Role of CYP2B6 in Stereoselective Human Methadone Metabolism

(RS )-(6-Dimethyamino-4,4-diphenyl-heptan-3-one) hydrochloride (RS -methadone), (S )-(6-dimethyamino-4,4-diphenyl-heptan-3-one) hydrochloride (S -methadone, 99% enantiomeric excess), troleandomycin, glucose-6-phospate, glucose-6-phospate dehydrogenase, and β-NADP were purchased from Sigma (St. Louis, MO). (R )-(6-Dimethyamino-4,4-diphenyl-heptan-3-one) hydrochloride (R -methadone, 99% enantiomeric excess) was purchased from Research Biochemicals Inc. (Natick, MA). (RS )-6-Dimethylamino-4,4-dipheny-1,1-trideuteroheptan-3-one (d3-methadone) was purchased from Cerilliant (Austin, TX). (±)-2-Ethyl-1,5-dimethyl-3,3-diphenylpyrrolidine perchlorate(EDDP) and [ethyl-2′,2′,2′-²H³]-3,3-diphenyl-2-ethyl-5-methyl-1-pyrroline hydrochloride (d3-EDDP) were obtained from the National Institute on Drug Abuse through Research Triangle Institute (Research Triangle Park, NC). Clopidogrel hydrogen sulfate was obtained from ChemPacific (Baltimore, MD), and (+)-N -3-benzyl-nirvanol was a generous gift from Allan Rettie, Ph.D. (Professor and Chairman, Department of Medicinal Chemistry, University of Washington, Seattle, WA). Chromatography-grade methanol, zinc sulfate, glacial acetic acid, and ammonium formate were purchased from Fisher Scientific (Pittsburgh, PA). Outdated human plasma was obtained from Puget Sound Blood Center (Seattle, WA) and was pooled from several donors. Baculovirus-insect cell microsomes (Supersomes®) containing expressed CYP3A4, CYP2B6, or CYP2C19 enzymes with coexpressed human cytochrome P-450 reductase and coexpressed human cytochrome b  ⁵were purchased from BD Gentest Corporation (Woburn, MA). Human liver microsomes (HLMs) were prepared as described previously.23Specific anti-CYP2B6 immunoquantitation (catalog No. 458226) antibody was also from BD Gentest Corporation. All stock drug solutions, buffers, and mobile phase were prepared using Milli-Q–grade water (Millipore, Bedford, MA).

Human liver samples (n = 60) were obtained from the University of Washington School of Pharmacy Human Tissue Bank. Relevant information about the bank tissues has been published previously.23Microsomes were prepared as described previously.24Protein concentrations were determined by the method of Lowry using bovine serum albumin as a reference. Immunoquantitation of CYP2B6 was performed by Western blot analysis using a specific CYP2B6 antibody (BD Gentest Corporation) as described.24Control experiments independently confirmed that the CYP2B6 antibody did not cross react with CYP1A2, CYP2C9/19, CYP3A4, or CYP3A5. The secondary antibody was an alkaline phosphatase–conjugated antibody, and quantitation was performed by integrating the optical density for each alkaline phosphatase substrate 5-bromo-4-cholor-3-indolyl phosphate/nitro blue tetrazolium developed band using a ChemiDoc Imager (Bio-Rad, Hercules, CA) and Quantity One Software (Bio-Rad). Unknown CYP2B6 protein levels were determined by comparison to a standard curve using CYP2B6 Supersomes® (BD Gentest; specific content 140 pmol/mg protein). Reported CYP2B6 protein contents are the average of three individual determinations measured on different occasions which did not vary by more than 20%. CYP3A4 and CYP3A5 protein content for the same HLMs were measured and reported elsewhere.23 

To assess the effect of CYP2B6 genotype on methadone metabolism, an array of liver samples was analyzed as follows: DNA was isolated from liver samples using commercially available kits (QIAGEN GmbH, Hilden, Germany). All single nucleotide polymorphism detection assays were performed in the DNA Sequencing and Gene Analysis Center at the University of Washington using an Applied Biosystems HT SDS Real-Time PCR System (Foster City, CA). Samples were genotyped using 5′-nuclease assays using mgb-probes 5′ labeled with either VIC or FAM and a black hole quencher (BHC) on the 3′ terminus. Samples were analyzed for the 785 A>G (CYP2B6*4, *6, or *7), 516 G>T (CYP2B6*9, *6, or *7), and 1459C>T (CYP2B6*5 or *7) variants. The primer and probe sets for 516 G>T and 1459C>T were purchased from Applied Biosystems under their Drug Metabolizing Enzymes genotyping product line. Genotyping for CYP2B6 785 A>G variant was performed using polymerase chain reaction followed by DNA sequencing (Big Dye version 3.1) on a capillary sequencing instrument (Applied Biosystems 3100) as described by Lang et al.  25 

Typical incubation conditions were as follows: All incubations were performed in 96-well plates using HLMs, CYP2B6, CYP3A4, or CYP2C19 Supersomes®. Incubations were performed in 100 mm potassium phosphate buffer (pH 7.4) and contained enzyme (0.1 mg/ml HLM, 10 pmol/ml CYP3A4 or CYP2B6, or 20 pmol/ml CYP2C19) and RS -, R -, or S -methadone. Incubations were equilibrated in a shaking water bath (37°C) for 5 min before the reaction was initiated with a nicotinamide adenine dinucleotide phosphate (NADPH) regenerating system (final concentrations: 10 mm glucose-6-phosphate, 1 mm β-NADP, 1 U/ml glucose-6-phosphate dehydrogenase, and 5 mm magnesium chloride, preincubated at 37°C for 10 min), unless otherwise indicated. Typically, incubations were quenched after 10 min (CYP2B6, CYP3A4) or 20 min (CYP2C19) with 15% zinc sulfate (for chiral analysis) or 20% trichloroacetic acid (for achiral analysis), and internal standard (d3-EDDP, 250 ng/ml) was added and then placed on ice before centrifugation for 10 min at 1,800g , unless other conditions are specified. Both the quench and the internal standard were added at 10% of the final incubation volume. The supernatant was transferred to a clean 450 μl conical bottom 96-well plate for EDDP analysis. Results were background subtracted (due to 0.2% and 0.4% EDDP in R - and S -methadone, respectively) using controls lacking NADPH.

To determine the contribution of individual isozymes to methadone metabolism in microsomal preparations, experiments using specific inhibitors for CYP2B6, CYP3A4, and CYP2C19 were performed. Incubations (final volume 0.25 ml), containing HLM (0.1 mg/ml), RS -methadone (2 μm) as substrate, and (+)-N -3-benzyl-nirvanol (final concentration 1 μm), were preincubated at 37°C for 5 min before initiation of the reaction by adding the NADPH regenerating system. Similar mixtures containing troleandomycin (final concentration 100 μm) and/or clopidogrel (final concentration 1 μm) were preincubated with enzyme, inhibitor, and NADPH regeneration system for 15 min before initiation of the reaction by adding methadone. Incubations containing all three inhibitors were performed similar to incubations containing troleandomycin alone. Control incubations (containing 0.17% methanol) were treated identically to those with inhibitor and were performed for all experiments. After 10 min reaction with substrate, incubations were quenched and prepared for chiral EDDP analysis as described above in General Incubation Conditions.

Incubations with HLMs (final volume 0.2 ml) were performed as described under General Incubation Conditions using a range of RS -methadone concentrations (0.5–1,500 μm). Experiments were performed under conditions where the rate of EDDP formation was linear with protein concentration (0.1 mg/ml) and time of incubation (10 min). Reactions were initiated by adding NADPH regenerating system and terminated after 10 min. Control incubations contained 100 mm potassium phosphate buffer instead of the NADPH regenerating system and were performed for all the concentrations used to generate the K^mand V^max. Determination of R - and S -EDDP by chiral analysis was performed as described below in Chiral Analysis of EDDP.

EDDP formation from R - or S -methadone metabolism was measured by liquid chromatography–mass spectrometry using selected ion monitoring after solid phase extraction, as described previously.26Incubations were quenched with 20% trichloroacetic acid and internal standard added as above. The quenched mixture was then applied to an MCX extraction cartridge for cleanup.26Samples (10 μl) were applied on to a Zorbax Eclipse XBD column (2.1 × 50 mm, 5 μm) with a Zorbax Eclipse C8 guard column (2.1 × 12.5 mm, 5 μm) (Agilent, Palo Alto, CA) and analyzed on an Agilent 1100 high-pressure liquid chromatograph–mass spectrometer. The mobile phase gradient (0.25 ml/min) started at 32% acetonitrile in 0.05% trifluoroacetic acid for 4 min, increased sharply to 50% at 5 min and to 90% at 5.5 min, and was held at 90% for 1.5 min before reequilibrating at 32% acetonitrile for 3.5 min. Under these conditions, EDDP and d3-EDDP eluted at 3.6 min and were monitored at m/z 278 and 281, respectively. The nitrogen drying gas was at 325°C and 6 l/min, fragmentor at 100 V, nebulizer pressure 20 psi, and the capillary at 2000 V. Standard curves were linear (r  ²> 0.99) over the range of 1–1,000 ng/ml EDDP. The intraday and interday coefficients of variation were 15% or less for all quality control standards.

Chiral analysis of R - and S -EDDP formed from RS -methadone metabolism was achieved using a liquid chromatography–mass spectrometry method, developed for measuring EDDP in plasma,27and modified as described under Clinical Investigation, with the exception that the supernatant from the deproteinized incubation mixture (20-85 μl injection) was analyzed. Analytes were quantified using standard curves of peak area ratios. Standard curves were linear (r  ²> 0.99) over the range of 2.5–500 ng/ml for each EDDP enantiomer. The intraday and interday variations were 15% or less for all quality control standards.

Plasma samples from a previous clinical investigation of methadone disposition (stored at −20°C), originally analyzed using an achiral assay, were reanalyzed using a chiral assay. Healthy volunteers received simultaneous oral deuterium-labeled and intravenous unlabeled methadone, in a randomized, balanced, four-way crossover after pretreatment with rifampin (hepatic/intestinal CYP induction), troleandomycin (hepatic/intestinal CYP3A inhibition), grapefruit juice (selective intestinal CYP3A inhibition), or nothing.13Plasma methadone and EDDP enantiomer concentrations were determined by deproteination, on-line solid phase extraction, chiral chromatography, and mass spectrometry as previously described27with the following changes: Samples were reconstituted with 225 μl aqueous ammonium formate (20 mm, adjusted to a pH of 5.7 with formic acid), the injection volume was 75 μl, and the final concentration of methanol in the gradient was 35%. Retention times were 12.5 and 14 min for R - and S -EDDP and 13.0 and 13.9 min for R - and S -methadone, respectively. Chiral concentrations were obtained by application of enantiomeric ratios to the original achiral quantitation.

Analytes were quantified using standard curves of peak area ratios. Minimum validation criteria were as follows: calibration curves r  ²> 0.95, quality control standards ≤15% deviation and <20% approaching the limit of quantitation. Incubations were performed in triplicate and results reported as the mean ± SD unless otherwise indicated. Differences between groups were determined by analysis of variance followed by the Student-Newman-Keuls test (SigmaStat 3.5.1; Systat, Point Richmond, CA). Microsomal velocity–versus –substrate concentration data were analyzed by nonweighted nonlinear regression analysis (SigmaPlot 9.01; Systat) using several models (two-site, two-substrate model, Hill cooperative model, etc. ) based on the recognition that CYPs 3A4 and 2B6 contain at least two binding sites, as described previously.21,26,28–31Choice of model and goodness of fit were guided by the Eadie-Hofstee curves and r  ², distribution of residuals, SE of the parameter estimates, and corrected Akaike information criterion. When Eadie-Hofstee plots were linear, velocity-versus -substrate data were analyzed using a simple Michaelis-Menten model:

Results are expressed as the parameter estimate ± SE of the estimate. Correlations were analyzed using the Pearson product–moment test, or the Spearman rank order test for nonnormally distributed data. Terminal slopes of plasma concentration–versus –time data were determined using WinNonlin (Pharsight Corp., Mountain View, CA).

N -Demethylation of (RS ) and R - and S -methadone to EDDP at therapeutic concentrations (2 or 1 μm) was determined. Stereoselectivity of EDDP enantiomer formation from (RS )-methadone and methadone enantiomers catalyzed by recombinant CYPs 2B6, 3A4, and 2C19 was S > R , S =R , and S << R  EDDP (fig. 1A). There was variability between various lots of recombinant CYP2B6 in the absolute rates of EDDP formation but consistency in the stereoselectivity of methadone N -demethylation (S > R ). There was relative consistency between various lots of CYP3A4 in the absolute rates of EDDP formation and in the lack of stereoselectivity. There was no evidence for metabolic interconversion (i.e. , no S -EDDP was formed from R -methadone, and no R -EDDP was formed from S -methadone).

Human liver microsomal methadone metabolism generally showed greater EDDP formation from S -methadone, whether as the single enantiomer or in the racemate (fig. 1B). There was, however, variability between livers in the degree of enantioselectivity. Microsomes with CYP2B6 content which was comparable to or higher than that of CYP3A4 (for protein content, see table 1) generally metabolized S -methadone at higher rates than R -methadone, whether from the racemate or from individual enantiomers. Conversely, microsomes with lower CYP2B6 content compared with CYP3A4 generally metabolized both R - and S -methadone enantiomers at similar rates and without stereoselectivity (i.e. , HLM-111 and -120).

To assess further the role of CYP2B6 in methadone metabolism, the entire University of Washington School of Pharmacy human liver bank (n = 60) was screened for CYP2B6 content. CYP2B6 protein was detected in all liver samples (fig. 2). Median CYP2B6 protein content was 14.2 pmol/mg protein; however, expression was highly variable (6.1–95.6 pmol/mg protein). CYP2B6 content was positively and significantly correlated with CYP3A4 protein content (r = 0.58, P < 0.001; fig. 2, inset).

From the entire set of HLMs shown in figure 2, a subset of HLM pairs was identified in which each pair contained similar CYP3A4 protein content but disparate (high and low) levels of CYP2B6. Rates of EDDP formation from R - and S -methadone (1 μm) were then determined to assess the significance of CYP2B6 in microsomal methadone metabolism (table 1). In general, in the livers containing greater amounts of CYP2B6, N -demethylation of both methadone enantiomers was greater, and greater stereoselectivity (S > R ) in methadone metabolism was observed. Furthermore, in this limited data set, a better correlation was observed between methadone enantiomer N -demethylation and CYP2B6 content (Pearson r = 0.63 and 0.66 for R - and S -methadone, respectively; both P < 0.02) than with CYP3A4 (r = 0.38 and 0.16 for R - and S -methadone, respectively; not significant) (not shown). Stepwise forward or reverse regression analysis of EDDP formation versus  CYP2B6 protein content was not improved by including CYP3A4 in the analysis. CYP3A5 protein content was not included in the analysis because CYP3A5 did not catalyze methadone N -demethylation (not shown), as observed previously.13 

The entire liver bank was analyzed for CYP2B6 *4, *5, *9,*6, and *7 alleles, which represent the most common polymorphisms.32There was no consistent relation between CYP2B6 genotype and methadone N -demethylation (table 1). Only one HLM sample (HLM-148) carried the *6/*6 genotype, which has been associated with slow metabolizer status for some CYP2B6 substrates.33Therefore, no definitive conclusions could be drawn from this limited data set on CYP2B6 genotype and methadone metabolism in vitro .

The role of CYP2B6, CYP3A4, and CYP2C19 in methadone enantiomer metabolism by HLMs was investigated using isoform-selective chemical inhibitors and 2 μm RS -methadone (fig. 3). Each pair of HLMs contained comparable CYP3A4 protein content and CYP2C19 activity (characterized using S -mephenytoin as a CYP2C19 probe; data not shown) but disparate CYP2B6 protein contents. The CYP2B6-, CYP3A4-, and CYP2C19-selective chemical inhibitors (used at concentrations and conditions described previously) were clopidogrel,34troleandomycin,35and (+)-N -3-benzyl-nirvanol,36respectively. N -Benzyl-nirvanol nonsignificantly inhibited less than 15% of R - and S -EDDP formation, although control reactions showed that greater than 90% of cDNA-expressed CYP2C19 activity was inhibited under similar experimental conditions (not shown). Troleandomycin inhibited 27–66% of EDDP formation and had comparable effects on the two methadone enantiomers. Clopidogrel caused the greatest decrease in methadone metabolism, inhibiting 14–54% of R -EDDP and 34–80% of S -EDDP formation. Inhibition was greatest in HLMs containing more CYP2B6 compared with CYP3A4, with greater inhibition of S - versus  R -EDDP formation. Together, clopidogrel and troleandomycin inhibited 60–90% of R - and S -EDDP formation, strongly suggesting that CYP3A4 and CYP2B6 are the major isozymes responsible for methadone metabolism in HLMs.

RS -Methadone metabolism was determined in two HLM samples (139 and 148) containing contrasting amounts of CYP2B6 (fig. 4). Both HLMs exhibited saturation kinetics and apparently linear Eadie-Hofstee curves; therefore, kinetic data were analyzed using a single enzyme Michaelis-Menten model (table 2).#Metabolism of both methadone enantiomers by HLM-139 was twofold to threefold greater than by HLM-148. The CYP2B6 content of HLM-139 was substantially greater than that of HLM-148, and CYP3A content was less than half. In HLM-139, with a high CYP2B6 content compared with CYP3A4, S -methadone intrinsic clearance was greater than that of R -methadone. Conversely, HLM-148, with relatively little CYP2B6 relative to CYP3A4, metabolized both methadone enantiomers similarly.

In a previous clinical investigation in healthy human volunteers, simultaneous intravenous and oral (deuterium-labeled) racemic methadone was dosed after pretreatment with rifampin (CYP induction), troleandomycin or grapefruit juice (CYP3A inhibition), or nothing, and plasma was analyzed using an achiral assay.13There was no effect of CYP3A inhibition, whereas CYP3A/2B6 induction increased methadone clearance. Nevertheless, this study did not evaluate potential enantioselective effects of CYP2B6 and CYP3A4 modulation. Therefore, the samples were reanalyzed using a chiral assay to quantify R - and S -methadone enantiomers. There were no significant differences between controls and troleandomycin– or grapefruit juice–treated subjects in the plasma concentration–time profile of either R - or S -methadone after either intravenous or oral administration (fig. 5). However, rifampin pretreatment caused a significant decrease in both R - and S -methadone concentrations after both oral and intravenous administration. Careful inspection of figure 5shows differences in the slopes of the elimination curves for R - and S -methadone, and it is known that the plasma R /S -methadone ratio increases over time.6Replotting results from figure 5as the R /S -methadone ratio showed the expected time-dependent increase in the ratio (fig. 6). This increase was augmented by rifampin pretreatment compared with controls, whereas troleandomycin and grapefruit juice had no significant effect.

One major finding of this investigation is that CYP2B6 is a major contributor to human liver microsomal N -demethylation of methadone at steady state therapeutic concentrations (1–2 μm).6CYP2B6 protein was quantifiable in all liver samples examined from a panel of 60 HLMs. EDDP formation rates from R - and S -methadone were significantly correlated with CYP2B6 but not CYP3A4 protein contents. The CYP2B6-selective inhibitor clopidogrel exhibited the largest inhibitory effect on methadone N -dealkylation in HLMs. Recombinant CYP2C19 catalyzed significantly less metabolism than CYP2B6 or CYP3A4, and CYP2C19 inhibition had little effect on HLM methadone metabolism, suggesting that CYP2C19 is relatively nonparticipatory in HLM methadone metabolism.

An equally major finding is that CYP2B6, but not CYP3A4, is responsible for the stereoselective N -demethylation of methadone in HLMs and by recombinant enzymes. At clinically relevant substrate concentrations, CYP2B6 N -dealkylated S -methadone at higher rates than R -methadone, consistent with the higher in vivo  clearance of S -methadone.37CYP2C19 had the opposite stereoselectivity, and CYP3A4 showed no enantiomeric preference. Microsomal samples with a high ratio of CYP2B6 to CYP3A4 protein content metabolized S -methadone at a higher rate than R -methadone, whereas microsomes high in CYP3A4 compared with CYP2B6 metabolized both enantiomers similarly. Furthermore, inhibition of HLM methadone N -demethylation by CYP2C19- and CYP3A4-selective chemical inhibitors was not stereoselective, whereas CYP2B6 inhibition by clopidogrel was clearly stereoselective.

The current results agree with previous studies on methadone metabolism by expressed CYP isoforms and HLM.11,13Both previous investigations found that expressed CYP2B6 catalyzed methadone N -demethylation at rates approximating or greater than those for CYP3A4. P450s (without coexpressed cytochrome b  ⁵) showed methadone stereoselectivities, as assessed by substrate depletion (S > R , S =R , and S << R  for CYPs 2B6, 3A4, and 2C19) or EDDP formation (S > R  and S << R  for CYPs 2B6 and 2C19),11similar to those herein. In microsomes from two human livers with high and low CYP2B6 content, the former generated more EDDP, despite equivalent CYP3A4 contents.11Anti-CYP2B6 antibody inhibited EDDP formation in microsomes from a liver with high, but not low, CYP2B6 content.13Nevertheless, most previous investigations evaluated racemic methadone metabolism, using either achiral or chiral assays, whereas the current investigation is the first comprehensive evaluation of both racemic and methadone enantiomer metabolism, using a large panel of HLMs, P450s with coexpressed cytochrome b  ⁵, and CYP-specific chemical inhibitors. All results demonstrate a significant role for CYP2B6 in methadone N -demethylation and in its stereoselectivity.

The third major finding of this investigation is that CYP3A does not seem to be a major CYP isoform responsible for clinical methadone enantiomer metabolism and elimination in humans. The CYP3A-selective inhibitor troleandomycin (which has no effect on CYP2B6 activity38) was previously shown to have no effect on plasma RS -methadone or EDDP concentrations or RS -methadone clearance.13That investigation, however, used an achiral assay that measured total methadone, rather then individual enantiomers. The current investigation showed that troleandomycin had no effect on either R - or S -methadone enantiomer plasma concentrations or clearance. Additional evidence refuting CYP3A as a major determinant of stereoselective methadone elimination is that neither of the CYP3A inhibitors (troleandomycin and grapefruit juice) altered plasma R /S -methadone concentration ratios.

The fourth major finding is an apparent role for CYP2B6 as a major determinant of stereoselective methadone elimination in humans, as articulated previously,11,13evidenced by rifampin induction of methadone elimination together with the lack of CYP3A inhibitor effects. Rifampin (which induces hepatic CYP2B6 threefold18in addition to inducing CYP3A4 and other P450s) decreased plasma concentrations of both methadone enantiomers and altered the plasma R /S -methadone concentration ratio. This clinical result is consistent with previous model predictions, based on methadone metabolism by CYPs 2B6 and 3A4 and scaling factors, that the rate of increase with time of the plasma methadone R /S  ratio would be proportional to the hepatic CYP2B6 content, that CYP2B6 induction would further increase the R /S -methadone ratio, and that CYP3A inhibition would have minimal effect on the R /S -methadone ratio.21Although HLM CYP2B6 protein content (1–10% of total CYP) is generally lower than CYP3A (as much as 50% of total CYP), these data nonetheless strongly implicate CYP2B6 in methadone disposition. Rifampin can induce CYP2C19, but CYP2C19 is unlikely to be responsible for rifampin alterations in the R /S  ratio because the stereoselectively of CYP2C19 (R > S  in vitro ) is opposite that observed clinically. Stereoselective methadone transport by P-glycoprotein and/or other influx or efflux transporters and induction by rifampin cannot, however, be excluded by the current results.

Previous evidence supporting a clinical role for CYP3A4 in methadone disposition has been variable. Correlations between CYP3A4 activity and either methadone clearance or doses required for maintaining opioid abstinence yielded inconsistent results. Some studies suggested that patients with higher CYP3A4 activity required a higher methadone dose to prevent withdrawal symptoms.4,14Attempts to correlate methadone clearance or maintenance dose with CYP3A4 activity (using xenobiotic probes such as midazolam or endogenous probes such as the 6-OH/17-OH corticosteroid ratio in urine) were unsuccessful.13,20The CYP3A4 inhibitors atazanavir, saquinavir, and ritonavir did not inhibit human methadone clearance, also suggesting that CYP3A4 is not the main enzyme responsible for methadone disposition.13,20,39,40Although fluconazole increased methadone plasma concentrations, this cannot be attributed solely to CYP3A4 inhibition because fluconazole is not a specific CYP3A4 inhibitor.41A major role for CYP2B6 in methadone disposition may help to explain these inconsistent and perplexing observations.

Why were extrapolations of in vitro  and liver microsomal data misleading and suggestive of a clinical role for CYP3A4? Why do some clinical reports suggest a role for CYP3A4?4,16In the current investigation, CYP2B6 protein content was significantly correlated with that of CYP3A4. In addition, CYP2B6 is affected by many of the same inhibitors (i.e. , protease inhibitors) and inducers (i.e. , phenobarbital and rifampin)17,18as CYP3A4.9,42–44Therefore, CYP3A4 may simply serve as a reporter for CYP2B6 activity in vivo  under certain conditions. Apparent correlations between methadone disposition and CYP3A4 may instead represent a dependence on CYP2B6 activity. The possibility of coordinate regulation of CYPs 2B6 and 3A4 has been suggested previously.38This may explain lower methadone plasma concentrations in patients with greater CYP3A activity (midazolam metabolic ratio).16Alteration in the plasma methadone R /S  ratio by rifampin but not by CYP3A inhibitors, the failure of CYP3A inhibitors such as troleandomycin and antiretroviral drugs to increase R - or S -methadone plasma concentrations, and the lack of meaningful correlations between methadone disposition and CYP3A are consistent with this hypothesis. Recent evidence also supports a predominant role for CYP2B6 in clinical methadone disposition.16,45 

Potential limitations of this investigation merit address. Most, but not all, microsome pairs showed greater methadone N -demethylation and stereoselectivity with higher CYP2B6 content. This may be attributable to potential differences in apoprotein versus  holoprotein in the Western blot analysis, which may also influence CYP correlations. Plasma samples were reanalyzed from a previous study.13Quality control samples from that study were not available to formally validate this approach. Nonetheless, much evidence supports the results and conclusions from these data. Subsequent to the original study using achiral analysis, we conducted several studies of methadone disposition using chiral analysis. Results (for controls) from the chiral reanalysis are similar to those from the subsequent chiral studies13and values published by others.37In addition, the time course of plasma R /S  ratios (for controls) from the chiral reanalysis is nearly identical to that published previously.6Last, differences between control and rifampin-treated subjects (and lack of difference in troleandomycin– or grapefruit juice–treated subjects) in plasma methadone enantiomer concentrations and R /S  ratios cannot be explained by analytical confounding.

The potential clinical significance of these finding is manifold, and results may have implications for clinical methadone dosing guidelines. Despite decades of research, the mechanism of interindividual variability in methadone pharmacokinetics has remained largely unknown. There is reticence in the clinical use and dosing of methadone because of uncertainty in the pharmacokinetics and fear of toxicity. This is due in part to concerns about drug interactions which have been classically attributed to CYP3A4.4The current results suggest that statements identifying methadone as a clinical CYP3A4 substrate and warnings in the product label about clinical effects of CYP3A4 induction and/or inhibition, which are based on extrapolations of in vitro  data, require reevaluation. Such apparently incorrect extrapolations reinforce the need to confirm in vitro  observations with clinical investigations.

A role for CYP2B6 in methadone disposition may help to explain the large interindividual variation observed in methadone plasma concentrations and methadone dosing. CYP2B6 protein is variably expressed in the liver, said to account for 1–10% of total hepatic P450s, and mirrored by the variable content (6–95 pmol/mg protein) among livers in this investigation. Furthermore, genetic polymorphisms in CYP2B622could also contribute to variable methadone clearance. To date, 53 haplotypes and 28 alleles have been identified for CYP2B6, with 7 variant alleles occurring at a frequency greater than 1%. There is considerable ethnic and racial variability in the occurrence of variant alleles. For example, CYP2B6*6, whose carriers have up to fourfold lower CYP2B6 expression and activity, occurs in approximately 25% of whites, one third of Africans and African-Americans, and up to 40% of Chinese.22,46CYP2B6 polymorphisms have been shown to affect the pharmacokinetics of other CYP2B6 substrates, such as efavirenz.32,46There is also a potential for CYP2B6-mediated drug–drug interactions, causing either induction or inhibition of methadone metabolism and clearance, exemplified by methadone interactions with antiretrovirals, antifungals, and antiepileptics.4,47,48 

In summary, these results show that CYP3A4 and CYP2B6 but not CYP2C19 contribute to the overall metabolism of methadone in HLMs, that CYP2B6 is responsible for stereoselective methadone metabolism in HLMs, and that CYP2B6 seems responsible for clinical methadone metabolism and clearance. Future studies with specific CYP2B6 probes are needed to validate the clinical role of CYP2B6 in methadone disposition.

The authors thank Kristi Stubbert, B.S. (Senior Research Technician, Department of Anesthesiology, Washington University in St. Louis, St. Louis, Missouri), for technical assistance.