Abstract
Hepatic cytochromes P-450 (CYP) are well characterized drug and xenobiotic metabolizing enzymes that are extensively regulated by genetic and environmental factors. Inflammatory mediators, including interleukins (ILs), interferons (IFNs), and tumor necrosis factor-α (TNF-α), have been shown to down-regulate several CYP isoforms; however, elucidation of the inflammatory mediators that are responsible for specific CYP down-regulation is difficult. The purpose of this experiment was to evaluate the role endogenous TNF-α plays in the regulation of liver CYP expression after endotoxin administration. Mice deficient in the p55 and p75 TNF receptors and wild-type mice were given Gram-negative bacterial lipopolysaccharide (LPS) and killed 24 h after administration. CYP analysis indicates that LPS decreases CYP1A, CYP2B, CYP3A, and CYP4A independently of TNF-α. CYP2D9 and CYP2E1 activities show differential responses to LPS between wild-type and TNF p55/p75 receptor knockout mice, indicating the down-regulation of CYP2D9 and CYP2E1 is differentially modulated by TNF-α expression. Furthermore, TNF-α appears to affect the constitutive expression of CYP2D9 and CYP2E1. To date, this is the first evidence suggesting that a proinflammatory cytokine is involved in the constitutive regulation of drug-metabolizing enzymes.
The cytochrome P-450 (CYP) superfamily is a family of enzymes that catalyze numerous reactions, including fatty acid metabolism, xenobiotic biotransformation, and endogenous substrate synthesis and metabolism (Ioannides, 1996). The various CYP isoforms primarily involved with xenobiotic biotransformation and drug metabolism are grouped into the CYP1, CYP2, and CYP3 subfamilies. Several factors contribute to the levels and activity of these various isoforms, including genetic factors, environmental factors, and xenobiotic substances that can induce or repress the activity of these and other isoforms (Halpert et al., 1994).
Lipopolysaccharide (LPS), also known as Gram-negative bacterial endotoxin, has been shown by numerous studies to decrease CYP drug metabolism in several species, including rodents (Chen et al., 1992;Morgan, 1993; Sewer et al., 1996) and humans (Shedlofsky et al., 1994); however, LPS has also been shown to increase the activity of CYP4A (Sewer et al., 1996) and nitric oxide synthase (Khatsenko et al., 1993). Several studies have shown that various inflammatory cytokines are up-regulated after LPS administration, including tumor necrosis factor-α (TNF-α), interleukin (IL)-1, IL-6, IL-8, and various interferons (IFNs) (Cassatella et al., 1993; Evans et al., 1993;Crawford et al., 1997). Macrophages are the primary source of these inflammatory cytokines and have been shown to mediate the effects of LPS on hepatocytes (Freudenberg et al., 1986; Bertini et al., 1989) as well as clearance of LPS from the bloodstream (Freudenberg et al., 1992). The administration of cytokines such as IL-1, IL-6, TNF-α, IFN, or a combination has shown that these cytokines mimic the effects of LPS on hepatic CYP mRNA, protein, and activity levels (Morgan et al., 1994; Nadin et al., 1995; Carlson and Billings, 1996; Sewer and Morgan, 1997). Studies have further shown that LPS has no toxic effect on primary hepatocytes (Sauer et al., 1996), whereas cytokines elicit decreases in various CYP isoforms depending on the cytokine administered (Morgan et al., 1994; Carlson and Billings, 1996; Sewer and Morgan, 1997).
TNF-α has been implicated as an important mediator of the physiological effects of LPS. Some studies indicate that TNF-α has a protective role against the lethal effects of LPS (Beutler et al., 1985; Freudenberg et al., 1986). TNF-α has also been implicated in the effects of LPS on drug metabolism (Ghezzi et al., 1986; Memon et al., 1993). The administration of TNF-α has been shown to decrease total CYP as well as CYP1A, CYP2B, CYP2C, CYP2E, and CYP3A subfamilies (Pous et al., 1990; Chen et al., 1992; Nadin et al., 1995; Monshouwer et al., 1996; Sewer and Morgan, 1997). In addition, TNF-α stimulates the production and secretion of other cytokines, including IL-1, IL-6, and IFN, which have also been shown to down-regulate the same CYP subfamilies (Ghezzi et al., 1986; Pous et al., 1990; Chen et al., 1992;Morgan et al., 1994; Carlson and Billings, 1996; Sewer and Morgan, 1997). However, elucidation of the role of endogenous TNF-α in the regulation of CYP and other metabolic enzymes has been difficult.
TNF-α is known to interact with two specific receptors: the p55 receptor, also known as TNF receptor 1, and the p75 receptor, also known as TNF receptor 2 (Gruss and Dower, 1995; Darnay and Aggarwal, 1997). Previous studies using the administration of exogenous TNF-α cannot delineate the specific role of TNF-α in the regulation of various CYP isoforms because TNF-α is known to induce IL-1, IL-6, and IFN. The purpose of this experiment was to elucidate the role of TNF-α in the regulation and constitutive expression of liver CYP enzymes. Using mice lacking both receptors (TNF p55/p75 receptor knockout) and wild-type C57BL/6 mice, we were able to block the effects of TNF-α. LPS was administered in sufficient quantity to induce the acute phase response with subsequent reduction of CYP enzymes in both groups of mice. Results indicate differences in the response of the TNF p55/p75 receptor knockout mice and differences between constitutive levels of specific CYP isoforms in the TNF p55/p75 receptor knockout mice compared with wild-type mice.
Methods
Animals and Treatments.
Unless otherwise specified, all chemicals were obtained from Sigma Chemicals (St. Louis, MO). Adult male (8 to 10 weeks old) wild-type and TNF receptor double knockout mice (p55−/−/p75−/−) were generated and maintained as described previously (Zheng et al., 1995; Bruce et al., 1996). Mice were maintained on a random C57BL/6 × 129 hybrid background. TNF p55/p75 receptor double knockout mice exhibit no overt phenotypes under normal conditions, although a variety of inflammatory responses are altered (Zheng et al., 1995). Animals were allowed water and food ad libitum. LPS derived from Escherichia coli (LD50 28.7 mg/kg; DIFCO Laboratories, Detroit, MI) was dissolved in 0.9% saline and injected intraperitoneally at a dose of 2 mg LPS/kg b.wt. in treated animals. Control animals received an equivalent volume of sterile saline. Animals administered LPS appeared visibly sick and became lethargic. Animals were euthanatized 24 h after LPS or saline administration. Food was removed 10 h before euthanasia, which occurred between 8:00 and 10:00 a.m. Animals were anesthetized with halothane before euthanasia. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Kentucky. Animals were divided into four groups: saline-treated wild-type mice (CC), LPS-treated wild-type mice (CE), saline-treated TNF p55/p75 receptor knockout mice (TC), and LPS-treated TNF p55/p75 receptor knockout mice (TE). Six animals were used for each group (n = 6).
Microsomal and Spectral Methods.
Livers were excised and perfused with cold 0.9% saline. Liver samples were placed in homogenization buffer (0.154 M KCl, 0.25 M potassium phosphate buffer, pH 7.4) with the addition of BHT as an antioxidant just before homogenization. Livers were homogenized using a Teflon grinder and spun to separate the microsomal and cytosolic fractions. The resultant microsomal pellet was resuspended in 0.25 M sucrose and 0.02 M Tris buffer, pH 7.4. Spectral analysis of total P-450 content was performed according to the method of Omura and Sato (1964). Total protein content was determined according to the method of Lowry et al.(1951).
Analysis of CYP Activity.
CYP activities were determined using the formation of monohydroxylated product from substrates associated with specific CYP isoforms. The formation of monohydroxylated products from testosterone was determined according toSonderfan et al. (1987). The 6-hydroxylation of chlorzoxazone was determined according to Peter et al. (1990) as modified by Jayyosi et al. (1995). The dealkylation of ethoxyresorufin and pentoxyresorufin to resorufin was performed according to Burke et al. (1985). The ω-hydroxylation of lauric acid according was measured according toGiera and van Lier (1991). CYP1A is strongly associated with ethoxyresorufin dealkylation in many species (Ioannides, 1996). CYP1A induction by polynuclear aromatic hydrocarbons has shown similar induction profiles for CYP1A mRNA, protein, and ethoxyresorufinO-dealkylation in mouse liver (Chaloupka et al., 1995; Jeong et al., 1995). CYP2A12 has been attributed to the formation of 7α-hydroxytestosterone (Iwasaki et al., 1993), and CYP2B10 has been shown to be the major catalyst of pentoxyresorufin dealkylation in mouse liver (Honkakoski et al., 1992). 16α-Hydroxytestosterone formation is catalyzed primarily by CYP2D9 in male mice (Wong et al., 1987, 1989), but induced CYP2B may be responsible for up to 30% of the metabolism of testosterone to the 16α-hydroxy metabolite (Honkakoski et al., 1992). The 6-hydroxylation of chlorzoxazone has been shown to have single enzyme kinetics with specific anti-CYP2E1 inhibition in mouse liver (Court et al., 1997), but Jayyosi et al. (1995)demonstrated that 6-hydroxychlorzoxazone formation can also be catalyzed by CYP1A and CYP3A in the rat. Formation of 6β-hydroxytestosterone has been attributed to the CYP3A subfamily (Yanagimoto et al., 1994), and CYP4A has been associated with the hydroxylation of lauric acid in mouse kidney and liver (Hiratsuka et al., 1996a, 1996b). Therefore, these substrate activities are assumed to be primarily catalyzed by associated CYP isoforms as described above.
Analysis of CYP Proteins by Enzyme-Linked Immunosorbent Assay.
Individual microsomal sample liver P-450 isoform content was quantified by noncompetitive enzyme-linked immunosorbent assay (ELISA). Liver microsomes were diluted in phosphate carbonate/bicarbonate-buffered saline, pH 9.6. Subsequently, 0.25 to 1 μg of total microsomal sample protein and known concentrations of microsomal standard (Gentest, Woburn, MA) were plated onto 96-well flat-bottomed plates (Corning, NY). Proteins were blocked and incubated with 50% horse serum and 50% Tris-buffered saline with 0.1% Tween 20. Plates were then washed and incubated with polyclonal goat anti-rat antibody for either CYP1A, CYP2B, CYP3A1/2, CYP2E1, or CYP4A (Gentest). Next, plates were washed and incubated with alkaline phosphatase-conjugated monoclonal rabbit anti-goat IgG antibody. After this incubation, plates were washed, and p-nitrophenol phosphate substrate (ELISA Technologies, Lexington, KY) was added. Plates were analyzed at 405 nm over 30 min at 37°C with a Biotek EL340 microplate reader for color formation. Isoform-specific CYP content in samples were quantified by extrapolation from a fitted standard curve. Although the specificity of these antibodies has not been definitively established, the pattern of CYP alteration obtained in the ELISA are supported by both Western blot and substrate specific activity determinations for each CYP analyzed.
Statistical Analysis.
Multiple comparisons were performed using SAS. Statistical comparisons were based on two-way analysis of variance with Fisher’s LSD post-hoc determination. Statistical differences were observed using p < .05.
Results
Effects of LPS Administration on Microsomal Spectral P-450.
LPS caused similar decreases on spectral P-450 levels in livers from both wild-type and TNF p55/p75 receptor knockout mice. Table1 shows the effects of LPS on spectral P-450 content. LPS significantly decreased spectral P-450 concentrations by 25.0% in wild-type mice and by 28.1% in TNF p55/p75 receptor knockout mice. No significant difference was observed in spectral P-450 concentrations between groups CC (0.780 ± 0.086 nmol/mg) and TC (0.787 ± 0.191 nmol/mg) or between groups CE (0.585 ± 0.033 nmol/mg) and TE (0.566 ± 0.076 nmol/mg).
Effects of LPS Administration on Microsomal CYP Activity and Protein.
Table 1 shows the effects of LPS administration on the dealkylation of ethoxyresorufin and pentoxyresorufin and CYP1A and CYP2B protein in wild-type and TNF p55/p75 receptor knockout mice. LPS significantly decreased ethoxyresorufin O-dealkylase (EROD) activity in both CE (to 49.2% of CC) and TE (to 49.0% of CC), but there was no significant difference between CC and TC. LPS significantly decreased 7-pentoxyresorufin O-dealkylase (PROD) activity in both CE (to 46.7% of CC) and TE (to 48.4% of CC), but no significant difference was observed between CC and TC. LPS had no significant effect on CYP1A protein in wild-type mice but significantly decreased CYP1A protein in TNF p55/p75 receptor knockout mice. CYP2B protein was not significantly decreased in TNF p55/p75 receptor knockout mice after LPS administration, but LPS administration significantly decreased CYP2B in wild-type mice.
The 6-hydroxylation of chlorzoxazone was used as a marker of CYP2E1 activity. LPS significantly decreased the 6-hydroxylation of chlorzoxazone in group CE (to 71.8% of CC) and TE (to 65.1% of CC), but LPS had no significant effect on TE compared with TC (Fig.1). 6-Hydroxychlorzoxazone formation was significantly lower in TC relative to CC (74.5% of CC) and not significantly different from either CE or TE. Analysis of CYP2E1 proteins by ELISA showed no significant difference between groups CC and TC, but LPS administration significantly decreased CYP2E1 proteins in both CE and TE. Combined with the results from 6-hydroxychlorzoxazone formation, these data suggest that the effects of LPS on chlorzoxazone hydroxylation and CYP2E1 protein are dependent on endogenous TNF-α.
LPS caused a significant decrease in 6β-hydroxytestosterone (OHT) formation and CYP3A protein to similar proportions in both TNF-α p55/p75 receptor knockout mice and wild-type mice (Fig.2). LPS had no significant effect on 7α-hydroxytestosterone formation in wild-type or TNF-α p55/p75 receptor knockout mice. Antibodies to CYP2A were not available for protein analysis.
LPS significantly decreased 16α-hydroxytestosterone formation in wild-type mice and in TNF-α p55/p75 receptor knockout mice (Fig.3). Constitutive 16α-hydroxytestosterone formation in TNF-α p55/p75 receptor knockout mice was significantly lower than that in wild-type mice. Antibodies specific to human CYP2D6 from Gentest were not cross-reactive with proteins in the mouse; therefore, analysis of CYP2D9 protein was not performed.
LPS caused a significant decrease in lauric acid ω-hydroxylase (LAH) formation in both wild-type and TNF p55/p75 receptor knockout mice (Fig. 4). A proportional decrease in CYP4A protein was observed after LPS administration for both CE and TE. No significant difference was observed between CC and TC for either LAH formation or CYP4A protein.
Western blots of pooled microsomes were performed for each group. Results using anti-rat CYP2E1 and CYP3A antibodies show a single band for CYP2E1 and two bands for CYP3A, indicating two CYP3A isoforms (data not shown). Western blots using anti-rat CYP4A antibodies show three bands, one of which is significantly decreased after LPS administration (data not shown). All Western blot migration patterns were consistent with a 55-kDa molecular mass protein. Cumulative ELISA concentrations from CYP1A, CYP2B, CYP2E, CYP3A, and CYP4A are greater than total spectral P-450 concentrations, but ELISA concentrations are relative measures of specific isoform content and should not be used as absolute markers of specific protein content.
Discussion
Analysis of cytokines using ELISA showed significant elevations of constitutive TNF-α in the plasma of TNF p55/p75 receptor knockout mice (data not shown) with no changes in phenotype. The levels of constitutive TNF-α in TNF p55/p75 receptor knockout mice was comparable to LPS-induced TNF-α levels in normal C57BL/6 mice. The constitutive elevation of TNF-α in TNF p55/p75 receptor knockout mice and lack of response to significantly elevated TNF support the lack of TNF p55/p75 receptor function in the TNF p55/p75 receptor knockout mice.
It should be noted that the characterization of CYP-specific substrates in the mouse is poorly understood compared with the rat and further studies using purified mouse CYP proteins are required to validate the specificity of substrate markers for specific CYP isoforms. Further reference regarding the association of substrates for CYP isoforms is detailed in Methods.
LPS administration to mice lacking TNF p55/p75 receptors and in wild-type mice produced decreases in OHT formation, indicating decreases in CYP3A activity. The observed similarity between wild-type and TNF p55/p75 receptor knockout mice for both CYP3A protein and activity indicate TNF-α is not a significant factor contributing to the decreases in CYP3A by LPS. These results are in agreement with previous experiments showing that pentoxifylline, an inhibitor of TNF-α production after LPS administration, has no significant effect on the down-regulation of CYP3A by LPS (Monshouwer et al., 1996). Other studies indicate decreases in CYP3A by LPS are mediated primarily by IL-1 and IL-6 (Chen et al., 1992; Morgan et al., 1994).
LAH activity and CYP4A protein significantly decreased after LPS administration. Sewer et al. (1996) showed increased CYP4A activity in Fischer rats but showed decreased CYP4A activity in Sprague-Dawley rats after LPS administration. The decrease in CYP4A activity and protein presented in this experiment contrast with the increases observed bySewer et al. (1996) in Fischer rats, but these differences could be due to species-specific responses or lack of specificity of lauric acid (Amet et al., 1994; Jayosi et al., 1995). Furthermore, the data presented here represent cumulative LAH. Sewer et al. (1996) showed a decrease in ω1-hydroxylase activity for both Fischer and Sprague-Dawley rats but observed a decrease in ω-hydroxylase activity only in Sprague-Dawley rats with an increase in Fischer rats after LPS administration. The decreases observed in CYP4A proteins by ELISA indicate a net down-regulation of CYP4A protein; however, specific CYP4A isoforms may be differentially regulated in response to LPS. A differential response is supported by the observation that the middle band intensity in the CYP4A immunoblot appeared to decrease after LPS administration, whereas LPS administration had no apparent effect on the intensity of the upper and lower bands in the CYP4A immunoblot (data not shown).
There were no significant differences in the responses of ethoxyresorufin and pentoxyresorufin dealkylation after LPS administration between wild-type and TNF p55/p75 receptor knockout mice, indicating TNF-α does not have a significant role in the regulation of CYP1A or CYP2B activity after LPS administration. Other studies have indicated that TNF-α is directly involved in the regulation of CYP1A and CYP2B. Inhibition of TNF-α production by pentoxyfylline after LPS administration partially prevented the inhibitory effects of LPS on CYP1A and CYP2B (Monshouwer et al., 1996); however, several studies show that IL-1, IL-2, IL-6, and IFN can decrease CYP1A and CYP2B (Chen et al., 1992; Cantoni et al., 1995;Carlson and Billings, 1996; Monshouwer et al., 1996). In addition, EROD and PROD may not be sufficiently specific for CYP1A and CYP2B at non-induced CYP concentrations. Consequently, EROD and PROD activity may partially reflect the activities of nonspecific enzymes. The lack of effect of LPS on CYP1A protein in wild-type mice and on CYP2B protein in TNF p55/p75 receptor knockout mice further supports the possible lack of specificity of EROD and PROD for CYP1A and CYP2B, respectively, at noninduced levels.
The most interesting and most significant observations presented are the constitutive decreases in 16α-hydroxytestosterone and 6-hydroxychlorzoxazone formation in the TNF p55/p75 receptor knockout mice. The decrease in 6-hydroxychlorzoxazone formation in the TNF receptor knockout mice with no change in CYP2E1 protein, compared with wild-type mice, indicates a post-translational mechanism of constitutive CYP2E1 down-regulation in the TNF receptor knockout mouse. There are no CYP2D9 antibodies yet available to help support this conclusion, but the patterns between apparent CYP2E1 and CYP2D9 activity are similar, indicating common alterations in expression for both CYP2D9 and CYP2E1. To date, this is the first evidence that cytokines may play a role in the constitutive regulation of drug-metabolizing enzymes. Further studies will be required to elucidate the specific pathway by which TNF-α or other cytokines may regulate the constitutive expression of various enzymes.
Several components have been implicated as possible factors contributing to the down-regulation of P-450 by LPS or TNF. TNF is known to activate nuclear factor-κB through activation of sphingomyelinase by the TNF receptor on binding TNF, sphingomyelin hydrolysis, and subsequent ceramide production (Heller and Kronke, 1994). Sphingomyelinase and ceramide have been shown to decrease CYP2C11 in the rat (Chen et al., 1995). TNF-α has also been shown to induce nitric oxide synthase and increase nitric oxide production (Harbrecht et al., 1994; Sewer and Morgan, 1997) and nitric oxide has been directly implicated in decreasing CYP (Wink et al., 1993; Carlson and Billings, 1996). Furthermore, nitric oxide synthase inhibitors have been shown to prevent depression of various CYP isoforms by TNF-α and LPS (Khatsenko et al., 1993; Carlson and Billings, 1996). Cellular energetics and glucocorticoid regulation also may be indirectly involved in the regulation of CYP isoforms. Increased cAMP has been shown to decrease TNF protein and mRNA (Feng et al., 1997). Pyruvate, dexamethasone, and cAMP have separately been shown to inhibit nitric oxide synthesis (Liang and Akaike, 1997), and even at low concentrations, dexamethasone is known to decrease CYP (Morgan et al., 1994). Consequently, decreased CYP could result from LPS or TNF-α administration leading to changes in cellular energetics, nitric oxide production, glucocorticoid secretion, or the production of cytokines, which can subsequently alter CYP enzyme expression. Elucidation of the actual pathways from the possible pathways will require analysis using different knockout and in vitro methodologies.
In conclusion, these results indicate that TNF-α plays a role in the constitutive regulation of CYP2D9 and CYP2E1. Furthermore, these data suggest that TNF-α does not play a significant role in the down-regulation of CYP1A, CYP2B, CYP3A, and CYP4A after LPS administration. The down-regulation of various CYP isoforms could result from pretranslational mechanisms including decreased transcription or mRNA stability, or post-translational mechanisms such as altered CYP protein degradation and turnover; however, further studies will be required to ascertain which mechanisms are involved in the constitutive and LPS induced down-regulation of specific CYP isoforms.
Acknowledgments
We thank Dr. Larry Robertson for technical assistance and Mr. Robert Whelan for surgical assistance.
Footnotes
-
Send reprint requests to: Dr. Robert A. Blouin, College of Pharmacy, 907 Rose Street, University of Kentucky, Lexington, KY 40536-0082. E-mail: rbloul{at}pop.uky.edu.
-
↵1 This work was supported by National Institutes of Health Grants NS29001, AG14554, and NS35253 (M.P.M.) and Kentucky Spinal Cord Head Injury Research Trust Grant BB-9502-K (R.A.B.). G.W.W. was supported by Institutional NIEHS Training Grant ES07266 and a Quality Achievement Award by the University of Kentucky. S.M.P. was supported by the American Foundation for Pharmaceutical Education and the College of Pharmacy at the University of Kentucky.
- Abbreviations:
- CYP
- cytochrome P-450
- LPS
- lipopolysaccharide
- TNF-α
- tumor necrosis factor-α
- IL
- interleukin
- ELISA
- enzyme-linked immunosorbent assay
- OHT
- hydroxytestosterone
- LAH
- lauric acid ω-hydroxylase
- IFN
- interferon
- Received May 18, 1998.
- Accepted September 29, 1998.
- The American Society for Pharmacology and Experimental Therapeutics