Abstract
The metabolic activation of a drug to an electrophilic reactive metabolite and its covalent binding to cellular macromolecules is considered to be involved in the occurrence of idiosyncratic drug toxicity (IDT). As a cellular defense system against oxidative and electrophilic stress, phase II enzymes are known to be induced through a Kelch-like ECH-associated protein 1/nuclear factor E2-related factor 2/antioxidant response element system. We presumed that it is important for the risk assessment of drug-induced hepatotoxicity and IDTs to observe the biological responses evoked by exposure to reactive metabolites, and then investigated the mRNA induction profiles of phase II enzymes in human hepatocytes after exposure to problematic drugs associated with IDTs, such as ticlopidine, diclofenac, clozapine, and tienilic acid, as well as safe drugs such as levofloxacin and caffeine. According to the results, the problematic drugs exhibited inductive effects on heme oxygenase 1 (HO-1), which contrasted with the safe drugs; therefore, the induction of HO-1 mRNA seems to be correlated with the occurrence of drug toxicity, including IDT caused by electrophilic reactive metabolites. Moreover, glutathione-depletion and cytochrome P450 (P450)-inhibition experiments have shown that the observed HO-1 induction was triggered by the electrophilic reactive metabolites produced from the problematic drugs through P450-mediated metabolic bioactivation. Taken together with our present study, this suggests that HO-1 induction in human hepatocytes would be a good marker of the occurrence of metabolism-based drug-induced hepatotoxicity and IDT caused by the formation of electrophilic reactive metabolites.
The occurrence of idiosyncratic drug toxicity (IDT) is a major problem in drug development and is of great concern to the pharmaceutical industry. A number of drugs have been withdrawn from the market or severely restricted in their use because of unexpected toxicities that become apparent only after the launch of new drug entities (Kaplowitz, 2005). Currently, it is suggested that the metabolic activation of a drug to a reactive metabolite and its covalent binding to cellular macromolecules are involved in the occurrence of IDTs (Walgren et al., 2005; Zhou et al., 2005). Bioanalytical techniques, including the detection of glutathione (GSH) adducts to reactive metabolites by liquid chromatography/tandem mass spectrometry (LC/MS/MS) and covalent binding assessment using radiolabeled chemical entities, have been used during lead optimization to minimize chemically reactive metabolite formation (Baillie and Davis, 1993; Evans et al., 2004; Gan et al., 2005). Because reactive metabolite formation and covalent binding to macromolecules are thought to be necessary yet insufficient processes for the generation of an idiosyncratic reaction, it is also important to observe the biological responses evoked by exposure to reactive metabolites for the risk assessment of IDTs (Liebler and Guengerich, 2005).
The electrophilic stress derived from reactive metabolites has been reported to cause an immediate adaptive defense response in cells. This involves various mechanisms, including the nuclear translocation of redox-sensitive transcription factors, such as nuclear factor E2-related factor 2 (Nrf2), which sense electrophilic stress and oxidative stress and protect against cellular damage via Kelch-like ECH-associated protein 1 (Keap1) (Wakabayashi et al., 2004; Dinkova-Kostova et al., 2005). An increasing number of studies have identified the genes regulated by Nrf2. These include genes involved in phase II drug metabolism, such as heme oxygenase 1 (HO-1), γ-glutamylcysteine synthetase (γ-GCS), glutathione-S-transferases (GSTs), NAD-(P)H:quinone reductase 1 (NQO1), and UDP-glucuronosyltransferases (UGTs) (Talalay et al., 2003). The cooperative activity of these enzymes serves as a cellular defense against electrophiles and oxidative stress products. In addition, the comparative analysis of gene expression changes using keap1 or Nrf2-deficient mice has also facilitated the identification of numerous new genes regulated by Nrf2 such as carboxylesterase 1 (CES1) (Thimmulappa et al., 2002). We postulate in this report that the induction of Nrf2-related genes would be a good indicator of electrophilic cell stress by reactive metabolites and would be useful as a risk marker for metabolism-based drug toxicity and IDTs that might not be expressed during drug development. Although intensive studies on Nrf2-mediated induction have been carried out using human cultured cell lines such as HepG2 or rodent hepatocytes (Cantoni et al., 2003; Nioi et al., 2003; Shinkai et al., 2006), there are only a few reports regarding this induction in human hepatocytes (Keum et al., 2006). For a better assessment of the potential risks of metabolism-based drug toxicities, it is necessary to detect the cellular responses to electrophilic stresses generated through drug metabolism in human hepatocytes.
In the present study, we have investigated the induction profiles of Nrf2-related genes in human hepatocytes after they were exposed to known “problematic” drugs associated with IDTs, such as ticlopidine, diclofenac, clozapine, and tienilic acid. The induction profiles after exposure to safe drugs such as levofloxacin and caffeine were also examined for the purposes of comparison.
Materials and Methods
Materials. Butylated hydroxyanisole (BHA), tert-butyl hydroquinone (tBHQ), l-buthionine-(S,R)-sulfoximine (BSO), aminobenzotriazole (ABT), diclofenac, clozapine, furosemide, acetaminophen, acetylsalicylic acid, and dimethyl sulfoxide (DMSO) were purchased from Sigma Chemical Co. (St. Louis, MO), and caffeine was purchased from Fluka (Buchs, Switzerland). Ticlopidine and levofloxacin were synthesized at Daiichi Pharmaceutical Co., Ltd. (Tokyo, Japan). 14C-Labeled acetylsalicylic acid and caffeine were purchased from American Radiolabeled Chemicals, Inc. (St. Louis, MO). Pooled human (n = 50, mix gender) liver microsomes were purchased from XenoTech (Lenexa, KS). β-NADP+ and glucose-6-phosphate dehydrogenase were purchased from Oriental Yeast Co. Ltd. (Tokyo, Japan), and glucose 6-phosphate was from Sigma. Fetal bovine serum, TRIzol, and custom primers were purchased from Invitrogen (Carlsbad, CA). TaqMan Gold reverse transcription-polymerase chain reaction (PCR) kits, TaqMan Universal PCR Master Mix, and custom TaqMan probes were purchased from Applied Biosystems (Foster City, CA). All the other chemicals and reagents were of analytical grade and were available from commercial sources.
Human Hepatocyte Culture. Cryopreserved human hepatocytes (lot number FEP, KCT, and ZCA) were obtained from In Vitro Technologies (Baltimore, MD). Hepatocytes were seeded onto 48-well collagen I-coated plates (Asahi Techno Glass, Chiba, Japan) in Lanford medium containing 5% v/v fetal bovine serum at a density of 3.0 × 105 viable hepatocytes/well/0.5 ml. They were placed into a 37°C, 5% CO2, saturated humidity incubator and incubated for 4 h to allow the cells to attach. Then the medium containing dead or unattached cells was removed from each well and replaced with 0.5 ml of warmed (37°C) Lanford medium. The hepatocyte cultures were maintained for 48 h after plating, with incubation medium changed daily. Stock solutions of the test compounds were prepared in DMSO and stored at -20°C until dilution with incubation medium to the final concentration. The final concentration of DMSO in each solution was 0.1% v/v. Plated hepatocytes were treated in triplicate for each treatment for 24 h at 37°C, 5% CO2, saturated humidity with incubation media containing each test compound. For a negative control, plated hepatocytes were also incubated with the medium containing DMSO (0.1% v/v). The hepatocyte cultures were treated with 0.25 ml of ice-cold TRIzol and harvested into separate 1.5-ml centrifuge tubes on ice. Total RNA was extracted by the acid guanidinium thiocyanate-phenol-chloroform method. For each RNA sample, the total RNA extracted was diluted to 5 ng/μl with diethyl pyrocarbonate-treated water and added to the reaction mixture. The concentrations and purity of the isolated RNA samples were determined using UV spectrophotometry conducted on a GeneQuant pro (Amersham Biosciences, Piscataway, NJ).
Reverse Transcription Reaction. The reaction mixture for reverse transcription was prepared with TaqMan Gold reverse transcription-PCR kits. The reverse transcription was performed on a GeneAmp PCR System 9700 (Applied Biosystems) for 10 min at 25°C, 30 min at 48°C, and 5 min at 95°C.
Quantitative Real-Time PCR. The PCR reaction mixture was prepared with TaqMan PCR Universal Master Mix. An inventoried Taqman MGB probe and primer mixture (Applied Biosystems) were used in the PCR reaction for CES1, γ-GCS, GSTA1, GSTP1, HO-1, NQO1, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and β-actin. The codes of the TaqMan MGB probes used were as follows: Hs00275607_m1 (CES1), Hs00155249_m1 (γ-GCS), Hs00272272_m1 (GSTA1), Hs00168310_m1 (GSTP1), Hs00157965_m1 (HO-1), Hs00168547_m1 (NQO1), 4316317E (GAPDH), and 4326315E (β-actin). The primers and probe for UGT1A1 (Accession number NM000463) were obtained from Invitrogen, and the positions of their sequences are as follows: 490 to 516 (forward primer), 603 to 627 (reverse primer), and 453 to 522 (probe). Quantitative real-time PCR was performed in a MicroAmp Optical 96-Well Reaction Plate (Applied Biosystems) with an ABI PRISM 7000 (Applied Biosystems). An initial step occurred for 2 min at 50°C, subsequently followed by heating to 95°C for 10 min, and followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. The relative quantification of the gene expression levels was determined by a standard curve generated from a dilution series (from 0.0061–400 ng/μl) of a standard RNA sample obtained from the control hepatocytes and normalized to the amount of GAPDH or β-actin mRNA. The results are expressed as mean ± S.D., and significant differences were evaluated using t test.
Lactate Dehydrogenase Release. To determine the toxic effects of the compounds, the levels of lactate dehydrogenase (LDH) in the cell culture medium were analyzed according to the recommendations of the manufacturer of the Cytotoxicity Detection Kit (LDH) (Roche Diagnostics Corporation, Indianapolis, IN).
In Vitro Covalent Binding Assay. The experimental procedure was based on that previously reported (Masubuchi et al., 2007). Briefly, the complete incubation of the 14C-labeled test compounds with human liver microsomes contained the following: 10 μM 14C-labeled test compound (substrate), 2 mg/ml human liver microsomes, 100 mM potassium phosphate buffer, pH 7.4, 25 mM glucose 6-phosphate, 2 units/ml glucose-6-phosphate dehydrogenase, and 10 mM MgCl2. This mixture was preincubated for 3 min at 37°C. The reaction was initiated by the addition of β-NADP+ to reach a final concentration of 2.5 mM. The final incubation volume was 0.5 ml. As the substrates were dissolved in acetonitrile, the final incubation mixture contained acetonitrile at the concentration of 1%. After 1-h incubation, the reaction was terminated by the addition of 0.5 ml of ice-cold acetonitrile. After vortexing and centrifugation, precipitated protein was washed with 80% (v/v) aqueous methanol containing 10% (w/v) trichloroacetic acid, diethyl ether/methanol (1:1, v/v), and 80% (v/v) aqueous methanol (twice for each solvent). The resulting precipitated protein was dissolved in 0.5 ml of 1.0 N NaOH, and the aliquots were taken for protein assay by DC protein assay kit (Bio-Rad Laboratories, Hercules, CA) and determination of radioactive content by liquid scintillation counting with the scintillation fluid mixture Hionic-Fluor (PerkinElmer, Boston, MA). The amount of test compounds covalently bound to the microsomal protein was determined, and covalent binding (pmol/min/mg protein) was calculated. As a background, radioactivity of the samples without incubation was also determined.
Results
To examine the cellular responses to electrophilic stress by a Keap1/Nrf2/antioxidant response element system in human hepatocytes, the induction profiles of typical Nrf2-responsive genes after exposure to BHA, a classical activator of Nrf2, were observed. The mRNA expression levels of HO-1, γ-GCS, UGT1A1, CES1, NQO1, GSTA1, and GSTP1 after 24-h exposure of human hepatocytes (lot FEP, KCT, and ZCA) to 100 μM BHA are shown in Table 1. The concentration and time of exposure were set according to the previous report (Keum et al., 2006). The mRNA expression levels of HO-1, CES, UGT1A1, and NQO1 were increased after 24-h exposure to BHA. The ratios of BHA-induced mRNA levels relative to the vehicle control were HO-1, 2.02 to 4.77; CES1, 1.67 to 4.49; UGT1A1, 1.77 to 10.31; and NQO1, 1.32 to 3.88. Lot KCT hepatocytes exhibited the highest induction levels of these enzymes among the three lots of hepatocytes used. The γ-GCS and GSTA1 mRNA expression levels were barely affected by the exposure to BHA, accounting for 0.85 to 1.22- and 1.16 to 1.46-fold changes relative to the vehicle control, respectively. GSTP1 mRNA expression levels were not determined because they were below the lower limit of quantification of real-time PCR.
We next investigated the induction profiles of Nrf2-related genes by problematic drugs that are known to be associated with severe IDTs, as well as safe drugs. The structures of the drugs tested are shown in Fig. 1. Human hepatocytes (lot KCT) were exposed for 24 h to increasing concentrations (30, 100, and 300 μM) of ticlopidine, clozapine, diclofenac, acetaminophen, and tienilic acid as the problematic drugs, and caffeine, levofloxacin, furosemide, and acetylsalicylic acid as the safe drugs, and BHA as a positive control, followed by the determination of the mRNA expression levels of HO-1, UGT1A1, CES1, and NQO1, which were all observed to be responsive to BHA (Fig. 2). Hepatocytes exposed to 300 μM BHA, ticlopidine, and clozapine exhibited increased LDH release (>2-fold compared with the vehicle control) in the culture medium, showing the cellular toxicity by these treatments. The mRNA levels of HO-1 increased after exposure to ticlopidine, clozapine, diclofenac, tienilic acid, and BHA in a concentration-dependent manner, and the induction ratios relative to the vehicle control were determined to be 2.56 for 100 μM ticlopidine, 3.90 for 100 μM clozapine, 2.71 for 300 μM diclofenac, 2.95 for 300 μM tienilic acid, and 4.77 for 100 μM BHA. In contrast, the increase in HO-1 mRNA levels after exposure to the safe drugs was determined to be less than 2-fold relative to the control, suggesting that HO-1 is not induced by safe drugs. These results showed a good contrast in HO-1 mRNA induction between the problematic and safe drugs. UGT1A1 was highly induced by exposure to the problematic drugs, showing a more than 10-fold increase in the mRNA level. Because the mRNA levels were also increased 2.90 to 6.94-fold after exposure to the safe drugs, the contrast in UGT1A1 induction between the problematic and safe drugs was not clear. In the cases of CES1 and NQO1, neither the problematic nor safe drugs caused significant increase in their mRNA expression levels. The BHA treatment as a positive control exhibited mRNA induction (>2-fold) in HO-1, UGT1A1, CES1, and NQO1. Next the observed -fold changes of induction of the drugs were compared with their intrinsic covalent binding yields in human liver microsomal systems reported previously (Masubuchi et al., 2007). Furthermore, we additionally conducted the covalent binding study of acetylsalicylic acid and caffeine using their 14C-labeled tracers to obtain the covalent binding data of the safe drugs by the same method. As a result, the safe drugs acetylsalicylic acid and caffeine exhibited trace levels of covalent binding (0.07 and 0.16 pmol/min/mg protein, respectively) in a human liver microsomal system. The -fold changes of induction of the drugs were plotted against their covalent binding yields as shown in Fig. 2E. HO-1 and UGT1A1 showed weak positive correlations between the induction and covalent binding, and their correlation coefficients were calculated to be 0.43 and 0.57, respectively. On the other hand, no correlation was observed in the case of CES1 and NQO1.
To investigate whether HO-1 induction in human hepatocytes by BHA and problematic drugs was affected by depletion of GSH, a major endogenous scavenger of electrophiles, the mRNA levels of HO-1 were determined after cotreatment of hepatocytes with BSO, a specific inhibitor of GSH synthesis. When human hepatocytes (lot KCT) were exposed to 100 μM BHA, 100 μM ticlopidine, 300 μM diclofenac, and 300 μM acetaminophen, they were cotreated with increasing concentrations (0, 10, and 50 μM) of BSO for 24 h, followed by the determination of HO-1 mRNA levels as shown in Fig. 3A. According to the results, cotreatment with BSO enhanced the HO-1 mRNA induction by BHA, ticlopidine, diclofenac, and acetaminophen in a concentration-dependent manner. Control incubations [(-) substrate] of hepatocytes with BSO and without substrate showed that mRNA levels of HO-1 were not changed solely by BSO at the concentrations tested. The effect of cotreatment with BSO and the safe drugs on the mRNA level of HO-1 was also examined as shown in Fig. 3B. The mRNA levels of HO-1 after exposure to 300 μM levofloxacin and caffeine with BSO were not different from those of the (-) substrate control, whereas those after exposure to acetylsalicylic acid were slightly enhanced under GSH depletion conditions.
Because BHA and the problematic drugs tested are known to be metabolized into electrophilic reactive species by cytochrome P450 (P450) enzymes, metabolism via P450s is presumed to contribute to the HO-1 induction by these drugs. Then, during 24-h exposure of human hepatocytes (lot KCT) with 100 μM BHA, ticlopidine, clozapine, and 300 μM diclofenac, they were cotreated with increasing concentrations (0, 300, and 1000 μM) of ABT, a nonspecific and irreversible inhibitor of P450s (Mico et al., 1988). In this experiment, the hepatocytes were also challenged with 50 μM BSO to detect the changes in HO-1 mRNA level with a wider dynamic range. As shown in Fig. 4A, ABT treatment significantly suppressed the HO-1 induction by BHA and problematic drugs in a concentration-dependent fashion. The mRNA levels in the control hepatocytes not exposed to any substrate were not changed by ABT, showing that ABT itself has no effect on HO-1 induction. In addition, we also examined the effect of ABT on HO-1 induction by tBHQ (Fig. 4B), which is a demethylated metabolite of BHA and is known to be an activator of Nrf2 (Keum et al., 2006). Hepatocytes exposed to 30 μM tBHQ for 8 h exhibited an approximately 2-fold increased mRNA level, which was not suppressed by the addition of ABT.
Discussion
Several in vitro studies using human cultured cell lines such as HepG2 or rodent hepatocytes, as well as in vivo studies in rodents, have shown that phase II enzymes are induced by electrophiles and that some drugs are known to produce electrophiles via a Keap1/Nrf2 system (Cantoni et al., 2003; Nioi et al., 2003; Shinkai et al., 2006), whereas there are only a few studies examining the Keap1/Nrf2-mediated enzyme induction profiles by the electrophilic stress in human hepatocytes (Keum et al., 2006). Induction studies of typical Nrf2-related genes by BHA have shown that the mRNA level induction of HO-1, UGT1A1, CES, and NQO1 is readily observable in human hepatocytes and that the mRNA levels induced were varied among the enzymes observed. This observed variation in the inductive response to electrophilic stress would be because there are different regulatory features in the gene expression process after Nrf2 activation. For example, the chromatin-remodeling molecule Brahma-related gene 1 has recently been identified as a specific coactivator of HO-1 gene expression that interacts with Nrf2 to selectively mediate HO-1 induction in response to oxidative and electrophilic stress (Zhang et al., 2006).
In the comparative induction study between the problematic and safe drugs shown in Fig. 2, we chose clozapine, ticlopidine, tienilic acid, diclofenac, and acetaminophen as the problematic drugs because these drugs are reported to be associated with IDT, including hepatotoxicity caused by the formation of the electrophilic reactive metabolites and covalent binding to cellular proteins. The proposed structures of the reactive metabolites generated from the problematic drugs are shown in Fig. 1. Clozapine causes fatal neutropenia, agranulocytosis, and hepatotoxicity and received a warning regarding its use (Alvir et al., 1993; Macfarlane et al., 1997). Clozapine forms reactive metabolites such as nitrenium ion by P450s, activated neutrophils, and bone marrow cells (Pirmohamed et al., 1995). Ticlopidine also received a warning and is restricted in its use because of severe hepatotoxicity and agranulocytosis (Skurnik et al., 2003). The thiophene moiety of ticlopidine is mainly oxidized by CYP2C19 with the formation of electrophilic reactive metabolites, including thiophene S-oxide and thiophene epoxide, which can form covalent binding to cellular proteins including CYP2C19 (Ha-Duong et al., 2001). Tienilic acid was withdrawn from the market in 1980 because of severe and fatal liver injury (Zimmerman et al., 1984). Tienilic acid has a thiophene moiety, as well as ticlopidine, and is metabolized mainly by CYP2C9 to form thiophene-derived electrophilic reactive metabolites (Lopez-Garcia et al., 1994). Diclofenac is associated with hepatotoxicity, which is considered to be because of the formation of electrophilic reactive metabolites, including acyl glucuronide by UGTs and benzoquinone-imines by CYP2C9 and CYP3A4, followed by their covalent binding to liver proteins (Boelsterli, 2003). As shown in Fig. 2, these problematic drugs exhibited inductive effects on HO-1, which contrasted with the safe drugs; therefore, the induction of HO-1 mRNA seems to be correlated with the occurrence of drug toxicity including IDT because of electrophilic reactive metabolites. Acetaminophen was categorized as a problematic drug but did not increase the mRNA level of HO-1 in the concentration range tested. Because acute and fatal liver injury of acetaminophen resulting from the formation of an electrophilic reactive metabolite, N-acetylbenzoquinone-imine, is caused by its overdose (e.g., 4000 mg/day) (Hinson et al., 1981), higher exposure concentrations of acetaminophen may be necessary for the induction of HO-1. Although acetaminophen is the most thoroughly studied problematic drug and there is some body of evidence of toxicity as a result of bioactivation particularly in mice, it is known that human hepatocytes are less responsive to acetaminophen than mice (Tee et al., 1987; Cheung et al., 2005). This species difference in hepatotoxic effects of acetaminophen is considered to be caused by the difference in the rate of CYP2E1-mediated formation of reactive metabolites. Thus, the no induction of HO-1 by acetaminophen observed could be explained by the low sensitivity of human hepatocytes to acetaminophen. The parallel gene induction studies with hepatocytes from animal species would be helpful to understand the species difference in metabolic bioactivation and toxicity.
UGT1A1 was induced by exposure to the safe drugs and the problematic drugs as shown in Fig. 2. Previous studies have shown that the regulation of UGT expression is targeted by a number of xenobiotic and steroid receptors such as PXR, CAR, and PPAR in response to xenobiotics, carcinogens, stress signals, and hormones other than the Keap1/Nrf2 signaling pathway (Soars et al., 2004; Yueh and Tukey, 2007). Therefore, the observed inductive effects of the safe drugs on UGT1A1 are thought to be caused by the induction through these receptors, and UGT1A1 induction would not be a specific marker for electrophilic stress responses.
The -fold changes of induction for HO-1 were shown to be positively and weakly correlated with the covalent binding yields. Nrf2 activation is known to be regulated by the covalent modification pattern of Keap1, which is dependent on the structures of the electrophilic reactive species (Hong et al., 2005). Therefore, the Nrf2-mediated gene induction would be dependent not only on the quantity of covalent binding but also on the structures of reactive metabolites and the pattern of covalent binding. This may be one reason for no strong correlation between covalent binding and Nrf2-mediated gene induction. For further consideration, however, it would be necessary to determine the covalent binding yields in the induction experiments using human hepatocytes. In the case of UGT1A1, the drugs with trace levels of covalent binding yield showed inductive effects, supporting the idea that the observed induction of UGT1A1 was, at least partially, mediated by induction pathways other than the Keap1/Nrf2 pathway.
Because electrophilic reactive species can be scavenged by GSH, a major intracellular nucleophile, the GSH depletion experiment using BSO represents that the formation of electrophilic reactive species should be involved in the HO-1 induction by BHA and the problematic drugs. Furthermore, GSH is also a major antioxidant; therefore, it was suggested that the hepatocellular redox status could affect the HO-1 induction by electrophilic stress. Although acetaminophen did not affect the mRNA level of HO-1 up to 300 μM under normal conditions as shown in Fig. 2A, it induced HO-1 under GSH-depleted conditions (Fig. 3A), indicating that acetaminophen also has a potential to induce HO-1 via reactive metabolite formation and which would be masked by GSH under normal conditions. The safe drugs levofloxacin and caffeine had no inductive effect on HO-1 even under the GSH-depleted conditions, showing that GSH depletion can sharpen the contrast in HO-1 induction between the problematic and safe drugs. Acetylsalicylic acid is believed to be relatively safe at therapeutic doses and thus is categorized as a safe drug in this study. However, it has been reported that its overdose results in hepatotoxicity, probably because of the formation of electrophilic acyl glucuronide; thus, it can be regarded as a potentially problematic drug (Dickinson et al., 1994; Bjorkman, 1998). The slight induction by acetylsalicylic acid observed under GSH-depleted conditions might represent its potential risk of hepatotoxicity.
Moreover, P450 inhibition experiments using ABT show that P450 metabolism is an essential pathway for the HO-1 induction by BHA and the problematic drugs. As mentioned above, ticlopidine, clozapine, and diclofenac are transformed to electrophilic reactive metabolites by P450 enzymes. BHA is also known to be typically metabolized by P450s to a demethylated metabolite, tBHQ, which is in turn converted to the electrophilic reactive species tert-butylquinone through auto-oxidation, as shown in Fig. 1 (Cummings et al., 1985). Thus, the observed HO-1 induction is considered to be triggered by the electrophilic reactive metabolites produced through P450 metabolism. The lack of an effect of ABT on HO-1 induction by tBHQ suggests that tBHQ is capable of inducing HO-1 without P450-mediated metabolic bioactivation, which seems to be in reasonable agreement with this conclusion because tBHQ can be auto-oxidized to tert-butylquinone.
HO-1 catalyzes degradation of heme, resulting in production of biliverdin and carbon monoxide. Biliverdin can suppress oxidative stress by scavenging reactive oxygen species (Baranano et al., 2002). Carbon monoxide has strong anti-inflammatory effects and protective effects against immune-mediated liver injury (Otterbein et al., 2000). Because HO-1 has some defensive effects against cellular damages caused by electrophilic stress, HO-1 induction by the problematic drugs would be reasonable from a functional point of view. Future studies to investigate changes in level of HO-1 activity and yields of HO-1 products after mRNA level induction will provide further support for the concept in this study.
In summary, we have investigated the induction profiles of typical Nrf2-related genes after the exposure of human hepatocytes to BHA and a series of the marketed drugs and have revealed that the mRNA levels of HO-1 can be increased specifically by problematic drugs associated with severe hepatotoxicity including IDT. Moreover, GSH-depletion and P450-inhibition experiments have shown that the observed HO-1 induction was triggered by the electrophilic reactive metabolites produced from the problematic drugs through P450-mediated metabolic bioactivation. Taken together, our present study suggests that HO-1 induction in human hepatocytes would be a good marker for the cellular stress response to electrophilic reactive metabolites and metabolism-based drug toxicity and IDT.
Footnotes
-
Article, publication date, and citation information can be found at http://dmd.aspetjournals.org.
-
doi:10.1124/dmd.107.018002.
-
ABBREVIATIONS: IDT, idiosyncratic drug toxicity; GSH, glutathione; LC/MS/MS, liquid chromatography/tandem mass spectrometry; Nrf2, nuclear factor E2-related factor 2; Keap1, Kelch-like ECH-associated protein 1; HO, heme oxygenase (EC: 1.14.99.3); GCS, glutamylcysteine synthetase (EC: 6.3.2.2); GST, glutathione-S-transferase (EC: 2.5.1.18); GSTA1, glutathione-S-transferase A1; GSTP1, glutathione-S-transferase P1; NQO, NAD(P)H:quinone oxidoreductase (EC: 1.6.5.2); UGT, UDP-glucuronosyltransferase (EC: 2.4.1.17); CES, carboxylesterase; BHA, butylated hydroxyanisole; tBHQ, tert-butyl hydroquinone; BSO, l-buthionine-(S,R)-sulfoximine; ABT, aminobenzotriazole; DMSO, dimethyl sulfoxide; PCR, polymerase chain reaction; GAPDH, glyceraldehyde-3-phosphate dehydrogenase (EC: 1.2.1.12); LDH, lactate dehydrogenase (EC: 1.1.1.27); P450, cytochrome P450.
- Received August 1, 2007.
- Accepted January 25, 2008.
- The American Society for Pharmacology and Experimental Therapeutics