Abstract
In vitro studies were conducted to identify the hepatic enzyme(s) responsible for the oxidative metabolism of linezolid. In human liver microsomes, linezolid was oxidized to a single metabolite, hydroxylinezolid (M1). Formation of M1 was determined to be dependent upon microsomal protein and NADPH. Over a concentration range of 2 to 700 μM, the rate of M1 formation conformed to first-order (nonsaturable) kinetics. Application of conventional in vitro techniques were unable to identify the molecular origin of M1 based on the following experiments: a) inhibitor/substrates for various cytochrome P-450 (CYP) enzymes were unable to inhibit M1 formation; b) formation of M1 did not correlate (r2< 0.23) with any of the measured catalytic activities across a population of human livers (n = 14); c) M1 formation was not detectable in incubations using microsomes prepared from a baculovirus insect cell line expressing CYPs 1A1, 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A5, and 4A11. In addition, results obtained from an in vitro P-450 inhibition screen revealed that linezolid was devoid of any inhibitory activity toward the following CYP enzymes (CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4). Additional in vitro studies excluded the possibility of flavin-containing monooxygenase and monoamine oxidase as potential enzymes responsible for metabolite formation. However, metabolite formation was found to be optimal under basic (pH 9.0) conditions, which suggests the potential involvement of either an uncharacterized P-450 enzyme or an alternative microsomal mediated oxidative pathway.
Development of bacterial resistance has become a serious clinical problem for many classes of antibiotics (Lyytikainen et al., 1996). The 3-aryl-2-oxazolidinones are a relatively new class of synthetic antibacterial agents with a new mechanism of action that involves early inhibition of bacterial protein synthesis (Dresser and Rybak, 1998; Swaney et al., 1998). Linezolid1 (Zyvox) [(S)-N-[[3-[3-fluoro-4-(4-morpholinyl)phenyl]-2-oxo-5-oxazolidinyl]methyl]-acetamide] (Fig. 1) is a potent synthetic oxazolidinone currently in clinical development for the treatment of Gram positive bacterial infections (Burghardt et al., 1998; Rybak et al., 1998).
In humans, linezolid circulates mainly as parent drug and is eliminated via renal and nonrenal pathways. In a human radiolabel study using [14C]linezolid, the total recovery of drug-related radioactivity was near quantitative in 48 h, with the majority of the radioactive dose recovered in the urine (>65%) excreted primarily as parent drug and a major inactive, morpholine ring-opened carboxylic acid metabolite (Fig.1).2 The initial metabolic event leading to the formation of the ring-opened carboxylic acid metabolite would appear to occur via oxidation of the morpholine ring, which suggests a potential role for the monooxygenase enzyme system. However to date, the specific drug-metabolizing enzyme(s) responsible for the oxidation of linezolid have not been identified. The objective of the present study was to investigate the biotransformation of linezolid by human liver microsomes and to identify the hepatic enzyme(s) responsible for its metabolism. In addition, in vitro experiments were conducted to evaluate possible drug/enzyme inhibitory interactions between linezolid and the major constitutively expressed hepatic cytochrome P-450 (CYP) enzymes (CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4) found in human (Wrighton and Stevens, 1992). It is anticipated that this information will provide a rational means to assess the potential for linezolid to be involved in clinically relevant drug-drug interactions in vivo.
Experimental Procedures
Materials.
Linezolid (Fig. 1) synthesized with a carbon-14 radiolabel (46.54 mCi/mmol); in the acetyl moiety of the molecule, hydroxylinezolid (M1 authentic standard) and [14C]delavirdine were obtained from Pharmacia & Upjohn (Kalamazoo, MI). The radiochemical purity of [14C]linezolid was >98% as determined by HPLC with radiochemical detection. [14C](S)-mephenytoin, [14C]chlorzoxazone and [14C]diclofenac were purchased from Amersham Corp. (Arlington Heights, IL); [14C]testosterone was obtained from DuPont NEN (Boston, MA); [14C]para-nitrophenol, α-naphthoflavone (ANF), lauric acid (LAUR), coumarin (COUM), orphenadrine (ORPH), sulfaphenazole (SULF), retinoic acid (RETN),para-nitrophenol (pNITR), quinidine (QUIN), ketoconazole (KETO), and NADPH were purchased from Sigma Chemical Co. (St. Louis, MO). (S)-Mephenytoin (MEPH) was a gift from Dr. W. F. Trager, Department of Medicinal Chemistry, University of Washington (Seattle, WA). UltimaFlo M liquid scintillant was purchased from Packard Instrument Co. (Downers Grove, IL). All other reagents and solvents were of analytical grade.
Microsomes.
Human livers were acquired from the International Institute for the Advancement of Medicine (IIAM; Exton, PA). Liver microsomal protein isolation and the specific catalytic activity of individual isoforms of P-450 were determined as previously described (Wienkers et al., 1996). Microsomes from a baculovirus insect cell line expressing CYPs 1A1, 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A5, and 4A11 were purchased from Gentest (Woburn, MA).
Incubation Conditions.
A typical incubation (final volume 0.2 ml) consisted of 0.1 mg of microsomal protein in 100 mM potassium phosphate buffer (pH 7.4). Stock solutions of linezolid were prepared in methanol (final concentration of methanol was less than 0.3% v/v) by combining appropriate amounts of radiolabeled (approximately 0.1 μCi/incubation) and nonradiolabeled drug. The drug, buffer, and microsomes were mixed and preincubated at 37°C for 4 min. Incubations were initiated by the addition of the NADPH, and incubations were conducted at 37°C for 30 min. For control incubations, NADPH was omitted. Reactions were terminated upon addition of 200 μl of acetonitrile, after which samples were vortex mixed and centrifuged for 15 min at 14,000g. The subsequent supernatants were transferred to a HPLC autosampler vial, capped, and samples were kept refrigerated until radio-HPLC or mass spectrometric analysis.
Radio-HPLC.
Analytical separation of linezolid and its metabolite were achieved using a binary gradient HPLC system equipped with a Perkin-Elmer Series 200 pump and autosampler (Perkin-Elmer, Norwalk, CT) equipped with a chilled sample tray maintained at 4°C. The analytical column was a reverse-phase Zorbax SB-CN (250 × 4.6 mm, 5 μm particle size) (Mac-Mod Analytical, Chadds Ford, PA). The mobile phase consisted of solvent A (90%:10%:0.2% water:methanol:acetic acid) and solvent B (10%:90%:0.2% water:methanol:acetic acid). Initial mobile phase conditions (100% solvent A) at a rate of 1 ml/min were held for 5 min, followed by a step gradient to 40% solvent B in 10 min, followed by a second step gradient to 90% solvent B in 5 min, the final conditions were held for 5 min, then returned to the original starting conditions. Quantitation of linezolid and M1 were detected using a FLO-ONE/Beta Series A500 flow-through radioactivity detector (Packard/Radiomatic, Meriden, CT). UltimaFlo M liquid scintillant was introduced postcolumn at a rate of 3 ml/min. The fractional contribution of each metabolite to total radioactivity was used to calculate the rates of metabolite formation.
Liquid Chromatography (LC)/Electrospray/Mass Spectrometry (MS) and Metabolite Confirmation.
Analytical separation and atmospheric pressure chemical ionization (APCI)-MS detection and verification of linezolid and M1 was accomplished using the method described below. A Finnigan TSQ7000 triple quadrupole mass spectrometer (Finnigan-MAT, San José, CA), utilizing APCI, was used in the third quadrupole (Q3) full scan mode for the identification and semiquantitative analysis. Metabolite identification was accomplished while operating the MS in positive ion mode scanning the Q3 from 150 to 450 amu at a scan rate of 150 amu/s. The capillary and tube lens voltages were 32 and 60 V, respectively. Nitrogen was used as a drying gas at a sheath pressure of 80 psi with auxiliary flow. The HPLC conditions used for the analysis are identical with the method described above.
Correlation Analysis.
The rate of formation for M1 was determined across a panel of liver microsomes prepared from 14 different human organ donors. The rates of formation of hydroxylinezolid were compared to the catalytic activities previously characterized for specific P-450 substrates (Wienkers et al., 1996). Incubations and sample work-up were carried out as described above. Correlation of determination (r2) for enzyme activities were determined by linear regression using the graphical/statistical program Prism 2.01 (GraphPad, San Diego, CA).
Chemical Inhibition Experiments.
Linezolid was incubated in pooled human liver microsomes in the presence of a panel of compounds that interacted selectively with various cytochrome P-450 enzymes. The following P-450 enzyme substrates/inhibitors were examined for their ability to alter the microsomal metabolism of linezolid: ANF (50 μM), COUM (50 μM), ORPH (50 μM), RETN (50 μM), SULF (9 μM), MEPH (250 μM), QUIN (5 μM), pNITR (100 μM), KETO (5 μM), and LAUR (50 μM). All the inhibitors were dissolved in methanol, and were added to the incubations such that the final amount of methanol was 1%. Control incubations (minus inhibitor) also contained 1% methanol.
Metabolism by cDNA-Expressed Microsomes.
The metabolism of linezolid was assessed in microsomes prepared from a baculovirus insect cell line expressing CYPs 1A1, 1A2, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1 3A4, 3A5, and 4A11. The incubations were conducted in a manner similar to the experiments described above with 50 μM [14C]linezolid and equivalent concentrations P-450 (40 pmol of CYP/ml) for each P-450 isoform tested.
P-450 Inhibition Screen.
The ability of linezolid to inhibit P-450 enzymes was investigated against six different cDNA-expressed human cytochrome P-450 enzyme systems (CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4). Incubation reactions, sample work-up, and quantitation of CYP marker metabolite formation using HPLC/radiochemical detection was conducted as previously described (Wynalda and Wienkers, 1997).
pH Study.
The effect of varying hydrogen ion concentrations on P-450 activity and M1 formation was investigated. Incubations (final volume 0.2 ml) consisted of 0.1 mg of microsomal protein, 100 μM linezolid in 100 mM potassium phosphate buffer adjusted to either pH 4.0, 7.4, or 9.0. Reaction conditions, sample preparation, and storage were identical with procedures described above. In addition, parallel incubations were performed with human liver microsomes and P-450 marker substrates, [14C]testosterone, lsqb]14C]chlorzoxazone, and [14C]diclofenac under the same buffer conditions (i.e., pH 4.0, 7.4, and 9.0). Incubation conditions, sample work-up, and analysis of the marker substrates were identical with procedures previously described (Wynalda and Wienkers, 1997).
Results and Discussion
Metabolite profiles obtained from human liver microsomes indicate that [14C]linezolid was oxidized to a single metabolite (Fig. 2). Cochromatography coupled with LC/MS using an authentic standard identified the metabolite, M1, as a hemiacetal metabolite arising via morpholine ring oxidation (Fig. 1). Formation of M1 was dependent upon coincubation with NADPH and was proportional with time (up to 60 min at 0.3 mg of protein) and protein concentration (up to 0.5 mg/ml protein for 30 min) at a substrate concentration of 100 μM (results not shown). In addition, linezolid incubation conducted in the presence of denatured (boiled) human liver microsomes supplemented with cofactor did not generate measurable amounts of M1 (data not shown). The effects of substrate concentration on the rate of M1 formation was determined in pooled human liver microsomal preparations. Under the previously described incubation conditions, formation of M1 did not undergo saturable kinetics [i.e., across all concentrations tested (up to the limit of solubility for linezolid, 700 μM) a linear relationship existed between V (the rate of M1 formation) and linezolid concentration], which suggests that the enzyme(s) responsible for M1 formation have a high apparent Km value (>700 μM).
The effects of various P-450 substrates and inhibitors on the metabolism of linezolid in human liver microsomes were investigated. Although some of the inhibitors used in this experiment interact with more than one P-450 isoform (Newton et al., 1995), they do so with differing enzyme affinities, such that with appropriate inhibitor concentrations it is possible to interact predominantly with the target CYP enzyme. Inhibitor concentrations chosen in the current study were selected to produce greater than 80% inhibition of total enzyme activity based on literature Ki values for each chemical and a substrate affinity constant for linezolid of 700 μM for M1 formation. The data presented in Table1 are expressed as a percentage of control (minus inhibitor) activity. From the data presented in Table 1, it appears that only coumarin, an inhibitor of CYP2A6 (Pearce et al., 1992), inhibited the microsomal oxidation of linezolid. In addition, in vitro studies aimed to correlate the sample to sample variation in M1 formation across a panel of human liver revealed that the rates of formation for M1 did not correlate (r2 < 0.23) with any of the previously measured CYP activities. Moreover, incubations with linezolid (100 μM) in the presence of microsomes selectively containing cDNA-expressed human CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP3A4, CYP3A5, and CYP4A11 were devoid of any discernable linezolid oxidase activity. In the case of linezolid, concentrations up to 100 μM were found not to have any inhibitory effect on specific P-450 enzyme marker activities (Table2). This observation conservatively suggests that if linezolid were an inhibitor of one of the P-450 enzymes tested, the inhibition would be characterized by an apparentKi of greater than 500 μM. Therefore, as long as the presence of linezolid does not drastically alter the dispositional characteristics of a second drug, linezolid should not substantially alter the metabolism of a second drug whose clearance is mediated by the major hepatic P-450 enzymes.
Many hypotheses are plausible for the failure of conventional in vitro techniques to identify the metabolic origin of M1. The field of possible enzymes found in microsomes capable of oxidizing linezolid is narrowed as formation of M1 was found to be NADPH-dependent and therefore excludes the potential involvement of monoamine oxidase (Dixon et al., 1994). It is possible that an alternative NADPH-dependent enzyme system [e.g., flavin-containing monooxygenase (FMO)] metabolizes linezolid. The likelihood of this enzyme system being the linezolid hydrolase is minimal, because the oxidation of linezolid occurs at the carbon adjacent to the morpholine oxygen, and FMO primarily oxidizes nitrogen and sulfur atoms (Poulsen and Zigler, 1995). However, to exclude the possibility of FMO involvement in linezolid oxidation, incubations were performed to differentiate P-450 activity from FMO activity using heat-inactivated microsomes (Grothusen et al., 1996). The results of these experiments revealed that pretreatment of microsomes with elevated temperatures had no effect on linezolid hydroxylase activity compared to that of control (e.g., untreated) microsomes (data not shown). In addition, reactions conducted with human liver microsomes pretreated with CO and incubated under an atmosphere of CO resulted in a marked reduction of M1 formation (23% M1 formation compared to control), which suggests at least an intermediary role of P-450 in M1 formation (Koley et al., 1994).
To clarify the molecular basis for M1 formation in vitro, the possible role of reactive oxygen species (ROS) that arises via the uncoupling the electron flux between oxidoreductase and cytochrome P-450 was investigated. The electron flux from NADPH to cytochrome P-450 by P-450 reductase is fairly constant from pH 7.0 to 9.5 (Ahmed, 1996). However the P-450 catalytic activity has a very restricted pH optima (∼7.1–7.6) (McManus et al., 1987). Therefore, to probe the role of ROS and not P-450 metabolism in the generation of M1, incubations were conducted at three pH values, 4.0, 7.4, and 9.0. In addition to microsomal incubations with linezolid, parallel incubations were conducted under identical pH conditions using testosterone, chlorzoxazone, and diclofenac as surrogate markers of P-450 metabolic activities for CYP3A4, CYP2E1, and CYP2C9, respectively. Under the various pH environments linezolid was oxidized to M1 at both pH 7.4 and 9.0 and not at pH 4.0 (data not shown). However, the three P-450 marker substrates were selectively metabolized in the incubations performed at pH 7.4 (data not shown). This data suggests that the observed oxidation of linezolid at higher pH conditions may be the result of noncatalytic interactions between microsomal P-450 enzymes and P-450 reductase, which results in the formation of ROS.3 The notion that linezolid may be susceptible to chemical oxidation via reactive oxygen species in part reflects the electrochemical potential observed for other compounds that possess a morpholino moiety and their ability to serve as efficient reactive oxygen scavengers (Gilbert et al., 1974;Dan et al., 1989; Tani et al., 1994).
Linezolid clearance appears to be via formation of M1, which serves as a precursor to the major urinary metabolite, the morpholine ring-opened carboxylic acid.2 The current in vitro findings suggest that the production of ROS represents one avenue for the oxidation of linezolid in human liver microsomes. Extrapolation of these in vitro results to the in vivo condition makes it difficult to predict the magnitude to which this process contributes to linezolid clearance. Coincubation of linezolid did not inhibit the metabolic activity of the following P-450 enzymes: CYP1A2, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4. In terms of predicting potential drug-drug interactions, given the lack of any discernable interaction (inhibitory or catalytic) toward the human hepatic cytochrome P-450 enzymes tested, clinically important interactions between linezolid and coadministered drugs that are metabolized by these enzymes appear unlikely.
Footnotes
-
Send reprint requests to: Larry C. Wienkers, Ph.D., Drug Metabolism Research, Pharmacia Corporation, 7265-300-313, 301 Henrietta St., Kalamazoo, MI. E:mail: larry.c.wienkers{at}am.pnu.com
-
↵2 Feenstra KL, Slatter JG, Stalker DJ, Welshman IR, Sams JP, Hauer MJ, Cathcart KS, Verburg MT, Johnson MG, Bothwell BE, Koets MD, Peng GW, Stryd RP and Fagerness PE (1998) Metabolism and excretion of the oxazolidinone antibiotic linezolid (PNU-100766) following oral administration of [14C]PNU-100766 to healthy human volunteers, 8th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC). Session 92-A, Poster A-53, 1998 Sept 24–27, San Diego, CA.
-
↵3 Additional studies aimed to elucidate the mechanism of linezolid microsomal oxidation are currently underway.
- Abbreviations used are::
- linezolid (Zyvox)
- (S)-N-[[3-[3-fluoro-4-(4-morpholinyl)phenyl]- 2-oxo-5-oxazolidinyl]methyl]-acetamide
- CYP or P-450
- cytochrome P-450
- QUIN
- quinidine
- KETO
- ketoconazole
- MEPH
- (S)-mephenytoin
- ANF
- α-naphthoflavone
- LAUR
- lauric acid
- SULF
- sulfaphenazole
- COUM
- coumarin
- ORPH
- orphenadrine
- RETN
- retinoic acid
- pNITR
- para-nitrophenol
- LC
- liquid chromatography
- MS
- mass spectrometry
- ROS
- reactive oxygen species
- FMO
- flavin-containing monooxygenase
- APCI
- atmospheric pressure chemical ionization
- Received December 3, 1999.
- Accepted May 17, 2000.
- The American Society for Pharmacology and Experimental Therapeutics