In vitro Analysis and Quantitative Prediction of Efavirenz Inhibition of Eight Cytochrome P450 (CYP) Enzymes: Major Effects on CYPs 2B6, 2C8, 2C9 and 2C19

Summary

In order to quantitatively predict drug interactions associated with efavirenz-based anti-HIV therapy, we evaluated reversible and time-dependent inhibitions of efavirenz on eight cytochrome P450 (CYP) enzymes in vitro. Present study showed that efavirenz was a potent competitive inhibitor of CYP2B6 in HLMs (average K_i = 1.68 μM) and expressed CYP2B6 (K_i = 1.38 μM). Moderate inhibition of CYP2C8 by efavirenz was observed in pooled HLMs (K_i = 4.78 μM) and HLMs with CYP2C8*3/*3 genotype (K_i = 4.80 μM). Efavirenz was a moderate inhibitor of CYP2C9 (K_i= 19.46 μM) and CYP2C19 (K_i= 21.31 μM); and a weak inhibitor of CYP3A (K_i= 40.33 μM). No appreciable inhibition was observed on CYP1A2, CYP2A6 or CYP2D6. No time-dependent inhibition of the CYPs by efavirenz was observed in this study. Quantitative predictions showed that single dose of efavirenz may substantially slow the elimination of drugs predominately cleared by CYP2B6, CYP2C19 or by both enzymes and may also lower the area under the plasma concentration time curve (AUC) of active metabolites of some pro-drugs (e.g. clopidogrel and proguanil ) by up to 30%. Depending on substrates, chronic administration of efavirenz may increase the AUC of CYP2C8 and CYP2C9 substrates about 3.5 ~ 4.4-fold and 1.7 ~ 2.0-fold at steady state.

Introduction

Efavirenz, a potent non-nucleoside reverse transcriptase inhibitor (NNRTI), remains preferred component of highly active antiretroviral therapy (HAART) for treatment naïve patients despite an emergence of second generation of NNRTIs and new classes of antiretroviral agents (Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. Department of Health and Human Services, March 27, 2012). It is the only antiretroviral agent currently on the market that has been combined with other two nucleoside/nucleotide reverse transcriptase inhibitors in a single tablet and administered once-daily, which significantly improve adherence ¹⁾. However, its propensity for clinically important drug-drug interactions is high. Efavirenz is always used in combination therapy and in concert with drugs directed at the treatment of opportunistic infections, cancers and other HIV-related co-morbidities. The alterations in pharmacokinetics of co-administered drugs by efavirenz lead to either lack of efficiency or adverse drug reactions of the victim drug ¹⁾. In order to predict and avoid adverse drug interactions with efavirenz-based therapy, it is important to identify mechanisms underlying those drug interactions.

Efavirenz is known to alter the pharmacokinetics of a long list of co-administered drugs [Product Information of Efavirenz (Sustiva), Bristol-Myers Squibb Company, June 2012], probably by modulating the activities of multiple drug metabolizing enzymes and/or drug transporters. Efavirenz is a substrate of cytochrome P450s (CYPs) and UDP-glucuronosyltransferases (UGTs). It is predominantly cleared mainly by CYP2B6-mediated 8-hydroxylation, with small contribution from other CYPs (e.g., CYP2A6, CYP3A and CYP1A2) ^2-3). Two minor pathways, efavirenz 7-hydroxylation and N-glucuronidation, are predominantly catalyzed by CYP2A6 ³⁾ and UGT2B7 ⁴⁾, respectively. Efavirenz, through activation of constitutive androstane receptor (CAR) and/or pregnane X receptor (PXR), enhances the expression of multiple enzymes regulated by these nuclear receptors, including CYP2B6, CYP2C19 and CYP3A ^5-6). Therefore, many drug interactions associated with efavirenz at steady state including decreased exposures of methadone ⁷⁾, statins ⁸⁾, omeprazole ⁹⁾, voriconazole ¹⁰⁾, proguanil ¹¹⁾ and etravirine ¹²⁾, and protease inhibitors ¹³⁾ can be primarily explained by the inductive effect of efavirenz. Furthermore, efavirenz enhances its own metabolism (auto-induction) upon repeated administration compared to a single dose ¹⁴⁾, probably through induction of CYP2B6 and other enzymes involved in its metabolism ^2-4). Besides induction, there is in vitro evidence that efavirenz may directly inhibit the activities of certain CYPs ^15-18). Indeed, scattered clinical cases of adverse drug interactions, e.g., with amodiaquine ¹⁹⁾, warfarin ²⁰⁾ and phenytoin ²¹⁾ suggest that efavirenz may alter the pharmacokinetics of co-administered drugs through inhibition of CYPs.

A comprehensive inhibitory analyses that encompass all major drug-metabolizing CYPs are important because: a) not all pharmacokinetic drug interactions involving efavirenz can be explained by the known inductive effect of efavirenz and by the CYPs studied so far; and b) the in vitro studies describing inhibition of CYPs by efavirenz provide only qualitative information, without generating in vitro inhibition parameters that will allow quantitative prediction of in vivo condition and without taking the contribution of time-dependent inactivation into account. In addition, the net effect on the pharmacokinetics of co-administered drugs seems to depend on its varied potencies of inhibition and induction on individual CYP isoform. In order to better predict in vivo drug-drug interactions associated with such mixed mechanisms, it is necessary to simultaneously take reversible inhibition, time-dependent inhibition, and induction into account.

The purpose of present study was to systematically evaluate the in vitro inhibitory potency of efavirenz on eight major human CYP isoforms and determine the mechanisms involved. For those isoforms that were inhibited in pilot experiments, inhibition constants (K_i values) were estimated with which the extent of in vivo drug interactions was quantitatively predicted.

Materials and Methods

Chemicals

Efavirenz, 8-hydroxyefavirenz, 7-hydroxycoumarin, bupropion, 4-hydroxybupropion, desethylamodiaquine, S-mephenytoin, 4-hydroxymephenytoin, R-omeprazole, R-hydroxyomeprazole and ritonavir were purchased from Toronto Research Chemicals (North York, Ontario, Canada). Acetaminophen, chloroquine, coumarin, dextromethorphan, dextrorphan, desmethyldiazepam, 8-methoxypsolaren, phenacetin, tolbutamide, 4-hydroxytolbutamide, testosterone, 6-β hydroxytestosterone, glucose-6-phosphate, NADP and glucose-6-phosphate dehydrogenase were purchased from Sigma-Aldrich (St. Louis, MO). Amodiaquine and levallorphan were purchased from the United States Pharmacopeia (Rockville, MD). All the other chemicals were of high performance liquid chromatography (HPLC) grade.

Microsomal preparations

Pooled human liver microsomes (HLMs) from 24 individual donors, HLMs with CYP2C8*3/*3 genotype, and other HLMs were obtained from BD Biosciences (Woburn, MA). Human CYP2B6 and CYP2C8 expressed in baculovirus infected insect cells with oxidoreductase and without co-expression of cytochrome (Cyt) b5 were obtained from BD Biosciences. All microsomal preparations were stored at −80°C until analysis.

General incubation conditions

Using incubation conditions specific to each isoform that were linear for time, substrate and protein concentrations, isoform selective substrate probes were incubated in duplicate at 37°C with HLMs (or with expressed CYP when required), 200 mM sodium phosphate reaction buffer (pH 7.4) and NADPH-regenerating system (1.3 mM NADP, 3.3 mM glucose-6-phosphate, 3.3 mM MgCl₂, and 1 μl/ml glucose-6-phosphate dehydrogenase) in the absence or presence of varying concentrations of efavirenz (or 8-hydroxyefavirenz). The test inhibitors were dissolved and diluted in methanol to required concentrations and methanol was removed by drying in speed vacuum before the addition of the incubation components. The following HLMs concentrations were used: 1 mg protein/ml for CYP1A2, CYP2C9, CYP2C19 and CYP2D6; 0.5 mg protein/ml for CYP2A6, 0.25 mg protein/ml for CYP2B6 and CYP3A; and 0.1 mg/ml for CYP2C8 incubations. Inhibition constants (K_i values) were determined in expressed CYP2C8 (26 pmol) and CYP2B6 (5 pmol).

Enzyme activity assays

The inhibitory effects of efavirenz on the activities of different CYP isoforms were studied in HLMs, expressed CYP2B6 and CYP2C8 using the following selective reaction probes: phenacetin O-deethylation (CYP1A2); coumarin 7-hydroxylation (CYP2A6); bupropion 4-hydroxylation (CYP2B6); amodiaquine N-desethylation (CYP2C8); tolbutamide 4-hydroxylation (CYP2C9); S-mephenytoin 4-hydroxylation/R-omeprazole 5-hydroxylation (CYP2C19); dextromethorphan O-demethylation (CYP2D6); and testosterone 6-β hydroxylation (CYP3A). The methods to measure the activity of CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2D6 and CYP3A in the absence or presence of the test inhibitors were adopted from previous studies ^22-23). LC/MS/MS assays were developed for quantification of bupropion 4-hydroxylation (CYP2B6) as described elsewhere²⁴. The MS/MS system was an API2000 MS/MS triple quadruple system (Applied Biosystems, Foster City, CA) equipped with a turbo ion spray and was coupled with a Shimadzu (Columbia, MD) HPLC system consisting of an LC-20AB pump and SIL-20A HT autosampler (Applied Biosystems/MDS Sciex, Foster City, CA). In Brief, bupropion, 4-hydroxybupropion, and the internal standard (nevirapine) were separated using Zorbax SB-C18 column (100 × 2.00 mm, 3-μm particle size), a Luna C18 guard column (30 × 4.6 mm, 5 μm), and an isocratic mobile phase that consisted of 75% formic acid (0.1% in H₂O) and 25% acetonitrile (flow rate, 0.3 ml/min). 4-Hydroxybupropion and nevirapine were detected using multiple reactions monitoring at a m/z of 256.1/238.0 and 267.2/224.4 in positive ion mode, respectively. R-omeprazole 5-hydroxylation (CYP2C19) was assayed as described previously ²⁵⁾. The MS/MS system was the same as that for bupropion hydroxylation assay above. R-Hydroxyomeprazole, and the internal standard (R-lansoprazole) were separated by Chiral-AGP (150 × 4.60 mm; 5 μM; Phenomenex). A gradient elution profile was used: initial mobile phase, 95% of 20 mM ammonium acetate (adjusted to pH 6.5) and 5% of acetonitrile (v/v); and the secondary mobile phase consisted of 90% of acetonitrile and 10% of 20 mM ammonium acetate (adjusted to pH 6.5) (v/v). The secondary mobile phase was increased from 0% to 40% linearly between 0 and 8 min; the initial mobile phase conditions were resumed after 9 min and remained constant for an additional 6 min, allowing the column to equilibrate. The selected reaction-monitoring transitions of the precursor ions to selected product ions were m/z 362.13/214.10 for R-5-hydroxyomeprazole and m/z 370.25/252.30 for the internal standard (R-lansoprazole).

Kinetic analysis

Kinetic analysis was performed for each substrate probe reaction before initiation of the inhibition experiments, and the data generated were used as a guide for selection of the appropriate concentrations of the substrate probes in the subsequent inhibition experiments. Thus, the kinetic parameters for the metabolism of each probe substrate were determined by incubating a range of different concentrations of the substrate (without the inhibitor) at 37°C in duplicate with HLMs (or expressed enzymes) and the NADPH-generating system. Phenacetin (5 to 1000 μM), coumarin (0.1 to 50 μM), bupropion (1 to 1000 μM), amodiaquine (0.1 to 100 μM), tolbutamide (5 to 500M), S-mephenytoin (5 to 100 μM), R-omeprazole (1 to 200 μM), dextromethorphan (1 to 200 μM), and testosterone (1 to 200 μM) were used. Formation rates of metabolite versus the substrate concentrations were fit to appropriate enzyme kinetic equations to estimate the apparent kinetic parameters.

Screening of inhibition on multiple CYPs

A single isoform-specific substrate concentration at about the respective K_m value (50 μM phenacetin, 10 μM coumarin, 50 μM bupropion, 25 μM amodiaquine, 150 μM tolbutamide, 50 μM S-mephenytoin, 25 μM R-omeprazole, 10 μM dextromethorphan, and 10 μM testosterone) was incubated at 37°C in duplicate with HLMs and the NADPH-generating system in the absence or the presence of efavirenz concentrations at 10 and 50 μM, which covered the range of average steady-state C_max receiving 600mg once daily [Product Information of Efavirenz (Sustiva), Bristol-Myers Squibb Company, June 2012]. Processing of the incubation mixture and HPLC analysis of the metabolites formed were performed as described above.

Positive control experiments were run in parallel by incubating each probe substrate at 37°C in duplicate with HLMs and the NADPH-generating system in the absence (control) and the presence of the following isoform-specific inhibitors: furafylline (20 μM; specific for CYP1A2), pilocarpine (50 μM; specific for CYP2A6), thioTEPA (50 μM; specific for CYP2B6), ticlopidine (5 μM; specific for CYP2B6), quercetin (10 μM; specific for CYP2C8), sulfaphenazole (25 μM; specific for CYP2C9), ticlopidine (5 μM; specific for CYP2C19), quinidine (1 μM; specific for CYP2D6), and ketoconazole (1 μM; specific for CYP3A). The substrate probes and concentrations that were used in the screening experiments (see above) were used for these positive control experiments. Formation rates of the metabolites in the presence of the isoform-specific inhibitor were compared with that for controls in which the inhibitor was replaced with vehicle.

Determination of inhibition constants (K_i values)

In pilot experiments, efavirenz showed inhibition on the activities of CYP2B6, CYP2C8, CYP2C9, CYP2C19 and CYP3A (by >20% at 50 μM); inhibition on the other CYPs (CYP1A2, 2A6, and 2D6) was minimal. Therefore, Dixon plots for the inhibition of CYP2B6, CYP2C8, CYP2C9, CYP2C19 and CYP3A were determined by incubating multiple concentrations of the respective substrate probe in the presence and absence of multiple concentrations of efavirenz with HLMs and cofactors. K_i values in expressed enzymes were determined for two CYP isoforms that showed potent inhibition in pooled HLMs with K_i < 10μM, i.e. CYP2B6 and CYP2C8. The following isoform-specific probe substrate concentrations were used: 25 to 75 μM bupropion for CYP2B6; 10 to 100 μM amodiaquine for CYP2C8; 50 to 250 μM tolbutamide for CYP2C9; 15 to 75 μM S-mephenytoin for CYP2C19 and 5 to 50 μM testosterone for CYP3A. The concentrations of efavirenz used were 0 to 100 μM. The inhibition data obtained from the pilot experiments were used as a guide to generate appropriate probe substrate and test inhibitor concentrations for the determination of the K_i values for each isoform.

Assessment of time-dependent inhibition

Efavirenz was reported to be a time-dependent inhibitor of expressed CYP2B6, with a K_I value of 30 μM ¹⁸⁾. Therefore, time-dependent inhibition was tested on eight major CYPs except for CYP2B6 using pooled HLMs. Efavirenz (50 μM) was pre-incubated in duplicate with HLMs and 200 mM sodium phosphate reaction buffer (pH 7.4) (without or with the NADPH-generating system) in the absence of a substrate probe for 0, 5, 10, 15 and 30 min at 37°C. The preincubation reaction was started by adding the NADPH-generating system. Controls were pre-incubated for 0 min without efavirenz and without the NADPH generating system. The total volume of the preincubation mixture was 600 μl. After 0, 5, 10, 15 and 30 min of preincubation, 50 μl of preincubation mixture was added to a glass tube containing 950 μl mixture that consisted of a substrate (final concentration corresponding to the V_max), phosphate reaction buffer and NADPH-generating system. The mixture was further incubated for the time specific for each assay. The reaction was stopped and processed as described above for the co-incubation experiments.

Data analysis

Apparent kinetic constants (K_m, V_max) were estimated by fitting formation rates of metabolites versus substrate concentrations to simple single-site Michealis-Menten equation by nonlinear regression analysis using SigmaPlot 11.0 (Systat Software Inc., Richmond, CA). To calculate K_i values, the inhibition data were fit to different models of enzyme inhibition (competitive, noncompetitive, and uncompetitive) by nonlinear least-squares regression analysis with the Prism Version 5.0 software (GraphPad software Inc., San Diego, CA, USA). The final model for each data set was selected on the basis of visual inspection of Lineweaver-Burk, Dixon, and Eadie-Hofstee plots, as well as the size of the sum of squares of residuals, the Akaike information criterion, and Schwartz criterion values.

Prediction of in vivo drug interactions

Predictions of in vivo drug interaction potential of efavirenz were made using the following equations, which has been described previously ²⁶⁾ :

$$\frac{{\mathrm{AUC}}_{\mathrm{I}}}{\mathrm{AUC}}=\frac{1}{\frac{{\mathrm{f}}_{\mathrm{m}}}{1+\frac{\left[\mathrm{I}\right]}{{\mathrm{K}}_{\mathrm{i}}}}+(1-{\mathrm{f}}_{\mathrm{m}})}$$ (1)

where: $\frac{{\mathrm{AUC}}_{\mathrm{I}}}{\mathrm{AUC}}$ is the ratio of the AUC of substrate after inhibition to the AUC of the uninhibited substrate; f_m is the fraction of substrate metabolized by the inhibited CYP pathway. The utilities of four different values for in vivo inhibitor concentrations, which are the total systemic C_max, free systemic C_max, total hepatic inlet C_max estimated after oral administration and free hepatic inlet C_max in the prediction of drug interactions have been compared before and estimation using free hepatic inlet C_max yielded the most accurate predictions of the magnitude of drug interactions ²⁶⁾. But efavirenz is highly protein bound with fraction unbound in plasma estimated to be 0.029 ²⁷⁾. In order to avoid underestimation of potential risk for drug interaction, total hepatic inlet C_max (C_hep,inlet) of efavirenz were used in present study (equation 2), which also showed a reasonably good prediction in the previous study ²⁶⁾.

$${\mathrm{C}}_{\mathrm{hep},\mathrm{inlet}}={\mathrm{C}}_{\max }+\frac{{\mathrm{K}}_{\mathrm{a}}{\mathrm{F}}_{\mathrm{a}}\mathrm{D}}{{\mathrm{Q}}_{\mathrm{h}}}$$ (2)

Where: C_max is maximum plasma concentration; K_a is the absorption rate constant, which is estimated to be 0.3 h^{-1 28)}; F_a is the fraction of the inhibitor passing through the intestine unchanged; D is the administered therapeutic dose (600 mg/day); and Q_h is hepatic blood flow (87 L/h) ²⁹⁾. Values of F_a can be estimated from oral bioavailability and hepatic extraction. Since both values are not available for efavirenz, a value of unity for F_a was assumed as the most cautious possibility ^29-30). There is no evidence that efavirenz enhances the activities of CYP2C8 and CYP2C9 in vivo. Therefore, C_max of efavirenz at steady state was used (9.2 ~ 16.6 μM) to predict the AUC changes of CYP2C8 and CYP2C9 substrates [Product Information of Efavirenz (Sustiva), Bristol-Myers Squibb Company, June 2012]. However, efavirenz has been shown to enhance the activities of CYP2B6, CYP2C19 and CYP3A upon multiple dosing, suggesting inhibition effect may masked by inductive effect of efavirenz in a time-dependent manner. Therefore, predicting AUC changes CYP2B6, CYP2C19 and CYP3A substrates was estimated using C_max (4.6 ~ 8.4 μM) obtained after the administration of a single 600 mg oral dose of efavirenz to 20 healthy volunteers (unpublished data). Specifically, we focused on substrates that exhibit narrow therapeutic range and thus initiation of efavirenz to patients stabilized on these drugs may increase the risk of adverse effects (methadone) or failure of therapy (clopidogrel and proguanil). Of note, clopidogrel and proguanil are pro-drugs that require conversion to pharmacologically active metabolites primarily by CYP2C19 ^31-33). For these prodrugs that require conversion by CYPC19 to active metabolites, the following equation published previously ³⁴⁾ was slightly modified to predict the extent of inhibition of bioactivation:

$$\frac{{\mathrm{AUC}}_{\mathrm{m}}}{{\mathrm{AUC}}_{\mathrm{p}}}=\frac{{\mathrm{f}}_{\mathrm{u},\mathrm{p}}\times {\mathrm{Cl}}_{\mathrm{f}}}{{\mathrm{f}}_{\mathrm{u},\mathrm{m}}\times {\mathrm{Cl}}_{\mathrm{m}}}$$ (3)

where $\frac{{\mathrm{AUC}}_{\mathrm{m}}}{{\mathrm{AUC}}_{\mathrm{p}}}$ is the ratio of the AUC of metabolite to the AUC of parent compound. The f _u,p and f _u,m are the plasma fraction unbound of the parent and metabolite. Cl _f and Cl _m are the formation and elimination clearance of metabolite.

The majority of clopidogrel is metabolized by an esterase ³⁵⁾ and relative importance of CYP2C19 for the overall elimination of proguanil is much lower than it is for the formation of cycloguanil ³²⁾. Assuming that efavirenz does not affect the elimination of metabolite and the change in the AUC of parent is negligible, the equation 4 can be derived from equation 1 and 3 as:

$$\frac{{\mathrm{AUC}}_{\mathrm{m}}}{{\mathrm{AUC}}_{\mathrm{m}},\mathrm{I}}=\frac{{\mathrm{Cl}}_{\mathrm{f}}}{{\mathrm{Cl}}_{\mathrm{f}},\mathrm{I}}=\frac{1}{\frac{{\mathrm{f}}_{\mathrm{m}}}{1+\frac{\left[\mathrm{I}\right]}{{\mathrm{K}}_{\mathrm{i}}}}+(1-{\mathrm{f}}_{\mathrm{m}})}$$ (4)

where $\frac{{\mathrm{AUC}}_{\mathrm{m}}}{{\mathrm{AUC}}_{\mathrm{m}},\mathrm{I}}$ is the ratio of the uninhibited AUC of metabolite to the AUC of metabolite with inhibition. $\frac{{\mathrm{Cl}}_{\mathrm{f}}}{{\mathrm{Cl}}_{\mathrm{f}},\mathrm{I}}$ is the ratio of the uninhibited formation clearance to the formation clearance after inhibition. f_m for clopidogrel and proguanil is defined as the fraction of active metabolites formed by CYP2C19. Since f_m for proguanil is not available in the literature, it was estimated using a pharmacogenetic method modified from a recent study for CYP2D6 substrate ³⁶⁾. The original method was based on the observation that the ratio of AUC in poor metabolizer (PM), AUC^PM, to the AUC in extensive metabolizer (EM), AUC^EM. Assuming that genetic polymorphisms do not affect the elimination of metabolite and has limited effects on the AUC of the parent compound, the value of f_m for proguanil metabolized to cycloguanil by CYP2C19 was estimated from a clinical study ³²⁾ using the following equation:

$$\frac{{{\mathrm{AUC}}_{\mathrm{m}}}^{\mathrm{EM}}}{{{\mathrm{AUC}}_{\mathrm{m}}}^{\mathrm{PM}}}=\frac{1}{1-{{\mathrm{f}}_{\mathrm{m}}}^{\mathrm{EM}}}$$ (5)

where $\frac{{{\mathrm{AUC}}_{\mathrm{m}}}^{\mathrm{EM}}}{{{\mathrm{AUC}}_{\mathrm{m}}}^{\mathrm{PM}}}$is the ratio of metabolite AUC in poor metabolizer (PM) to the metabolite AUC in extensive metabolizer (EM).

Results

Screening for inhibition of multiple CYPs by efavirenz

The inhibitory effect of efavirenz at 10 and 50 μM on the activities of eight CYP isoforms in pooled HLMs is shown in Figure 1. Efavirenz was a potent inhibitor of CYP2B6 (by 90% at 10 and 50 μM). It also showed inhibition of CYP2C8, CYP2C9 and CYP2C19 by >20% at 10 μM and by >50% at 50 μM. Efavirenz only showed weak inhibition on the activity of CYP3A (testosterone β-hydroxylation) by 10% at 10 μM and by 40% at 50 μM.

Figure 1. Inhibitory effects of efavirenz on human CYP activities in pooled HLMs.

A substrate probe of each specific CYP isoform at a single concentration was incubated with pooled HLMs and cofactors in the absence (control) or the presence of efavirenz (10 and 50 μM) for times and with protein concentrations that were linear for the respective reaction described in detail in Material and Methods. The specific concentrations of each probe used are illustrated in Materials and Methods. Each point represented the average of duplicate incubations. EFV, efavirenz.

The activity of CYP2C19 was assessed using two substrates (R-omeprazole and S-mephenytoin) since substrate-dependent effect on CYP2C19 inhibition profile was observed previously ³⁷⁾. In the present study, efavirenz inhibited CYP2C19 activity by 35% and 70% at 10 and 50 μM, when S-mephenytoin was used as a substrate (Figure 1), but its effect on CYP2C19-mediated R-omeprazole 5-hydroxylation was marginal (by 10% at 50 μM efavirenz) (data not shown). This result is consistent with a previous report that S-mephenytoin is more sensitive to CYP2C19 inhibition than R-omeprazole ³⁷⁾. Therefore, S-mephenytoin was used as a substrate of CYP2C19 in subsequent inhibition experiment.

We have in vivo evidence that efavirenz reduces CYP1A2 activity, as measured by caffeine metabolism: compared to a single efavirenz dose (600 mg orally), pretreatment with efavirenz (600 mg/day for 17 days) significantly decreased 5-7hrs concentration ratio of paraxanthine/caffeine (from 0.52+0.28 single dose to 0.31+0.18 multiple dose; P<0.0001) ³⁸⁾. However, the present in vitro data derived from pooled HLMs shown in Figure 1 did not indicate that efavirenz inhibits CYP1A2 activity (Figure 1). Therefore, we tested whether the major metabolite of efavirenz, 8-hydroxyefavirenz, contributes to inhibition of CYP1A2 and showed that 8-hydroxyefavirenz inhibited CYP1A2 by ~20% up to 10μM (Figure 2), which suggests that other alternative mechanisms should account for the reduced CYP1A2 activity that we observed in vivo.

Figure 2. Inhibition of CYP1A2 by 8-hydroxyefavirenz in pooled HLMs.

Phenacetin (50 μM) was incubated with pooled HLMs (1 mg/ml) and cofactors in the absence (control) or the presence of 8-hydroxyefavirenz (0.5 to 10 μM) at 37°C for 30 min. Each point represented the average of duplicate incubations. 8OHEFV, 8-hydroxyefavirenz.

The inhibitory effect of efavirenz on the activity of CYP2A6 and CYP2D6 was negligible (less than 10% at both efavirenz concentrations) (Figure 1).

Estimation of K_i values

In order to obtain quantitative prediction of magnitude of drug interaction in vivo, further experiments were performed to determine the K_i values for the inhibition of CYP2B6, 2C8, 2C9 and 2C19 by efavirenz. Although relatively weak inhibition of CYP3A was observed by efavirenz, K_i value was determined in pooled HLMs because a previous study reported that the value of IC₅₀ is around 20 μM using midazolam as a substrate ¹⁶⁾.

Of all the CYPs tested, CYP2B6 was the most sensitive to efavirenz inhibition (Table 1). Visual inspection of the Dixon plot and further analysis of the parameters of the enzyme inhibition models suggested that the inhibition data fit well to a competitive type of inhibition. The K_i values estimated by using a nonlinear regression model for competitive enzyme inhibition of CYP2B6-catalyzed bupropion 4-hydroxylation in pooled HLMs and CMV negative HLMs were 2.96 ± 0.67 μM and 0.39 ± 0.10 μM, respectively. K_i value determined in expressed CYP2B6 was 1.38 ± 0.09 μM. Representative Dixon plots for the inhibition of CYP2B6 in CMV negative HLMs and expressed CYP2B6 are shown in Figure 3A and 3B, respectively.

Table 1.

K_i values of efavirenz for the inhibition of CYPs in HLMs and expressed CYPs.

CYP Isoform	Substrate	Systems	Ki value (μM) (inhibition model)
CYP2B6	Bupropion	pooled HLMs	2.96 ± 0.67 (competitive)
		CMV negative HLMs	0.39 ± 0.10 (competitive)
		expressed CYP2B6	1.38 ± 0.09 (competitive)
CYP2C8	Amodiaquine	pooled HLMs	4.78 ± 2.24 (competitive)
		HLMs with CYP2C8*3/*3	4.80 ± 0.35 (competitive)
		expressed CYP2C8	6.05 ± 2.86 (competitive)
CYP2C9	Tolbutamide	pooled HLMs	19.46 ± 2.78 (non-competitive)
CYP2C19	S-mephenytoin	pooled HLMs	21.31±2.57 (competitive)
CYP3A	Testosterone	pooled HLMs	40.33 ± 0.33 (competitive)

Figure 3. Dixon plots for the inhibition of bupropion 4-hydroxylation by efavirenz in CMV Negative HLMs (A) and expressed CYP2B6 (B).

Bupropion (25 to 75 μM) was incubated with CMV negative HLMs (0.25 mg/ml) or expressed CYP2B6 (5 pmol) and cofactors at 37°C for 15 min with or without efavirenz (0-1 μM). Each point represented the average of duplicate incubations.

Inhibition of CYP2C8 by efavirenz was determined in two HLMs and expressed CYP2C8. As shown in Table 1, efavirenz showed potent competitive inhibition of CYP2C8 activity in pooled HLMs (K_i = 4.78 ± 2.24 μM). The second HLMs was obtained from human liver tissues with the CYP2C8*3/*3 genotype and the K_i value (4.80 ± 0.35 μM) derived from this HLM was not different from that derived from pooled HLMs (Table 1). Efavirenz exhibited similar competitive inhibition in expressed CYP2C8 with an estimated K_i value of 6.05 ± 2.86 μM (Table 1). In Figure 4, Dixon plots for the inhibition of CYP2C8-catalyzed N-desethylation of amodiaquine by efavirenz in pooled HLMs (Figure 4A), HLMs with CYP2C8*3/*3 genotype (Figure 4B) and expressed CYP2C8 (Figure 4C) are shown.

Figure 4. Dixon plots for the inhibition of amodiaquine desethylation by efavirenz in pooled HLMs (A), HLMs with CYP2C8*3/*3 genotype (B) and expressed CYP2C8 (C).

Amodiaquine (10 to 100 μM) was incubated with pooled HLMs (0.1 mg/ml) or HLMs with CYP2C8*3/*3 genotype (0.1 mg/ml) and cofactors at 37°C for 15 min with or without efavirenz (0-50 μM). Expressed CYP2C8 (26 pmol) was used in the inhibition study. Each point represented the average of duplicate incubations.

Efavirenz was found to be a moderate inhibitor of CYP2C9 (K_i =19.46 ± 2.78 μM; Table 1 and Figure 5) and CYP2C19 (21.31±2.57 μM; Table 1 and Figure 6), and a weak inhibitor of CYP3A (K_i = 40.33 ± 0.33 μM; Table 1 and Figure 7).

Figure 5. Dixon plot for the inhibition of tolbutamide 4-hydroxylation by efavirenz in pooled HLMs.

Tolbutamide (50 to 250 μM) was incubated with pooled HLMs (1 mg/ml) and cofactors at 37°C for 15 min with or without efavirenz (0-100 μM).

Figure 6. Dixon plot for the inhibition of S-mephenytoin 4-hydroxylation by efavirenz in pooled HLMs.

S-mephenytoin (15 to 75 μM) was incubated with pooled HLMs (1 mg/ml) and cofactors at 37°C for 15 min with or without efavirenz (0-100 μM).

Figure 7. Dixon plot for the inhibition of testosterone β-hydroxylation by efavirenz in pooled HLMs.

Testosterone (5 to 50 μM) was incubated with pooled HLMs (0.25 mg/ml) and cofactors at 37°C for 15 min with or without efavirenz (0-50 μM).

Time-dependent inactivation

As shown in Figure 8, efavirenz preincubation for 30 min only marginally inhibited the activity of those CYPs tested. Shorter preincubation times (5-15 min) were also tested, but did not show any indication of time-dependent inactivation.

Figure 8. Time-dependent inhibition of CYP isoforms by efavirenz in pooled HLMs.

Efavirenz (50 μM) was preincubated in duplicate with HLMs and phosphate reaction buffer (pH 7.4) (without or with the NADPH-generating system) in the absence of a substrate probe for 0, 5, 10, 15 and 30 min at 37°C. Controls were preincubated for 0 min without efavirenz and without the NADPH generating system. Protein concentrations and the specific concentrations of each probe used are illustrated in Materials and Methods. Each point represents the average of duplicate incubations.

Quantitative prediction of in vivo drug interactions

The predicted ratios of $\frac{{\mathrm{AUC}}_{\mathrm{I}}}{\mathrm{AUC}}$ for each substrate co-administered with a single dose or multiple doses of efavirenz are listed in Table 2. Compared to control (without efavirenz), a single 600 mg oral dose of efavirenz was predicted to result in ~3-fold changes in the exposure of methadone (CYP2B6 substrate). Also, based on the inhibition data generated using S-mephenytoin hydroxylation as a marker of CYP2C19, we predicted lower AUCs of active metabolites of clopidogrel and proguanil (by 17% ~ 29% and 29% ~ 33% respectively), and higher omeprazole AUC (by 1.4- to 1.6-fold) in extensive metabolizer of CYP2C19. However, when data generated using R-omeprazole 5-hydroxylation is used, no inhibition could be predicted in vivo. Based on our in vitro data, a single dose of efavirenz is unlikely to alter the pharmacokinetics of CYP3A substrates.

Table 2.

Prediction of changes in AUC of CYP2B6, 2C9 and 2C19 substrates in vivo by efavirenz.

	fm	Predicted AUC Ratio	Reported AUC Ratio
Methadone	0.75 27)	2.9 -3.1	N.A.
Amodiaquine	0.93 56)	3.5 -4.4	2.15 —4.02 19)
Phenytoin	0.90 57)	1.7 -2.0	N.A.
S-Warfarin	0.9126)	1.7 -2.0	N.A.
Omeprazole	0.87 26)	1.4 -1.6	N.A.
Proguanila	0.84	29% ~33%
Clopidogrela	0.56-0.6458)	17% ~29%	N.A.

Plasma concentrations of efavirenz after a single dose were used to predict its effect on AUC change of methadone, omeprazole and the active metabolites of proguanil and clopidogrel. Plasma concentrations upon multiple doses were used for predicting AUC change of amodiaquine, phenytoin and S-warfarin.

^a, The value of percentage change in the AUC of active metabolite (file) was predicted. N.A.: not available

After multiple doses of efavirenz, the AUC of CYP2C8 and CYP2C9 substrates was predicted to be ~3.5- to 4.4-fold higher (CYP2C8 substrate: amodiaquine) and 1.7- to 2.0-fold higher (CYP2C9 substrates: phenytoin and S-warfarin) compared to controls (without efavirenz).

Discussion

The primary aim of this study is to characterize the inhibition constant (K_i) of efavirenz in CYPs that showed inhibition and quantitatively predict its inhibition effect on the AUC of clinically important co-administered drugs. Previous in vitro studies only provided qualitative information (IC₅₀ values). In the present study, we have shown that efavirenz is a potent competitive inhibitor of CYP2B6 (average K_i= 1.68 μM in HLMs and K_i= 1.38 μM in expressed CYP2B6) and CYP2C8 (K_i = 4.78 μM in pooled HLMs and K_i = 4.80 μM in HLMs with CYP2C8*3/*3 genotype). In pooled HLMs, efavirenz showed moderate inhibition of CYP2C9 (K_i= 19.46 μM) CYP2C19 (K_i= 21.31 μM),and a weak inhibitor of CYP3A (K_i= 40.33 μM). Inhibition of CYP1A2, CYP2A6 and CYP2D6 by efavirenz was marginal. No time-dependent inactivation of the CYP isoforms tested was observed. Based on the in vitro to in vivo quantitative prediction, efavirenz is expected to : a) increase the AUC of methadone (CYP2B6 substrate) by 2.9- to 3.1-fold, omeprazole by 1.4- to 1.6-fold (CYP2C19 substrate),and may also lower the AUC of active metabolites of some pro-drugs (e.g. clopidogrel and proguanil ) by up to 30% during initiation of efavirenz-based anti-HIV therapy; and b) increase the AUC of amodiaquine by ~3-fold (CYP2C8 substrate) and phenytoin and warfarin by 1.7- to 2.0-fold (CYP2C9 substrates) during a single dose or multiple doses of efavirenz. Our data suggest that efavirenz may increase the risk for adverse effects by increasing the exposure of the parent drug or reduce efficacy by diminishing the formation of pharmacologically active metabolites from prodrugs.

Of the CYPs tested, CYP2B6 was most sensitive to efavirenz inhibition with K_i value of ~1.7 μM in HLMs and ~1.38 μM in expressed CYP2B6. Although the ability of efavirenz to inhibit CYP2B6 was previously reported ^17-18), the present data provide key information that allowed in vivo quantitative prediction of the magnitude of interaction. The high inhibition potency of efavirenz in our study is worth commenting. Efavirenz has a higher binding affinity to CYP2B6 with K_m values of 13 ~ 20 μM ^2-3) than bupropion with K _m values of 90 ~ 130 μM ^39-40). Thus, it is plausible that the high inhibition potency of efavirenz on bupropion hydroxylation could be due to the fact that efavirenz has higher binding affinity to CYP2B6 than bupropion. Similar mechanism contributing to high inhibition potency has been reported for CYP2D6 ⁴¹⁾. To put the in vitro inhibition data on CYP2B6 into perspective, it is important to point out that efavirenz enhances its own metabolism upon multiple doses preferentially through CAR-mediated induction of CYP2B6 ⁴²⁾. Efavirenz also enhances the metabolism of co-administered CYP2B6 substrates, including methadone ^{7, 43)} and bupropion ⁴⁴⁾. Considering the high inhibition potency of efavirenz, a substantial increase in AUC of CYP2B6 substrates and potentially the risk to adverse effects may be expected, when efavirenz-based therapy is initiated in patients who are stabilized on CYP2B6 substrates. We predicted approximately 2.9- to 3.1-fold increase in methadone AUC, when a single 600 mg oral dose of efavirenz is co-administered. However, during chronic administration, inhibition of CYP2B6 by efavirenz appears to be masked by its marked induction and the net effect becomes induction.

Our study demonstrates that efavirenz inhibits CYP2C8-mediated amodiaquine desethylation with K_i values of 4.78 and 6.05 μM in pooled HLMs and expressed CYP2C8 respectively, which broadly concurs with an IC₅₀ of 4μM reported in expressed CYPs ¹⁵⁾. The inhibition potency of efavirenz in HLMs with the CYP2C8*3/*3 genotype, the most frequent and functionally relevant variant in Caucasians ⁴⁵⁾, was not different from that observed in pooled HLMs. The possibility of substrate-dependent interaction cannot be fully excluded ⁴⁶⁾, but our data suggest that the CYP2C8*3 allele does not seem to alter susceptibility to efavirenz inhibition. We expect ~3-fold higher AUC of amodiaquine and probably other substrates such as chloroquine, certain anti-diabetics, montelukast and rosiglitazone ^47-48), when co-administered with efavirenz. A clinical study that was designed to evaluate drug interactions between anti-malarials and efavirenz-based anti-HIV therapy was prematurely discontinued after the first two subjects developed hepatotoxicity ¹⁹⁾. A 2.2- to 4-fold increase in amodiaquine AUC was also noted ¹⁹⁾ and it is highly likely that this interaction occurred through inhibition of CYP2C8, as predicted from our in vitro data.

Our data showed that efavirenz inhibits CYP2C9 activity (K_i= 19.46 μM) with 1.7- to 2-fold predicted increase in AUC of drugs mainly cleared by CYP2C9, consistent with an in vitro study reporting an IC₅₀ value of ~15μM ¹⁶⁾. This enzyme is involved in the metabolism of more than 100 currently used drugs, including drugs with narrow therapeutic range, e.g., oral anticoagulants, oral hypoglycemic agents and phenytoin ⁴⁹⁾. Therefore, co-administration of efavirenz may likely increase the risks to adverse effects of these drugs, which is supported by clinical cases of inhibition drug interactions of efavirenz with the CYP2C9 substrate phenytoin ²¹⁾ and warfarin ²⁰⁾ .

We found that the extent of CYP2C19 inhibition by efavirenz was substrate-dependent: modest inhibition of S-mephenytoin 4-hydroxylation (K_i= 21.31 μM) and marginal inhibition of R-omeprazole 5-hydroxylation consistent with a previous study reporting that R-omeprazole is less sensitive to CYP2C19 inhibitors than S-mephenytoin ³⁷⁾. The clinical relevance of efavirenz inhibition on CYP2C19 and the mechanism of substrate-dependent inhibition remain unclear, but this interaction may be important for prodrugs that require bioactivation by CYP2C19. For example, at a single dose or acute dosing, efavirenz may inhibit the formation of active metabolite of clopidogrel by CYP2C19 and to some extent by CYP2B6 and CYP2C9 ⁵⁰⁾. Provided that efavirenz is a potent inhibitor of CYP2B6 and a moderate inhibitor of CYP2C19 and CYP2C9 (present data), it is likely that administration of efavirenz to patients stabilized on clopidogrel may result in substantially reduced formation of active metabolite and lack of efficacy. At steady state, efavirenz may be a mixed inhibitor and inducer of CYP2C19 ⁹⁾, although induction appears to predominate as shown with the enhanced metabolism of CYP2C19 substrates including voriconazole ¹⁰⁾, omeprazole ⁹⁾, etravirine ¹²⁾. Contradictory effects were observed in another CYP2C19 substrate proguanil, with decreased proguanil exposure co-administered with 600 mg/day efavirenz for 1 month ¹¹⁾ and increased proguanil exposure co-administered with 400 mg/day efavirenz for 9 days ⁵¹⁾. This discrepancy may be due to differences in dose regimens, duration of treatment and study populations.

Efavirenz-mediated in vivo inhibition of CYP3A seems unlikely given the high K_i value (K_i= ~40μM) observed in the present study. Using triazolam as a substrate, another study reported lower IC₅₀ values (17-20μM) ¹⁶⁾, but the significance of in vivo inhibition of CYP3A, if any, is likely marginal. Efavirenz, through activation of PXR/CAR, induces CYP3A in vitro ^5-6) and in vivo ^{9, 52)}. Hence, efavirenz enhances the elimination of many CYP3A substrates, including protease inhibitors, statins ⁸⁾, calcium channel blockers and anti-fungals [Product Information of Efavirenz (Sustiva), Bristol-Myers Squibb Company, June 2012]. Therefore, induction drug interactions between efavirenz and CYP3A substrates appear to predominate at steady-state.

CYP1A2, CYP2A6 and UGT2B7 all have been shown to be involved in efavirenz metabolism ^{2, 4)}. Efavirenz is not only a substrate but also a moderate inhibitor of UGT2B7 ⁴⁾, while no reversible inhibition or time-dependent inhibition by efavirenz was observed for CYP1A2 and CYP2A6 in the present study. The drug interaction with CYP2A6 substrates mediated by efavirenz inhibition seems very unlikely, but the possibility of efavirenz inhibiting the metabolism of CYP1A2 substrates can't be excluded, because the preliminary results from our laboratory showed that efavirenz reduces CYP1A2 activity as measured by caffeine metabolism in vivo.

The use of efavirenz is made difficult by high interindividual variability in its disposition and by often unpredictable and complex drug interactions. The extent of drug interactions with efavirenz varies greatly among individuals, and interpatient differences in efavirenz exposure contribute to this variability. Efavirenz exposure is governed by complex factors: efavirenz is mainly cleared by CYP2B6, with some contribution from accessory pathways catalyzed by enzymes that include CYP2A6, CYP1A2, CYP3A and UGT2B7 ^{2-4, 53-55)}; efavirenz activates CAR and PXR, induces CYP2B6 (and other drug disposition genes) and autoinduces its own metabolism ⁴²⁾ upon repeated administration; and efavirenz is a potent inhibitor of CYP2B6 with no effect on CYP2A6 and CYP1A2 (present study) and a moderate inhibitor of UGT2B7 in vitro ⁴⁾. These complex inhibition/induction processes and genetic variations of CYP2B6 would likely contribute to variable efavirenz exposure and drug interactions. The net effect of efavirenz on drug interactions (induction versus inhibition) in vivo is likely to be dependent on: duration of efavirenz administration (acute versus chronic); genetic and nongenetic factors; and the potency with which efavirenz induces or inhibits drug metabolizing enzymes. Although predicting the extent and direction of drug interactions with efavirenz in vivo might be difficult for the individual patient, general comments could be made based on our data and the literature. During initiation of efavirenz-based therapy, it is expected that efavirenz reduces the elimination of CYP2B6, CYP2C8, CYP2C9 and CYP2C19 substrates. Mixed induction/inhibition occurs upon repeated administration of efavirenz, but the net effect of efavirenz appears to be induction for CYP2B6 and CYP2C19, while inhibition appears to dominate for CYP2C8 and CYP2C9. Induction appears the main mechanism for efavirenz interactions involving CYP3A. Together, efavirenz's complex interaction with enzymes involved in its own metabolism and the metabolism of co-administered drugs may contribute to the large interindividual variability of efavirenz exposure and unpredictable drug interactions associated with it.

Abbreviations

P450

cytochrome P450

HLMs

human liver microsomes

Cyt b5

cytochrome b5

HPLC

high performance liquid chromatography

LC/MS/MS

liquid chromatography/tandem mass spectrometry

EFV

efavirenz

8-OHEFV

8-hydroxyefavirenz

BUP

bupropion

SNP

single nucleotide polymorphism

K_m

Michaelis-Menten constant

V_max

maximum enzyme velocity

AUC

area under the plasma concentration-time curve

EM

extensive metabolizer

PM

poor metabolizer