cytochrome p450 2e1 is the principal catalyst of human oxidative

Cytochrome P450 2E1 is the Principal Catalyst of HumanOxidative Halothane Metabolism in Vitro1

DOUGLAS K. SPRACKLIN, DOUGLAS C. HANKINS, JEANNINE M. FISHER, KENNETH E. THUMMEL andEVAN D. KHARASCH

Departments of Anesthesiology (D.K.S., D.C.H., E.D.K.), Pharmaceutics (J.M.F., K.E.T.) and Medicinal Chemistry (E.D.K.), University ofWashington, Seattle, Washington

Accepted for publication December 6, 1996

ABSTRACTThe volatile anesthetic halothane undergoes substantial bio-transformation generating metabolites that mediate hepatotox-icity. Aerobically, halothane undergoes cytochrome P450-cat-alyzed oxidation to trifluoroacetic acid (TFA), bromide and areactive intermediate that can acetylate liver proteins. Theseprotein neo-antigens stimulate an immune reaction that medi-ates severe hepatic necrosis (“halothane hepatitis”). This inves-tigation identified the human P450 isoform(s) that catalyze ox-idative halothane metabolism. Halothane oxidation by humanliver microsomes was assessed by TFA and bromide formation.Eadie-Hofstee plots of TFA and bromide formation were bothnonlinear, suggesting the participation of multiple P450s. Mi-crosomal TFA and bromide formation were inhibited 45 to 66%and 21 to 26%, respectively, by the P450 2A6 inhibitors 8-me-thoxypsoralen and coumarin, 84 to 90% by the P450 2E1inhibitor 4-methylpyrazole and 55% by diethyldithiocarbamate,an inhibitor of both P450 2A6 and 2E1. Selective inhibitors ofP450s 1A, 2B6, 2C9/10, 2D6 and 3A4 did not affect halothaneoxidation. At saturating halothane concentrations (2.4 vol%)only cDNA-expressed P450 2A6 and 2B6 catalyzed significantrates of TFA and bromide formation, and P450 2E1 catalyzedcomparatively minimal oxidation. Conversely, at subsaturatinghalothane concentrations (0.30 vol%), metabolism by P450 2E1

exceeded that by P450 2A6. Among a panel of human livermicrosomes, there were significant linear correlations betweenhalothane oxidation and P450 2A6 activity and protein contentat saturating halothane concentrations (2.4 vol%), and a signif-icant correlation between metabolite formation and P450 2E1activity (but not P450 2A6 activity) at subsaturating concentra-tions (0.12 vol%). These experiments suggested P450 2A6 and2E1 as the predominant catalysts at saturating and subsaturat-ing halothane concentrations, respectively. Further kinetic anal-ysis using cDNA-expressed P450 and liver microsomes clearlydemonstrated that P450 2E1 is the high affinity/low capacityisoform (Km 5 0.030-0.053 vol%) and P450 2A6 is the lowaffinity/high capacity isoform (Km 5 0.77-1.2 vol%). Evidencewas also obtained for substrate inhibition of P450 2E1. The invitro clearance estimates (Vmax/Km) for microsomal P450 2E1(4.3-5.7 ml/min/g) were substantially greater than those formicrosomal P450 2A6 (0.12-0.21). These clearances, as well asrates of apparent halothane oxidation predicted from kineticparameters in conjunction with plasma halothane concentra-tions measured during clinical anesthesia in humans, demon-strated that both P450 2E1 and P450 2A6 participate in humanhalothane metabolism, and that P450 2E1 is the predominantcatalytic isoform.

Halothane is one of the more widely used volatile anesthet-ics in the world. Halothane undergoes extensive biotransfor-mation with approximately 50% of an administered dosemetabolized by reductive and oxidative pathways, both evi-dent during routine anesthesia (Carpenter et al., 1986). Nu-merous epidemiological, clinical and laboratory investiga-tions have established that halothane metabolism mediatesboth a mild and severe form of hepatic toxicity (for recent

reviews see Cousins et al., 1989; Ray and Drummond, 1991;Gut et al., 1995). Mild hepatic reactions occur in up to 20% ofhalothane anesthetics (Ray and Drummond, 1991) and areevidenced clinically by elevated postoperative liver enzymes(AST, ALT, GST) (de Groot and Noll, 1983; Akita et al., 1989;Sato et al., 1990). Current theories suggest that this mild,subclinical hepatotoxicity is attributable to anaerobic reduc-tive halothane metabolism that results in free radical gener-ation and lipid peroxidation (de Groot and Noll, 1983; Sato etal., 1990; Awad et al., 1996). The severe form of hepatictoxicity is a rare but often fatal fulminant hepatic necrosis,commonly known as “halothane hepatitis” (Ray and Drum-

Received for publication July 23, 1996.1 This work was supported by grants from the National Institutes of Health

(R01 GM48712, GM 48349) and by a Pharmaceutical Research and Manufac-turers of America Foundation Faculty Development Award to E.D.K.

ABBREVIATIONS: cDNA, complementary deoxyribonucleic acid; GC/MS, gas chromatography-mass spectrometry; HLM, human liver micro-somes; HPLC, high performance liquid chromatography; IOD, integrated optical density; NADPH, b-nicotinamide adenine dinucleotide, reducedform; P450, cytochrome P450; TFA, trifluoroacetic acid; vol%, headspace concentration (v/v); AST, aspartate aminotransferase; ALT, alanineaminotransferase; GST, glutathione S-transferase.

0022-3565/97/2811-0400$03.00/0THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 281, No. 1Copyright © 1997 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A.JPET 281:400–411, 1997

400

at ASPE

T Journals on A

pril 10, 2018jpet.aspetjournals.org

Dow

nloaded from

http://jpet.aspetjournals.org/

mond, 1991). Clinically, halothane hepatitis occurs in ap-proximately 1:6,000 to 35,000 halothane anesthetics (Rayand Drummond, 1991) but is fatal in 75% of these cases(Cousins et al., 1989). It is manifested by fever, jaundice andgrossly elevated serum transaminase concentrations. Patho-logically, it is characterized by massive centrilobular necro-sis. The cytotoxicity associated with halothane hepatitis isconsistent with an immunological reaction to trifluoroacety-lated (TFA) liver protein neo-antigens. These TFA-antigensderive from acylation of native liver proteins that, in suscep-tible individuals, serve as neo-antigens that stimulate theformation of anti-TFA-protein antibodies. On reexposure tohalothane or certain other volatile anesthetics, these anti-bodies initiate an immunological cascade that ultimately re-sults in “halothane hepatitis.” TFA-protein formation resultsfrom oxidative halothane metabolism. Investigations suggestthat the amount of antigen formation, and thus, the rate andextent of halothane metabolism may be a critical regulatoryfactor in the onset of halothane hepatitis (Christ et al., 1988a,1988b; Pohl et al., 1989; Kenna et al., 1990). For example, inimmunoblot analysis, the sera from six patients with halo-thane hepatitis cross-reacted with neo-antigens of halothane-and enflurane-treated rats but not with isoflurane-treatedanimals (Christ et al., 1988a). The relative amount of cross-reaction (halothane.enflurane.isoflurane) correlates withthe extent of metabolism of the three anesthetics. Therefore,the seminal event in immune-based halothane hepatitis isP450-catalyzed oxidative halothane metabolism (Kenna etal., 1987).

The anaerobic reductive pathway of halothane metabolismhas been well-described (Ray and Drummond, 1991; Sprack-lin et al., 1996). The oxidative pathway is shown in figure 1.Under sufficient oxygen tensions, halothane undergoes P450-catalyzed oxidation to trifluoroacetyl chloride, with concom-itant loss of bromine. This unstable intermediate undergoesfurther reactions, including: 1) hydrolysis to yield the non-toxic metabolite TFA; 2) binding to phospholipids (Mullerand Srier, 1982) and 3) acetylation of tissue proteins to formthe TFA-protein adducts (Ray and Drummond, 1991). Todate, a number of the TFA-modified proteins have been iden-tified. These include protein disulfide isomerase, microsomalcarboxylesterase, calreticulin, stress protein ERp72 andERp99/endoplasmin/GRP 94 in microsomes (Gut et al., 1995),and glutathione-S-transferase in cytosol (Brown and Gan-dolfi, 1994). However, the precise antigens that cause halo-thane hepatitis are unknown.

The mechanism and consequences of oxidative halothanemetabolism have been well studied; nonetheless, cliniciansare still unable to identify which patients will develop anti-TFA proteins, or which patients will develop halothane hep-atitis. However, it is clear that P450-catalyzed oxidativehalothane metabolism is the critical initiating event, andinhibition of P450-catalyzed halothane oxidation is a prom-ising potential clinical strategy to prevent halothane hepati-tis (Kharasch et al., 1996). Nevertheless, the exact identity ofthe P450 isoform(s) that catalyze oxidative metabolism ofhalothane in humans is unknown. Previous investigationshave suggested a role for P450 2E1 with possible involvementof P450 2A6 (Brown et al., 1995; Kharasch et al., 1996).Disulfiram, an effective P450 2E1 inhibitor, significantly di-minished halothane oxidation in humans (Kharasch et al.,1996), however disulfiram may inhibit P450 2A6 as well asP450 2E1. Therefore, the purpose of this investigation was toclarify the human liver P450 isoform(s) responsible for oxi-dative halothane metabolism.

Materials and MethodsHalothane was purchased from Halocarbon Laboratories (N. Au-

gusta, SC). Furafylline and sulfaphenazole were generous gifts fromDrs. Kent Kunze and William F. Trager, respectively (University ofWashington, Seattle, WA). Sodium trifluoroacetate, sodium bromideand chlorodifluoroacetic acid were purchased from Fluka ChemicalCo. (Ronkonkoma, NY) and were of the highest purity available.Microsomes containing individual cDNA-expressed cytochrome P450isoforms were purchased from Gentest (Woburn, MA). Unless spec-ified, all other reagents were purchased from Sigma Chemical Co.(St. Louis, MO) and were of the highest purity available. All buffersand reagents were prepared with high-purity ($18.2 MVzcm) water(Milli-Q, Millipore, Bedford, MA).

Microsomes were prepared from human livers as described previ-ously (Kharasch and Thummel, 1993). Microsomal protein concen-trations were determined by the method of Lowry et al. (1951). Totalmicrosomal cytochrome P450 content was determined from the re-duced minus oxidized carbon monoxide difference spectrum (Es-tabrook et al., 1972).

Halothane metabolism was determined in scintillation vials (24.4ml) containing human liver microsomes (2 mg/ml), halothane andNADPH (2 mM) in a final volume of 1.0 ml potassium phosphatebuffer (50 mM, pH 7.4). The reaction was initiated by the addition ofhalothane that was added either undiluted (2 ml, producing a head-space concentration of 2.4 vol% at 37°C) for experiments at saturat-ing substrate concentrations, or diluted in methanol (final aqueousmethanol concentration 0.2%) for experiments at subsaturating sub-

Fig. 1. Mechanism of TFA and bromide formation from P450-catalyzed halothane oxidation.

1997 Halothane Oxidation in Vitro by CYP 2E1 401

at ASPE

T Journals on A


Dow

nloaded from


strate concentrations. Preliminary experiments showed that thismethanol concentration did not affect the rate of halothane oxida-tion. Incubations were routinely carried out for 60 min at 37°Cunless indicated otherwise. Reactions were quenched by quantita-tively transferring the reaction mixture to a 2-ml polypropylene vialcontaining trichloroacetic acid (10 ml of a 6 N aqueous solution; 60mmol) and mixing thoroughly for 10 sec. The internal standardchlorodifluoroacetic acid (25 ml of a 1 mM aqueous solution; 25 nmol)was then added and the solution vortexed for another 10 sec. Thevials were placed on ice for 15 min and then centrifuged at 16,000 3g for 30 min. The resulting supernatant was filtered through a 0.2mm DIMEX syringe filter (Millipore) directly into an autosamplervial. Experiments using cDNA-expressed protein were carried outsimilarly using typical protein concentrations of 1 mg/ml and incu-bation times of 60 or 90 min.

Metabolites and internal standards were analyzed by ion HPLCwith conductivity detection. Analyses were performed using a DX-300 HPLC-IC system (Dionex Corp. Sunnyvale, CA), consisting of anAGP gradient pump, LCM-3 chromatography module, CDM-3 con-ductivity detector, ASM-3 autosampler, IonPac AS11 analytical col-umn and AG11 guard column and an ASRS-1 anion self-regenerat-ing suppressor operating in the autosuppression recycle mode.Dionex AI-450 software was used to control the hardware, detectorsignal acquisition and chromatographic peak integration. Injectionsof 50 ml were made via an autosampler utilizing 0.5-ml polypro-pylene vials equipped with 20-mm filter caps (PolyVial, DionexCorp.). The sodium hydroxide concentration was initially 0.75 mMfor 5 min, linearly decreased to 0.5 mM at 0.1 mM/min, linearlyincreased to 3.0 mM over 10.5 min, increased to 80 mM at 25mM/min and held at 80 mM for 5 min. The concentration was thenlinearly decreased to 0.75 mM at 14.5 mM/min, and the columnallowed to reequilibrate at this concentration for 7 min. The eluantflow rate was 2.0 ml/min and the detector sampling rate was 0.20 sec.

Under these conditions, the retention times for trifluoroacetate,bromide and chlorodifluoroacetate were 6.1, 7.8 and 11.2 min, re-spectively. Trichloroacetic acid did not interfere with analyte quan-titation. Standards were prepared by adding aqueous TFA (0.25-40nmol), sodium bromide (0.25-40 nmol) and the internal standard tomicrosomal mixtures prepared similarly to the incubation mixturesexcept for the omission of halothane. Standard curves for TFA andbromide were constructed from peak area ratios of metabolite tointernal standard. Standard curves were linear over the concentra-tion range 0 to 40 mM (r2 5 0.99). The lower limit of quantitation (0.1mM) was defined as a signal to noise ratio of 3:1. Quantitation ofmetabolic TFA and bromide was accomplished by comparing samplepeak area ratios to those of the standard curve.

For kinetic experiments measuring TFA and bromide formation asa function of substrate concentration, headspace halothane concen-trations were measured as previously described (Spracklin et al.,1996). Briefly, incubation mixtures were prepared identically tothose used to assess TFA and bromide formation except that NADPHwas omitted and the vial was sealed with a rubber septum instead ofa screw cap. After 10 min at 37°C, an aliquot of the headspace gasfrom the reaction vial was transferred to another sealed vial. Head-space GC/MS analysis was used to quantitate halothane. Microsomalhalothane concentrations were measured by a previously reportedmethod used to measure whole blood halothane concentrations(Kharasch et al., 1996). After 10 min at 37°C, an aliquot of thereaction mixture was added to heptane. After centrifugation, analiquot of the heptane was analyzed by gas chromatography to quan-titate halothane.

Experiments with isoform-selective P450 inhibitors were con-ducted at the following final concentrations: 7,8-benzoflavone (P4501A, 16 mM), furafylline (P450 1A2, 20 mM), 8-methoxypsoralen (P4502A6, 28 mM), coumarin (P450 2A6, 36 mM), orphenadrine (P450 2B6,5 mM), sulfaphenazole (P450 2C9/10, 3.6 mM), (S)-mephenytoin(P450 2C19, 100 mM), quinidine (P450 2D6, 45 nM), 4-methylpyra-zole (P450 2E1, 540 mM), diethyldithiocarbamate (P450 2E1, 100

mM), troleandomycin (P450 3A4, 100 mM), ketoconazole (P450 3A4,90 nM), n-octylamine (P450, 3 mM). All inhibitors were added inpotassium phosphate buffer except 7,8-benzoflavone, 8-methoxypso-ralen, (S)-mephenytoin, troleandomycin, ketoconazole and n-oc-tylamine which were diluted in methanol (final methanol concentra-tion 0.2%). Substrate and inhibitor concentrations were chosen totheoretically suppress more than 80% of isoform activity based onpublished Ki values. In experiments using the competitive inhibitors7,8-benzoflavone, coumarin, 8-methoxypsoralen, sulfaphenazole, (S)-mephenytoin, quinidine, 4-methylpyrazole and ketoconazole, the in-hibitor was added followed by a solution of halothane [2 ml of 10%(v/v) halothane/methanol, equivalent to 0.2 ml halothane or 0.24vol%, final methanol concentration 0.2%], and the reaction was ini-tiated by the addition of the NADPH. Reactions were carried out at37°C for 60 min and then quenched with trichloroacetic acid asdescribed above. Incubations containing the mechanism-based inhib-itors furafylline, orphenadrine, diethyldithiocarbamate, troleando-mycin and n-octylamine were first preincubated at 37°C for 15 minwith NADPH under aerobic conditions, after which time halothanewas added (2 ml; 2.4 vol%). Reactions were carried out at 37°C for 60min and then quenched as described previously. Kinetic and inhib-itor experiments were performed in a single liver that was competentin all the major drug metabolizing P450 isoforms.

Microsomal catalytic activities of P450s 1A2, 2A6, 2C9, 2D6, 2E1and 3A4 were measured by (R)-warfarin 6-hydroxylation, coumarin7-hydroxylation, (S)-warfarin 7-hydroxylation, metoprolol a-hy-droxylation, chlorzoxazone 6-hydroxylation and midazolam 19-hy-droxylation, respectively (Miles et al., 1990; Kharasch and Thummel,1993; Kunze et al., 1996; Thummel et al., 1996). Microsomal P450isoform content was determined by Western blot analysis as de-scribed previously (Kharasch and Thummel, 1993; Thummel et al.,1993). In addition to P450 2E1, anti-P450 2E1 antibody also detecteda separately migrating lower molecular weight protein recently iden-tified as P450 2A6 (K. E. Thummel, unpublished results). This an-tibody was used to quantitate P450 2A6 content using cDNA-ex-pressed P450 2A6 as the standard.

All results are expressed as the mean 6 S.D.of three experiments.Typical coefficients of variation in a set of triplicate measurementswere # 10%. Statistical analyses were carried out with SigmaStat(version 1.02) and nonlinear regression analyses were carried outwith SigmaPlot (version 5.01) (Jandel Scientific, San Rafael, CA).

A clinical investigation was conducted to determine blood halo-thane concentrations during anesthesia. The investigation was ap-proved by the institutional Human Subjects Committee and all pa-tients provided written informed consent. Twenty normal weightmales, without hepatic or renal disease, ethanol abuse or current useof medications known to alter hepatic drug metabolism who wereundergoing anesthesia for elective surgery that did not significantlyalter hepatic blood flow were studied. All patients were anesthetizedwith 1.0% end-tidal halothane (determined by an infrared detector;Capnomax, Datex Medical, Tewksbury, MA) for 3 hr. The inspiredhalothane concentration was adjusted to maintain 1% end-tidal halo-thane concentration. Venous blood samples for determination ofblood halothane concentration were obtained before anesthesia andat hourly intervals thereafter. Whole blood halothane concentrationswere determined by gas chromatography. Details have been pub-lished previously (Kharasch et al., 1996).

ResultsDetails of oxidative halothane metabolism by human liver

microsomes have not been reported previously. Therefore,initial experiments characterized the NADPH-, time- andprotein-dependence of the reaction. Ion HPLC chromato-grams showing trifluoroacetate (RT 5 6.1 min), and bromide(RT 5 7.8 min) produced in a 60-min incubation of halothaneand human liver microsomes in the presence (fig. 2A) or

402 Spracklin et al. Vol. 281

at ASPE

T Journals on A


Dow

nloaded from


absence (fig. 2B) of NADPH illustrate that no TFA or bro-mide formation was observed in the absence of NADPH. TFAand bromide formation increased linearly with time for 60min (fig. 2C). Formation of both metabolites was also linearwith protein concentration up to 5 mg/ml (fig. 2D). Adequatesensitivity was obtained at 2.0 mg/ml which was used forsubsequent incubations.

Anesthetic doses delivered during surgery are measured bythe concentration (in volumes percent) in inspired gas. Tis-sue anesthetic concentrations are most closely reflected bythose in expired alveolar gas (end-tidal concentration), butactual hepatic concentrations corresponding to those in re-spiratory gas are unknown. Therefore halothane was addedto microsomes in sealed vials and the concentration of halo-thane in both the headspace gas and the microsomal suspen-sions was measured (fig. 3). There was a linear relationshipbetween halothane concentrations (vol%) in the headspacegas and the amount of halothane added up to 3 ml. Similarly,

there was a linear relationship between halothane concen-trations (mM) in the microsomal suspension and the amountof halothane added, for halothane additions up to 1 ml. Aque-ous solubility was limited at higher halothane concentra-tions. The relationship between headspace and solution con-centrations is shown in the inset to figure 3. For example, aheadspace concentration of 0.6 vol% corresponded to a sus-pension concentration of 400 mM, resulting from 0.5 ml ofhalothane. Calculation of a microsome:gas partition coeffi-cient (l) from the measured microsomal halothane concen-trations yielded a value of 2.2. This is in good agreement withthe blood:gas partition coefficient (l 5 2.3) reported for halo-thane (Eger, 1985). Thus the blood:gas partition coefficientappears to be a reasonable approximation for the microsome:gas partition coefficient.

The substrate concentration dependence of halothane oxi-dation in human liver microsomes was examined over therange of 0 to 3 ml added, producing a headspace concentrationof 0 to 3.6 vol% (fig. 4). Eadie-Hofstee plots for both TFA andbromide formation were nonlinear, suggesting the participa-tion of multiple enzymes in halothane oxidation. Experimen-tal data were fit to a two-enzyme Michaelis-Menten model bynonlinear regression analysis. The parameters obtained aresummarized in table 1. For TFA formation, the apparentparameters obtained were Vmax(1) 5130 pmol/min/mg pro-tein, Km(1) 5 0.045 vol% (30 mM); Vmax(2) 5 94 pmol/min/mgprotein, Km(2) 5 1.2 vol% (800 mM). For bromide, the appar-ent parameters were Vmax(1) 5 170 pmol/min/mg protein,Km(1) 5 0.045 vol% (30 mM); Vmax(2) 5 130 pmol/min/mgprotein, Km(2) 5 0.94 vol% (630 mM).

To identify the P450 isoforms responsible for oxidativehalothane metabolism, the effect of isoform-selective P450inhibitors on rates of TFA and bromide formation were de-termined (fig. 5). The P450 2A6-selective inhibitor 8-me-thoxypsoralen (Maenpaa et al., 1994), used as a competitiveinhibitor, decreased the rate of TFA and bromide formationby 45 and 66%, respectively. The P450 2A6 substrate couma-

Fig. 2. NADPH-, time- and protein-de-pendence of TFA and bromide formationfrom P450-catalyzed halothane oxida-tion. Ion HPLC chromatograms showingTFA and bromide formation in a 60 minincubation of halothane with human livermicrosomes as described in “Methods.”Incubations were carried out in a singleliver that was competent in all the majordrug metabolizing P450 isoforms. Incu-bations were performed in the (A) pres-ence or (B) absence of NADPH. Alsoshown are rates of TFA and bromideformation as a function of (C) time (2mg/ml protein) and (D) protein concen-tration (incubation time was 60 min).

Fig. 3. Relationship between halothane addition and measured halo-thane concentrations. Halothane concentrations were measured in theheadspace (triangles) and in the microsomal suspension (circles) afteraddition of halothane to sealed vials containing human liver micro-somes (2 mg/ml) and equilibrated at 37°C for 10 min. The inset showsthe relationship between halothane concentrations measured in theheadspace and the microsomal suspension (r 5 0.99; P , .001).


at ASPE

T Journals on A


Dow

nloaded from


rin decreased the rate of TFA and bromide formation by 26and 21%, respectively. Rates of TFA and bromide formationwere both inhibited 84 to 90% by the P450 2E1-selectiveinhibitor 4-methylpyrazole. Additionally, TFA and bromideformation were each inhibited 55% by diethyldithiocarbam-ate, an inhibitor of P450 2E1 and 2A6. In contrast, the P4501A-, 2B6-, 2C9/10-, 2C19-, 2D6-, 3A4-selective inhibitors 7,8-benzoflavone, furafylline, orphenadrine, sulfaphenazole, (S)-mephenytoin, quinidine, troleandomycin and ketoconazolehad no significant effect on rates of TFA or bromide forma-tion. Additionally, the nonselective P450 inhibitor n-oc-tylamine (Jefcoate et al., 1969) almost completely inhibitedTFA and bromide formation. These results suggested theinvolvement of both P450 2E1 and 2A6 in halothane oxida-tion, and did not provide evidence for the participation ofother P450 isoforms.

To identify further the isoforms responsible for oxidativehalothane metabolism, the rates of TFA and bromide forma-tion by cDNA-expressed P450 isoforms were examined (fig.6). At saturating halothane concentrations (2 ml 5 2.4 vol%),P450 2A6 (0.45 pmol/min/pmol P450) and P450 2B6 (0.47pmol/min/pmol P450) catalyzed significant and comparablerates of TFA formation. P450 1A2 (0.050 pmol/min/pmolP450), P450 2C9 (,0.001 pmol/min/pmol P450), P450 2D6(0.036 pmol/min/pmol P450), P450 2E1 (0.084 pmol/min/pmolP450) and P450 3A4 (,0.001 pmol/min/pmol P450) catalyzedmuch lower amounts of TFA formation. A similar pattern forbromide formation was observed. At saturating halothaneconcentrations, P450 2A6 (0.34 pmol/min/pmol P450) andP450 2B6 (0.37 pmol/min/pmol P450) catalyzed significantand comparable rates of bromide formation. P450 1A2 (0.028pmol/min/pmol P450), P450 2C9 (,0.001 pmol/min/pmol

P450), P450 2D6 (0.0024 pmol/min/pmol P450), P450 2E1(0.032 pmol/min/pmol P450) and P450 3A4 (0.010 pmol/min/pmol P450) catalyzed much less bromide formation. How-ever, different results were obtained at lower, subsaturatinghalothane concentrations (0.25 ml 5 0.30 vol%). At theselower substrate concentrations (fig. 6, inset), the rates of TFAformation catalyzed by P450 2E1 (0.19 pmol/min/pmol P450)markedly exceeded those catalyzed by 2A6 (0.054 pmol/min/pmol P450). Similarly, the rates of bromide formation cata-lyzed by P450 2E1 (0.28 pmol/min/pmol P450) exceeded thosecatalyzed by 2A6 (0.14 pmol/min/pmol P450). These resultssuggested a predominant role in halothane oxidation forP450 2E1 at subsaturating halothane concentrations, andP450 2A6 at saturating halothane concentrations.

Halothane metabolism by microsomes from a panel of sev-eral human livers was examined to assess the relationshipbetween metabolite formation and P450 content. This rela-tionship was examined at both high (2 ml 5 2.4 vol%) and low(0.10 ml 5 0.12 vol%) halothane concentrations. At saturatinghalothane concentrations (2.4 vol%), among a panel of 20livers, there was a highly significant linear correlation be-tween coumarin hydroxylase (P450 2A6) activity and bothTFA (r 5 0.90; P , .001) (fig. 7A) and bromide formation (r 50.85; P , .001) (fig. 7D). Similarly, there was a significantlinear correlation between P450 2A6 protein content andboth TFA (r 5 0.71; P , .001) and bromide (r 5 0.63; P ,.003) formation (fig. 7, B and E). Conversely, at these halo-thane concentrations, the correlation between chlorzoxazonehydroxylase (P450 2E1) activity and either TFA or bromideformation was less significant (P , .11 and .05, respectively)and exhibited considerably greater scatter (fig. 7, C and F).As a further refinement of the analysis, multiple linear re-

Fig. 4. Substrate dependence of TFAand bromide formation from halothaneoxidation. Reactions were carried out asdescribed in “Methods” with varyinghalothane concentrations (0-3 ml, pro-ducing a gas phase concentration of0-3.6 vol%). Incubations were carriedout in a single liver that was competentin all the major drug metabolizing P450isoforms. Symbols denote observed me-tabolite formation. Lines represent ratespredicted using Michaelis-Menten ki-netic parameters derived from nonlinearregression analysis of the experimentaldata (TFA, r 5 0.99; Br, r 5 0.99, P ,.05). The inset shows Eadie-Hofsteeplots for TFA and bromide formation as afunction of halothane concentration(TFA, r 5 0.94; Br, r 5 0.99, P , .05).

TABLE 1Kinetic Parameters for TFA and bromide formation from P450-catalyzed halothane oxidation by HLM and cDNA-expressed P450

Enzyme

TFA Br

Vmaxa Km

(vol%)Km

(mM)Ki

(vol%)Vmax/Km

(ml/min/g) VmaxKm

(vol%)Km

(mM)Ki

(vol%)Vmax/Km

(ml/min/g)

HLMb High affinity 130 0.045 30 4.3 170 0.045 30 5.7cDNAc 2E1 0.30 0.053 35 0.36 0.52 0.030 20 1.3HLM Low affinity 94 1.2 800 0.12 130 0.94 630 0.21cDNA 2A6 0.59 1.2 800 0.87 0.77 510a Rates of halothane oxidation by human liver microsomes expressed as pmol/min/mg protein; Rates of halothane oxidation by cDNA-expressed P450 expressed

as pmol/min/pmol P450.b Human liver microsomes.c cDNA-expressed P450.


at ASPE

T Journals on A


Dow

nloaded from


gression analysis for P450 2A6 and 2E1 activity yieldedslightly improved correlation coefficients compared with lin-ear regression analysis for P450 2A6 alone. The correlationcoefficients (r) obtained using multiple linear regression were0.93 (P , .001) for TFA formation and 0.91 (P , .001) forbromide formation. There was no significant correlation be-tween either TFA or bromide formation and P450 1A2, 2C9or 2D6 activities. There was a significant correlation betweenTFA and bromide formation and P450 3A4 activity. However,this was due to coexpression of P450s 2A6 and 3A4 activities(r 5 0.59; P , .006). The relationship between metaboliteformation and P450 content was also examined at subsatu-rating halothane concentrations (fig. 8). For reactions at thislower halothane concentration (0.12 vol%), the relationshipbetween metabolite formation and P450 2A6 and 2E1 activ-ities was opposite to that observed at saturating halothaneconcentrations. Among a panel of 15 livers, there was apositive correlation between chlorzoxazone hydroxylase(P450 2E1) activity and both TFA (r 5 0.48; P , .08) andbromide (r 5 0.63; P , .02) formation (fig. 8, B and D).Conversely, there was no significant correlation between cou-marin hydroxylase (P450 2A6) activity and either TFA orbromide formation (fig. 8, A and C). There was not a signif-

icant correlation between TFA or bromide formation andP450 1A2, 2C9, 2D6 or 3A4 activities. These results furthersupported the predominant catalytic participation of P4502E1 at low, subsaturating halothane concentrations, andP450 2A6 at higher halothane concentrations.

To define further the roles of P450 2A6 and 2E1 in oxida-tive halothane metabolism, the substrate dependence of halo-thane oxidation was examined using cDNA-expressed P450s2A6 and 2E1 (figs. 9 and 10). Eadie-Hofstee plots for TFA andbromide formation by cDNA-expressed P450 2A6 were linear,and the data were fit to a one enzyme Michaelis-Mentenmodel using nonlinear regression analysis (fig. 9). Kineticparameters are summarized in table 1. For TFA formation,Vmax was 0.59 pmol/min/pmol P450 and Km was 1.2 vol%(800 mM). For bromide formation, Vmax was 0.87 pmol/min/pmol P450 and Km was 0.77 vol% (510 mM). By comparison,the oxidation of halothane by P450 2E1 was more complex(fig. 10). At very low halothane concentrations (0.012-0.25vol%), TFA and bromide formation exhibited single-enzymeMichaelis-Menten kinetics and linear Eadie-Hofstee plots.However, at higher halothane concentrations TFA and bro-mide formation decreased, consistent with substrate inhibi-tion. Therefore, experimental data for metabolite formationby cDNA-expressed P450 2E1 were fit by nonlinear regres-sion analysis to a one enzyme Michaelis-Menten model thatincorporated a term for substrate inhibition (Andersen et al.,1987). The parameters obtained are summarized in table 1.For TFA, the apparent parameters were: Vmax 5 0.30 pmol/min/pmol P450; Km 5 0.053 vol% (35 mM); Ki 5 0.36 vol%(240 mM). For Br, the apparent parameters were: Vmax 50.52 pmol/min/pmol P450; Km 5 0.030 vol% (20 mM); Ki 5 1.3vol% (870 mM). These results demonstrated that among theP450 isoforms that catalyze oxidative halothane metabolism,P450 2E1 is the high affinity/low capacity P450 isoform andP450 2A6 is the low affinity/high capacity isoform.

The relationship between rates of TFA and bromide forma-tion by microsomes from each liver within a panel of humanlivers was examined (fig. 11). The results are consistent withthe known mechanism of halothane oxidation in which TFA

Fig. 5. Effects of isoform-selective inhibitors and competitive sub-strates on TFA and bromide formation. Rates of TFA (solid) and bro-mide (cross-hatch) formation are expressed as a percentage of controlmixtures lacking the inhibitor. Reactions were carried out as describedin “Methods.” Incubations were carried out in a single liver that wascompetent in all the major drug metabolizing P450 isoforms. Finalconcentrations of each inhibitor were: 7,8-benzoflavone, 16 mM;furafylline, 20 mM; 8-methoxypsoralen, 28 mM; coumarin, 36 mM; or-phenadrine, 5 mM; sulfaphenazole, 3.6 mM; (S)-mephenytoin, 100 mM;quinidine, 45 nM; 4-methylpyrazole, 540 mM; diethyldithiocarbamate(DDC), 100 mM; troleandomycin, 100 mM; ketoconazole, 90 nm; n-octylamine, 3 mM.

Fig. 6. Rates of TFA and bromide formation from halothane oxidationcatalyzed by cDNA-expressed P450. Shown are rates of TFA (solid) andbromide (cross-hatch) formation from incubations of halothane at sat-urating concentrations (2 ml 5 2.4 vol%) and cDNA-expressed P450, asdescribed in “Methods.” The inset shows rates of TFA and bromideformation catalyzed by cDNA-expressed P450s 2A6 and 2E1 at sub-saturating halothane concentrations (0.25 ml 5 0.30 vol%).


at ASPE

T Journals on A


Dow

nloaded from


and bromide are formed in equimolar amounts. At both sat-urating (2 ml 5 2.4 vol%) and subsaturating (0.10 ml 5 0.12vol%) halothane concentrations, there was a highly signifi-cant linear correlation between TFA and bromide formationrates among the liver microsomes examined.

DiscussionThe results of this investigation demonstrate that human

liver microsomal oxidative halothane metabolism is cata-lyzed by multiple P450 isoforms. Near complete inhibition ofTFA and bromide formation by the nonselective P450 inhib-itor n-octylamine suggests that oxidative halothane metabo-lism is catalyzed exclusively by P450. Eadie-Hofstee plots ofhuman liver microsomal TFA and bromide formation wereboth nonlinear, consistent with a reaction catalyzed by two ormore P450 isoforms. TFA and bromide formation were eachdecreased by inhibitors of two different isoforms. Experi-ments using cDNA-expressed P450 proteins showed thatmultiple P450 isoforms catalyzed oxidative halothane metab-olism. Finally, in a panel of human livers, there was a cor-relation of TFA and bromide formation with one isoform at

saturating halothane concentrations and a correlation with adifferent isoform at subsaturating halothane concentrations.

Several lines of investigation were used to determine thatP450s 2A6 and 2E1 are the isoforms which catalyze oxidativehalothane metabolism in human liver microsomes. Theseincluded: 1) Effects of isoform-selective inhibitors and com-petitive substrates on rates of TFA and bromide formation, 2)rates of TFA and bromide formation by cDNA-expressedP450 isoforms and 3) correlation of TFA and bromide forma-tion with isoform activity in a panel of human liver micro-somes. Moreover, a similar experimental strategy using dif-ferent halothane concentrations suggested that P450 2E1 isthe predominant catalyst at subsaturating halothane concen-trations and 2A6 is the predominant catalyst at saturatinghalothane concentrations. Inhibitors of P450 2A6 (8-me-thoxypsoralen and coumarin) and P450 2E1 (4-methylpyra-zole and DDC) significantly decreased the rates of both TFAand bromide formation. At saturating halothane concentra-tions, high rates of TFA and bromide formation were cata-lyzed by cDNA-expressed P450 2A6. However, at subsaturat-ing halothane concentrations, the rate of metabolite

Fig. 7. Correlation of TFA and bromide formation with P450 2A6 and 2E1 activities and protein contents at saturating halothane concentrations.P450 2A6 and 2E1 activities were measured by coumarin 7-hydroxylation and chlorzoxazone 6-hydroxylation, respectively. The correlationcoefficients and statistical significance are indicated (n 5 20). For reactions at saturating halothane concentrations (2 ml 5 2.4 vol%), the upperpanels show correlation of TFA formation with: A, P450 2A6 activity; B, P450 2A6 protein content; C, P450 2E1 activity. At the same halothaneconcentrations, the lower panels show correlation of bromide formation with: D, P450 2A6 activity; E, P450 2A6 protein content; F, P450 2E1activity.


at ASPE

T Journals on A


Dow

nloaded from


formation by P450 2E1 exceeded that of 2A6. At saturatinghalothane concentrations, cDNA-expressed P450 2B6 cata-lyzed rates of metabolite formation comparable to that of2A6. However, the P450 2B6-selective inhibitor orphena-drine had no effect on halothane oxidation, and P450 2B6 isminimally expressed in human livers (Yamano et al., 1989;Mimura et al., 1993; Shimada et al., 1994). The cumulativeevidence suggests that P450 2B6 does not play a significantrole in oxidative halothane metabolism in human liver mi-crosomes. Finally, in a panel of human liver microsomes, forreactions at saturating halothane concentrations, there wasa significant linear correlation between both TFA and bro-mide formation and P450 2A6 activity and content. The cor-relation was improved by a multiple linear regression anal-

ysis using both P450 2A6 and 2E1 activities. Atsubsaturating halothane concentrations, there was a signif-icant linear correlation between both TFA and bromide for-mation and P450 2E1 activity. A previous investigation sug-gested that P450 2E1 is not the principal catalyst of oxidativemetabolism of halothane by human liver microsomes (Brownet al., 1995). The present data do not support this conclusionand clearly demonstrate that both P450s 2A6 and 2E1 arethe isoforms that catalyze oxidative halothane metabolism inhuman liver microsomes. These results suggest further thatat low halothane concentrations, P450 2E1 is the predomi-nant catalyst and at high concentrations, P450 2A6 is thepredominant catalyst of halothane oxidation.

Kinetic analysis of halothane oxidation by human liver

Fig. 8. Correlation of TFA and bromide formationwith P450 2A6 and 2E1 activities and protein con-tents at subsaturating halothane concentrations.The correlation coefficients and statistical signifi-cance are indicated (n 5 15). For reactions at sub-saturating halothane concentrations (0.10 ml 5 .12vol%), the upper panels show correlation of TFAformation with: A, P450 2A6 activity; B, P450 2E1activity. At the same halothane concentrations, thelower panels show correlation of bromide forma-tion with: C, P450 2A6 activity; D, P450 2E1 activ-ity.

Fig. 9. TFA and bromide formation from halothaneoxidation catalyzed by cDNA-expressed P450 2A6.Reactions were carried out as described in “Methods”with varying halothane concentrations (0-4 ml, 0-4.9vol%). Symbols denote observed metabolite forma-tion. Lines represent rates predicted using Michaelis-Menten kinetic parameters derived from nonlinear re-gression analysis of the experimental data (TFA, r 50.99; Br, r 5 0.98, P , .05). The inset shows Eadie-Hofstee plots for TFA and bromide formation as afunction of halothane concentration (TFA, r 5 0.99; Br,r 5 0.91, P , .05).


at ASPE

T Journals on A


Dow

nloaded from


microsomes and cDNA-expressed P450 clearly showed thatP450 2E1 is the high affinity, low capacity enzyme and P4502A6 is the low affinity, high capacity enzyme (table 1) thatcatalyzes halothane oxidation. There was excellent agree-ment between the high affinity Km values for TFA formationobtained for both microsomes and cDNA-expressed P450 2E1(0.045 and 0.053 vol%, respectively). Similarly, the low affin-ity Km values derived from both enzyme sources are in su-perb agreement (1.2 vol% for both). There was also excellentagreement between the high affinity Km values for bromideformation obtained for both microsomes and cDNA-expressedP450 2E1 (0.045 and 0.030 vol%, respectively). Similarly, thelow affinity Km values derived from both sources are in ex-cellent agreement (0.94 and 0.77 vol%, respectively). The Km

values obtained for TFA and bromide formation were compa-rable in both microsomes and expressed P450 in every case,consistent with the known mechanism of halothane oxidationwhereby TFA and bromide are formed in equimolar amounts.Most impressively, the Km value for the high affinity isoformcatalyzing microsomal halothane oxidation (0.045 vol%) is inremarkable agreement with that reported (Km 5 0.029%) forhuman halothane metabolism in vivo, in which halothanewas administered over the alveolar concentration range0.0007 to 0.13% (Cahalan et al., 1982). These investigationsexplicitly demonstrated that P450 2E1 is the high affinity,

low capacity catalyst and P450 2A6 is the low affinity, highcapacity catalyst of halothane oxidation.

Compelling evidence was obtained to suggest that humanliver microsomal and cDNA-expressed P450 2E1 are subjectto substrate inhibition at high concentrations of halothane.Furthermore, this substrate inhibition occurs at halothaneconcentrations that occur during anesthesia. Simple Michae-lis-Menten kinetics did not adequately model the kinetic datafor cDNA-expressed P450 2E1. Rather, the experimentaldata for cDNA-expressed P450 2E1-catalyzed halothane ox-idation to both TFA and bromide were best modeled by in-corporating a term for substrate inhibition. Evidence for sub-strate inhibition of P450 2E1 in human liver microsomes wasalso obtained. At halothane concentrations exceeding 4 vol%,the rates of microsomal TFA and bromide formation de-creased (data not shown). Furthermore, the two enzymeMichaelis-Menten model of the microsomal data predictedthat unlike the data for cDNA-expressed enzymes, the highaffinity isoform was the high capacity isoform. Although theapparent parameters correctly modeled the data, the abso-lute values for Vmax were inaccurate because the simplemodel did not include a term for substrate inhibition. At-tempts to model microsomal halothane oxidation including acomponent for substrate inhibition were unsuccessful be-cause there was not a unique solution to the equation. Sub-strate inhibition has also been observed previously for theP450-catalyzed metabolism of 2,2-dichloro-1,1,1-trifluoroeth-ane (HCFC-123), a halothane congener (Vinegar et al., 1994).HCFC-123 metabolism could only be accurately describedwhen the physiological based pharmacokinetic model in-cluded a term for substrate inhibition (Andersen et al., 1987).

These results provide a biochemical rationale for observa-tions in patients where halothane metabolism decreased asadministered halothane concentrations increased (Cascorbiet al., 1970; Cahalan et al., 1981; Carpenter et al., 1986).Patients receiving 0.11% halothane metabolized a greaterproportion (55%) of halothane than did those receiving 0.44%halothane (41%) (Cahalan et al., 1981). Also, in a valiantstudy by Cascorbi and co-workers (1970), the authors wereinjected with radioactive halothane, both with and withoutconcomitant halothane anesthesia. Halothane metabolismwas greater when the subjects were not anesthetized. Twoinvestigations in animals (Eckes and Buch, 1985; Lind andGandolfi, 1993) have also demonstrated that at high halo-thane concentrations, halothane metabolism was inhibited;however, as halothane concentrations decreased after cessa-tion of anesthesia, metabolism increased. The present results

Fig. 10. TFA and bromide formation from halothaneoxidation catalyzed by cDNA-expressed P450 2E1.Reactions were carried out as described in “Methods”with varying halothane concentrations (0-2 ml 5 0-2.4vol%). Symbols denote observed metabolite forma-tion. Lines represent predicted rates using Michaelis-Menten kinetic parameters derived from nonlinear re-gression analysis of the experimental data, including aterm for substrate inhibition (TFA, r 5 0.98; Br, r 50.98, P , .05).

Fig. 11. Comparison of TFA and bromide formation from P450-cata-lyzed halothane oxidation by a panel of human liver microsomes. Re-actions were performed at saturating (2 ml 5 2.4 vol%) halothaneconcentrations (n 5 20, r 5 0.99; P , .001) and (inset) at subsaturatinghalothane (0.10 ml 5 0.12 vol%) concentrations (n 5 15, r 5 0.98; P ,.001) as described in “Methods.”


at ASPE

T Journals on A


Dow

nloaded from


provide a biochemical explanation for all these observations.In vivo halothane metabolism decreases as halothane con-centrations increase because the P450 isoform that catalyzes

the majority of halothane oxidation, P450 2E1, is subject tosubstrate inhibition.

The ultimate objective of human microsomal investiga-

Fig. 12. Predicted halothane oxidationcatalyzed by P450s 2E1 and 2A6. A,Blood halothane concentrations for pa-tients undergoing a 3-hr halothane anes-thetic were measured up to 9 hr postan-esthesia and extrapolated to 4 days(inset). B, Rates of predicted TFA forma-tion from halothane oxidation catalyzedby microsomal P450s 2E1 and 2A6. Therates were calculated using blood halo-thane concentrations in (A) and kineticparameters obtained from nonlinear re-gression analysis of microsomal TFAformation. C, Rates of predicted TFA for-mation from halothane oxidation cata-lyzed by cDNA-expressed P450s 2E1and 2A6. For various relative contents ofP450s 2E1 and 2A6, the rates of halo-thane oxidation were calculated usingblood halothane concentrations in (A)and kinetic parameters obtained fromnonlinear regression analysis of TFA for-mation catalyzed by cDNA-expressedP450s 2E1 and 2A6.


at ASPE

T Journals on A


Dow

nloaded from


tions in vitro is the understanding and prediction of humandrug metabolism in vivo. The kinetic analysis of halothaneoxidation demonstrates that at clinically relevant concentra-tions, both P450 2E1 and 2A6 participate in halothane oxi-dation. Typical inspired halothane concentrations during an-esthesia are 0.4 to 3 vol%, which exceeds the apparent Km

values for both P450 2E1 and 2A6 predicted from both mi-crosomes (0.045 and 1.2 vol%, respectively) and from cDNA-expressed P450 (0.053 and 1.2 vol%, respectively). Further-more, the following analysis predicts that P450 2E1 catalyzesthe majority of halothane oxidation during and after anes-thesia, while P450 2A6 catalyzes meaningful amounts ofhalothane oxidation only at the high concentrations presentduring halothane administration. The blood halothane con-centrations for patients undergoing a 3-hr halothane anes-thetic were measured during anesthesia and for 9 hr postan-esthesia, and were extrapolated to 4 days (fig. 12A). Based onthe metabolism calculated using the apparent kinetic param-eters derived from microsomal data, P450 2E1 is predicted tocatalyze the majority of oxidative halothane metabolism (fig.12B). Additionally, the kinetic parameters obtained fromanalysis of halothane oxidation by cDNA-expressed P450also predict that P450 2E1 is the major catalyst of halothaneoxidation (fig. 12C). The metabolism calculated for variousrelative P450 2E1 and 2A6 contents indicated that P450 2E1is the predominant catalytic isoform, and that a 2E1:2A6ratio of . 1 most closely reflected the observed microsomaldata. In good agreement, the liver that was used for thekinetic experiments ranked second in P450 2E1 content andsixth in P450 2A6 content (by relative IOD) among the 20human livers examined, thus suggesting a 2E1:2A6 ratio of .1 in this liver. In further support for the predominant role ofP450 2E1 in vivo was the observation that the in vivo Km isidentical to that for P450 2E1. Finally, the in vitro clearanceestimates (Vmax/Km) for microsomal P450 2E1 (4.3-5.7 ml/min/g) were substantially greater than those for microsomalP450 2A6 (0.12-0.21). Notably, these in vitro predictions areconcordant with in vivo results. The present in vitro modelspredict a predominant role for P450 2E1 in halothane oxida-tion and in vivo results also suggested that P450 2E1 was apredominant catalyst of halothane oxidation (Kharasch et al.,1996). In vivo, disulfiram, an inhibitor of P450 2E1 (Khara-sch et al., 1993), and possibly P450 2A6, significantly de-creased urinary TFA and bromide excretion after halothaneanesthesia. Thus, in vitro studies using human liver micro-somes appear to be a useful model for understanding in vivohuman halothane oxidation.

One final point deserving mention is the relevance of ani-mal models. P450 2E1 is highly conserved among variousanimal species which suggests that animal results concern-ing P450 2E1-catalyzed metabolism may be extrapolated tohumans. In animals, numerous investigations have estab-lished the role of P450 2E1 in halothane oxidation both invitro (Loesch et al., 1987; Gruenke et al., 1988) and in vivo(Eckes and Buch, 1985; Rice et al., 1987; Lind et al., 1990). Inrats, the role of P450 2E1 in TFA-neoantigen formation(Kenna et al., 1990) has also been suggested. These animalstudies are consistent with the present results describingP450 2E1 as the major catalyst of human halothane oxida-tion in vitro, and with those showing a predominant role forP450 2E1 in human halothane oxidation in vivo (Kharasch etal., 1996). Therefore, animal models describing the role of

P450 2E1 in halothane hepatotoxicity may be relevant tohumans. Animal models have also demonstrated a role forphenobarbital-induced P450 isoforms in halothane oxidationin vitro (Gruenke et al., 1988) and in vivo (Jenner et al.,1990), suggesting the participation of non-2E1 isoforms.P450 2B enzymes are the major phenobarbital-induced iso-forms in rats. Experiments with cDNA-expressed P450 2B6,as well as previous experiments (Gruenke et al., 1988; Jenneret al., 1990), demonstrated the activity of P450 2B towardhalothane oxidation. Indeed, other P450 2E1 substrates canalso be metabolized by P450 2B at high substrate concentra-tions (Nakajima et al., 1990; Nakajima et al., 1992). However,P450 2B6 is not routinely expressed in human liver, andthus, does not appear to be a significant catalyst of humanhalothane oxidation. Therefore, animal models describingthe role of phenobarbital-induced P450s in halothane oxida-tion must be interpreted with caution.

The mild and severe forms of hepatotoxicity mediated byhalothane metabolism are believed to arise from differentroutes of biotransformation, with different clinical sequale.The present investigation, coupled with recent findings(Spracklin et al., 1996), provide biochemical evidence that themild and severe halothane hepatotoxicities indeed arise fromdifferent routes of biotransformation. Furthermore, thesedisparate pathways of metabolism are catalyzed by differentP450 isoforms. P450 2A6 and P450 3A4 catalyze reductivehalothane metabolism while P450 2E1 and 2A6 catalyzeoxidative halothane metabolism. Halothane is a unique sub-strate in that it readily undergoes both oxidative and reduc-tive metabolism. The underlying basis for this apparent ox-ygen-dependence of isoform specificity is currently unknown.

The identification of the enzymes that catalyze halothaneoxidation is an important step in understanding halothanehepatitis. These investigations have identified P450 2E1 asthe major catalyst of oxidative halothane metabolism usingan in vitro model. This model has been used to rationalize invivo results in humans, where P450 2E1 was identified as thepredominant catalyst of oxidative halothane metabolism.Thus, the microsomal in vitro model is an accurate predictorof in vivo human halothane oxidation.

Acknowledgments

The authors thank Dr. Douglas S. Mautz for the measurement ofblood halothane concentrations.

References

AKITA, S., KAWAHARA, M., TAKESHITA, T., MORIO, M. AND FUJII, K.: Halothane-induced hepatic microsomal lipid peroxidation in guinea pigs and rats.J. Appl. Toxicol. 9: 9-14, 1989.

ANDERSEN, M. E., GARGAS, M. L., CLEWELL III, H. J. AND SEVERYN, K. M.:Quantitative evaluation of the metabolic interactions between trichloroeth-ylene and 1,1-dichloroethylene in vivo using gas uptake methods. Toxicol.Appl. Pharmacol. 89: 149-157, 1987.

AWAD, J. A., HORN, J.-L., ROBERTS II, J. AND FRANKS, J. J.: Demonstration ofhalothane-induced hepatic lipid peroxidation in rats by quantification ofF2-isoprostanes. Anesthesiology 84: 910-916, 1996.

BROWN, A. P. AND GANDOLFI, A. J.: Glutathione-S-transferase is a target forcovalent modification by a halothane reactive intermediate in the guinea pigliver. Toxicology 89: 35-47, 1994.

BROWN, R. M., TATEISHI, T., GUENGERICH, F. P., WOOD, M. AND WOOD, A. J. J.:Halothane microsomal oxidation is inhibited by diethyldithiocarbamate, aninhibitor of CYP2E1. Anesth. Analg. 80: S59, 1995.

CAHALAN, M. K., JOHNSON, B. H. AND EGER II, E. I.: Relationship of concentra-tions of halothane and enflurane to their metabolism and elimination inman. Anesthesiology 54: 3-8, 1981.

CAHALAN, M. K., JOHNSON, B. H., EGER II, E. I., SHEINER, L. B., RICHARDSON, C.A., VARNER, J. K. AND SEVERINGHAUS, J. W.: A noninvasive in vivo method of


at ASPE

T Journals on A


Dow

nloaded from


assessing the kinetics of halothane metabolism in humans. Anesthesiology57: 298-302, 1982.

CARPENTER, R. L., EGER II, E. I., JOHNSON, B. H., UNADKAT, J. D. AND SHEINER, L.B.: The extent of metabolism of inhaled anesthetics in humans. Anesthesi-ology 65: 201-205, 1986.

CASCORBI, H. F., BLAKE, D. A. AND HELRICH, M.: Differences in the biotransfor-mation of halothane in man. Anesthesiology 32: 119-123, 1970.

CHRIST, D. D., KENNA, J. G., KAMMERER, W., SATOH, H. AND POHL, L. R.: Enflu-rane metabolism produces covalently bound liver adducts recognized byantibodies from patients with halothane hepatitis. Anesthesiology 69: 833-838, 1988a.

CHRIST, D. D., SATOH, H., KENNA, J. G. AND POHL, L. R.: Potential metabolic basisfor enflurane hepatitis and the apparent cross-sensitization between enflu-rane and halothane. Drug Metab. Dispos. 16: 135-140, 1988b.

COUSINS, M. J., PLUMMER, J. L. AND HALL, P. M.: Risk factors for halothanehepatitis. Aust. N.Z. J. Surg. 59: 5-14, 1989.

DE GROOT, H. AND NOLL, T.: Halothane hepatotoxicity: Relation between meta-bolic activation, hypoxia, covalent binding, lipid peroxidation and liver celldamage. Hepatology 3: 601-606, 1983.

ECKES, L. AND BUCH, H. P.: Influence of disulfiram, diethyldithiocarbamate andcarbon disulfide on the metabolic formation of trifluoroacetic acid fromhalothane in the rat. Arzneim. Forsch. 35: 1447-1451, 1985.

EGER, R. AND EGER II, E. I.: Effect of temperature and age on the solubility ofenflurane, halothane, isoflurane, and methoxyflurane in human blood.Anesth. Analg. 64: 640-642, 1985.

ESTABROOK, R. W., PETERSON, J., BARON, J. AND HILDEBRANDT, J.: The spectro-photometric measurement of turbid suspensions of cytochromes associatedwith drug metabolism. Methods Pharmacol. 2: 303-350, 1972.

GRUENKE, L. D., KONOPKA, K., KOOP, D. R. AND WASKELL, L. A.: Characterizationof halothane oxidation by hepatic microsomes and purified cytochromesP-450 using a gas chromatographic mass spectrometric assay. J. Pharmacol.Exp. Ther. 246: 454-459, 1988.

GUT, J., CHRISTEN, U., FREY, N., KOCH, V. AND STOFFLER, D.: Molecular mimicryin halothane hepatitis: Biochemical and structural characterization of lipoy-lated autoantigens. Toxicology 97: 199-224, 1995.

JEFCOATE, C. R. E., GAYLOR, J. L. AND CALABRESE, R. L.: Ligand interactions withcytochrome P-450. I. Binding of primary amines. Biochemistry 8: 3455-3463,1969.

JENNER, M. A., PLUMMER, J. L. AND COUSINS, M. J.: Influence of isoniazid,phenobarbital, phenytoin, pregnenolone 16-a-carbonitrile, and b-naph-thoflavone on halothane metabolism and hepatotoxicity. Drug Metab. Dis-pos. 18: 819-822, 1990.

KENNA, J. G., MARTIN, J. L., SATOH, H. AND POHL, L. R.: Factors affecting theexpression of trifluoroacetylated liver microsomal protein neoantigens inrats treated with halothane. Drug. Metab. Dispos. 18: 788-793, 1990.

KENNA, J. G., NEUBERGER, J. AND WILLIAMS, R.: Specific antibodies to halothane-induced liver antigens in halothane-associated hepatitis. Br. J. Anaesth. 59:1286-1290, 1987.

KHARASCH, E. D., HANKINS, D., MAUTZ, D. AND THUMMEL, K. E.: Identification ofthe enzyme responsible for oxidative halothane metabolism: implications forprevention of halothane hepatitis. Lancet 347: 1367-1371, 1996.

KHARASCH, E. D., THUMMEL, K., MHYRE, J. AND LILLIBRIDGE, J.: Single-dosedisulfiram inhibition of chlorzoxazone metabolism: A clinical probe for P4502E1. Clin. Pharmacol. Ther. 53: 643-650, 1993.

KHARASCH, E. D. AND THUMMEL, K. E.: Identification of cytochrome P450 2E1 asthe predominant enzyme catalyzing human liver microsomal defluorinationof sevoflurane, isoflurane and methoxyflurane. Anesthesiology 79: 795-807,1993.

KUNZE, K. L., WIENKERS, L. C., THUMMEL, K. E. AND TRAGER, W. F.: Warfarin-fluconazole I. Inhibition of the human P450-dependent metabolism of war-farin by fluconazole: In vitro studies. Drug Metab. Dispos. 24: 414-421, 1996.

LIND, R. C. AND GANDOLFI, A. J.: Concentration-dependent inhibition of halo-thane biotransformation in the guinea pig. Drug Metab. Dispos. 21: 386-389,1993.

LIND, R. C., GANDOLFI, A. J. AND HALL, P. D. L. M.: Isoniazid potentiation of aguinea pig model of halothane-associated hepatotoxicity. J. Appl. Toxicol. 10:161-165, 1990.

LOESCH, J., SIEGERS, C.-P. AND YOUNES, M.: Influence of cimetidine and dieth-yldithiocarbamate on the metabolism of halothane and methoxyflurane invitro. Pharmacol. Res. Commun. 19: 395-403, 1987.

LOWRY, O. H., ROSEBROUGH, N. J., FARR, A. L. AND RANDALL, R. J.: Proteinmeasurement with the Folin phenol reagent. J. Biol. Chem. 193: 265-275,1951.

MAENPAA, J., JUVONEN, R., RAUNIO, H., RAUTIO, A. AND PELKONEN, O.: Metabolicinteractions of methoxsalen and coumarin in humans and mice. Biochem.Pharmacol. 48: 1363-1369, 1994.

MILES, J. S., MCLAREN, A. W., FORRESTER, L. M., GLANCEY, M. J., LANG, M. A. AND

WOLF, C. R.: Identification of the human liver cytochrome P-450 responsiblefor coumarin 7-hydroxylase activity. Biochem. J. 267: 365-371, 1990.

MIMURA, M., BABA, T., YAMAZAKI, H., OHMORI, S., INUI, Y., GONZALEZ, F. J.,GUENGERICH, F. P. AND SHIMADA, T.: Characterization of cytochrome P-4502B6 in human liver microsomes. Drug Metab. Dispos. 21: 1048-1056, 1993.

MULLER, R. AND SRIER, A.: Modification of liver microsomal lipids by halothanemetabolites; a multi nuclear NMR spectroscopic study. Naunyn-Schmiede-berg Arch. Pharmacol. 321: 234-237, 1982.

NAKAJIMA, T., WANG, R.-S., ELOVAARA, E., PARK, S. S., GELBOIN, H. V. AND VAINIO,H.: A comparative study on the contribution of cytochrome P450 isozymes tometabolism of benzene, toluene and trichloroethylene in rat liver. Biochem.Pharmacol. 43: 251-257, 1992.

NAKAJIMA, T., WANG, R.-S., MURAYAMA, N. AND SATO, A.: Three forms of trichlo-roethylene-metabolizing enzymes in rat liver induced by ethanol, phenobar-bital, and 3-methylcholanthrene. Toxicol. Appl. Pharmacol. 102: 546-552,1990.

POHL, L. R., KENNA, J. G., SATOH, H., CHRIST, D. D. AND MARTIN, J. L.: Neoan-tigens associated with halothane-hepatitis. Drug. Metab. Rev. 20: 203-217,1989.

RAY, D. C. AND DRUMMOND, G. B.: Halothane hepatitis. Br. J. Anaesth. 67:84-99, 1991.

RICE, S. A., MAZE, M. R., SMITH, C. M., KOSEK, J. C. AND MAZZE, R. I.: Halothanehepatotoxicity in Fischer 344 rats pretreated with isoniazid. Toxicol. Appl.Pharmacol. 87: 411-419, 1987.

SATO, N., FUJII, K., YUGE, O. AND MORIO, M.: The association of halothane-induced lipid peroxidation with the anaerobic metabolism of halothane: Anin vitro study in guinea pig liver microsomes. Hiroshima J. Med. Sci. 39: 1-6,1990.

SHIMADA, T., YAMAZAKI, H., MIMURA, M., INUI, Y. AND GUENGERICH, F. P.: Inter-individual variations in human liver cytochrome P-450 enzymes involved inthe oxidation of drugs, carcinogens and toxic chemicals: Studies with livermicrosomes of 30 Japanese and 30 Caucasians. J. Pharmacol. Exp. Ther.270: 414-423, 1994.

SPRACKLIN, D. K., THUMMEL, K. E. AND KHARASCH, E. D.: Human ReductiveHalothane Metabolism in vitro is Catalyzed by P450 2A6 and 3A4. DrugMetab. Dispos. 24: 976-983, 1996.

THUMMEL, K. E., KHARASCH, E. D., PODOLL, T. AND KUNZE, K.: Human livermicrosomal enflurane defluorination catalyzed by cytochrome P-450 2E1.Drug Metab. Dispos. 21: 350-357, 1993.

THUMMEL, K. E., O’SHEA, D., PAINE, M. F., SHEN, D. D., KUNZE, K. L., PERKINS, J.D. AND WILKINSON, G. R.: Oral first-pass elimination of midazolam involvesboth gastrointestinal and hepatic CYP3A-mediated metabolism. Clin. Phar-macol. Ther. 59: 491-502, 1996.

VINEGAR, A., WILLIAMS, R. J., FISHER, J. W. AND MCDOUGAL, J. N.: Dose-dependent metabolism of 2,2-dichloro-1,1,1,-trifluoroethane: A physiologi-cally based pharmacokinetic model in the male Fischer 344 rat. Toxicol.Appl. Pharmacol. 129: 103-113, 1994.

YAMANO, S., NHAMBURO, P. T., AOYAMA, T., MEYER, U., INABA, T., KALOW, W.,GELBOIN, H. V., MCBRIDE, O. W. AND GONZALEZ, F. J.: cDNA cloning andsequence and cDNA-directed expression of human P450 IIB1: Identificationof a normal and two variant cDNAs derived from the CYP2B locus onchromosome 19 and differential expression of the IIB mRNAs in humanliver. Biochemistry 28: 7340-7348, 1989.

Send reprint requests to: Dr. Evan Kharasch, University of Washington,Department of Anesthesiology, Box 356540, Seattle, WA 98195.


at ASPE

T Journals on A


Dow

nloaded from


cytochrome p450 2e1 is the principal catalyst of human oxidative

Documents