Abstract
NADPH-cytochrome P-450 oxidoreductase (CPR) is essential for the catalytic activity of cytochrome P-450 (P-450). On a molar basis, the amount of P-450 exceeds that of CPR in human liver. In this study, we investigated whether drug-drug interactions can occur as a result of competition between P-450 isozymes for this ancillary protein. For this purpose, combinations of P-450 isozymes were coexpressed together with P-450 reductase in Escherichia coli. We show that testosterone inhibited the CYP2D6-mediated bufuralol 1′-hydroxylase activity in bacterial membranes containing both CYP2D6 and CYP3A4 but not in membranes containing CYP2D6 alone. Conversely, bufuralol inhibited the CYP3A4-mediated testosterone 6β-hydroxylase activity in bacterial membranes containing both CYP3A4 and CYP2D6 but not in membranes containing only CYP3A4. In each case, inhibition was seen even at a P-450 to P-450 reductase ratio of 1.9:1, which is more favorable than the ratio of 4 reported for human liver. The physiological significance of this mechanism was demonstrated by the observation that testosterone inhibited several prototypical P-450 enzyme activities, such as bufuralol 1′-hydroxylase, coumarin 7-hydroxylase, and 7-ethoxyresorufin O-dealkylase, in human liver microsomes, but not if tested against a panel of bacterial membranes containing the human P-450 isozymes that mainly catalyze these reactions.
Drug metabolism determines the pharmacokinetic properties of pharmaceuticals and is involved in the toxication and detoxication of natural and human-made compounds (Parkinson, 1996). It can be divided into two phases. In phase I, reactive groups are introduced into the molecules. These reactions are mainly catalyzed by the cytochrome P-450 (P-450) monooxygenase system (Gonzalez, 1992; Friedberg and Wolf, 1996). In phase II, the metabolites are conjugated to small, hydrophilic endogenous molecules such as glucuronic acid (Burchell and Coughtrie, 1992). Because phase I of drug metabolism determines the subsequent metabolic events, considerable effort has been directed at its characterization.
The P-450 monooxygenase system consists of cytochrome P-450 and ancillary proteins that supply the P-450 with reducing equivalents. P-450 genes belong to a gene superfamily (CYP), which, based on sequence homologies, has been subgrouped into several CYP families (Nelson et al., 1996). The various P-450 isozymes differ in their catalytic properties. The cellular localization of P-450s is either in the endoplasmic reticulum or in mitochondria. Drug metabolizing P-450s are almost exclusively localized in the endoplasmic reticulum and receive the reducing equivalents from the flavoprotein NADPH-P-450 oxidoreductase (CPR; Porter and Coon, 1991). In the general reaction of P-450-catalyzed oxidation, the ferric form of P-450 binds the substrate and is subsequently reduced to the ferrous molecule, which then captures molecular oxygen. With the possible involvement of cytochromeb 5, a second electron is introduced via the CPR into the oxygenated form of the hemoprotein. This results in the activation of the bound oxygen, one atom of which is introduced into the substrate, with the other atom being reduced to water. Dissociation of the oxidized metabolite restores the ferric form of P-450.
In human liver microsomes, P-450s are in molar excess of CPR, the ratio of the two components being 4:1, respectively, as calculated from the cytochrome c reductase activity reported for this subcellular fraction (Forrester et al., 1992) and the known turnover number of CPR for this substrate (approximately 3000 nmol cytochromec reduced/nmol CPR/min as calculated from literature; Nadler and Strobel, 1991). Upon induction of P-450s by xenobiotics, the ratio may become even less favorable, as has been shown in animals (Okey, 1992). The fact that CPR is limiting for P-450-mediated reactions in human liver is also reflected by the observation that incorporation of extra CPR in liver microsomes led to an increase of P-450 enzyme activity (Kitada et al., 1979). Maximal activity was observed at a P-450/CPR ratio of 1:1, which points to the formation of functionally active binary complexes between P-450s and their reductase. However, there is no unified view about the molecular organization of the monooxygenase complex in the lipid bilayer of the endoplasmic reticulum. NADPH-dependent reduction in the membrane displays biphasic kinetics, most of the hemoprotein being reduced in the fast phase. At least three possible explanations for this effect have been presented. 1) The fast and the slow phase correspond to the reduction within P-450/CPR clusters and within separate P-450 and CPR molecules, respectively (Peterson et al., 1976; Engelke et al., 1993). 2) The two phases correspond to the reduction of different conformational states of the P-450, having different abilities to interact with CPR (Backes and Eyer, 1989). 3) The biphasic reduction properties are due to some inherent property of the reductase, such as the multiple oxidation-reduction states involved in electron transfer (Oprian et al., 1979). However a definitive answer on the mechanisms responsible for the biphasic reduction kinetics is complicated by the observation that these kinetics are dependent on the type of P-450 isozyme and the membrane system used for the kinetic assays (e.g., hepatic microsomes versus membranes isolated from a baculovirus expression system;Guengerich and Johnson, 1997). Although there are several potential explanations for the biphasic reduction of P-450, when P-450 is in excess, only a portion of it can be reduced in the initial phase. The remaining P-450 enzymes not initially complexed with CPR will be reduced more slowly but only after formation of a functional complex with the CPR. Formation of clusters between P-450 and CPR has been also found by rotational diffusion analysis, which indicated that hepatic P-450s from phenobarbital-treated rats incorporated into phospholipid vesicles formed immobile complexes (Yamada et al., 1995). The incorporation of CPR into reconstituted vesicles decreased the rotational relaxation time of hemoprotein most likely due to the dissociation of P-450 aggregates.
Dependent on the nature of the structural organization of the P-450 monooxygenase system in the membrane, it can be hypothesized that different P-450 isozymes in the presence of their substrates compete for electrons supplied by P-450 reductase. Strong competition is unlikely if collision of P-450 and CPR proceeds with high frequency concomitant with rapid electron transfer and vice versa. Because competition between P-450s for CPR would be of major physiological significance, we investigated the effects of CYP3A4 substrates on CYP2D6-mediated metabolism and vice versa. To study these interactions in a natural membrane environment and not in an artificially reconstituted system, we isolated membranes from bacteria coexpressing P-450s together with CPR. The results of our study demonstrated that P-450s compete for P-450 reductase. The physiological relevance of this finding was extended by showing that the competition between P-450s for CPR also occurred in human liver microsomes, as evidenced by the effects of testosterone on several prototypical P-450 enzyme activities in these subcellular fractions compared with the effects seen with the relevant recombinant P-450s.
Experimental Procedures
Materials.
Bactopeptone, bactotryptone, yeast extract, and bactoagar were purchased from Difco Laboratories (East Molesey, UK). Ampicillin (Penbritin) was obtained from Beecham Research (Welwyn Garden City, UK), and isopropyl β-d-thiogalactopyranoside (IPTG) was purchased from Melford Laboratories (Ipswich, UK). Aprotinin, leupeptin, and NADP (disodium salt) were purchased from Boehringer Mannheim (Lewes, UK). Restriction and other DNA-modifying enzymes were purchased from Gibco-BRL (Paisley, UK) and Promega (Southampton, UK). All other chemicals were purchased from Sigma (Poole, UK). (±)Bufuralol-HCl and 1′-hydroxybufuralol maleate were a kind gift of Dr. Steve Clarke (SmithKline Beecham, Welwyn, UK). Human liver microsomes were a representative pool from 30 donors and were obtained from the International Institute for the Advancement of Medicine (Leicester, UK).
Construction of CYP3A4/CYP2D6 Coexpression Plasmid.
The strategy for the coexpression of CYP2D6 fused to the bacterialompA leader sequence with CPR has been described recently (Pritchard et al., 1998). Originally CYP3A4 was also expressed with this leader and yielded spectrally active CYP3A4 (Pritchard et al., 1997). However, this construct did not couple efficiently with CPR, most likely because it was not processed by the bacterial signal peptidase. Subsequently, the leader sequence was modified by optimization of the signal peptidase cleavage site to allow efficient processing (M. P. Pritchard, manuscript in preparation). The modified CYP3A4 displayed a similar 6β-testosterone hydroxylase activity (Table 3) to the CYP3A4, which had been modified for expression within its membrane anchor sequence (Gillam et al., 1993;Blake et al., 1996). For construction of the CYP2D6/CYP3A4 coexpression vector, the modified CYP2D6 cDNA was excised from the CYP2D6 expression plasmid together with the Ptactac promoter usingBclI and BglII and ligated into the uniqueBglII site of the CYP3A4 expression plasmid (Fig.1).
Effects of bufuralol on testosterone 6β-hydroxylase activities of membranes containing different combinations of CYP3A4, CYP2D6, and CPR
Construction of expression plasmids: Left, generation of construct encoding CYP3A4 and CYP2D6. Plasmid pCW-ompA-CYP2D6 (Pritchard et al., 1997) was restricted with BglII/BclII to release a fragment containing the P(tac)2 promoter andompA-CYP2D6 cDNA, which was subsequently cloned intoBglII-digested pCW-CYP3A4 plasmid. Resulting construct pCW-CYP3A4/CYP2D6 contains both P-450s under control of separate IPTG inducible P(tac)2 promoters. Right, construction of a plasmid directing low levels of CPR expression. Fragment containing thelacZ promoter was isolated from plasmid pCW by PCR and subcloned into CvnI and NdeI sites of pJR7 (Blake et al., 1996).
Construction of a Low-Level Reductase Expression Plasmid.
The expression plasmid pJR7, which contains the reductase fused to apelB leader sequence under the control of the Ptac tac promoter, has already been described (Blake et al., 1996). The promoter of pJR7 was exchanged for the lacZpromoter to construct a plasmid that should direct low expression of reductase (Fig. 1). The lacZ promoter was isolated from the plasmid pCW by PCR using a 5′ primer (5′ AGTATCGGCCTGAGGCGCAACGCAATTAATGTG AGTTAGC 3′) introducing aCvnI restriction site at the initiation codon, and a 3′ primer (5′ AATCATGGTCATATGTGTTTCCTGTGTGAAATTG TTATCCGC 3′) introducing a NdeI restriction site at the 3′ end of the promoter.
Constructs Used for Coexpression of CYP1A2 and CYP2A6 with CPR.
The P-450 isozymes were expressed from the plasmid pCW and the P-450 reductase from the plasmid pJR7, as previously described for the coexpression of CYP2D6 and CPR in Escherichia coli(Pritchard et al., 1998). CYP1A2 was expressed fused to theompA leader sequence and CYP2A6 was expressed modified within its N-terminus to achieve optimal expression in E. coli (Barnes et al., 1991; Gillam et al., 1993; Blake et al., 1996).
Coexpression of Recombinant P-450 Isozymes with P-450 Reductase in E. coli.
Expression conditions were a modification of those described elsewhere (Gillam et al., 1993;Pritchard et al., 1997). Expression was carried out inE.coli strain JM109. Expression cultures were shaken at 30°C until the optical density (O.D.) at 600 nm reached 0.7, when IPTG (1 mM) and δ-aminolevulinic acid (0.5 mM) were added. Expression was allowed to proceed for 20 to 24 h.
Harvesting Cultures, Membrane Preparation, and Determination of Expression Levels.
Bacterial cells were harvested and fractionated as described previously (Gillam et al., 1993) The P-450 content of the membranes was determined spectrally (Omura and Sato, 1964). The level of P-450 reductase was estimated in membranes using a spectrophotometric assay to measure cytochrome c reduction (Strobel and Dignam, 1978).
Immunodetection of Recombinant CYP3A4, CYP2D6, and P-450 reductase.
Proteins were separated by SDS-polyacrylamide gel electrophoresis in 9% acrylamide gels (Laemmli, 1970) and then transferred onto Hybond-enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham, Buckinghamshire, UK), essentially as previously described (Towbin et al., 1979). Blots were probed with either rabbit anti-CYP2D6 or sheep anti-CYP3A4 or rabbit anti-human P-450 reductase antibody, followed by incubation with the appropriate secondary antibody coupled to horseradish peroxidase (Scottish Antibody Production Unit, Carluke, UK). Detection was by ECL (Amersham).
Testosterone 6β-Hydroxylase Assay .
The assay was carried out with membranes isolated from E. coli expressing the various recombinant proteins (Blake et al., 1996). Incubations contained 10 pmol recombinant P-450 or human liver microsomes (200 pmol P-450) and 30 mM MgCl2 in 50 mM phosphate buffer, pH 7.4. The final testosterone concentration was 0.1 mM. The reaction was started by adding a NADPH-generating system (final concentration 1 mM NADP, 5 mM glucose 6-phosphate, 1 unit glucose 6-phosphate dehydrogenase). Reactions were carried out at 37°C for 10 min and stopped by the addition of 100 μl ice-cold methanol plus 5 μl of 60% perchloric acid and placed on ice for 10 min. Following centrifugation, the metabolites in the supernatant were separated by HPLC on a Spherisorb ODS-2 (5 μm) 250 × 4.6-mm column (Hewlett Packard, Strathaven, UK) using a gradient based on water, methanol, and acetonitrile at a flow rate of 1 ml/min and detection at 240 nm. The yield of the 6β-hydroxytestosterone was calculated by reference to a known concentration of this metabolite.
Bufuralol 1′-Hydroxylase Assay.
The assay was carried out with the E. coli-derived membrane fraction (10 pmol P-450) or human liver microsomes (200 pmol P-450) in 50 mM phosphate buffer, pH 7.4 (Pritchard et al., 1998). The final concentration of bufuralol was 10 μM. The reaction was started by adding a NADPH-generating system (see above). Reactions were carried out at 37°C for 10 min, then stopped by addition of 15 μl of 60% perchloric acid. Following centrifugation, the metabolites in the supernatant were separated by HPLC on a Spherisorb ODS-2 (5 μm) 250 × 4.6-mm column using a gradient based on aqueous ammonium acetate and acetonitrile at a flow rate of 1 ml/min. Metabolites were detected fluorometrically at λex 252 nm, λem 302 nm. The yield of the 1′-hydroxybufuralol was calculated by reference to a known concentration of this metabolite.
Coumarin 7-Hydroxylase Assay.
Coumarin 7-hydroxylase was assayed fluorometrically as previously described (Yun et al., 1991). The assay was carried out at 37°C in 100 mM Tris-HCl, pH 7.4, containing 5 μM coumarin, 10 pmol CYP2A6 in bacterial membranes, or 200 pmol P-450 in human liver microsomes and a NADPH-generating system (see above) in a total volume of 500 μl.
7-Ethoxyresorufin O-Deethylase (EROD) Assay.
EROD was determined by a fluorescence assay as described elsewhere (Burke and Mayer, 1975). The assay was carried out at 37°C in 100 mM potassium phosphate, pH 7.8, containing 5 μM 7-ethoxyresorufin, 10 pmol CYP1A2, or human liver microsomes (200 pmol P-450) and a NADPH-generating system (see above) in a total volume of 500 μl.
Results
Establishment of Functional P-450 Monooxygenase Systems inE. coli.
The strategy for the coexpression of either CYP2D6 or CYP3A4 together with CPR in E. coli has been outlined recently (Blake et al., 1996; Pritchard et al., 1998). To obtain strains that coexpressed both P-450 isozymes together with CPR, the ompA-CYP2D6 and the ompA+2-CYP3A4 were expressed from two separate Ptac tac promoters from the vector pCW (Fig. 1, left). The CPR was coexpressed from a separate plasmid either under the control of the (tac tac) or of the weaker lacZ promoter (Fig. 1, right). Membranes isolated from the different strains had a P-450 content of 260 to 430 pmol P-450/mg membrane protein (Table 1). The CPR activity of membranes isolated from bacteria in which the expression of this protein was under the control of the (tac tac) or of the weaker lacZ promoter was approximately 450 or 100 nmol cytochrome c reduced/min/mg membrane protein, respectively. The difference in the expression level of CPR was also seen by immunoblotting (Fig. 2), which in addition verified that both CYP2D6 and CYP3A4 were coexpressed in strains carrying the relevant expression constructs. The ratio of these P-450 isozymes was estimated by comparison with a calibration curve obtained either with purified CYP2D6 or with purified CYP3A4 (Fig. 3). From this analysis it can be concluded that membranes contained these enzymes in an approximate ratio of 1:2.
P-450 content and reductase activity in membranes isolated fromE. coli-expressing P-450s and CPR
Immunoblot analysis of CPR, CYP2D6, and CYP3A4 inE. coli membranes. Blots A, B, and C were probed with anti-P-450 reductase, anti-CYP2D6, and anti-CYP3A4 antibodies, respectively; subsequently blots were incubated with horseradish peroxidase-coupled secondary antibodies and developed using ECL system (Amersham). The following samples were analyzed: lane 1, membranes from cells carrying empty pCW; lane 2, human liver microsomes; lane 3, membranes from cells carrying CYP3A4/pJR7; lane 4, membranes from cells carrying CYP3A4/CYP2D6/pJR7; lane 5, membranes from cells carrying CYP3A4/CYP2D6/pDRlacZ, lane 6, membranes from cells carrying CYP2D6/pJR7. Twenty micrograms of microsomal protein and 5 μg of bacterial membrane protein were analyzed.
Quantitation of CYP2D6 and CYP3A4 in E. coli membranes. Blots A and B were probed with anti-CYP3A4 and anti-CYP2D6 primary antibody, respectively. As standards, decreasing amounts of purified CYP3A4 (blot A) or CYP2D6 (blot B) were used. A dilution row of membranes isolated from bacteria coexpressing CYP2D6 and CYP3A4 together with CPR was analyzed. Numbers indicate amount (μg) of purified CYP3A4 or CYP2D6 loaded per track.
Competition Between P-450s for CPR in a Recombinant System.
The effects of testosterone on bufuralol 1′-hydroxylase activity were studied with bacterial membranes containing CYP2D6 alone or in combination with CYP3A4 (Table 2). In addition, these membranes had either high or low levels of CPR. No significant effect of the steroid on the enzyme activity of membranes containing CYP2D6 alone was seen. However a pronounced inhibition was noticed upon the additional presence of CYP3A4. The magnitude of inhibition was moderately influenced by the CPR level (57% and 36% in the presence of high and low levels, respectively, of CPR). The inhibition was not caused by 6β-hydroxytestosterone, which is formed in the presence of CYP3A4, because this metabolite did not affect the activity.
Effects of testosterone on bufuralol 1′-hydroxylase activities of membranes containing different combinations of CYP3A4, CYP2D6, and CPR
In an analogous manner to the inhibition of the bufuralol 1′-hydroxylase activity by testosterone described above, bufuralol also inhibited the testosterone 6β-hydroxylase activity in membranes containing CYP3A4 and CYP2D6 with CPR (Table3). Bufuralol did not alter the activity of membranes containing only CYP3A4 with CPR. The inhibition of the testosterone hydroxylase activity by bufuralol was less pronounced than the inhibition of the bufuralol hydroxlase activity by testosterone (cf. Tables 2 and 3). Again, the P-450 reductase level had a moderate effect on the inhibition of testosterone hydroxylase activity. The inhibition of the testosterone hydroxylase was not caused by 1′-hydroxlase bufuralol, which is formed in the presence of CYP2D6, because this metabolite had no influence on the activity of CYP3A4 (Table 3).
In membranes containing CYP2D6 and CYP3A4 with either low or high levels of CPR, the inhibition of bufuralol 1′-hydroxlase activity increased with increasing concentrations of testosterone in the assay (Fig. 4A) concomitant with an increased testosterone 6β-hydroxylase activity (Fig. 4B). Clear inhibition was observed at a testosterone concentration of 40 μM, which is below theK m (56 μM) of CYP3A4 for testosterone 6β-hydroxylase activity (Ding et al., 1997). The effect of the CPR level on the inhibition became more pronounced at higher concentrations of testosterone. Testosterone 6β-hydroxylase was not saturated at the highest concentration of testosterone, which, however, was only 2-fold above K m (Fig. 4A). A similar effect was seen with membranes that contained only CYP3A4 and CPR (data not shown).
Correlation between testosterone concentration and bufuralol 1′-hydroxylase and testosterone 6β-hydroxylase activity. Membranes were isolated from E. coli expressing CYP2D6 and CYP3A4 with either high (plasmid pJR7, ‘) or low (plasmid pDRlacZ, ⋄) levels of CPR. Inhibition of bufuralol 1′-hydroxylase activity (A) and effects on testosterone 6β-hydroxylase activity (B) was determined as described in Experimental Procedures in the presence of various concentrations of testosterone as indicated. Assay was performed in triplicate with S.D. of enzyme activity being less than 5% of means.
The effects of testosterone on different P-450-mediated enzyme activities were also determined in human liver microsomes (pooled from 30 donors) and for comparison in membranes containing the recombinant P-450s, which are mainly responsible for these reactions. Testosterone had no significant effect on bufuralol 1′-hydroxylase, coumarin 7-hydroxylase, and EROD activity of recombinant CYP2D6, CYP2A6, and CYP1A2, respectively. However, in liver microsomes testosterone inhibited these activities by 38%, 20%, and 30%, respectively (Table4). It should be noted that the substrate turnover numbers determined for the various recombinant isozymes were much higher than those obtained with human liver microsomes, because in microsomes each isozyme represents only a fraction of the total P-450 content.
Effects of testosterone on several P-450-mediated enzyme activities in bacterial membranes containing recombinant human P-450s or in human liver microsomes
Discussion
Drug-drug interactions frequently complicate multiple-drug therapy. Pharmacokinetic interactions are often due to one drug modulating the P-450- mediated metabolism of another either by binding to the same P-450 isozyme or by altering its cellular levels (Nies and Spielberg, 1996). However, in this article we demonstrated that they can also occur by competition between P-450s for CPR. This was evidenced by the inhibition of the CYP3A4- mediated testosterone 6β-hydroxylase activity by the protoypical CYP2D6 substrate bufuralol and by the inhibition of the CYP2D6-mediated bufuralol 1′-hydroxylase activity by testosterone in membranes containing both isozymes together with CPR but not in membranes containing only one of these isozymes. Inhibition of bufuralol 1′-hydroxylase by testosterone was more pronounced than the inhibition of testosterone 6β-hydroxylase activity by bufuralol (cf. Tables 2 and 3). This is presumably a result of the higher specific activity in these membranes toward testosterone compared with bufuralol, requiring more electrons being transferred to CYP3A4 than to CYP2D6-mediated reactions. The higher activity toward testosterone was due to the higher turnover number of this substrate by CYP3A4 (Tables 2 and 3) and the higher level of this enzyme compared with CYP2D6 in membranes containing both P-450 isozymes (Fig. 3). The relationship between the level of testosterone 6β-hydroxylase activity and the inhibition of the bufuralol 1′-hydroxylase activity was also seen in an experiment in which increasing concentrations of testosterone were added to membranes containing both P-450s (Fig. 4). This led to an increase of testosterone 6β-hydroxylase concomitant with a linearly increased inhibition of the bufuralol hydroxylase activity. However an additional explanation for the more pronounced inhibition of the bufuralol 1′-hydroxylase activity by testosterone as compared with the inhibition of the testosterone 6β-hydroxylase activity could be that CYP2D6 and CYP3A4 have different affinities for CPR that are differentially altered by the presence of xenobiotics.
One may argue that the effects described above were not due to competition of P-450 isozymes for P-450 reductase but that the presence of a second P-450 (e.g., CYP3A4) facilitated the interaction of a substrate with a P-450 metabolizing it (e.g., bufuralol for CYP2D6) and that this facilitation was inhibited by the presence of the substrate for the second P-450 (e.g., testosterone). However, this mechanism is unlikely to explain the drug-drug interactions described in this work, because, first, K m values obtained with membranes containing a recombinant P-450 isozyme are usually in agreement with values obtained for that isozyme in hepatic microsomes that contain several additional P-450 isozymes (Lee et al., 1995), and, second, these interactions were influenced by the ratio of P-450s/CPR in the present study. The magnitude of the inhibition (57%) of testosterone 6β-hydroxylase by bufuralol seen in membranes containing high levels of CPR was surprisingly high, because these preparations had a ratio of P-450 to CPR of 1.9:1 (Table 1) which is much more favorable than the ratio (4:1) found in human liver microsomes (Forrester et al., 1992). Increasing this ratio to 13 resulted in an increased inhibition (Tables 2 and 3).
Previously, it had been reported that the rabbit CYP2B4-mediated pentoxyresorufin O-dealkylase activity was decreased by the presence of rabbit CYP1A2 in a reconstituted system in the absence of a second substrate (Cawley et al., 1995). It was concluded that the CPR was physically sequestered by CYP1A2. In our experiments, the testosterone 6β-hydroxylase turnover number (expressed as nmol product/min/nmol P-450) determined in membranes containing CYP3A4 only and membranes containing CYP2D6 in addition differed 1.8-fold (Table3). Only a 1.5-fold decrease would have been expected upon expression of CYP2D6, based on the contribution of CYP3A4 (66%) and CYP2D6 (34%) to the total content of P-450 in the membranes (Fig. 3). However, because the determination of each isozyme was only semiquantitative and did not distinguish between P-450 holoenzyme and apoprotein, we cannot decide whether the P-450 dependent sequestration of P-450 reductase seen by others (Cawley et al., 1995) occurred in our model.
A very recent report (Tan et al., 1997) described the inhibition of CYP2E1 catalyzed dealkylation of N-nitrosodimethylamine by the CYP2A6- mediated hydroxylation of coumarin in baculovirus membranes containing both P-450s and P-450 reductase. From these and our data, it is evident that, at least in recombinant models, drug-drug interaction can take place due to competition between P-450s for CPR. However, the significance of this finding for hepatic metabolism is debatable, because membranes isolated from the recombinant models have a different phospholipid and protein composition than hepatic microsomes. To address this important question, we determined the effects of testosterone on several P-450 enzyme activities in human liver microsomes. To exclude direct inhibition of these reactions by testosterone, we also measured its effect on the activity of the major P-450 isozymes catalyzing these reactions. For this determination, several human P-450 isozymes were coexpressed together with CPR inE. coli. As mentioned previously, testosterone did not significantly affect the bufuralol 1′-hydroxylase activity of recombinant CYP2D6. However, this activity was inhibited by 38% in microsomes upon addition of the steroid (Table 4). Similarly, significant inhibition of coumarin 7-hydroxylase and EROD activities was also observed in microsomes. However, testosterone did not inhibit the former reaction in bacterial membranes, which contained the major coumarin hydroxylase, namely CYP2A6. The steroid had also no effect on the EROD activity in bacterial membranes containing the P-450, which mainly catalyzes this reaction in human liver, namely CYP1A2. It is important to note that the testosterone 6β-hydroxylase activity of recombinant CYP3A4 in the different buffer systems employed for the various P-450 assays differed by less than 2-fold, indicating that the degree of competition between CYP3A4 and the other P-450 isozymes for reducing equivalents was similar under these conditions. We also tried to investigate whether testosterone would inhibit the CYP2C9-mediated 4′-hydroxylation of diclofenac. However, in this case, we found that the hormone would already inhibit the activity of the recombinant CYP2C9.
From our data it is evident that competition between several P-450 isozymes for P-450 reductase can occur in human liver microsomes. Therefore, these effects are likely to take place in vivo, particularly if the P-450-mediated metabolism of two drugs that are metabolized by two different P-450 isozymes is high.
Footnotes
-
Send reprint requests to: T. Friedberg, Biomedical Research Centre, University of Dundee, Ninewells Hospital and Medical School, Dundee, DD1 9SY, United Kingdom. E-mail:t.h.friedburg{at}dundee.ac.uk
-
↵1 This work was sponsored by the United Kingdom Biological Sciences Research Council, the United Kingdom Department of Trade and Industry and the LINK consortium of pharmaceutical companies: Astra, Glaxo-Wellcome, Janssen Pharmaceutica, Lilly, Novo-Nordisk, Parke-Davis, Pfizer, Roche Products, Sanofi, Servier, Smith-Kline Beecham, Wyeth-Ayerst and Zeneca.
- Abbreviations:
- CPR
- NADPH-cytochrome P-450 oxidoreductase
- P-450
- cytochrome P-450
- IPTG
- isopropyl β-d-thiogalactopyranoside
- EROD
- 7-ethoxyresorufinO-deethylase
- Received July 7, 1998.
- Accepted December 14, 1998.
- The American Society for Pharmacology and Experimental Therapeutics
