Cytochrome P450 2D6 and treatment of codeine dependence

Oral opioid analgesics such as codeine are used extensively worldwide and are frequently misused. Codeine is a substrate of CYP2D6, a genetically polymorphic P450 enzyme, and is metabolized to the more potent drug morphine. CYP2D6 activity can be inhibited by fluoxetine, and the inhibition of morphine formation may help individuals reduce their use of codeine. Fourteen long-term users of oral opiates (principally codeine) were assessed for an open-label pilot treatment study of fluoxetine 20 mg/day combined with a brief behavioral intervention and structured tapering of the opiate. Eight subjects entered and completed the 8-week treatment. Opiate use decreased by 30% to 100% of baseline use (p < 0.0001) in parallel with a decrease in CYP2D6 activity. Fluoxetine may have a role in the treatment of opiate dependence by decreasing opiate-reinforcing properties.

Inhibitors of cytochrome P450 differentially modify discriminative-stimulus and antinociceptive effects of hydrocodone and hydromorphone in rhesus monkeys

The present study was conducted to investigate the role of cytochrome P450 in the discriminative-stimulus and antinociceptive effects of hydrocodone (HC) and hydromorphone (HM) in rhesus monkeys. In morphine-deprived monkeys, morphine dose-dependently reversed naltrexone-lever responding, an effect also produced by HC and HM. HC and HM also produced antinociception in a warm-water tail withdrawal procedure. Budipine and naltrexone shifted the dose-effect curves for the discriminative-stimulus effects of HC and HM to the right. In contrast, naltrexone, but not budipine (10.0 mg/kg) or quinidine (10.0 mg/kg), dose-dependently antagonized the antinociceptive effects of HC. Budipine and quinidine decreased the concentration of HM in plasma without significantly affecting the levels of HC, suggesting that these CYP2D6 inhibitors decreased the conversion of HC HM. Thus, some behavioral effects of HC are not modified by a marked inhibition of CYP2D6, suggesting that these effects of HC are not due to its conversion to HM but, rather, that both HC and HM act directly on mu receptors.

Comparison of three CYP2D6 probe substrates and genotype in Ghanaians, Chinese and Caucasians

The ability to metabolize CYP2D6 substrates sparteine, debrisoquine, and dextromethorphan was studied in healthy Caucasian (n = 20), Ghanaian (n = 21), and Chinese (n = 22) CYP2D6 extensive metabolizers. Genotype analysis for the CYP2D6*1, *3, *4, *5, *9, *10, and *17 alleles was performed. Interethnic differences in the disposition of the probe drugs were found among the extensive metabolizers; extensive metabolizer status was confirmed by phenotype and genotype analysis. The mean metabolic rate was lower for Caucasians than for Ghanaians for sparteine (P < 0.02) and for both Ghanaians and Chinese for debrisoquine (P < 0.02). Correlation comparisons resulted in lower pairwise correlation coefficients in Ghanaians compared with Chinese and Caucasians for every combination of probe substrates. In addition, in Chinese and Caucasians, metabolic rates for each pair of probe drugs were significantly correlated (P < 0.002), but in Ghanaians the dextromethorphan metabolic rates were not correlated to either sparteine or debrisoquine (P < 0.05). Even when only those with a CYP2D6*1/*1 genotype were included in the correlation calculations, the Ghanaians had very low correlation coefficients (r(s) - 0.02-0.2, n = 9); much lower than those found in Caucasian (r(s) 0.78-0.92, n = 14) or Chinese (r(s) 0.54-0.96, n = 7) individuals. Quinidine had significantly less affect on sparteine metabolic rates in Ghanaians than both Caucasians and Chinese (P < 0.02). In addition, five of the 21 Ghanaian individuals had dextromethorphan metabolic ratios which were unaffected by quinidine. These individuals also had differences in urinary recovery of dextromethorphan and its metabolites when compared to the other Ghanaian individuals. These results confirm the large ethnic differences in probe drug metabolism and quinidine sensitivity among these ethnic groups. They also suggest that the Ghanaians have an additional unidentified allele(s) with altered substrate specificity and quinidine sensitivity which is currently genotyped as CYP2D6*1.

Effect of CYP2D1 inhibition on the behavioural effects of d-amphetamine

In rats, amphetamine (AMP) conversion to 4-OH-AMP is metabolized by CYP2D1, the rat equivalent of the human enzyme CYP2D6. To determine the impact of impaired AMP metabolism on its behavioural effects, AMP-induced hyperactivity, AMP discrimination and AMP self-administration were examined in male Wistar rats with or without pretreatment with the CYP2D1 inhibitors quinine and budipine. In vivo, quinine (20 mg/kg) and budipine (10 mg/kg) increased the plasma area under the curve of AMP 4-fold and 3.6-fold respectively, and decreased the plasma levels of 4-OH-AMP, 3-fold and 8.6-fold, confirming that the doses used suppressed CYP2D1 activity. Both inhibitors prolonged AMP-induced hyperactivity (0.3 mg/kg) and prolonged the duration of AMP-appropriate responding for periods of up to 90 min post-AMP administration in a drug discrimination procedure. In rats given a preload dose of AMP (0.8 mg/kg) 3 h prior to the self-administration test session, CYP2D1 inhibition resulted in fewer AMP infusions being taken compared with rats receiving the AMP preload dose alone. These studies indicate that AMP is responsible for the behavioural effects seen in rats and that a rat phenocopy model of the human CYP2D6 deficiency state can be produced by CYP2D1 inhibitors.

Interactions of amphetamine analogs with human liver CYP2D6

The interaction of fifteen amphetamine analogs with the genetically polymorphic enzyme CYP2D6 was examined. All fourteen phenylisopropylamines tested were competitive inhibitors of CYP2D6 in human liver microsomes. The presence of a methylenedioxy group in the 3,4-positions of both amphetamine (Ki = 26.5 microM) and methamphetamine (Ki = 25 microM) increased the affinity for CYP2D6 to 1.8 and 0.6 microM, respectively. Addition of a methoxy group to amphetamine in the 2-position also increased the affinity for CYP2D6 (Ki = 11.5 microM). The compound with the highest affinity for CYP2D6 was an amphetamine analog (MMDA-2) having both a methoxy group in the 2-position and a methylenedioxy group (Ki = 0.17 microM). Mescaline did not interact with CYP2D6. O-Demethylation of p-methoxyamphetamine (PMA) by CYP2D6 was characterized (Km = 59.2 +/- 22.4 microM, and Vmax = 29.3 +/- 16.6 nmol/mg/hr, N = 6 livers). This reaction was negligible in CYP2D6-deficient liver microsomes, was inhibited stereoselectively by the quinidine/quinine enantiomer pair, and was cosegregated with dextromethorphan O-demethylation (r = 0.975). The inhibitory effect of methylenedioxymethamphetamine (MDMA) was enhanced by preincubation with microsomes, suggesting that MDMA may produce a metabolite complex with CYP2D6. These findings suggest that phenylisopropylamines as a class interact with CYP2D6 as substrates and/or inhibitors. Their use may cause metabolic interactions with other drugs that are CYP2D6 substrates, and the potential for polymorphic oxidation via CYP2D6 may be a source of interindividual variation in their abuse liability and toxicity.

CYP2D6 inhibition in patients treated with sertraline

Sertraline, a selective serotonin reuptake inhibitor used to treat depression, inhibits CYP2D6 in vitro (Ki = 1.2 microM) less potently than fluoxetine (Ki = 0.15 microM). To determine the extent and time course of CYP2D6 inhibition in patients, six males (mean age: 40 years, range: 29-64 years), who were starting treatment for depression with sertraline, were phenotyped on five occasions (once before treatment and approximately 3, 7, 14, and 21 days later). Phenotype status was determined using oral dextromethorphan (30 mg) by calculating the urinary ratio of O-demethylated metabolites to parent drug (i.e., log ODMR). CYP2D6 genotype was determined by leukocyte DNA analysis using polymerase chain reaction amplification. Compliance was confirmed by sertraline plasma levels. Daily sertraline dosages ranged from 50 to 150 mg. Genotype results indicated all subjects were extensive metabolizers (four homozygous wild type [wt], two heterozygous wt/B mutation). Phenotype results showed that CYP2D6 inhibition in patients treated with sertraline appeared to be related to baseline CYP2D6 activity and sertraline dosage. Some patients with high CYP2D6 activity can demonstrate inhibition with sertraline dosages as low as 50 mg.

Inhibition of cytochrome P450 2D6 metabolism of hydrocodone to hydromorphone does not importantly affect abuse liability

Enzymatic conversion of hydrocodone to hydromorphone is catalyzed by cytochrome P450 2D6, which is inactive in about 7% of Caucasians [poor metabolizers (PMs)] and can be inhibited by quinidine pretreatment in the remainder [extensive metabolizers (EMs)]. If hydromorphone, having a substantially higher mu-receptor affinity than hydrocodone, contributes importantly to the physiological and subjective effects of oral hydrocodone, then PMs should be less responsive to the same doses, and quinidine pretreatment should cause EMs to temporarily respond as PMs. Seventeen EMs and 8 PMs who previously responded positively to hydromorphone s.c. received placebo and hydrocodone (10 mg, 15 mg and 22.5 mg p.o.) and were retested with their favorite dose after placebo or quinidine (100 mg) pretreatment; physiological and subjective measures were collected at base line and four times after drug administration, and urine was collected for 8 hr. EMs and PMs were equally responsive to oral hydrocodone, and quinidine had no consistent effect on their responses, even though quinidine abolished the pre-existing metabolic differences in hydromorphone production, as measured in urine. These data suggest only a small role of hydromorphone in eliciting abuse-related responses to oral hydrocodone.

Simultaneous gas chromatographic determination of methamphetamine, amphetamine and their p-hydroxylated metabolites in plasma and urine

We report a method for the simultaneous determination of methamphetamine, amphetamine and their hydroxylated metabolites in plasma and urine samples using a GC-NPD system. The analytical procedures are: (1) adjust the sample to pH 11.5 with bicarbonate buffer, saturate with NaCl and extract with acetate; (2) back-extract the amines in the ethyl acetate fraction with 0.1 M HCl; (3) adjust the pH of the acid fraction to 11.5 and follow by extraction in ethyl acetate; (4) reduce the volume of ethyl acetate under nitrogen and derivatize the concentrate with trifluoroacetic anhydride or heptafluorobutyric anhydride before the GC analysis. The derivatives were separated on a GC-NPD system equipped with a HP-5 column of 25 m x 0.32 m I.D. and a 0.52 micron film of 5% phenylmethylsilicone. The detection limit (taking a signal-to-noise ratio of 2) of heptafluorobutyl derivatives of methamphetamine and its metabolites in plasma and the trifluoroacetyl derivatives in urine was 1 ng/ml (22 pg on column). The limit of quantitation of the heptafluorobutyl derivatives in the plasma was 1 ng/ml (22 pg on column), and that of the trifluoroacetyl derivatives in urine was 20 ng/ml (73 pg on column). The between-day variation was from 0.9 to 17.4% and within-day variation from 0.9 to 8.3%. This method was used successfully in the quantitative determination of methamphetamine and its p-hydroxylated metabolites in the plasma and urine of human subjects.

Effect of cytochrome P450 2D1 inhibition on hydrocodone metabolism and its behavioral consequences in rats

Humans that lack cytochrome P450 2D6 (CYP2D6) activity may have an altered risk of drug dependence or abuse because this enzyme is important in the metabolism of some drugs of abuse, including hydrocodone. In rats, hydrocodone conversion to hydromorphone is catalyzed by CYP2D1, the rat homolog of the human CYP2D6. To determine the impact of impaired hydromorphone formation on the behavioral effects of the parent compound, hydrocodone-induced analgesia and hyperactivity, hydrocodone discrimination and self-administration were examined in male Wistar rats, with or without pretreatment with CYP2D1 inhibitors (quinine and budipine). In vivo, quinine (20 mg/kg) and budipine (10 mg/kg) produced a marked suppression in brain and plasma hydromorphone levels detected after the peripheral administration of hydrocodone, thus confirming that the doses used suppressed CYP2D1 activity. In contrast, CYP2D1 inhibition had no impact on the analgesic or discriminative stimulus effects of hydrocodone, nor did this type of manipulation alter hydrocodone self-administration. The effects of quinine on the locomotor activating effects of hydrocodone were subtle at best. Because inhibition of CYP2D1 in this rat strain is proposed to be a useful animal counterpart for studying the impact of CYP2D6 polymorphism in humans, these data suggest that differences in CYP2D6 phenotype will have limited influence on the drug response to hydrocodone after nonoral administration. This has recently been verified in a study showing that inhibition of hydrocodone biotransformation to hydromorphone does not affect measures of abuse liability. Therefore, hydrocodone's behavioral effects are most likely attributable to its own intrinsic effects at mu opioid receptors.

Venlafaxine oxidation in vitro is catalysed by CYP2D6

1. Several selective 5-HT reuptake inhibitors (SSRIs) are inhibitors of the genetically polymorphic drug metabolizing enzyme, CYP2D6. We studied the interaction of venlafaxine, a new SSRI, with CYP2D6 in human liver microsomes. 2. Venlafaxine was a less potent inhibitor of this enzyme activity in vitro than other SSRIs tested. The average apparent Ki values determined using CYP2D6-dependent dextromethorphan O-demethylation were: 33, 52 and 22 microM for rac-venlafaxine, R(+)-venlafaxine and S(-)-venlafaxine, respectively, vs 0.065 to 1.8 microM for paroxetine, fluoxetine, norfluoxetine, fluvoxamine and sertraline. 3. Microsomes from human livers (n = 3) and from yeast transformed with an expression plasmid containing human CYP2D6 cDNA catalyzed the O-demethylation of venlafaxine, which is the major metabolic pathway in vivo. Intrinsic metabolic clearance values (Vmax/Km) indicated that S(-)-venlafaxine was cleared preferentially via this pathway. 4. In microsomes from CYP2D6-deficient livers (n = 2), Vmax/Km of O-demethylation of venlafaxine was one to two orders of magnitude lower and was similar to the rate of N-demethylation. 5. Studies with chemical probes which preferentially inhibit P450 isoforms suggested that CYP3A3/4 is involved in venlafaxine N-demethylation. 6. These in vitro findings predict phenotypic differences in the kinetics of venlafaxine in vivo, although the clinical importance of this is unclear as O-demethylvenlafaxine is pharmacologically similar to the parent drug. The findings also predict relatively limited pharmacokinetic interaction between venlafaxine and other CYP2D6 substrates.

Effects of route of administration on dextromethorphan pharmacokinetics and behavioral response in the rat

One of the potential problems of using dextromethorphan as a neuroprotective or anticonvulsant agent is the phencyclidine-like behavioral effects that have been attributed to its major metabolite dextrorphan. Because previous behavioral studies of dextromethorphan have generally failed to consider metabolic conversion to this metabolite, the present studies were conducted to examine the effects of route of administration on dextromethorphan pharmacokinetics and locomotor activity in the rat. The bioavailability of dextromethorphan was 1.3-fold lower and the formation of dextrorphan and other metabolites was 3-fold greater after i.p. injection of 30 mg/kg of dextromethorphan as compared to the s.c. route, indicating substantial effect of first-pass metabolism. Plasma dextromethorphan was correlated with brain dextromethorphan (r = 0.84, P < .001), and the brain/plasma concentration ratio was about 6.5. Plasma-free dextrorphan, but not conjugated dextrorphan, was correlated with brain dextrorphan (r = 0.97, P < .001). Tmax of brain dextrorphan was earlier, and Cmax was higher after i.p. injection of dextromethorphan than s.c. administration (60 min vs. 120 min and 1.0 nmol/g vs. 0.2 nmol/g). Dextromethorphan (60 mg/kg i.p.) increased locomotor activity in the rat 60 min postdose, whereas the same dose of dextromethorphan administered by s.c. injection was without effect. These data demonstrate the route-specific effects on the disposition of dextromethorphan and dextrorphan in rat plasma and brain, as well as the behavioral consequence of the difference.

Pharmacokinetics of dextromethorphan and metabolites in humans: influence of the CYP2D6 phenotype and quinidine inhibition

Dextromethorphan is primarily metabolized to dextrorphan by cytochrome P450 2D6 (CYP2D6), a genetically polymorphic enzyme in humans. Dextrorphan is an active metabolite that produces phencyclidine-like behavioral effects in animals and exhibits anticonvulsant and neuroprotective properties in a variety of experimental models. In these studies, we examined the effects of CYP2D6 phenotype and quinidine inhibition on the pharmacokinetics of dextromethorphan and its metabolites in humans. After a single oral dose of dextromethorphan HBr (30 mg), the major metabolites in the plasma of extensive metabolizers (N = 5) were conjugated dextrorphan and conjugated 3-hydroxymorphinan. Free dextrorphan concentrations were about 100-fold less than the conjugated dextrorphan, and dextromethorphan was not detectable. Pretreatment of these subjects with 100 mg of quinidine, a selective inhibitor of CYP2D6, significantly suppressed the formation of dextrorphan and elevated the concentrations of dextromethorphan (t1/2, 16.4 hours). In poor metabolizers (N = 4) given the same dose, dextromethorphan was the major component in the plasma with a t1/2 of 29.5 hours. Present at concentrations 5- to 10-fold less were conjugated dextrorphan and the other two metabolites. Urinary recovery studies indicated that the inhibition by quinidine was reversible and that the elimination of dextromethorphan primarily depends on CYP2D6 activity rather than renal elimination. These data demonstrated that the CYP2D6 phenotype and the concurrent administration of quinidine significantly affect the disposition of dextromethorphan and the formation of the active metabolite dextrorphan and are important factors to be considered in studies of the pharmacologic and behavioral effects of dextromethorphan.

CYP2D6 phenotype determines the metabolic conversion of hydrocodone to hydromorphone

The contribution of cytochrome P450 2D6 (CYP2D6) to the formation of hydrocodone's active metabolite, hydromorphone, was examined in vitro and in vivo. Human liver microsomes prepared from an individual homozygous for the D6-B mutation of the CYP2D6 gene catalyzed this reaction at a negligible rate. Urinary metabolic ratios of hydrocodone/hydromorphone were highly correlated with O-demethylation ratios for dextromethorphan, an established marker drug of CYP2D6 activity (rs = 0.85; n = 18). The kinetics of hydrocodone after a single oral dose and its partial metabolic clearance to hydromorphone were investigated in five extensive metabolizers of dextromethorphan, six poor metabolizers, and four extensive metabolizers after pretreatment with quinidine, a selective inhibitor of CYP2D6 activity. The mean values for partial metabolic clearance by O-demethylation in the three groups were 28.1 +/- 10.3, 3.4 +/- 2.4, and 5.0 +/- 3.6 ml/hr/kg, respectively. No statistically significant phenotypic differences in physiologic measures were observed. However, over the first hour after dosing, the extensive metabolizers reported more "good opiate effects" and fewer "bad opiate effects" than poor metabolizers and extensive metabolizers in whom CYP2D6 was inhibited by quinidine. These data establish the importance of CYP2D6 in the formation of hydromorphone from hydrocodone and suggest that the activity of this enzyme may limit the abuse liability of hydrocodone.

Human hepatic cytochrome P450 2D6-like activity in nonhuman primates: catalytic characterization in vitro

Previous studies have identified the monkey as an animal model for the genetic polymorphism affecting human hepatic cytochrome P450 2D6 enzyme. However, contrary to an earlier in vivo observation, the present study failed to find evidence of polymorphism of this enzyme activity in liver preparations of 84 African green (Cercopithecus aethiops) monkeys. The kinetics of dextromethorphan O-demethylation were similar in liver microsomes from African green (n = 21) and Crab eater (Macaca fascicularis, n = 7) monkeys (Km = 1.1 +/- 0.07 and 1.7 +/- 0.27 microM; Vmax = 215 +/- 71.7 and 152 +/- 21.1 nmol/mg/h, respectively). These Km values were lower and less variable than those in liver microsomes from 10 human extensive metabolizers (5.3 +/- 2.43 microM). Furthermore, the Vmax of the reaction in human liver microsomes was significantly lower (32 +/- 15.7 nmol/mg/h; P < .001). Inhibitor constants (Ki values) determined in monkey and human liver microsomes were highly correlated (r = 0.97), but two high-affinity inhibitors of the human enzyme (quinidine and lobeline) were approximately 40-fold less potent in monkey livers than in human livers. These data show that the monkey enzyme is functionally homologous, but is not identical to human hepatic cytochrome P450 2D6. Failure to observe poor metabolizer monkeys does not preclude their potential usefulness in evaluating the role of human hepatic cytochrome P450 2D6 activity in drug addiction and neurotoxicity because of the possibility of producing poor metabolizer phenocopies by potent hepatic cytochrome P450 2D6-like enzyme inhibitors in monkeys.

Inhibition by fluoxetine of cytochrome P450 2D6 activity

Potent inhibition of cytochrome P450 2D6 (CYP2D6) in human liver microsomes by fluoxetine and its major metabolite norfluoxetine was confirmed (apparent inhibition constant values, 0.2 mumol/L). Several other serotonergic agents were also found to be competitive inhibitors of this genetically polymorphic enzyme. The O-demethylation ratio of dextromethorphan that expressed CYP2D6 activity in 19 patients receiving fluoxetine fell in the region of the antimode separating the O-demethylation ratio values observed in 208 extensive metabolizers from 15 poor metabolizers of a control group of healthy subjects. Inhibition of CYP2D6 activity in patients undergoing treatment with fluoxetine or other serotonin uptake inhibitors could contribute to toxicity or attenuated response from concurrent medications that are substrates of this enzyme. Other in vitro studies indicated that CYP2D6 catalyzes the O-demethylation of oxycodone to form oxymorphone. This reaction was inhibited by fluoxetine and its normetabolite in liver microsomes from both extensive and poor metabolizer individuals, indicating that these compounds are not selective inhibitors of CYP2D6 activity.

Inhibition of human cytochrome P450 2D6 (CYP2D6) by methadone

1. In microsomes prepared from three human livers, methadone competitively inhibited the O-demethylation of dextromethorphan, a marker substrate for CYP2D6. The apparent Ki value of methadone ranged from 2.5 to 5 microM. 2. Two hundred and fifty-two (252) white Caucasians, including 210 unrelated healthy volunteers and 42 opiate abusers undergoing treatment with methadone were phenotyped using dextromethorphan as the marker drug. Although the frequency of poor metabolizers was similar in both groups, the extensive metabolizers among the opiate abusers tended to have higher O-demethylation metabolic ratios and to excrete less of the dose as dextromethorphan metabolites than control extensive metabolizer subjects. These data suggest inhibition of CYP2D6 by methadone in vivo as well. 3. Because methadone is widely used in the treatment of opiate abuse, inhibition of CYP2D6 activity in these patients might contribute to exaggerated response or unexpected toxicity from drugs that are substrates of this enzyme.

Catalytic and immunologic similarities between monkey and human liver cytochrome P-450db1 (human cytochrome P-450 2D6)

In vivo pharmacogenetic studies have suggested that the monkey may be an animal model for the human polymorphism of cytochrome P-450 2D6 (also called cytochrome P-450db1). In the present study, the catalytic, immunologic, and electrophoretic properties of cytochrome P-450db1 in liver microsomes from African green monkeys (Cercopithecus aethiops) were examined and compared with P-450db1 in human liver microsomes. Using sparteine as the substrate, the activity of microsomal P-450db1 from the two sources was indistinguishable in terms of the pattern of sparteine metabolites produced, the apparent Ki values of 8 competitive inhibitors (r = 0.94, p less than 0.001), and the extent of immunoinhibition by anti-rat P-450db1 antibody. Kinetic analyses demonstrated that the apparent KM values of the high affinity component of sparteine oxidation in monkey liver microsomes fell within the range observed in human livers; the Vmax of this component was as much as six times greater than the highest value reported for human liver. Western immunoblots showed a protein band in monkey liver microsomes that co-migrated with P-450db1 in human liver. The high degree of similarity observed here between P-450db1 of monkey and human liver microsomes suggests that the monkey will be a good animal model for P-450db1 enzyme studies, and possibly for studies of the role of this enzyme in drug abuse and dependence.

Drug metabolism and interactions in abuse liability assessment

The interpretation of acute abuse liability studies and drug interaction studies would be importantly strengthened by the routine inclusion of drug concentration measurements at appropriate sampling times. Reliance on mean kinetic data misrepresents the variation in drug kinetics and fails to take experimental advantage of the natural differences in the population which may represent the extremes of abuse risk. Pharmacokinetic-pharmacodynamic studies to better understand the relationship of plasma drug concentration, drug concentration in the receptor biophase and specific drug reinforced behaviour will ensure proper study design and yield useful theoretic information. Multi- and poly-drug abuse (including heavy smoking and heavy ethanol use) are very common. Such patterns of use can have quite large effects on drug kinetics. Because of the potentially large qualitative and quantitative differences in drug metabolism and kinetics between pre-clinical species and the human, data should be gathered at the earliest possible time with respect to human metabolic rates, patterns, and identification of inhibitors. The availability of human liver microsomes facilitates such studies.

Characterization of metronidazole metabolism by human liver microsomes

The metabolism of metronidazole was studied in microsomes isolated from livers of human kidney donors. The formation of the major in vivo metabolite, hydroxymetronidazole, proceeded according to biphasic kinetics, suggesting the involvement of at least two enzymatic sites. The affinity constant (Km) of the high affinity site ranged from 140 to 320 microM and metabolism at this site contributed more than 75% of the intrinsic clearance. Thus, at therapeutic doses of metronidazole most of the hydroxylation in vivo should be associated with this site. Antipyrine, cimetidine, alpha-naphthoflavone, caffeine, theophylline, mephenytoin, tolbutamide, quinidine, acetone and nifedipine were poor inhibitors of the formation of hydroxymetronidazole by human liver microsomes. Propranolol (500 microM) inhibited the hydroxylation rate by 70%. Phenacetin inhibited metronidazole hydroxylation with a competitive inhibition constant (Ki) of 4-5 microM. However, metronidazole did not inhibit the O-deethylation of phenacetin. It is concluded that cytochromes P450 IA2, IIC9, IIC10, IID6, IIE1 and IIIA3 do not contribute significantly to the high affinity hydroxylation of metronidazole in man.

Propranolol oxidation by human liver microsomes--the use of cumene hydroperoxide to probe isoenzyme specificity and regio- and stereoselectivity

1. Three oxidations of the enantiomers of propranolol were studied in human liver microsomes under two reaction conditions. Previous in vitro studies had established that two of the livers were from poor metaboliser (PM) phenotypes for the debrisoquine 4-hydroxylase (cytochrome P-450IID6) and the remaining seven were from extensive metaboliser (EM) phenotypes. 2. In the presence of NADPH and oxygen 4- and 5-hydroxylation of propranolol occurred in microsomes from all nine livers, as did propranolol N-desisopropylation. R(+)-propranolol was oxidized preferentially along the three pathways, although enantioselectivity observed for N-desisopropylation may have arisen not only from stereoselectivity in formation rates, but also from stereoselectivity in subsequent microsomal metabolism, possibly by monoamine oxidase. The involvement of monoamine oxidase in the further microsomal metabolism of N-desisopropylpropranolol was indicated by inhibition of the metabolism of this compound when incubated with phenelzine. 3. Cumene hydroperoxide has been proposed to support only the activity of cytochrome P450IID6. This is consistent with the observations that a) propranolol 4- and 5-hydroxylation occurred in microsomes from the EM livers only and b) side-chain oxidation was not observed under these conditions in either PM or EM livers. 4. Using cumene hydroperoxide to support the reactions, the 4-hydroxylation of propranolol showed little enantioselectivity, whereas S(-)-propranolol was 5-hydroxylated about twice as fast as the R(+)-enantiomer. There were highly significant correlations between the rates of 4- and 5-hydroxylation of R(+)-propranolol (r = 0.96, P less than 0.001, n = 7 livers) and of S(-)-propranolol (r = 0.98, P less than 0.001). Both oxidations were described by single-site Michaelis-Menten kinetics. 5. The findings suggest that P450IID6 is involved in both the 4- and 5-hydroxylations of propranolol, but that these metabolites can also be formed by other P450 isoenzymes. It is confirmed that P450IID6 does not contribute to the N-desisopropylation of propranolol. Furthermore, the finding that mephenytoin did not inhibit the appearance of this metabolite is not consistent with the results of in vivo studies suggesting the involvement of the same enzyme in the side-chain oxidation of propranolol and the 4-hydroxylation of mephenytoin.

Use of quinidine inhibition to define the role of the sparteine/debrisoquine cytochrome P450 in metoprolol oxidation by human liver microsomes

The oxidation of the beta adrenoceptor antagonist metoprolol exhibits genetic polymorphism of the sparteine/debrisoquine (SP/DB) type. The alpha-hydroxylation of metoprolol is absent in poor metabolizers, whereas metoprolol O-demethylation is only partially impaired, suggesting that an enzyme or enzymes other than cytochrome P450-SP/DB contribute to the latter reaction. Using inhibition by the quinidine/quinine isomer pair as a marker for the activity of cytochrome P450-SP/DB, the role of this enzyme in the in vitro oxidation of the enantiomers of metoprolol by human liver microsomes was examined. Unlike alpha-hydroxylation, only a portion of metoprolol O-demethylation showed the stereoselective inhibition by quinidine and quinine characteristic of in vitro reactions catalyzed by cytochrome P450-SP/DB. Furthermore, the kinetics of metoprolol O-demethylation were biphasic, the two components of O-demethylase activity being distinguishable by their enantioselectivity and sensitivity to inhibition by quinidine. Microsomes from one liver formed no detectable alpha-hydroxymetoprolol, and O-demethylation by these microsomes corresponded to the low affinity site observed in eight other livers. The rate of metoprolol O-demethylation by the quinidine-inhibitable high affinity component was directly proportional to the rate of alpha-hydroxylation. These findings support the hypothesis that cytochrome P450-SP/DB catalyzes the formation of alpha-hydroxymetoprolol, but is only partially responsible for metoprolol O-demethylation. Such a mechanism could explain the previously reported inability to detect polymorphism in the O-demethylation pathway in vivo.

In vitro evidence against the oxidation of quinidine by the sparteine/debrisoquine monooxygenase of human liver

Competitive inhibition studies using human liver microsomes have shown that quinidine (QD) has an exceptionally high affinity (60 nM) for the genetically variable cytochrome P-450 that catalyzes the formation of 4-hydroxydebrisoquine and dehydrosparteines from debrisoquine and sparteine. The present study examined the effect of sparteine and debrisoquine on the oxidation of QD by microsomes prepared from two human livers. QD and its major metabolite 3-hydroxy-QD were measured by quantitative TLC. QD 3-hydroxylation followed saturable single-site kinetics over a 1-250 microM range of QD concentrations. The Km and Vmax of the reaction in the two liver specimens were 47.5 +/- 3.5 microM and 58.7 +/- 5.9 microM, and 0.36 +/- 0.08 and 0.29 +/- 0.02 nmol of 3-hydroxy-QD/mg of protein/min. Sparteine and debrisoquine (250 microM) had no effect on this QD 3-hydroxylase activity. Furthermore, near-saturation of the sparteine/debrisoquine isozyme by 250 microM sparteine had no effect on the oxidation of QD by all routes (measured by QD disappearance from an initial level of 70 nM during an 8-hr incubation period). These observations indicate that none of the major oxidative reactions of QD are catalyzed by the sparteine/debrisoquine isozyme; QD may simply bind to this cytochrome P-450, without being oxidized by it.

Imipramine demethylation and hydroxylation: impact of the sparteine oxidation phenotype

Eighteen healthy volunteers, selected according to their ability to oxidize sparteine, took single oral doses of 100 mg imipramine and desipramine. For imipramine the following clearances (L X min-1) were found in six rapid extensive metabolizers (EM), six slow EM, and six poor metabolizers (PM), respectively (mean and range): apparent oral clearance: 2.55 (1.39 to 3.47), 2.28 (1.18 to 4.26), and 1.35 (0.96 to 1.64). Clearance via demethylation was: 1.42 (0.61 to 2.01), 1.60 (0.78 to 3.81), and 1.09 (0.76 to 1.64); clearance via other pathways was: 1.13 (0.74 to 1.75), 0.69 (0.40 to 1.59), and 0.26 (0 to 0.46). For desipramine the apparent oral clearance (L X min-1) was 0.19 (0.12 to 0.24) in PM compared with 1.64 (1.46 to 1.80) and 1.03 (0.77 to 1.13) in rapid EM and slow EM. Extremely long elimination t1/2s of desipramine were seen in PM: 81 to 131 hours compared with 13 to 23 hours in EM. 2-OH-imipramine and 2-OH-desipramine were detectable in plasma of only the 12 EM, where the ratio-to-parent compound was higher in rapid EM than in slow EM. This study confirms that 2-hydroxylation of imipramine and desipramine depends almost exclusively on the sparteine oxygenase, whereas the demethylation of imipramine depends mainly on a different isozyme.

Sparteine oxidation is practically abolished in quinidine-treated patients

In eight patients a sparteine-test was carried out immediately before and after 1 week of treatment with quinidine 600-800 mg day-1. Before treatment one patient was classified as a poor metaboliser (metabolic ratio: greater than or equal to 20), and seven patients as extensive metabolisers. During quinidine treatment, the formation of sparteine metabolites (2- and 5-dehydrosparteine) was practically abolished. Patients initially classified as extensive metabolisers thus exhibited the phenotype of poor metabolisers during quinidine treatment.

Sparteine oxidation polymorphism: a family study

Polymorphic oxidation of the pharmacogenetic probe drug sparteine was investigated in 35 parents and 29 siblings of 20 unrelated poor metabolizer (PM) probands. Phenotyping was carried out on the basis of metabolic ratio (MR) = sparteine/dehydrosparteines in the 12 h urine. The distribution of MR was bimodal: 47 relatives (20 siblings and 27 parents) had MR ranging from 0.22-12 and were defined as extensive metabolizers (EM) whereas MR ranged from 20-340 in nine siblings and eight parents thus defined as PM. The 20 pedigrees confirmed that poor metabolism of sparteine is inherited as an autosomal recessive character. The mean recovery of dehydrosparteines (% of dose) was 27% in 23 positively identified drug free heterozygotes compared with 37% in unrelated EM (both genotypes) (P less than 0.05) previously phenotyped. Degrees of dominance of 73% and 77% were calculated on basis of log excretion of dehydrosparteines (% of dose in 12 h urine) and log MR, respectively.

A simple borohydride/GC method for measuring sparteine metabolites in man

A simple borohydride/GC method was developed for phenotyping sparteine oxidation in man. The major metabolites of sparteine found in human urine, 2- and 5-dehydrosparteine, were converted quantitatively back to sparteine by sodium borohydride reduction. The amount of sparteine metabolites can be estimated from the difference of sparteine concentrations between the borohydride-treated and untreated urine samples. The coefficient of variation of this assay was estimated from repeated analyses to be +/- 3% within a day (intra-assay) and +/- 8% between days (inter-assay).

Steady-state concentrations of imipramine and its metabolites in relation to the sparteine/debrisoquine polymorphism

Thirty-five imipramine treated patients were phenotyped with regard to polymorphic drug oxidation using sparteine and/or debrisoquine. During treatment with 100 mg imipramine per day the mean steady-state concentrations and ratios in 28 extensive metabolizers were: imipramine 169 nmol/l; desipramine 212 nmol/l; 2-OH-imipramine/imipramine 0.25; 2-OH-desipramine/desipramine 0.57. The corresponding values in two poor metabolizers were: imipramine 455 and 302 nmol/l; desipramine 1148 and 1721 nmol/l; 2-OH-imipramine/imipramine 0.06 and 0.05; 2-OH-desipramine/desipramine: 0.09 and 0.04 respectively. The metabolic ratios (MR) sparteine/dehydrosparteine and debrisoquine/4-OH-debrisoquine (% of dose in 12-h urine samples) correlated poorly with the imipramine steady-state concentrations during administration of 100 mg per day, but quite well with the desipramine steady-state concentrations. Significant negative correlations were found between sparteine and debrisoquine MR and the 2-OH-imipramine/imipramine and 2-OH-desipramine/desipramine ratios. In most patients the initial dose was changed to obtain concentrations in the therapeutic range, and concentrations for imipramine + desipramine of (mean +/- SD) 713 +/- 132 nmol/l were achieved in 33 patients. The therapeutic dose was 50 mg per day in one poor metabolizer and ranged from 50-400 mg per day in 32 extensive metabolizers. There was a weak negative correlation between sparteine MR and daily dose. Treatment with imipramine inhibited metabolism of both sparteine and debrisoquine (MR values about doubled), but did not affect the interpatient correlations.

Sparteine oxidation polymorphism in Denmark

Sparteine oxidation was polymorphic among 301 healthy Danish volunteers. Hence 22 subjects or 7.3% were phenotyped as poor metabolizers (PM) whereas 279 subjects were classified as extensive metabolizers (EM). The metabolic ratio (MR) between sparteine and 2- and 5-dehydrosparteine (% of dose) in 12 hrs urine ranged from 0.11-12.6 in EM and from 30-394 in PM. Urinary excretion of 2- and 5-dehydrosparteine also discriminated between PM and EM. Age, sex, and smoking habits did not influence the MR. This study confirms that sparteine is a useful probe drug in pharmacogenetic investigations.

High affinity of quinidine for a stereoselective microsomal binding site as determined by a radioreceptor assay

The techniques of the radioreceptor binding assay were applied to detect stereoselective binding of quinidine and quinine to a site on human liver microsomes. Binding of 3H-dihydroquinidine was 50% inhibited by 20-100 nM quinidine, while its enantiomer quinine did not displace the 3H-ligand at concentrations up to 500 nM. This stereoselectivity agreed with the affinity values measured by functional enzyme assays of cytochrome P450 activity using sparteine or debrisoquine as substrates.

A human cytochrome P-450 characterized by inhibition studies as the sparteine-debrisoquine monooxygenase

The present study compares the debrisoquine monooxygenase and the sparteine monooxygenase activities of human liver microsomes. In the presence of 14 competitive inhibitors, apparent inhibition constants (Ki) as determined by these two activities ranged over four orders of magnitude with a correlation coefficient 0.99. These in vitro results represent the strongest evidence to date that the debrisoquine monooxygenase and the sparteine monooxygenase are identical and involve a single isozyme of cytochrome P-450.

Competitive inhibition of sparteine oxidation in human liver by beta-adrenoceptor antagonists and other cardiovascular drugs

The rate of oxidation of sparteine by the 9000 x g supernatant fraction of a human liver was measured in the presence of various drugs which exert cardiovascular effects. Hexamethonium, ouabain, caffeine and isoproterenol had no effect on this rate, while alprenolol, metoprolol, oxprenolol, propranolol, timolol, pindolol, lidocaine, mexiletine, 17-n-pentyl-sparteine, tolazoline, quinine, quinidine, cinchonine and cinchonidine inhibited the in vitro reaction competitively. Stereoselective inhibition was observed between quinine (Ki = 15 microM) and quinidine (Ki = 0.06 microM). Genetic evidence suggests that the primary metabolism of sparteine depends on a single species of cytochrome P450. In vitro competitive inhibition of sparteine oxidation by a drug indicates that this drug is capable of occupying the same enzymatic site as sparteine. This may mean that the competing drug is also metabolized at that site and thereby subject to the same genetic variation as sparteine's oxidation; absence of inhibition excludes this possibility.

Comparative pharmacogenetics of sparteine and debrisoquine

Capacities to oxidize sparteine and debrisoquine in healthy Canadian Caucasians were compared. The Spearman rank correlation between the conventional urinary metabolic ratios (drug/metabolite) was rs = 0.79 (P less than 0.001), but the sparteine metabolic ratio appears to be the more discriminating probe to distinguish metabolizers and nonmetabolizers. The urinary amount of oxidized sparteine alone may allow reliable detection of nonmetabolizers. From a total of 17 poor metabolizers observed in this study and in studies in Germany and Sweden, all were deficient in metabolizing capacity for both sparteine and debrisoquine.

Inhibition of sparteine oxidation in human liver by tricyclic antidepressants and other drugs

Testing for competitive inhibition of sparteine oxidation in the 9000 x g supernatant fraction from human liver provides an in vitro means to identify drugs which can bind to the same form of cytochrome P450 which oxidizes sparteine. There has so far been only two outcomes of this test: either the drug examined competed with sparteine for a common binding site, or it did not inhibit the reaction. The results of such in vitro testing implicated the involvement of guanoxan, nortriptyline, desipramine, imipramine, amitriptyline and chlorpromazine with this enzyme. Amobarbital, tolbutamide and guanethidine in therapeutic concentrations did not interfere with sparteine oxidation by this preparation.

In vitro metabolism of sparteine by human liver: competitive inhibition by debrisoquine

Population data indicate that the genetic control is the same for the oxidation of sparteine and debrisoquine, although whether the level of control is regulatory or enzymatic is not clear. Therefore, the influence of debrisoquine on the rates of in vitro formation of the two dehydrogenated metabolites of sparteine in the 9000 x g supernatant fractions of human liver was examined. The interaction of these two drugs was competitive, indicating that the same form of cytochrome P450 is responsible for their biotransformation. Antipyrine at concentrations as high as 4 mM had no effect on sparteine oxidation.

Sparteine metabolism in Canadian Caucasians

The capacity for sparteine (SP) metabolism was determined in 48 Caucasian subjects by measuring amounts of drug and dehydrogenated metabolites in urine after an oral dose of SP sulfate. Three phenotypes were recognized and were assumed to represent individuals homozygous for poor SP oxidation (group III) and those heterozygous (group II) and homozygous (group I) for extensive SP oxidation. Separation of groups I and II, although incomplete, was improved by alterations in the published analytic procedure. The pattern of deviations from the normal distribution was similar for both dehydrosparteine metabolites. This supports the hypothesis of a common intermediate, the formation of which is monogenically controlled. Correlation analysis of the two metabolites indicates the possibility of further metabolism of 5-dehydrosparteine.

Debrisoquine hydroxylation capacity: problems of assessment in two populations

Following the investigation of debrisoquine (D) metabolism to 4-hydroxydebrisoquine (4OHD) in two populations, problems with the current practice of assessing this capacity by urinary D:4OHD ratios presented themselves. The distributions of both components of this ratio, when examined separately, clearly displayed interethnic differences. The mean 0- to 8-hr recovery of 4OHD in the Caucasian group was 15.5 +/- 1.1% (SE) of a 20-mg dose and 8.0 +/- 1.5% in the Oriental group. The corresponding values for unchanged D excretion were 33.0 +/- 4.1% in Orientals and 18.2 +/- 2.3% in Caucasians, a difference that is overlooked when D:4OHD ratios alone are examined. Although D and 4OHD excretion rates were equally variable, they correlated poorly; that is, they tended to vary independently of each other. This finding, in conjunction with the two separate ethnic differences, indicates that the ratio D:4OHD does not solely reflect an individual's capacity towards D 4-hydroxylation, and that equating these is unwarranted. This limitation does not necessarily invalidate judicious use of the ratio at the edges of the distribution curve, but it does call for additional investigations into the means to assess D metabolizing capacity in most subjects.

Ethnic difference in drug metabolism: debrisoquine 4-hydroxylation in Caucasians and Orientals

The extent to which debrisoquine (D) is metabolized to 4-hydroxydebrisoquine (4OHD) was found to be influenced by the ethnic origin of the subjects. Following ingestion of 20 mg of debrisoquine sulfate, Orientals excreted, on the average, significantly more D and less 4OHD in 0- to 8-h urine than Caucasians; however, the sums of D + 4OHD did not differ.

Deficient metabolism of debrisoquine and sparteine

Genetic deficiencies of alicyclic hydroxylation of debrisoquine and of sparteine oxidation are independently discovered entities, each of clinical significance in its sphere. This paper reports evidence to indicate that these 2 deficiencies have the same cause. Previous investigation of one of the affected subjects had revealed normal oxidative metabolism of amobarbital and antipyrine in terms of both metabolic rates and urinary metabolite patterns. Thus the genetic defect in the metabolism of sparteine and debrisoquine is not a generalized deficiency of drug oxidation or of the cytochrome P450 system.