|Home | About | Journals | Submit | Contact Us | Français|
Flavin-containing monooxygenases often are thought not to be inducible but we recently demonstrated aryl hydrocarbon receptor (AHR)-dependent induction of FMO mRNAs in mouse liver by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Celius et al., Drug Metab Dispos 36:2499, 2008). We now evaluated FMO induction by other AHR ligands and xenobiotic chemicals in vivo and in mouse Hepa1c1c7 hepatoma cells (Hepa-1). In mouse liver, 3-methylcholanthrene (3MC) induced FMO3 mRNA 8-fold. In Hepa-1 cells, 3MC and benzo[a]pyrene (BaP) induced FMO3 mRNA >30-fold. Induction by 3MC and BaP was AHR-dependent but, surprisingly, the potent AHR agonist, TCDD, did not induce FMO3 mRNA in Hepa-1 cells nor did chromatin immunoprecipitation assays detect recruitment of AHR or ARNT to Fmo3 regulatory elements after exposure to 3MC in liver or in Hepa-1 cells. However, in Hepa-1, 3MC and BaP (but not TCDD) caused recruitment of p53 protein to a p53 response element in the 5'-flanking region of the Fmo3 gene. We tested the possibility that FMO3 induction in Hepa-1 cells might be mediated by Nrf2/antioxidant response pathways but agents known to activate Nrf2 or to induce oxidative stress did not affect FMO3 mRNA levels. The protein synthesis inhibitor, cycloheximide (which causes “superinduction” of CYP1A1 mRNA in TCDD-treated cells) by itself caused dramatic upregulation (>300-fold) of FMO3 mRNA in Hepa-1 suggesting that cycloheximide prevents synthesis of a labile protein that suppresses FMO3 expression. Although FMO3 mRNA is highly induced by 3MC or TCDD in mouse liver and in Hepa-1 cells, FMO protein levels and FMO catalytic function showed only modest elevation.
Flavin-containing monooxygenases (FMOs) metabolize a broad range of substrates that contain nitrogen or sulfur atoms, including therapeutic agents, environmental compounds and endogenous compounds derived from the diet such as trimethylamine (reviewed in: (Krueger and Williams, 2005; Cashman and Zhang, 2006)). For example, sequential FMO-dependent S-oxygenations of thiocarbamides/thioureas produce the mono-oxygenated sulfenic and di-oxygenated sulfinic acid metabolites. Both reactive metabolites have been isolated and confirmed for a number of thioureas in timecourse studies (e.g. (Poulsen, 1979; Francois et al., 2009)). Sometimes, only one of the oxygenated metabolites has been observed, presumably due to differences in the reaction rates that either favor accumulation of the mono- or di-oxygenated metabolites, or absence of detection favored by further fast non-FMO dependent reactions. In addition, sulfenic acids are known to bind glutathione and initiate futile redox cycling in the cell producing oxidative stress and toxicity in liver and other target organs (Krieter et al., 1984; Ruse and Waring, 1991; Smith and Crespi, 2002; Henderson et al., 2008).
Regulation of FMO expression by endogenous factors, including hormones is complex (Klick and Hines, 2007; Hines et al., 2008; Klick et al., 2008). FMOs traditionally have been viewed as not being inducible by xenobiotic chemicals (Krueger and Williams, 2005; Cashman and Zhang, 2006; Hines, 2006). However, we recently found that the mRNAs for FMO1, FMO2 and FMO3 are highly induced in livers of male mice treated with the potent aryl hydrocarbon receptor (AHR) agonist, 2,3,7,8-tetraclorodibenzo-p-dioxin (TCDD) (Tijet et al., 2006; Celius et al., 2008). Induction of FMO2 and FMO3 mRNAs by TCDD occurs in mice that have wildtype AHR but not in Ahr-null mice; this confirms that the AHR is essential for FMO upregulation by dioxin-like chemicals in vivo (Tijet et al., 2006). Furthermore, we demonstrated recruitment of AHR and AHR nuclear translocator (ARNT) to regulatory regions in the mouse Fmo3 gene indicating that induction by TCDD is an AHR-dependent process (Celius et al., 2008). Two other AHR agonists also have been reported to induce FMO mRNAs in rodent livers in vivo: FMO3 in mouse liver by ß-naphthoflavone (BNF) (Patel et al., 2007) or benzo[a]pyrene (BaP) (Yauk et al., 2010) and FMO1 in rat liver by 3-methylcholanthrene (3MC) (Chung et al., 1997).
The Hepa-1 cell model has been exceptionally useful for defining the mechanism of cytochrome P4501A1 (CYP)1A1) induction by AHR agonists (Okey et al., 1980; Denison and Whitlock, 1995; Hankinson, 1995). Thus, in the current research, to facilitate mechanistic studies of FMO regulation, we employed the Hepa-1 cell model to investigate induction of FMOs by various chemical agents. By using Hepa-1 variant cell lines deficient in AHR or ARNT, our experiments indicate that the AHR and its dimerization partner, ARNT, are required for induction of FMO3 mRNA by 3MC and BaP in Hepa-1 cells but that additional mechanisms must also be involved in upregulation of FMO3 by these carcinogenic polycyclic aromatic hydrocarbons.
3MC, BaP, butylated hydroxylanisole (BHA), sulforaphane (SUL), menadione, tert-butylhydroquinone (t-BHQ), MDL28170, chloroquine, MG132, sulindac sulfide (SS) and sulindac sulfoxide (SOX) as well as antibodies to IgG used in the ChIP assay were obtained from Sigma-Aldrich (St. Louis, MO). Ethionamide sulfoxide (ETASO) was a kind gift from Dr. Paul R. Ortiz de Montellano (University of California, San Francisco, CA). Mouse anti-AHR antibody (SA-210) was from BIOMOL Research Laboratory (Plymouth Meeting, PA); human anti-p53 antibody (FL-393) and human anti-ARNT were from Santa Cruz Biotechnology, Inc (Santa Cruz, CA). The Hepa-1c1c7 (Hepa-1) mouse hepatoma cell line, the AHR-deficient cell line c35 (B16GBi1c3), and the ARNT-deficient cell line c4 (B13NBii1) were obtained from the American Type Culture Collection (Manassas, VA). TCDD was from Wellington Laboratories (Guelph, ON, Canada); BNF and PCB126 from Accustandard (New Haven, CT); Power SYBR Green PCR Master Mix, TaqMan Universal PCR Master Mix, and 96-well real-time PCR plates were from Applied Biosystems (Foster City, CA). Reagents for cell culture were provided by Invitrogen (Carlsbad, CA). All other reagents were of the highest quality available from commercial sources.
Adult (8-week old) male C57BL/6J mice from Charles River Laboratories (Montreal, QC, Canada) were housed under a 12-h light/12-h dark cycle and given food and water ad libitum. Mice were injected intraperitoneally with 3MC (80 mg/kg body weight) or corn oil vehicle. Livers were harvested 6 h after treatment and were snap-frozen in liquid nitrogen, then stored at −80 °C for later use in chromatin immunoprecipitation (ChIP) assays, RNA isolation and real-time PCR analysis, proteomic analysis and measurements of catalytic activity.
All hepatoma cell lines were cultured in nucleoside-free α-minimal essential medium containing 10% heat-inactivated fetal bovine serum (FBS) (HyClone, Logan, UT) and 1% penicillin/streptomycin (PEST). Cells were maintained at 37°C in a 5% CO2/air incubator with 90% humidity. Medium was changed every second day and cells were sub-cultivated once a week. For FMO induction experiments cells were seeded in 6-well plates (200,000 cells/well) and cultured for 24 h prior to treatment. Cells were exposed to the known AHR agonists: TCDD (10 nM), 3MC (1 μM), BaP (1 μM), BNF (1 μM), or PCB 126 (1 μM) for 6, 12, or 24 h. We also exposed cells to several compounds known to activate Nrf2 or oxidative stress pathways: BHA (50 μM), SUL (20 μM), menadione (10 μM), and t-BHQ (30 μM). Potential superinduction of FMO3 and CYP1A1 was studied by exposing Hepa-1 cells to10 μg/ml cycloheximide (CHX) for 1 h prior to exposure of the cells to 1 μM 3MC for 5 h or 23 h.
Probes and primers for mouse Fmo1 (NM_010231), Fmo2 (NM_018881), Fmo3 (NM_008030), Fmo4 (NM_) and Fmo5 (NM_010232) were designed as reported in Celius et al., (Celius et al., 2008). Total RNA from Hepa-1 cells was isolated using RNeasy mini kits (Qiagen, Mississauga, ON) according to the manufacturer's instructions. Real-time RT-PCR was performed using 1 μg of total RNA and a random hexamer primer in a standard procedure. Two μl of RT-reaction was used for each real-time qPCR reaction. Each reaction (25 μl) contained optimized probe and primer concentrations as well as the TaqMan universal mastermix. Expression levels of target mRNAs were normalized to the levels of 18S rRNA and analyzed using the comparative CT (ΔΔ CT) method.
ChIP assays were performed according to Matthews et al. (Matthews et al., 2005) except that 100–120 mg of snap-frozen liver was homogenized and cross-linked in 1% formaldehyde and the DNA was isolated using a PCR purification kit (BioBasic) and eluted in 50 μl of elution buffer provided in the kit. When Hepa-1 cells were used for preparation of chromatin for ChIP assays, cells (1.8×106) were plated in 10-cm dishes and cultured for 24 h prior to exposure to 1 μM 3MC or 1 μM BaP for 1, 6, 15 or 24 h. Cell pellets were cross-linked in 1% formaldehyde and the in vitro ChIP assay followed the same protocol as for in vivo ChIP assays. ChIP DNA (1 μl) was amplified by PCR with primers 5'AAGCCAAGCCAGAAAATCAA3' and 5' TCGAGGAACAGAGTGCAATG3' for the Fmo3 AH response element (AHRE); 5'-GAGGATGGAGCAGGCTTACG-3' and 5'-GGGCTACAAAGGGTGATGCTT-3' for the mouse Cyp1a1 AHRE; 5`-ATGGAAGGGTCTTGACACCA3` and 5`-CAAATTCTGCCCCATCTGTT3` for the mouse Fmo3 p53RE; and 5′-CCCGAAACCCAGGATTTTAT-3′ and 5′-GGTCTGTCCCTGACCAACT-3′ for the p53 response element in the Cdnk1a gene (also known as p21). For real-time PCR, Power SYBR green PCR master mix was used to amplify the DNA fragments.
Hepa-1 cells were plated (1.8×106) in a 10-cm dish and pre-exposed to protease inhibitors (5 μM MG132; 25 μM MDL28170 or 50 μM chloroquine) for 1 h followed by 23 h exposure to DMSO or 3MC. Cells were scraped and washed once with PBS and pellets were resuspended in buffer (10 mM phosphate buffer, pH 7.4, 20% glycerol, 1.0 mM EDTA, 0.2 mM PMSF) and frozen at −80°C. Cells were subsequently lysed after adding 1 μl/ml CalBiochem protease inhibitor cocktail III, by alternately freezing the cell suspension in liquid nitrogen and thawing on ice, three times. The homogenate was further degraded by passage through a 28-gauge needle 5 to 10 times. Protein concentration was measured using Pierce's Coomassie Plus assay.
Mouse liver microsomes (100 μg for ethionamide (ETA); 200 μg for SS) were incubated for 15 min at 37° C with 200 μM substrate and 1 mM NADPH in 0.1 M Tricine buffer at pH 8.5. Reactions were stopped by the addition of 75 μl CH3CN (ETA) or 2 ml ethyl acetate (SS) and chilling on ice. For ETA, samples were transferred to microcentrifuge tubes and centrifuged for 30 min at 15,000 × g to remove protein and the supernatant analyzed by reverse phase HPLC (Henderson et al., 2008) for presence of ETA S-oxide (ETASO). We assessed activity toward SS as described previously (Koukouritaki et al., 2007): samples were vortexed for 1 min and the ethyl acetate layer removed to a new tube. The extraction was repeated and the samples taken to dryness under vacuum in a Speed Vac. After resuspension in 50:50 CH3CN:H2O, samples were analyzed by reverse phase HPLC for presence of sulindac sulfoxide (SOX). Statistical analysis was performed with GraphPad Prism 5 software unpaired two-tailed t-test. * = significant at p ≤ 0.05.
AQUA (absolute quantification) mass spectrometry quantifies proteins by tandem mass spectrometry and MRM (multiple reaction monitoring) (Gerber et al., 2003). AQUA employs synthetic peptide standards identical to unique tryptic peptides except that standards incorporate stable isotopes. A manuscript providing a detailed description of AQUA for quantification of mouse FMOs 1–3 and 5 will be forthcoming (MCH, SKK and DEW; in preparation); a brief description is included here.
An FMO3 standard candidate DSFPGL*NR (mouse Fmo3 amino acids 158–165) containing dual 13C- and 15N-labeled leucine (L*) was synthesized (Sigma, St. Louis, MO) after QTOF experiments confirmed that trypsinized microsomes from both Sf9 cells expressing mouse FMO3 and mouse liver contained the corresponding unlabeled peptide. Protein samples (25 μg/well) were separated on NuPAGE 3–8% Tris-Acetate gels (Invitrogen). Subsequently, FMO bands were excised, spiked with labeled peptide, and in-gel trypsinization performed (Shevchenko et al., 2006) with sequencing-grade trypsin (Promega, Madison, WI).
Peptides were separated by UPLC on a Waters Symmetry C18 (180 μm × 20 mm) trapping column, and eluted onto a Waters nano Acquity UPLC column (1.7 μm, BEH130 C18, 100 μm × 100 mm) connected to a nano Acquity UPLC (Waters Corporation, Milford, MA.). Solvents were 0.1% formic acid (A) and ACN/0.1% formic acid (B), (Burdick & Jackson) at a flow rate of 0.47 μl/min with a gradient from 5–35% B. Mass spectrometry was performed using an Applied Biosystems 4000 Qtrap (Applied Biosystems/MDS Analytical Technologies, Concord, ON, Canada) using the nanospray ion source. Data was acquired in MRM mode with a 50 mSec dwell time for each transition. Precursor and product ions were 453.2 (precursor), 703.4, and 556.3 for the native protein and 456.7 (precursor), 710.4 and 563.3 for the standard.
Values obtained for area under the curve of the first product ions of sample and standard were used to calculate protein content (Analyst Software, Applied Biosystems). The second product ion provided additional confirmation of identity. The standard curve range was 50 to 1000 fmol.
FMO gene expression is regulated by multiple agents including hormones, developmental factors and nutritional status (Krueger and Williams, 2005; Hines, 2006). Although FMOs generally have been considered not to be inducible by xenobiotic chemicals, recent studies indicate that several AHR agonists – including TCDD, 3MC and BNF – can upregulate FMO3 mRNA (Chung et al., 1997; Tijet et al., 2006; Patel et al., 2007; Celius et al., 2008; Novick et al., 2009). Induction of FMO3 mRNA by TCDD in mouse liver in vivo proceeds essentially via the same well-characterized AHR-dependent mechanism that regulates CYP1A1 induction (Celius et al., 2008). However, the mechanisms responsible for FMO3 induction by non-halogenated xenobiotics have not been explored in detail. The goals of our current study were to extend the range of compounds tested for FMO3 induction and to identify induction mechanisms, primarily using the Hepa-1 cell model.
3MC is a well-known inducer of the classic AHR-regulated gene, Cyp1a1 in a wide variety of rodent tissues in vivo (Okey, 2007). Thus, as a prelude to studies in the Hepa-1 cell culture model, we tested the non-halogenated AHR agonist 3MC to determine if it could induce FMO mRNA in mouse liver in vivo. Treatment of adult male C57BL/6 mice with 80 mg/kg of 3MC for 6 h significantly increased mRNA levels both for FMO3 (8-fold) and for FMO2 (2.5-fold) (Fig. 1). 3MC had no effect on mRNA levels for FMO1, FMO4 or FMO5 in mouse liver (data not shown).
Previously we found, by ChIP assays, that in vivo treatment with the potent AHR agonist, TCDD, causes significant recruitment of AHR and ARNT proteins to an AH response element (AHRE) in an upstream regulatory region of the mouse Fmo3 gene (Celius et al., 2008). Although 3MC treatment induced FMO3 mRNA 8-fold in mouse liver (Fig. 1), we did not detect any recruitment of AHR or ARNT to AHREs in the Fmo3 gene in livers of 3MC-treated mice (data not shown). The Cyp1a1 gene served as a positive control in the ChIP experiments. There was strong recruitment of AHR and ARNT to an AHRE regulatory region in the Cyp1a1 gene in mice treated with 3MC (Fig. 2), confirming that the 3MC treatment regimen we employed activates the AHR/ARNT pathway. Thus the lack of recruitment of AHR and ARNT to AHREs in the Fmo3 gene is not due to a general inability of 3MC to activate AHR pathways in vivo but reflects a selective failure of recruitment to the Fmo3 gene.
Given that a non-halogenated AHR agonist, 3MC, induced FMO3 mRNA in mouse liver in vivo, we tested additional well-characterized AHR agonists, including both non-halogenated and halogenated compounds in the Hepa-1 cell culture model. The non-halogenated polycyclic aromatic hydrocarbons, 3MC and BaP, both caused significant induction of FMO3 mRNA in Hepa-1 cells after a 24 h exposure (Fig. 3). In addition to FMO2 and FMO3, we examined FMO1, FMO4 and FMO5 but none of these mRNAs was induced by 3MC or BaP in Hepa-1 cells (data not shown).
Although we previously found the prototypical AHR agonist, TCDD, to be a very efficacious inducer of multiple species of FMO mRNA in mouse liver in vivo (Celius et al., 2008), surprisingly TCDD did not induce FMO3 mRNA in Hepa-1 cells (Fig. 3). The co-planar dioxin-like PCB126 and BNF, both of which are known AHR agonists for induction of CYP1A1 (Denison and Nagy, 2003) and Figure 4), also failed to induce FMO3 mRNA. FMO2 mRNA was not induced in Hepa-1 cells (data not shown) by any of the compounds illustrated in Figure 3, including 3MC and TCDD, in contrast to our previous in vivo study which showed TCDD to be a potent inducer of FMO1, FMO2 as well as FMO3 in mouse liver (Celius et al., 2008).
As described above, 3MC did not cause recruitment of AHR/ARNT to AHREs in the Fmo3 gene in mouse liver in vivo; thus FMO3 induction by 3MC does not appear to adhere to the canonical AHR induction model. Therefore we evaluated alternative mechanisms that might explain induction. We initially focused on the Nrf2/ARE pathway because this system mediates induction of multiple enzymes that protect cells from xenobiotic chemicals and oxidative stress (Kohle and Bock, 2006). It also has become increasingly apparent that the AHR pathway and the Nrf2 pathway are interconnected in a way that facilitates protection by synergistically regulating a partially-overlapping set of genes involved in cell defense from xenobiotics (Kohle and Bock, 2006). In fact, Nrf2 itself is induced by TCDD, in an AHR-dependent fashion providing a further link between these two regulatory systems (Miao et al., 2005; Franc et al., 2008). In addition, we gave attention to the Nrf2/ARE pathway because we previously found that the 5`-flanking region of the mouse Fmo3 gene contains multiple copies of the anti-oxidant response element (ARE) (Celius et al., 2008). Moreover, 3MC is categorized as a “bifunctional inducer” (Prochaska and Talalay, 1988; Kwak and Kensler, 2006) or “mixed activator” (Kohle and Bock, 2006) meaning that it activates not only the AHR but also the Nrf2/ARE pathway. Thus we tested other bifunctional inducers as well as “monofunctional inducers” (which activate only the Nrf2/ARE pathway) to determine whether the Nrf2/ARE system might be responsible for upregulation of FMO3 mRNA.
The bifunctional inducers/mixed-activators, BHA, BNF and t-BHQ (Kohle and Bock, 2006) had little effect on FMO3 mRNA levels in Hepa-1 cells (Fig. 3). Sulphoraphane (SUL) activates Nrf2 (Miao et al., 2004) and has been reported to also activate the AHR (Anwar-Mohamed and El-Kadi, 2009); however, SUL also failed to induce FMO3. If FMO3 induction in Hepa-1 cells occurs through combined activation of the AHR and Nrf2, then bifunctional inducers such as SUL and BNF should upregulate FMO3 mRNA but BNF was unable to accomplish this induction. t-BHQ is a Nrf2-activator that induces multiple “Phase II” enzymes such as glutathione-S-transferases, UDP-glucuronosyltransferases and NQO1 (Kohle and Bock, 2006; Kohle and Bock, 2007). Recently t-BHQ was reported to also act as an agonist for the AHR and to directly induce CYP1A1 in Hepa-1 cells; thus t-BHQ now is classed as a bifunctional inducer (Gharavi and El-Kadi, 2005; Kohle and Bock, 2006). Despite the fact that t-BHQ can act directly on both AHR and Nrf2 pathways, t-BHQ did not induce FMO3 mRNA in Hepa-1 cells in our experiments (Fig. 3). The results with BNF and t-BHQ indicate that combined activation of AHR and Nrf2 is not sufficient to upregulate FMO3. Nrf2 was a plausible candidate to regulate induction of FMO3 by 3MC and BaP in Hepa-1 cells but Nrf2 is ruled out by our experimental evidence.
TCDD is known to provoke oxidative stress in rodent liver in vivo (Hassoun et al., 2002) and in Hepa-1 cells (Puga et al., 2000) but TCDD did not induce FMO3 mRNA in Hepa-1 cells in our experiments (Fig. 3). We also subjected the Hepa-1 cells to oxidative stress by exposing them to menadione but this, too, failed to produce much elevation in FMO3 mRNA (Fig. 3). Thus it seems very unlikely that FMO3 induction reflects a component of the complex cellular response to oxidative stress.
In Hepa-1 cells there was no significant increase in FMO3 mRNA levels at 6 h after exposure to any of the AHR agonists tested (Fig. 4). FMO3 mRNA was significantly increased in cells exposed to 3MC or BaP for 12 h but the maximum response was not obtained until a 24-h exposure to 3MC (Fig. 4). CYP1A1 mRNA, which served as a benchmark for induction by AHR agonists, was significantly induced within 6 h by all 5 AHR agonists tested. CYP1A1 mRNA levels continued to increase up to 24 h for all compounds tested except for BNF which produced its maximal induction at 12 h (Fig. 4).
To further characterize FMO3 induction we obtained dose-response curves for induction of FMO3 mRNA by 3MC and found that 1 μM of 3MC was required to significantly elevate mRNA levels in Hepa-1 cells after 24 h of exposure (Fig. 5), whereas CYP1A1 was induced at 3MC concentrations as low as 0.01 μM (Fig. 5).
CHX has been reported to suppress AHR turnover in TCDD-exposed Hepa-1 cells (Ma et al., 2000) and to cause “superinduction” of CYP1A1 mRNA in these cells (Lusska et al., 1992; Ma et al., 2000) or in MCF10A breast cancer cells (Joiakim et al., 2004). Therefore we tested whether CHX would cause superinduction of FMO3 mRNA (as it does for CYP1A1) by pre-treating Hepa-1 cells with 10 μg/ml CHX for 1 h followed by 5 h or 23 h exposure to 1 μM 3MC (Fig. 6). Similar to what was shown in previous studies with TCDD (Lusska et al., 1992; Ma et al., 2000; Joiakim et al., 2004), CHX potentiated induction of CYP1A1 mRNA by 3MC, particularly after 24 h (Fig. 6; data not shown for 6 h). However, CHX did not potentiate induction of FMO3 mRNA by 3MC.
Unexpectedly, CHX, on its own, caused a very large increase in FMO3 mRNA levels (Fig. 6). Our ChIP studies did not show any recruitment of AHR or ARNT to Fmo3 regulatory regions in cells exposed to CHX, indicating that the AHR/ARNT complex is not directly responsible for upregulation of FMO3 mRNA by CHX (Fig. 7). Previously CHX, in the absence of any exogenous AHR agonist, had been found to cause only a small increase in CYP1A1 mRNA levels (Ma, 2007). Superinduction of CYP1A1 by the combination of CHX plus an AHR agonist has been attributed to the fact that CHX treatment increases the level of AHR protein in cells by blocking AHR proteolysis (Ma and Baldwin, 2000; Ma et al., 2000). The proteolytic process may require the presence of a labile protein whose synthesis is blocked by CHX (Ma, 2007). The dramatic upregulation of FMO3 mRNA that we observed in cells treated with CHX alone might occur because CHX prevents synthesis of a labile protein that acts as a repressor of Fmo3 gene transcription or that facilitates degradation of FMO3 mRNA. We postulate a labile protein as a negative regulator of constitutive FMO3 mRNA expression because our data indicate that this protein requires constant synthesis given that, in the presence of the protein-synthesis inhibitor CHX, levels of FMO3 mRNA increased rapidly and dramatically. This rise plausibly reflects decreased synthesis of a labile protein that normally suppresses basal expression of FMO3. If the protein was stable, we would expect that basal levels of FMO3 would be unchanged after CHX treatment, since the expression of a stable protein would be not affected by CHX within the time-course of this experiment.
We previously found that an intact AHR signaling pathway is essential to FMO3 mRNA upregulation by TCDD in mouse liver in vivo since induction does not occur in Ahr-null mice (Tijet et al., 2006). The wildtype Hepa-1 cells used in experiments up to this point contain high levels of the AHR (Okey et al., 1980). In order to determine whether the AHR and its dimerization partner, the ARNT protein, are essential to the observed upregulation of FMO3, we tested the ability of 3MC to induce FMO3 mRNA in variant cell lines derived from the Hepa-1 parent: an AHR-deficient cell line (c35), and the ARNT-deficient cell line (c4). 3MC was completely unable to induce FMO3 mRNA in variant cells that are deficient in either AHR or ARNT (data not shown); thus both the AHR and ARNT are required for FMO3 mRNA induction by 3MC.
In addition to the mouse wildtype Hepa-1 cell line and its AHR-deficient and ARNT-deficient derivatives, we also tested whether AHR agonists could induce FMO3 in the rat 5L hepatoma cell line and the human HepG2 hepatoma cell line. Both the 5L cell line and the HepG2 cell line contain AHR and both lines exhibit CYP1A1 induction when treated with AHR agonists such as TCDD (Reiners et al., 1999) or 3MC (Li et al., 1998). However, neither TCDD nor 3MC was able to induce FMO3 mRNA in 5L cells or in HepG2 cells despite the presence of an intact AHR mechanism in these cell lines (data not shown).
FMO3 induction by 3MC or BaP in Hepa-1 cells clearly requires the AHR pathway. However, induction by these non-halogenated PAHs does not seem to be mediated simply by the conventional mechanism involving binding of the ligand:AHR:ARNT complex to AH-responsive elements in the 5'-flanking region of the Fmo3 gene. We were unable to detect any binding of AHR:ARNT to the Fmo3 regulatory region by ChIP assays with DNA from cells treated with 3MC although the positive control (the Cyp1a1 gene) from the same cells demonstrated a strong signal of AHR:ARNT (Fig. 7). In liver of mice treated in vivo, TCDD produces robust binding of AHR:ARNT to an AHRE-I motif in the 5'-flank of the Fmo3 gene accompanied by elevation of FMO3 mRNA levels (Celius et al., 2008). Induction of FMO3 by TCDD in liver in vivo probably proceeds by a mechanism very similar to that which regulates CYP1A1 induction but this does not appear to be the case for FMO3 induction by 3-MC in vivo nor by 3-MC or BaP in Hepa-1 cells in culture.
There is precedent for an alternative mode of transcriptional regulation by 3MC and BaP in Hepa-1 cells. Mathieu and co-workers (Mathieu et al., 2001) found that p53 mediates induction of the multidrug resistance gene, mdr1, in Hepa-1 cells treated with 3MC or BaP and that 3MC and BaP require metabolic activation in order to stimulate the p53 pathway. 3MC and BaP induce their own metabolism by binding to the AHR and upregulating expression of CYP1A1 and other members of the AH gene battery (Okey and Vella, 1982; Okey et al., 1984; Riddick et al., 1994; Nebert et al., 2004).
To test whether 3MC could recruit p53 to p53-response elements in the Fmo3 gene we used ChIP assays with the Cyp1a1 gene as a positive control. As expected, 3MC caused robust recruitment of AHR and ARNT proteins to regulatory regions in the Cyp1a1 gene in Hepa-1 cells (Fig. 7, top panel). Recruitment to the Cyp1a1 regulatory region in 3MC-treated cells was significantly higher in cells that had been pre-incubated with CHX than in cells not exposed to CHX (Fig. 7, top panel). Despite the fact that 3MC caused a substantial increase in FMO3 mRNA in Hepa-1 cells, our ChIP analyses did not detect significant recruitment of AHR or ARNT to regulatory regions in the Fmo3 gene in either the presence or absence of CHX (Fig. 7, bottom panel). Nor did CHX, on its own, stimulate recruitment of AHR or ARNT to the regulatory regions of the Fmo3 gene (Fig. 7, bottom panel).
However, both 3MC and BaP caused a significant increase in binding of the p53 protein to a p53-response element in the 5'-flanking region of the Fmo3 gene in wildtype Hepa-1 cells that contain AHR and ARNT (Fig. 8a) as these PAHs did to the Cdnk1a gene, a known p53-regulated gene (Mathieu et al., 2001) used as a positive control (Fig. 8b). In ARNT-deficient cells, neither 3MC nor BaP caused recruitment of p53 to the p53 response element in the Fmo3 gene (data not shown).
The AHR pathway is the central mechanism for FMO3 mRNA induction by TCDD in mouse liver in vivo as well as being the extensively-characterized central mechanism for CYP1A1 induction in Hepa-1 cells. Despite abundant AHR, Hepa-1 cells fail to exhibit induction of FMO3 mRNA by TCDD. Both 3MC and TCDD are classic agonists for the AH receptor and both induce FMO3 mRNA in mouse liver in vivo. However, our experiments indicate that the mechanism of FMO3 induction is different for these two agents. TCDD, a halogenated polycyclic aromatic, upregulates FMO3 transcription by direct recruitment of AHR/ARNT to AH response elements in the 5'-flanking region of the Fmo3 gene. 3MC is unable to promote binding of AHR/ARNT to the Fmo3 AHRE.
The aggregate data from our experiments suggest a model (Fig. 9) in which 3MC and BaP initially bind to the AHR and upregulate CYP1 enzymes. In turn the CYP1 enzymes convert 3MC or BaP into reactive metabolites that damage DNA, leading to activation of the p53 protein and its subsequent binding to a p53-response element in enhancer regions of the Fmo3 gene. TCDD is poorly metabolized, forms few reactive metabolites (Poland and Glover, 1979), and does not induce FMO3 expression in Hepa-1 cells, again consistent with a mechanism in which the inducer must first be metabolically activated in order to upregulate FMO3. Fmo3 regulation in Hepa-1 cells differs from that in mouse liver in vivo. In liver in vivo, TCDD induces FMO3 mRNA expression in parallel with recruitment of TCDD-bound AHR to an upstream regulatory region of the Fmo3 gene. With 3MC as the inducer, we did not detect AHR binding to an AHRE-containing region of the Fmo3 gene in mice in vivo. The predominant mechanism of Fmo3 upregulation by TCDD in liver in vivo appears to be via the canonical AHR:AHRE system whereas induction of FMO3 mRNA by 3MC or BaP in Hepa-1 cells in culture requires the AHR/AHRE mechanism followed by activation of p53. There is a lag-time for elevation of FMO3 mRNA when compared with the rapid upregulation of CYP1A1 mRNA which is mediated by direct binding of activated AHR to AHREs. The requirement for AHR/ARNT, taken together with the delayed response for FMO3 induction, is consistent with a model in which 3MC and BaP first induce CYP1A1 which then activates these PAHs to metabolites that stimulate p53 binding to its regulatory elements.
There is another very recent example of induction of drug-metabolizing enzymes by reactive metabolites. Weems & Yost (2010) found that concentrations of 3-methylindole that damage DNA also induce CYP1A1 in human bronchial epithelial cells and that CYP1A1 induction was attenuated by the inhibitor of P450 catalytic activity, 1-aminobenzotriazole. The potential role of p53 in mediating induction of CYP1A1 by 3-methylindole was not examined in the study by Weems & Yost but seems worthy of testing.
We previously observed (Celius et al., 2008) with methimazole as substrate that the marked induction of FMO mRNA by TCDD in mouse liver was not accompanied by a comparable rise in catalytic activity. To ensure that this apparent lack of induction of catalytic activity was not peculiar to methimazole, we examined activity towards two additional FMO substrates of clinical importance, ETA and SS. ETA is a prodrug, activated by FMO-dependent S-oxygenation to a metabolite toxic to Mycobacterium tuberculosis (Shimizu et al., 2007; Henderson et al., 2008). SS is an nonsteroidal anti-inflammatory (NSAID) which is S-oxygenated by FMO to the pharmacologically inactive S-oxide (Hamman et al., 2000). Genetic polymorphisms in expression of FMO3 allelic variants are responsible for the disease known as trimethylaminuria (Yeung et al., 2007); these patients also exhibit altered activity towards other pharmaceuticals known to be FMO3 substrates (Shimizu et al., 2007). A clinical trial of SS in chemoprevention of Familial Adenomatous Polyposis demonstrated that this FMO3 genetic polymorphism significantly impacted the chemopreventive potency of SS (Hisamuddin et al., 2004).
There was a small but significant increase in S-oxygenation of ETA to the sulfenic acid in liver microsomes from TCDD-treated male but not female mice (Fig. 10). The basal activity of female liver microsomes toward ETA was higher than control or TCDD-treated males and was not further increased by TCDD. There was a slight increase in SS S-oxygenation in males, but this was not significant. No increase in S-oxygenation was observed in liver microsomes from TCDD-treated females. Based on these observations showing a lack of any marked impact on FMO activity by TCDD, it is not apparent that exposure to environmental levels of TCDD would impact drug metabolism even though there is such a large increase in FMO mRNA levels.
The constitutive FMO3 protein content of liver microsomes was approximately 15-fold higher in female mice than in male mice. However, despite robust induction of FMO3 mRNA, TCDD had no significant effect on FMO3 protein content in either male or female (Fig. 11). There is precedent with another drug-metabolizing enzyme for mRNA induction by 3MC and TCDD without an increase in the encoded protein. Deb and Bandiera (Deb and Bandiera, 2010) found that CYP2S1 mRNA was induced by 3MC and TCDD in several rat tissues, including liver, but that CYP2S1 protein levels were unaffected by these inducers. Multiple mechanisms can be postulated (see (Deb and Bandiera, 2009; Hrycay and Bandiera, 2009) to account for the failure of the induced mRNA to yield significant increases in protein levels or protein function. One possibility is that the induced protein is rapidly degraded by cellular proteases. However, chloroquine, an inhibitor of autosomal/lysosomal pathways (Lee et al., 2008), did not increase FMO3 protein levels over that observed in DMSO-treated control cells or cells treated with 3MC nor did MDL28170, an inhibitor of calpain-like activity (Dale and Eltom, 2006) (Supplemental Fig. 1). We also tested the proteasome inhibitor, MG132 (5 μM), but MG132 was toxic to the Hepa-1 cells (data not shown). Our measurements of FMO3 protein levels in Hepa-1 cells must be considered pilot experiments because only a single sample was able to be analyzed for each treatment condition. Nonetheless, it seems clear that the large elevation of FMO3 mRNA in 3MC-treated mouse liver in vivo or in Hepa-1 cells in culture is not accompanied by any meaningful increase in FMO3 protein or catalytic activity.
It is clear that FMO3 regulation is exceptionally complex. The discrepancy between mRNA levels and protein levels for FMO3 exemplifies this complexity. We also have found that upregulation of FMO mRNAs by AHR agonists does not occur uniformly between sexes or across different animal species. Our previous study revealed that FMO2 and FMO3 mRNAs are highly induced by TCDD in liver of adult male mice. In adult female mice the magnitude of hepatic FMO3 induction is much more modest than in males and there is no induction of FMO2 in females (Celius et al., 2008). Although strong upregulation of FMO3 mRNA by TCDD occurs in mouse liver (Celius et al., 2008; Novick et al., 2009), TCDD does not induce FMO3 in liver of adult rats (Celius et al., 2008). Transgenic mice that express wildtype rat AHR exhibit full FMO2 and FMO3 induction by TCDD (Celius et al., 2008); thus the rat AHR is competent to drive induction of these mRNAs. The species-difference in induction must reside at a regulatory level other than the AHR.
Our current study shows that FMO induction also differs between mouse liver in vivo and mouse hepatoma cells in culture; i.e., the prototypical AHR agonist, TCDD, is a potent inducer of mRNAs encoding FMO1, FMO2, and FMO3 in liver of male mice in vivo (Celius et al., 2008) but TCDD is completely inactive as an FMO3 inducer in the hepatoma-derived Hepa-1 cell line that has been extensively used as a model for regulation by the AHR (Okey, 2007). The non-halogenated AHR agonists, 3MC and BaP, induce FMO3 mRNA in mouse liver and in Hepa-1 cells. However, even though both 3MC and BaP are AHR ligands, the mechanism of FMO3 induction by these PAHs is distinctly different from the conventional AHR mechanism by which TCDD acts in vivo.
The profound variations in response between different agonists, between rodent species and between cells in culture versus in vivo indicates that multiple regulatory mechanisms are involved, even when all of the xenobiotic inducers tested are confined to compounds known to be AHR agonists. Our current study has revealed a new mechanism, p53, by which FMO3 mRNA is upregulated by 3MC and BaP. Clearly much more remains to be clarified regarding the multiple mechanisms that alter FMO3 expression, particularly the reason why the highly-induced FMO3 mRNA does not yield a proportional increase in FMO3 protein.
Supported by Grant MOP-57903 from the Canadian Institutes of Health Research to ABO and Paul C. Boutros, Grant MOP-82715 from the Canadian Institutes of Health Research to JM and Grants HL038650 and P30ES000210 from USPHS-NIH to DEW. JM is the recipient of a Canadian Institutes of Health Research New Investigator Award. The sponsors had no involvement in the study design, data collection, data analysis or interpretation, writing of the report or in the decision to submit the paper for publication.
1Abbreviations: AHR, aryl hydrocarbon receptor; AHRE, AH response element; ARE, anti-oxidant response element; ARNT, aryl hydrocarbon receptor nuclear translocator; BaP, benzo[a]pyrene; BHA, butylated hydroxyanisole; BNF, ß-naphthoflavone; ChIP, chromatin immunoprecipitation; CHX, cycloheximide; DMSO, dimethylsulfoxide; ETA, ethionamide; FMO, flavin-containing monooxygenase; 3-MC, 3-methylcholanthrene; Nrf2, NF-E2-related factor 2; PCB126, 3,3′,4,4′,5-pentachlorobiphenyl; SS, sulindac sulfide; t-BHQ, tert-butylhydroquinone; SOX, sulindac sulfoxide; SUL, sulphoraphane; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Potential conflicts of interest Allan Okey has received compensation as a member of the Dioxin Scientific Advisory Board of The Dow Chemical Company within the last three years. All other authors declare no potential conflicts of interest.