Abstract
Serum of rabbits with a turpentine-induced acute inflammatory reaction (RSINFLA) and serum of humans with a viral infection (HSINF) were previously shown to diminish hepatic cytochrome P450 (P450) content and activity. To document the role of reactive oxygen intermediates in the serum-mediated decrease in P450 content and activity, hepatocytes of rabbits with an acute inflammatory reaction (HINFLA) were incubated with RSINFLAand HSINF for 4 h, and total P450 content (spectrally measurable P450), P450 activity (assessed by estimating the formation of theophylline metabolites), and amount of CYP1A1, CYP1A2, and CYP3A6 proteins were measured. RSINFLA or HSINFdecreased P450 content and activity without affecting the amount of CYP1A1 and -1A2 HINFLA. Exposure of HCONT or HINFLA to hydrogen peroxide (0.01–1.0 mM) and sodium nitroprusside (0.01–1.0 mM) produced a dose-dependent decrease in P450 content and in the formation of theophylline metabolites without modifying the amount of CYP1A1 and CYP1A2, whereas lipid peroxidation increased. Incubation of l-NAME (0.05–1.0 mM), dimethylthiourea (6.25–50 mM), or N-acetylcysteine (0.01–1.0 mM) with HINFLA partially prevented the decrease in P450 content and activity and the increased lipid peroxidation induced by RSINFLA and HSINF. On the other hand, 3-amino-1,2,4-triazole (10–100 mM) or diethyldithiocarbamate (1.0–10 mM) potentiated RSINFLA- and HSINF-mediated decreases in P450 content and activity and the increase in lipid peroxidation, without affecting the amount of CYP1A1 or -1A2;dl-buthionine-(S,R)-sulfoximine (2.5–25 mM) potentiated only the inhibition of 1,3-dimethyluric acid formation. It is concluded that reactive oxygen intermediates are implicated in the decrease of HINFLA P450 content and activity induced by 4 h of exposure to RSINFLA or HSINF.
Endotoxins, sepsis, and acute inflammatory reactions down-regulate hepatic cytochrome P450 (P450)1 apoproteins through transcriptional and post-transcriptional mechanisms (Morgan, 1989). However, the cascade of events leading to the down-regulation of P450 remains poorly defined. Reactive oxygen intermediates (ROI), induced by the inflammatory reaction, have been implicated in the signaling pathway leading to P450 down-regulation in vivo (Galal and du Souich, 1999) and in vitro (Chen et al., 1995). Hydrogen peroxide (H2O2) is able to diminish the expression of selected isozymes of hepatic P450 (Barker et al., 1994), and H2O2 mediates the depression of P450 by interferon-γ (IFN-γ) (Moochhala and Renton, 1991). Moreover, nitric oxide synthase (NOS) inhibitors prevent the decrease in P450 activity mediated by cytokines or lipopolysaccharides in vivo (Khatsenko et al., 1993) and in vitro (Carlson and Billings, 1996). Furthermore, the addition of nitric oxide (NO⋅) generators to microsomes (Khatsenko et al., 1993) or cultured hepatocytes (Carlson and Billings, 1996) down-regulates hepatic P450.
Potentially conflicting with these reports, it has been shown that xanthine oxidase is not involved in the P450 down-regulation induced by IFN-γ or poly(IC) (Cribb and Renton, 1993). On the other hand, NOS inhibitors do not prevent cytokine-induced down-regulation of CYP2C11 protein and mRNA (Sewer and Morgan, 1997) or the reduction in P450 content (Hodgson and Renton, 1995). In addition, in vivo the administration of NO⋅ generators does not appear to down-regulate P450 (Hodgson and Renton, 1995).
Forty-eight hours after the production of an inflammatory reaction by the s.c. injection of turpentine, hepatic CYP1A1, -1A2, and -3A6 are down-regulated (Kurdi et al., 1999). The serum from these rabbits with a turpentine-induced inflammatory reaction (RSINFLA) and the serum from humans with an upper respiratory viral infection (HSINF) when incubated for 4 h with hepatocytes from rabbits with a turpentine-induced inflammatory reaction (HINFLA) reduce P450 content and activity. The RSINFLA- and HSINF-induced decrease in P450 content is directly associated with an increase in lipid peroxidation, suggesting that ROI are implicated in the decrease in P450 content (El-Kadi et al., 1997).
The present study aimed to investigate whether ROI are implicated in the RSINFLA- and HSINF-mediated decrease in hepatic P450 content and activity. For this purpose, we have 1) determined the effect of H2O2 and NO⋅ on the amount and activity of P450, 2) assessed whether the addition ofl-NAME and antioxidants to hepatocytes prevent the decrease in P450 content and activity mediated by RSINFLAand HSINF, and 3) examined whether the addition of inhibitors of antioxidant enzymes to hepatocytes potentiate the decrease in P450 content and activity mediated by RSINFLA and HSINF. P450 activity was assessed by measuring the metabolism of theophylline. In rabbits, theophylline is transformed to 1,3-dimethyluric acid (1,3-DMU), 3-methylxanthine (3-MX), and 1-methyluric acid (1-MU) essentially by CYP1A1, -1A2, and more marginally by CYP3A6 (Kurdi et al., 1999).
Materials and Methods
Experimental Protocol.
Male New Zealand rabbits (1.8–2.2 kg) were obtained from the Ferme Cunicole (St. Valérien, Québec). Rabbits were housed in separate cages and fed water and chow ad libitum for at least 7 days before being used. The inflammatory reaction was induced locally by injecting turpentine (5 ml) s.c. at two distinct sites on the back of the rabbits (Parent et al., 1992). Control rabbits received 5 ml of sterile NaCl 0.9% s.c. at two sites of the back. The severity of the inflammatory reaction was assessed by measuring the concentration of seromucoids. Seromucoids are acute phase proteins, mostly α1-glycoprotein acid, and in minor amounts α1-antitrypsin and haptoglobulin. They were assayed by means of a colorimetric reaction with copper and folin phenol reagents yielding a maximal absorbance at 540 nm (Parent et al., 1992). All the experiments were conducted according to the Canadian Council on Animal Care guidelines for use of laboratory animals.
Blood (10 ml) was withdrawn from the rabbits 48 h after the s.c. injection of turpentine or saline in a sterile Vacutainer Brand SST (Becton Dickinson, Mississauga, Ontario, Canada), left at room temperature for at least 2 h, and thereafter centrifuged at 2500 r.p.m. for 5 min to obtain the serum. Blood (10 ml) was withdrawn from four subjects (two females and two males, 23–45 years old) with an inflammatory reaction secondary to a viral infection, at the apex of clinical symptomatology, i.e., usually 24 h after the appearance of overt manifestations of upper respiratory tract viral infection, such as rhinorrhea, sneezing, nasal congestion, sore throat, cough, and systemic signs of malaise, including fever, in absence of purulent secretions. Blood was processed as described for the rabbits. Blood from the same subjects was withdrawn at least 6 weeks later to obtain control sera.
Rabbit hepatocytes were isolated 48 h after the induction of the inflammatory reaction according to the two-step liver perfusion method of Seglen (1976), with minor modifications (El-Kadi et al., 1997). Hepatocytes (4 × 106/ml) were placed into 12-well plastic culture plates (Falcon, Becton Dickinson Labware, Lincoln Park, NJ) coated with Type I rat-tail collagen (Sigma Chemical Company, St. Louis, MO); cells were suspended in William's medium E (Sigma) supplemented with 10% calf serum and insulin 1 μM (Boehringer Mannheim GmbH, Germany). The plastic culture plates were incubated at 37°C in a humidifier with 95% O2and 5% CO2. Viability was assessed before and after the incubation period by the trypan blue (0.2%) exclusion method, and in both instances the viability was over 90%.
In all experiments, the effect of the experimental conditions on hepatic P450 was characterized by culturing hepatocytes with serum for 4 h, and assessing the catalytic activity of the P450 using theophylline as a probe. The experiments were initiated by adding 100 μl of serum and 50 μl of theophylline (Sigma), dissolved in serum-free William's medium E, to 1 ml of incubation media containing freshly isolated 4 × 106 hepatocytes per well. The final concentration of theophylline was 176 μM. At time zero, 350 μl of the supernatant was collected from each well (control sample), and following 4 h of incubation, the remaining medium was collected and frozen at −20°C until the theophylline metabolites 3-MX, 1-MU, and 1,3-DMU (Sigma) were assayed by high performance liquid chromatography (du Souich et al., 1989). In addition, the effect of the experimental conditions was assessed by measuring P450 content (spectrally measurable P450) in the hepatocytes (Omura and Sato, 1964). The amount of proteins in the hepatocytes was determined by the method of Lowry et al. (1951). The lipid peroxidation induced by the turpentine-induced inflammation and the experimental conditions was assessed measuring the amount of malondialdehyde formed in the hepatocytes by the thiobarbituric acid reaction (Ohkawa et al., 1979).
To ensure that coincubation of HINFLA with RSINFLA or HSINF for 4 h with theophylline did not distort the results, two control experiments were conducted. First, the rate of formation of theophylline metabolites was assessed in microsomes from HCONT or HINFLApreincubated with RSINFLA or HSINF for 4 h. To 1 mg of microsomal proteins in 0.1 M phosphate buffer (pH 7.4) was added NADP (0.5 mM), glucose 6-phosphate (5 mM), MgCl2 (5 mM), and glucose-6-phosphate dehydrogenase (2 international units), and theophylline (176 μM) to a final volume of 1 ml. Following 4 h of incubation at 37°C, the incubation media was collected and frozen at −20°C until theophylline, 3-MX, 1-MU, and 1,3-DMU were assayed by HPLC (n = 3). Second, HCONT and HINFLA were incubated with RSINFLA or HSINF for 4 h, and the William's medium E was changed with fresh medium containing only theophylline (176 μM) and incubated for an additional 4 h (n = 4). Under both experimental conditions, the formation of 3-MX, 1-MU, and 1,3-DMU decreased by approximately 35, 30, and 45%, respectively. Simultaneous incubation of HINFLA, RSINFLA, or HSINF and theophylline resulted in decreases in the formation of 3-MX, 1-MU, and 1,3-DMU of the same order.
To demonstrate whether the addition of ROI to cultured hepatocytes inactivates or reduces the amount of selected apoproteins of the P450, HCONT (n = 5) and HINFLA (n = 5) were incubated with different concentrations of H2O2 (0.01, 0.05, 0.25, and 1.0 mM) and of sodium nitroprusside (0.01, 0.05, 0.25, and 1.0 mM). The content of P450, formation of theophylline metabolites, and lipid peroxidation were assessed 4 h later. CYP1A1 and -1A2 proteins were measured by Western blot analysis in two samples of HINFLA incubated with 1.0 mM H2O2 or sodium nitroprusside. H2O2 was selected because of its ability to diffuse into the hepatocytes, act as a source of hydroxyl radicals (⋅OH), produce cellular damage, and suppress the transcription of P450 apoproteins (Karuzina and Archakov, 1994). Sodium nitroprusside was selected because it can generate NO⋅, which is able to down-regulate different apoproteins of the P450 (Stadler et al., 1994).
The role of ROI in the decrease in P450 content and activity induced by an inflammatory reaction was further investigated by preincubating HINFLA with an inhibitor of NOS,Nω-nitro-l-arginine methyl ester (l-NAME; 0.05, 0.25, and 1.0 mM), dimethylthiourea (6.25, 12.5, and 50 mM), andN-acetylcysteine (0.05, 0.25, and 1.0 mM) (all from Sigma) for 30 min. Thereafter, 100 μl of RSINFLA(n = 5) and HSINF(n = 4) was added to HINFLA, and P450 content, theophylline metabolites formation, and lipid peroxidation were assessed 4 h later. CYP1A1 and -1A2 proteins were measured by Western blot analysis in samples of hepatocytes incubated with l-NAME (1.0 mM), dimethylthiourea (50 mM), or N-acetylcysteine (1.0 mM), and RSINFLA (n = 3) or HSINF (n = 3). Dimethylthiourea was used, because it is an effective ⋅OH scavenger able to penetrate into the cells (Heo et al., 1995), andN-acetylcysteine, a precursor of glutathione, is able to react with ROI (Moldeus and Cotgreave, 1994).
To investigate whether the inhibition of antioxidant enzymes potentiate the decrease in P450 content and activity induced by the inflammatory reaction, RSINFLA (100 μl, n = 6) and HSINF (100 μl, n = 4) were preincubated with HINFLA for 30 min with 3-amino-1,2,4-triazole (10, 25, and 100 mM) (Sigma), an inhibitor of catalase,dl-buthionine-(S,R)-sulfoximine (2.5, 6.25, and 25 mM) (Sigma), an inhibitor of glutathione peroxidase, and diethyldithiocarbamate (1.0, 2.5, and 10 mM) (Sigma), an inhibitor of superoxide dismutase. P450 content, formation of theophylline metabolites, and lipid peroxidation were assessed 4 h later. CYP1A1 and -1A2 proteins were measured by Western blot analysis in samples of HINFLA incubated with 3-amino-1,2,4-triazole (100 mM),dl-buthionine-(S,R)-sulfoximine (25 mM), or diethyldithiocarbamate (10 mM), and RSINFLA (n = 2) or HSINF (n = 4).
Western Blot Analysis.
Proteins were separated by polyacrylamide gel electrophoresis (7.5% polyacrylamide) under nonreducing conditions (Smith, 1994). Separated proteins were electrophoretically transferred to a nitrocellulose membrane using a semidry transfer process (Bio-Rad, Hercules, CA). CYP1A1 and -1A2 proteins were detected with a polyclonal anti-rabbit CYP1A1 (Oxford Biochemical Research, Oxford, MI) and visualized with an alkaline phosphatase-conjugated secondary goat antibody using nitro blue tetrazolium as the substrate (Kruger, 1994). CYP3A6 protein was detected with a monoclonal anti-rat CYP3A1 (Oxford Biochemical Research), using a secondary antibody conjugated to a chemiluminescence reagent (horseradish peroxidase enzyme), and visualized by autoradiography (Thorpe et al., 1985). The intensities of the bands were measured with the software UN-SAN-IT gel version 5.1 (Silk Scientific, Inc., Orem, UT). .
Statistical Analysis.
All results are reported as means ± S.E. The comparison of the results from the various experimental groups and their corresponding controls was carried out using a one-way ANOVA followed by Newman-Keul's post hoc test. The differences were considered significant when P < .05.
Results
Effect of Rabbit and Human Serum on P450 Content and Activity.
In rabbits with a turpentine-induced inflammatory reaction, seromucoid concentrations were 72.7 ± 2.4 mg/dl compared with 20.8 ± 2.6 mg/dl in control rabbits (P < .05). The content of P450 in HINFLA was lower than in HCONT, i.e., 0.22 ± 0.01 versus 0.34 ± 0.04 nmol/mg of protein (P < .05). By comparison to HCONT, the formation of 3-MX, 1-MU, and 1,3-DMU was reduced by 30, 50, and 28%, respectively, in HINFLA (Table 1). CYP1A1 and -1A2 levels were 43% and 48% lower, respectively, in HINFLA than in HCONT. The amount of CYP3A6 protein was almost undetectable in HINFLA (Fig. 1).
Incubation of RSINFLA with HINFLA for 4 h further decreased P450 content to 0.11 ± 0.01 nmol/mg of protein. RSINFLA reduced the formation of 3-MX, 1-MU, and 1,3-DMU by 38, 32, and 31%, respectively (Table2). On the other hand, after an incubation period of 4 h with RSINFLA, the expression of CYP1A1 and -1A2 proteins was not modified (Fig. 1). The effect of RSINFLA on CYP3A6 could not be assessed, because in HINFLA this apoprotein was barely detectable. Incubation of HSINF with HINFLA for 4 h reduced the formation of theophylline metabolites by 40% (Table3) without affecting the amounts of CYP1A1 and -1A2 proteins, i.e., the relative density of CYP1A1 and -1A2 in HINFLA was 20,570 ± 540 and 23,375 ± 540, respectively, and after incubation with HSINF these values were 19,787 ± 678 and 25,359 ± 940, respectively.
Effect of H2O2 and Sodium Nitroprusside on P450 Content and Activity.
Viability of hepatocytes was not affected by concentrations of ≤1 mM H2O2 or sodium nitroprusside. Exposure of HCONT to various concentrations of H2O2 or sodium nitroprusside produced a dose-dependent decrease in P450 content (Fig. 2). As a consequence, H2O2 and sodium nitroprusside diminished dose dependently the formation of theophylline metabolites, 3-MX, 1-MU, and 1,3-DMU (Table 1). Lipid peroxidation, as expressed in terms of the amount of malondialdehyde generated, was increased with the addition of H2O2 and sodium nitroprusside (Fig. 2). Total P450 content, biotransformation of theophylline, and hepatic lipid peroxidation were not significantly influenced by the lowest concentration of H2O2 tested (0.01 mM).
Changes of similar magnitude were observed when HINFLA were exposed to various concentrations of H2O2 or sodium nitroprusside, i.e., total P450 content was decreased by 15, 25, 30, and 45% (P < .05) or by 30, 40, 40, and 55% (P < .05) by the addition of 0.01, 0.05, 0.25, and 1 mM H2O2 or sodium nitroprusside, respectively. On the other hand, lipid peroxidation was increased by 10, 22, 39, and 42% (P < .05) or by 41, 100, 120, and 173% (P < .05) by the addition of 0.01, 0.05, 0.25, and 1 mM H2O2or sodium nitroprusside, respectively. The amounts of CYP1A1 and -1A2 proteins were not affected by the incubation of HCONT or HINFLA with H2O2 or sodium nitroprusside (Fig. 3).
Effect of l-NAME and Antioxidants on Rabbit and Human Serum-Mediated Decrease in P450 Content and Activity.
In preliminary studies (n = 5), HINFLA was incubated withl-NAME, dimethylthiourea, orN-acetylcysteine and with serum from control rabbits or healthy humans to assess whether these antioxidants affect P450 content, theophylline metabolism, or lipid peroxidation. The results indicated that at the concentrations equal or lower than 1, 50, and 1 mM, l-NAME, dimethylthiourea, orN-acetylcysteine did not affect the parameters assessed.
l-NAME, dimethylthiourea, or N-acetylcysteine partially prevented the decrease in P450 content and the inhibition of theophylline metabolism, as well as the increase in lipid peroxidation mediated by RSINFLA and HSINF (Figs. 4 and5, and Tables 2 and 3). At the highest concentration tested, all three compounds elicit a similar protection on the formation of theophylline metabolites, i.e., about 60%. The amount of CYP1A1 and -1A2 proteins in HINFLA were not modified by the presence of l-NAME, dimethylthiourea, or N-acetylcysteine. The relative densities of CYP1A1 in HINFLA and HINFLA in the presence of HSINF, l-NAME, dimethylthiourea, or N-acetylcysteine were 13,860, 13,230, 10,080, 9,200, and 8,820, respectively, and for CYP1A2 these values were 12,600, 13,860, 11,970, 12,560, and 9,454, respectively.
Effect of Inhibitors of Antioxidant Enzymes on the Decrease of P450 Content and Activity Induced by Rabbit and Human Serum.
In preliminary studies (n = 5), HINFLA was incubated with the inhibitors of the antioxidant enzymes and with serum from control rabbits or healthy humans to assess whether at the concentrations used 3-amino-1,2,4-triazole,dl-buthionine-(S,R)-sulfoximine, or diethyldithiocarbamate had a direct effect on P450 content, theophylline metabolism, or lipid peroxidation. The results indicated that, at the concentrations used, none of the inhibitors of the antioxidant enzymes affected the parameters assessed.
3-Amino-1,2,4-triazole and diethyldithiocarbamate significantly potentiated the reduction in P450 content, the inhibition of theophylline metabolism, as well as the increase in lipid peroxidation induced by RSINFLA or HSINFin a dose-dependent manner (Figs. 6 and7).dl-Buthionine-(S,R)-sulfoximine (25 mM) potentiated the inhibition of 1,3-DMU formation by both sera (Tables 4 and5) and potentiated the increase in lipid peroxidation mediated by HSINF. The addition of 3-amino-1,2,4-triazole,dl-buthionine-(S,R)-sulfoximine, or diethyldithiocarbamate to hepatocytes cultured with RSINFLA did not change the amounts of CYP1A1 or -1A2 proteins. In the presence of HSINF, 3-amino-1,2,4-triazole anddl-buthionine-(S,R)-sulfoximine did not change the amounts of CYP1A1 or -1A2 proteins, but diethyldithiocarbamate decreased the amounts of CYP1A1 and -1A2 proteins to half the control values. The relative densities of CYP1A1 in HINFLA alone, and HINFLAincubated with HSINF, 3-amino-1,2,4-triazole, anddl-buthionine-(S,R)-sulfoximine and diethyldithiocarbamate were 13,230, 11,760, 10,290, 10,290, and 4,410, respectively, and for CYP1A2, these values were 18,375, 16,170, 13,965, 13,965, and 4,410, respectively.
Discussion
The present results confirm that, in vivo in the rabbit, a turpentine-induced inflammatory reaction down-regulates hepatic CYP1A1, -1A2, and -3A6 proteins and show that the incubation of RSINFLA or HSINF with HINFLA for 4 h reduces P450 content and the formation of theophylline metabolites without decreasing the amounts of CYP1A1 and -1A2 proteins. To explain this apparent contradiction, several elements should be taken into account. First, the spectrophotometric measure of P450 is based on the fact that CO binds to reduced iron (Fe2+) of the heme moiety at the binding site of O2. Therefore, binding of CO to Fe2+, and spectrophotometric measure of P450, depends upon the availability of this binding site. On the other hand, catalytic activity of P450 isoforms also depends upon this binding site to form the P450-dioxygen complex necessary to transfer an oxygen atom to the substrate. Immunoquantitation by using an antibody to a specific sequence of amino acids will measure the amount of isoforms; however, it does not indicate whether the isoform measured is active. Because the formation of theophylline metabolites is essentially dependent upon CYP1A1 and -1A2 activity (Kurdi et al., 1999), we may postulate that the decrease in the rate of theophylline metabolism is due to a decrease in CYP1A1 and/or -1A2 activity, because their amounts are not modified. Supporting such hypothesis, spectrally measurable P450 in HINFLA was reduced by RSINFLA or HSINF.
Sodium nitroprusside, a source of NO⋅, reduces P450 content as well as the formation of theophylline metabolites without affecting the amounts of CYP1A1 and -1A2 proteins, while lipid peroxidation is increased. Further supporting a role for NO⋅ in the decrease of spectrally measurable P450 content and activity is the fact that the addition of l-NAME to the incubation media partially prevents RSINFLA- or HSINF-induced diminution in P450 content and activity. NO⋅ binds to the same binding site on the reduced iron of the heme as do CO and O2, and as a consequence, NO⋅ decreases spectrally measurable P450, as well as its activity (Khatsenko et al., 1993; Wink et al., 1993). In agreement with the present results, it has been shown that NO⋅is able to decrease the activity CYP1A1, -1A2, -2B1, -2B2, -2C11, and -3A2 (Khatsenko et al., 1993; Wink et al., 1993; Stadler et al., 1994). Moreover, NO⋅ can inactivate CYP2E1, and the inactivated form retains the epitope for its recognition when assayed by Western blot analysis (Gergel et al., 1997). Therefore, in the absence of changes in the amount of CYP1A1 and -1A2 proteins, and in agreement with the conclusions reached by Khatsenko et al. (1993), we propose that, under the present experimental conditions, RSINFLA and HSINF induce the release of NO⋅, which operates inactivating CYP1A1 and -1A2.
N-Acetylcysteine attenuates the decrease in P450 content and activity induced by the addition of RSINFLA or HSINF, and 3-amino-1,2,4-triazole, an inhibitor of catalase (Mian and Martin, 1995) potentiates the effect of RSINFLA or HSINF on P450 content and activity, without affecting CYP1A1 and -1A2. Exposure of HCONT or HINFLA to various concentrations of H2O2 for 4 h reduces P450 content as well as the formation of theophylline metabolites without affecting the amounts of CYP1A1 and -1A2 proteins, strongly supporting that H2O2 is contributing to RSINFLA- or HSINF-induced decrease in P450 content and activity. The possibility that H2O2 might directly decrease the activity of CYP1A1 was already proposed some years ago (Flowers and Miles, 1991). The exact mechanism by which H2O2 would decrease the activity of selected P450 isoforms remains unclear, but it has been shown that H2O2 formed in the hemoprotein active center can interact with the enzyme-associated Fe2+ leading to heme destruction and enzyme inactivation (Karuzina and Archakov, 1994; Archakov et al., 1998). On the other hand, H2O2 might simply be a second messenger in the signaling pathway leading to the inactivation and/or the down-regulation of selected P450 isoforms (Bae et al., 1997; Lowe et al., 1998).
The fact thatdl-buthionine-(S,R)-sulfoximine, an inhibitor of glutathione peroxidase (Masaki et al., 1998), does not potentiate the effect of RSINFLA or HSINF suggests that the effect ofN-acetylcysteine, a source of reduced glutathione the substrate of glutathione peroxidase (Singh et al., 1998), is due to its ability to scavenge H2O2(Vanderbist et al., 1996). On the other hand, diethyldithiocarbamate, an inhibitor of superoxide dismutase (Martin et al., 1994), potentiates RSINFLA- or HSINF-mediated reduction in P450 content and activity, without affecting the amounts of CYP1A1 and -1A2 proteins. Superoxide dismutase transforms superoxide (O⨪2) to H2O2, which is catalyzed to water by catalase and glutathione peroxidase (Southorn and Powis, 1988). Because it has been shown that O⨪2is able to reduce the activity of CYP1A1 (Flowers and Miles, 1991), we postulate that O⨪2 has contributed directly in the decrease in P450 activity, or indirectly as a source of H2O2. Despite the fact that dimethylthiourea is a rather specific scavenger of ⋅OH, a role for ⋅OH is less clear, because dimethylthiourea can also scavenge peroxynitrite (Hara et al., 1998). Nevertheless, the present results suggest that several species of ROI are implicated in the RSINFLA- or HSINF-mediated reduction in P450 content and activity. This multiplicity of ROI may explain why l-NAME and other antioxidants confer only a partial protection to the serum-mediated P450 decrease in activity.
Alternatively, ROI may have reduced the activity of selected isoforms of the P450 indirectly by inducing the phosphorylation of the isoforms. Effectively, activation of kinases can phosphorylate P450 resulting in its inactivation (Rhee, 1999). In vitro, several isoforms of rabbit and rat P450 can be phosphorylated on a serine residue at positions 127–129 by cAMP-dependent kinase A and protein kinase C (Moritz et al., 1998). ROI, specifically O⨪2 and H2O2, appear to play an important role as messengers as well as stimulators of protein tyrosine kinase (Bae et al., 1997; Lowe et al., 1998), protein kinase C (Boyer et al., 1995), protein kinase A (Suzuki et al., 1997), and mitogen-activated protein kinases (Goldstone and Hunt, 1997). The mechanism by which phosphorylation inactivates P450 is unclear; it has been shown that phosphorylation increases the uncoupling from NADPH-dependent hydroxylation of xenobiotics by P450, P450-reductase, and cytochrome b5 (Mkrtchian and Andersson, 1990). Indeed, further studies are required to understand the signal transduction pathways activated by RSSINF and HSINF leading to P450 inactivation.
In this laboratory we have shown that the decrease in P450 activity elicited by RSINFLA and HSINF is mediated by interleukin-6 (IL-6), and by IFN-γ, IL-1β, and IL-6, respectively (Bleau et al., 2000). Keeping in mind the diversity of serum mediators, multiple sources of ROI are possible: membrane and cytosolic NADPH oxidase, xanthine oxidase, the mitochondrial respiratory chain, and the P450 system. Concerning the NADPH oxidase system, IL-6, IL-1β and IFN-γ can activate phospholipase A2 (McPhail et al., 1993) to produce arachidonic acid, which will be transformed into leukotrienes by 5-lipoxygenase and generate ROI (Bonizzi et al., 1999). Alternatively, both arachidonic acid and leukotrienes may activate membrane or cytosolic NADPH oxidase, an important source of ROI (Rhee 1999). Cytokines such as IFN-γ, IL-1β, and IL-6 are capable to activate xanthine oxidase to generate O⨪2 (Ghezzi et al., 1985). Supporting that xanthine oxidase is a source of ROI, turpentine-induced inflammatory reaction increases hepatic xanthine oxidase activity (Proulx and du Souich, 1995). In response to multiple stimuli, such as sepsis, cytokines, or ceramide, Complex I to III of the mitochondrial respiratory chain will generate ROI, which may be used as signal transduction messengers to activate transcription factors, the expression of manganese superoxide dismutase, nitric oxide synthase, and/or lead to apoptosis (Degli Esposti and McLennan, 1998). Finally, in the hepatocyte, several P450 isoforms are able to generate ROI (Serino et al., 1993), specifically, CYP2E1 may generate H2O2(Gergel et al., 1997); CYP3A4, -1A1, -1A2, and -2B6 produce O⨪2(Puntarulo and Cederbaum, 1998); and CYP3A yields NO⋅ (Kuo et al., 1995).
In conclusion, 48 h after a turpentine-induced acute inflammatory reaction there is a down-regulation of CYP1A1, -1A2, and -3A6 proteins. When incubated for 4 h with HINFLA, the serum of rabbits with a turpentine-induced inflammatory reaction and the serum of humans with a viral infection decrease spectrally measurable P450 and the formation of theophylline metabolites without affecting the amounts of CYP1A1 and -1A2 proteins. This suggests that a short incubation of HINFLA with serum decreases the activity of CYP1A1 and -1A2. The fact that l-NAME and antioxidants are unable to completely abrogate the serum-induced decrease in activity of P450 indicates that several species of ROI contribute to the decrease in activity of hepatic P450, possibly NO⋅, H2O2, and O⨪2.
Acknowledgments
The authors are grateful to Mrs. Lucie Héroux for her skilful technical assistance. A.O.S El-Kadi is recipient of a scholarship from Novartis Pharma Inc. (Canada).
Footnotes
-
Send reprint requests to: Patrick du Souich, MD, PhD, Department of Pharmacology, Faculty of Medicine, University of Montréal, P.O. Box 6128, Stat. Centre-Ville, Montréal, Québec, Canada H3C 3J7. E-mail: patrick.du.souich{at}umontreal.ca
-
This study was supported by the Medical Research Council of Canada (Grant MT-14478).
- Abbreviations used are::
- P450
- cytochrome P450
- CYP
- apoprotein of the cytochrome P450
- 1,3-DMU
- 1,3-dimethyluric acid
- HCONT
- hepatocytes from control rabbits
- HINFLA
- hepatocytes from rabbits with a turpentine-induced acute inflammatory reaction
- HSINF
- serum of humans with a viral infection
- IFN-γ
- interferon-γ
- IL-
- interleukin
- l-NAME
- Nω-nitro-l-arginine methyl ester
- 1-MU
- 1-methyluric acid
- 3-MX
- 3-methylxanthine
- NO⋅
- nitric oxide radical
- NOS
- nitric oxide synthase
- ⋅OH
- hydroxyl radical
- O⨪2
- superoxide radical
- ROI
- reactive oxygen intermediates
- RSINFLA
- serum from rabbits with a turpentine-induced acute inflammatory reaction
- Received November 19, 1999.
- Accepted May 26, 2000.
- The American Society for Pharmacology and Experimental Therapeutics