% of that produced by TCDD (Fig. 4B). While these results would suggest that BT and MTBT are weak AhR agonists, their inability to induce AhR-dependent gene expression may result from lower potency/affinity, more rapid metabolism within the cell and/or other factors.

Activation of the AhR by BT derivatives

The above analysis demonstrates the ability of the OBT and MBT to bind to and activate the AhR and AhR-dependent gene expression. While a previous study identified a BT derivative as an AhR agonist [30], the widespread use of these chemicals in many materials and potential exposure to these compounds raises the question as to whether AhR agonist activity is a common property of other BT derivatives. Accordingly, we examined the ability of sixteen commercially available BT derivatives (Fig. 5) to stimulate AhR-dependent luciferase gene expression (Fig. 6A) as well as AhR transformation and DNA binding by gel retardation analysis (Fig. 6B). Although some compounds were inactive as activators of AhR-dependent gene expression in H1L1.1c2 cells after 4 h of incubation (compounds 2, 4–6, 11), several compounds (namely, compounds 3, 8–10, 13, 14) were moderately active, inducing luciferase reporter gene activity to between 40 and 100% of that obtained with a maximally inducing concentration of TCDD. To examine the metabolic stability of these novel AhR agonists, we further examined their ability to induce luciferase gene expression after 24 h incubation using mouse hepatoma (H1L6.1c2) cells. Similar to the results with OBT and MBT, most of the BT derivatives induced significantly less luciferase activity at 24 h as compared to 4 h (Fig. 6A), indicating that they are metabolically labile AhR agonists. Interestingly, compound 14 induced significantly more luciferase activity at 24 h than at 4 h, suggesting a somewhat different mechanism of action for this compound and/or its metabolism/conversion into a more potent AhR agonist. The ability of these BT derivatives to stimulate guinea pig AhR transformation and DNA binding in vitro (Fig. 6B,C) were relatively consistent with the mouse cell induction for compounds 1, 3, 7 to 10, 12 to 14 (Figs. 6A upper panel, 6C), indicating that they can directly activate the AhR; compound 11 was inactive in both assays. While compounds 15 and 16 induced reporter gene activity to 20 to 30 % of that of TCDD, they stimulated little or no AhR transformation or DNA binding (Fig. 6B,C). In contrast, compounds 2,4 to 6 were similar to BT and MTBT in that they could stimulate AhR DNA binding (to 10–60% of that of TCDD [Fig. 6B,C]), yet did not induce AhR-dependent gene expression (Fig. 6A). Previous studies have identified chemicals that can bind to and activate the AhR in vitro, but fail to induce AhR-dependent gene expression in cells in culture [15,16,18], most likely the result of rapid metabolic degradation of the chemicals within the cells.