Hydroxylation of warfarin by human cDNA-expressed cytochrome P

Poongavanam Vasanthanathan , Jozef Hritz , Olivier Taboureau , Lars Olsen , Flemming Steen Jørgensen , Nico P. E. Vermeulen and Chris Oostenbrink...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1992,5, 54-59

54

Hydroxylation of Warfarin by Human cDNA-Expressed Cytochrome P-450: A Role for P-4502C9 in the Etiology of (S)-Warfarin-Drug Interactions Allan E. Rettie,t Kenneth R. Korzekwa,t Kent L. Kunze,? Ross F. Lawrence,t A. Craig Eddy,§ Toshifumi Aoyama,*J Harry V. Gelboin,t Frank J. Gonzalez,f and William F. Trager*p+ Department of Medicinal Chemistry, School of Pharmacy, and Department of Surgery, School of Medicine, University of Washington, Seattle, Washington 98195, and Laboratory of Molecular Carcinogenesis, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892 Received August 9, 1991

Previous kinetic studies have identified a high-affinity (SI-warfarin 7-hydroxylase present in human liver microsomes which appears to be responsible for the termination of warfarin's biological activity. Inhibition of the formation of (S)-7-hydroxywarfarin, the inactive, major metabolite of racemic warfarin in humans, is known to be the cause of several of the drug interactions experienced clinically upon coadministration of warfarin with other therapeutic agents. In order to identify the specific form(s) of human liver cytochrome P-450 involved in this particular toxicity, we have determined the metabolic profiles of 11human cytochrome P-450 forms expressed in HepG2 cells toward both (R)-and (Wwarfarin. Of the 11forms examined only 2C9 displayed the regioselectivity and stereoselectivity appropriate for the high-affinity human liver microsomal (S)-7-hydroxylase. We further compared Michaelis-Menten and sulfaphenazole inhibition constants for (S)-warfarin 7-hydroxylation catalyzed by cDNA-expressed 2C9 and by human liver microsomes. Similar kinetic constants were obtained for each enzyme source. It is concluded that 2C9 is likely to be a principal form of human liver P-450 which modulates the in vivo anticoagulant activity of the drug. It is further concluded that those drug interactions with warfarin that arise as a result of decreased clearance of the biologically more potent S-enantiomer may have as their common basis the inhibition of P-450 2C9. 5). A number of human liver forms have also been purified

Introduction

and their metabolic profiles toward (R)- and @)-warfarin described (6). However, as discussed recently by Abernethy et al. (7),these forms cannot be reliably identified within the standardized P-450 nomenclature. The expression of cloned P-450's in systems where reliable reaction kinetics for a given pathway can be obtained may provide a means whereby specific forms involved in a particular oxidative pathway can be identified (8). In this paper we have examined the warfarin metabolite profiles generated by 11human cDNA-expressed P-450's, and we report the identification of a specific human liver form, P-450 2C9, which is responsible for the stereo- and regioselective hydroxylation of warfarin to its principal in vivo metabolite, (S)-7-hydroxywarfarinn.We further demonstrate that P-450 2C9 is not only the most likely but may in fact be the only form of P-450 to govern this oxidative pathway in vivo.

An important mechanism responsible for the occurrence

of clinically significant drug interactions is inhibition of the clearance of the primary drug by suppression of its biotransformation to inactive metabolites. This can often lead to serious adverse effects if the affected drug accumulates to toxic concentrations. Clearly, the most clinically relevant interactions of this type will involve drugs with a narrow therapeutic index (I), for example, the coumarin anticoagulants. We have previously documented that the interaction, in vivo, between the oral anticoagulant warfarin (Figure 1) and the uricosuric agent sulfinpyrazone ( 2 ) has, as its mechanistic basis, inhibition of the formation of the biologically inactive 7-hydroxywarfarin metabolite from biologically active (S)-warfarin. In vitro studies, performed with human liver microsomes, suggest that the formation of (S)-7-hydroxywarfarin is catalyzed by a restricted, high-affinity subset of human liver P-450's (3). The identification of the specific human form(s) of P-450 responsible for the in vivo formation of (S)-7-hydroxywarfarin, coupled to knowledge of ita substrate specificity, should provide predictive power for this drug interaction since potent inhibitors of this form(s) of P-450 might be expected to lead to a drug interaction, if administered concomittantly with warfarin. Extensive investigations have been carried out to elucidate the identity of specific rat and rabbit liver P-450 isozymes which are involved in warfarin metabolism (4,

Experimental Procedures Materials. Optically pure warfarin enantiomersand phenolic metabolites and their deuterated analogues were obtained as described previously (3). All other materials were obtained from Sigma Chemical Co. (St.Louis, MO), Merck & Co. (Rahway, NJ), or Boehringer-Mannheim Corp. (New York, NY). Warfarin Metabolism by Human P-450Forms Expressed in HepG2 Cells. HepG2 cells were infected with a recombinant vaccinia virus containing human c D N h for one of the following; P-450lA2,2A6,' 2B7,2C8,2C9? 2D6,2E1,2Fl, 3A4,3A5, or 4B1. P-4502A6 is the new designation for the coumarin 7-hydroxylase isozyme previously referred to as 2A3 (12). P-4502C9, used in this study, corresponds to the form termed Human-2 originally described by Dr. Kato's group (26). This form represents the major allelic form of 2C9 in Caucasiansand Japanese. A minor allelic form of 2C9, used in earlier studies (13),corresponds to the former 2C1

of Pharmacy, University of Washington. National Cancer Institute. School of Medicine, University of Washington. 11 Present address: Department of Biochemistry, Shinshu University School of Medicine, Asahi, Mataumoto, Japan. t School

*

0893-228x192 , ,12705-0054m3.00IO I

(29).

0 1992 American Chemical Societv -

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 55

Human P-450's: Warfarin-Drug Interactions

A (S)

Figure 1. Structures of (R)-and (S)-warfarin. Arrows denote the principal sites of oxidative attack by the human cDNA-expressed P-450 forms used in this study. Information detailing the construction of these recombinant viruses has been published elsewhere (9-12). Control HepG2 cells were infected with wild-type vaccinia virus. HepG2 cell lysate fractionation yielded low recoveries of total P-450 in the 1OOO0lOOOOOg ("microsomal") pellet. Therefore, whole-cell lysates, prepared as previously described (13),were routinely employed in this study. Incubations contained 3.5-12.5 mg of cell lysate protein, 0.1 M potassium phosphate buffer, pH 7.4, 1 pmol of NADPH, and 650 nmol of either (S)- or @)-warfarin in a final volume of 1.0 mL. Reactions were initiated by the addition of 1pmol of NADPH and terminated after 20 min by the addition of 0.6 mL of acetone. Phenolic metabolites of warfarin were analyzed by GC-MS3 after the addition of deuterium-labeled internal standards as described previously (14). In the case of the 3A4 and 3A5 isozymes, GC-MS analysis was also performed to determine the rates of formation of (S)- and (R)-10-hydroxywarfarin, according to the method of Lawrence et al. (15). Kinetics of Formation of (S)-7-Hydroxywarfarin and (S)-6-Hydroxywarfarin. Rates of formation of (S)-7hydroxywarfarin and (S)-&hydroxywarfarinwere determined over the concentration range 1-20 pM (S)-warfarin, as described above, using either P-450 2C9 cell lysate or human liver microsomes from donors 109, 110, and 111 as the source of the enzyme. cDNAexpressed 2C9 incubations contained 158 pmol of P-450 (6.7 mg of cell lysate protein), and human liver microsomal incubations contained 1nmol of total P-450 (1.6-2.2 mg of microsomal protein). Kinetic parameters were obtained by nonlinear regression analysis of the rate data (16)) using the Michaelis-Menten equation.

Sulfaphenazole Inhibition Kinetics in Human Liver Microsomes. Details concerning acquisition, handling, storage, and preparation of microsomes from human liver have been published (3). The kinetics of sulfaphenazole's inhibitory effect on the formation of (S)-7-hydroxywarfarin were assessed using human liver microsomes from donor 110. Incubations contained 1.6 mg of microsomal protein, 1pmol of NADPH, 0.1 M potassium phosphate buffer, pH 7.4, (@warfarin (2-18 pM), and sulfaphenazole (0-900 nM) added in an aqueous solution. Reactions were carried out and analyzed as above. The rate data were analyzed graphically to obtain approximate kinetic constanb and to determine the nature of the inhibition. Rate data were further analyzed by nonlinear regression using the standard equation for competitive inhibition kinetics. Effect of Sulfaphenazole and Tolbutamide on Human Liver Microsomal Metabolism of (R)-Warfarin and (S)Warfarin. Sulfaphenazole and tolbutamide, in 50 pL of ethanol, were added to glass scintillation vials to give final concentrations (in 1.0 mL) of 0.5 and 300 pM, respectively. Control incubations received 50 pL of ethanol alone. Vials were heated at 37 "C for 45 min to remove the alcohol. Liver microsomal protein, 1.6 mg, from donor 111 was added to each vial, together with 0.1 M potassium phosphate buffer, pH 7.4, and either 4 pM @)-warfarin or 400 r M (R)-warfarin. The vials were preincubated by shaking vigorously at 37 O C for 15 min, prior to initiation of the reaction with NADPH. Analysis of phenolic metabolite profiles was performed as described above. Effect of Sulfaphenazole on the Metabolism of ( S ) Warfarin Catalyzed by 2C9. Rates of formation of (S)-7hydroxywarfarin and (S)-6-hydroxywarfarin were determined as described above, in incubations which contained 4.6 mg of 2C9 cell lysate protein, a final concentration of 4 pM (S)-warfarin,and Abbreviation: GC-MS,gas chromatography-mass spectrometry.

E Lll

3A5 481

0

100

200

300

Formation rate (pmollmg cell lysatdhr)

Figure 2. Rates of formation of (S)-7-hydroxywarfarincatalyzed by human cDNA-expressed P-450. Eleven human cDNA-expressed forms were individually incubated with 650 pM (S)warfarin and the formation rates of (S)-7-hydroxywarfarin determined as described under Experimental Procedures. 0-200 nM sulfaphenazole. The same experiment was performed with human liver microsomes from donor 110, with the exceptions that incubations contained 1.6 mg of microsomal protein and 0-600 nM sulfaphenazole. Estimates of the apparent Ki for sulfaphenazole were made from Dixon plots. Other Assays. Protein concentrations were determined by the Lowry method (17). Cytochrome P-450 concentrations were determined from the reduced versus oxidized carbon monoxide difference spectra, as described by Estabrook et al. (18).

Results Formation of Warfarin Metabolites by cDNA-Expressed Human Liver P-450's. Individual P-450 cDNA's were expressed using vaccinia virus and HepG2 cells. Levels of holo-P-450 expression ranged from 10 to 20 pmol/mg of cell lysate. Rates of formation of (29-7hydroxywarfarin at substrate concentrations of 650 pM by P-450's lA2,2A6,2B7,2C8,2C9,2D6,2El, 2F1,3A4,3A5, and 4B1 are shown (Figure 2). Only 1A2,2C9, and 3A4 formed amounts of this metabolite in excess of the background rates obtained with control HepG2 cells. Metabolite profiles (at 650 NM substrate) obtained from P-450 1A2, 2C9, and 3A4 are shown in Figure 3. Form 1A2 was principally a @)-warfarin hydroxylase with a pronounced regiochemical preference for the 6-position. Form 3A4 was predominantly an (R)-10-hydroxylase but with significant activity as an (S)-4'-hydroxylase and an (S)-10-hydroxylase. Form 2C9 stereospecifically hydroxylated @)-warfarin, forming principally (S)-7hydroxywarfarin, about one-third as much (S)-6hydroxywarfarin, and a minor amount of (S)-4'-hydroxywarfarin. No quantifiable levels of a n y warfarin phenols were generated by either P-450 2A6, 2D6, 2E1, or 4B1. Form 2C8 exhibited low activity as an (S)-4'-hydroxylase [18 pmol/(mgh)]. Forms 2F1 and 2B7 were principally (R)-4'-hydroxylases,although 2F1 catalyzed this reaction much more rapidly than did 2B7 [61 and 17 pmol/(mg.h), respectively]. Only (R)-10-hydroxywarfarin was formed by 3A5, but at a much lower rate than that obtained using 3A4. Kinetic Parameters for the Formation of (5)-7- and (5)-6-Hydroxywarfarin by P-450 2C9 and by Human Liver Microsomes. Kinetic parameters for the formation of (S)-7-hydroxywarfarin and ($)-6-hydroxywarfarin catalyzed by 2C9 and by three separate human liver microsomal preparations are shown (Table I). V,,, values,

Rettie et al.

56 Chem. Res. Toxicol., Vol. 5, No. 1, 1992 R-Warfarin VA21

1,4

5-Warfarln [1M1

S-8-hydroxy R-8-hydroxy

1.2

S-7-hydrOxy R-7-hydroxy S-B-hydroxy 1.0

R-&hydroxy S4-hydroxy R4-hydroxy 1

2500

1500

500

500

1500

2500

Rate of metaboilte formatlon (pmollmglhr)

-.-

. E

-E

0.8

0

S-8-hydroxy

. E -.

R-6-hydroxy

?

-E0

0.6

r

r

S-7-hydroxy

0.4

R-7-hydroxy S-6-hydroxy R-&hydroxy

0.2

I

S-4'-hydroxy R-4'-hydroxy I

'

1

.

1

'

I

.

I

.

I

0.0 .SO0

R-Warfarln PA41

100

100

300

500

Rate of metabollte formation (pmollmglhr)

Figure 3. Warfarin metabolite profides generated using human cDNA-exprd P-450's 1A2,2C9,and 3A4. Enzyme incubations and metabolite analysis were carried out in duplicate as described under Experimental Procedures. Table I. Comparison of the Michaelis-Menten Kinetic Parameters Obtained for (S)-Warfarin Metabolism Catalyzed by Human cDNA-Expressed P-4502C9 and by Human Liver Microsomesa 7-hydroxylation 6-hydroxylation enzyme source ~

P-450 2C9

HLlO9M HLllOM HLlllM

Vmam

Vmam

pmol/(nmol. K,, rcM min) 4.1 (0.6)b 421 (21) 3.9 (0.3) 5.6 (0.2) 3.8 (0.1) 6.4 (0.1) 3.9 (0.1) 14 (0.2)

K,, JIM

4.0 (0.3) 3.6 (0.2) 3.7 (0.1) 4.3 (0.2)

100

300

500

700

BOO

Figure 4. Dixon plot for the inhibition, by sulfaphenazole, of @)-warfarin ?-hydroxylase activity catalyzed by human liver microsomes. Metabolite analysis was carried out in duplicate as described under Experimental Procedures. Human liver microsomes from donor 110 were used as the enzyme source.

R-8-hydroxy S-7-hydroxy R-7.hydroxy S-8-hydroxy R-&hydroxy S4-hydroxy R4-hydroxy 300

-100

Sulfaphenazole concentration (nM)

5-Warfarin PA41

S-1O-hydroxy R-1O-hydroxy S-8-hydroxy

500

I

-300

pmol/(nmol. min) 122 (3.0) 1.3 (0.0) 1.6 (0.0) 3.9 (0.1)

Incubations and kinetic analyses were performed as described under Experimental Procedures. *Values in parentheses are the SEs of the mean.

expressed on the basis of total P-450 content, were at least 30 times higher for the 2C9-catalyzed reactions than for the same reactions catalyzed by human liver microsomes. The ratio of V,, for (S)-7-hydroxylation to that of (SI6-hydroxylationwas 3.5 for 2C9 and varied from 3.6 to 4.3 in the microsomal preparations. K , values of 4 p M were obtained for both metabolic routes and with both 2C9 and

each human liver microsomal preparation. Inhibition of Human Liver Microsomal (S)-Warfarin 7-Hydroxylase by Sulfaphenazole and Tolbutamide. Sulfaphenazole, a potent and relatively selective inhibitor of human liver tolbutamide hydroxylase (19-21), was found to be a potent, competitive inhibitor of the formation of (S)-7-hydroxywarfarin from human liver microsomes, exhibiting a Ki of 0.206 f 0.004 pM for this process (Figure 4). The reaction in the absence of inhibitor proceeded with a K , of 3.92 f 0.09 pM. Both sulfaphenazole and tolbutamide were stereoselective inhibitors of (8-warfarin metabolism in human liver microsomes (Figure 5). In order to provide a basis for comparison of inhibitor efficiency, this experiment was performed with inhibitor concentrations equal to 2.5 times the Ki for sulfaphenazole (Le., 0.5 pM)or 2.5 times the K, for tolbutamide (i.e., 300 p M ) (21). As shown (Figure 5 ) , both inhibitors decreased the formation of (S)-6- and (S)-7-hydroxywarfarinto 40% of control values under the conditions of the experiment, but neither inhibitor decreased the rates of formation of any of the (R)-warfarin metabolites by more than 7 7%. Inhibition of 2C9-Catalyzed Metabolism of ( S ) Warfarin by Sulfaphenazole. A Dixon plot for the inhibition of the 2C9-catalyzed formation of (S)-6hydroxywarfarin and (S)-7-hydroxywarfarin by sulfaphenazole is shown (Figure 6). Under the conditions of the experiment, where the substrate concentration used (4 pM)was equal to the K,, the intercepts on the x-axis are equivalent to -Xi.Thus, Ki equals 0.17 p M and 0.18 pM for the formation of (S)-6-hydroxywarfarinand (S)7-hydroxywtufarin, respectively. This is in good agreement with the value of 0.21 p M obtained from a full kinetic analysis of (S)-7-hydroxywarfarin formation catalyzed by

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 57

Human P-450’s: Warfarin-Drug Interactions

S-6

S-7

R-4‘

R-6

R-7

R-8

.

#

Warfarin Phenol

Figure 5. Stereoselective inhibition of warfarin metabolism by sulfaphenazole and tolbutamide. Metabolite analysis was carried out in duplicate as described under Experimental Procedures. Human liver microsomes from donor 111were used as the enzyme source.

human liver microsomes (Figure 4). A full kinetic analysis was not performed with 2C9 due to limitations on the availability of the cloned material. In order to provide validation for the values derived from limited kinetic analysis of the cloned enzyme, the same experimental protocol used to derive estimates for the Ki from human liver microsomes was used. When this experiment was performed at a single substrate concentration (4 pM) equal to the K,, values of 0.13-0.17 pM were obtained for Ki (data not shown).

Discussion The primary focus of this report concerns the identification of the specific forms of human liver P-450 responsible for terminating the anticoagulant activity of racemic warfarin. It is well established that most of the biological activity elicited by the drug resides in the S-enantiomer and that all metabolites are essentially inactive. It is also known that greater than 80% of the S-enantiomer in a racemic dosage form is converted to two metabolites, (S)-7-hydroxywarfarinand (S)-6-hydroxywarfarin,which are formed in about a 3:l ratio (2). Thus, whichever form(s) of P-450 governs (govern) the biotransformation of (S)-warfarinto these two metabolites w i l l be the form(s) primarily responsible for termination of the anticoagulant effect. Analysis of warfarin metabolite profiles generated by 11forms of human P-450 indicated that a protein encoded by the CYP2C9 gene was the most active in the formation of both (S)-7-hydroxywarfarin and (S)-6hydroxywarfarin (Figures 2 and 3). Two additional forms of human liver P-450,1A2 and 3A4, also were active in the formation of these metabolites, but to a minor extent and at a substrate concentration at least 2 orders of magnitude higher than the peak plasma level attained by a therapeutic dose of the drug. These forms did, however, show significant activity toward the formation of other oxidative metabolites of the drug; (R)-6-hydroxywarfarinin the case of 1A2, and (R)-10-hydroxywarfarin and (S)-4’-hydroxywarfarin in the case of 3A4. However, from consideration of the in vivo metabolite profiles discussed above, the

1

-300

-200

-100

0

100

200

[I1 nM

Figure 6. Dixon plot for the inhibition, by sulfaphenazole, of (S)-7-hydroxywarfarin and (S)-6-hydroxywarfarin formation catalyzed by human cDNA-expressed P-4502C9. Incubations and metabolite analysis was carried out as described under Experimental Procedures.

metabolic activity of 1A2 and 3A4 is urilikely to contribute significantlyto the termination of the anticoagulant effect of the drug. Since K , and the ratio of Vmaxvalues for the formation of (S)-7-hydroxywarfarin and (S)-6-hydroxywarfarin are independent of enzyme concentration, they are the critical parameters in terms of being able to draw a correlation between the kinetic data obtained from the cDNA-expressed enzyme preparations relative to microsomal preparations. In this regard, a K , value of-4 pM was obtained for both the (S)-7-hydroxylase and (S)-6hydroxylase activities of 2C9 and three separate human liver microsomal preparations (Table I). In addition, despite the wide variation in individual V- values between the various enzyme preparations, the ratio of Vmaxfor (S)-7-hydroxywarfarin formation to Vmax for (S)-6hydroxywarfarinformation was relatively constant, ranging from 3.5 for 2C9 to 4.3 for the human liver 109M microsomes (Table I). These data are consistent with P-450 2C9 representing the principal form in human liver microsomes responsible for the formation of these two metabolites at substrate concentrations less than 10 pM. The formation of (S)-7-hydroxywarfarin and (S)-6hydroxywarfarin by human liver microsomes is strongly inhibited (Ki = 6-8 pM) by (R)-warfarin in a competitive manner (22). However, no (&warfarin hydroxylase with a correspondingly low Kmhas been identified in human liver microsomes, which suggests that binding of (R)warfarin is not catalytically productive. The finding that (R)-warfarin is not a substrate for 2C9 is consistent with this interpretation. Sulfaphenazole was found to be a potent, selective inhibitor of @)-warfarin metabolism catalyzed by human liver microsomes (Figures 4 and 5). Therefore, we also determined the Ki’s for sulfaphenazole’sinhibition of the

58 Chem. Res. Toxicol., Vol.5, No.1, 1992

formation of (S)-7-hydroxywarfarin by human liver microsomes and by 2C9. Inhibition kinetics in microsomes were competitive and displayed a Ki of 0.21 pM. Assuming competitive inhibition at a true K , of 4 pM for 2C9, a Ki of 0.174.18 pM was obtained for sulfaphenazole's inhibition of 2C9-mediated formation of (S)-7-hydroxywarfarin (Figure 6). Therefore, in addition to the congruent metabolite profiles and Michaelis-Menten kinetic data obtained for the microsomal (S)-7-hydroxylase and P-450 2C9, inhibition constants for sulfaphenazole's modulation of (S)-7-hydroxylase activity coincide for both enzyme sources. A variety of parallels are now clearly evident for (8warfarin and tolbutamide metabolism in humans. Sulfaphenazole is a highly potent, competitive inhibitor of both human liver (SI-warfarin hydroxylation (this study) and tolbutamide hydroxylation in vitro (19-21). Moreover, both sulfinpyrazone and phenylbutazone, inhibitors of (S)-warfarin metabolism in vivo, also inhibit tolbutamide metabolism in vivo (23,241. To explore this relationship further, the ability of both sulfaphenazole and tolbutamide to inhibit the human liver microsomal metabolism of (R)and @)-warfarin (Figure 5) was determined. Under the conditions of the experiment, where a concentration equivalent to 2.5 times the Ki (or K,) of the inhibitor (see Experimental Procedures) was used, both sulfaphenazole and tolbutamide decreased formation of (S)-7-hydroxywarfarin and (S)-6-hydroxywarfarinto about 40% of their control values, while all other metabolic pathways remained essentially unaffected. These data, in light of the findings of this report, necessitate that human tolbutamide hydroxylase and (@-warfarin 7-hydroxylase activity be closely related. P-450 2C8 and the 2C9 protein used in this study each catalyze the hydroxylation of tolbutamide in vitro (13,s). 2C8 is not an @)-warfarin 7-hydroxylase and need not be considered further for a possible in vivo role in warfarin metabolism. The sequence of the 2C9 protein used in this study is identical to the sequence originally reported for Human-2 (26) and more recently (27) to the protein sequence obtained from clone 65 of Romkes et al. However, in addition to this molecular species, several other very similar proteins, differing by only one or two amino acids from one another, have been described. These forms have been variously termed 2C1 (29), 2C9 ( 2 7 , 2 8 ) ,and 2C10 (20). While each of these forms possesses tolbutamide no data have, as yet, been hydroxylme activity (13,20, a), presented concerning their ability to metabolize warfarin. The 2C family in humans is clearly very complex, and although we cannot exclude the possibility that one of these variants (or other P-450 forms not tested in this study) might catalyze the formation of (S)-7-hydroxywarfarin, we believe that the congruity evident in this study for both the metabolite profiles and the kinetic constants obtained for the low K , microsomal (S)-7hydroxylase and P-450 2C9 is persuasive evidence for their identity. In addition to providing information on the likely role of different P-450 forms in the pathogenesis of (5')-warfarin-drug interactions, the metabolite profiles presented for the cDNA-expressed enzymes also provide insight into the isozyme selectivity involved in the inhibition of (R)warfarin metabolism reported for enoxacin (30) and diltiazem (7). Both drugs selectively inhibited the formation of (R)-6-hydroxywarfarinin humans. This is a very minor pathway for the termination of pharmacological activity of the racemate in vivo, and microsomal studies suggest that it is catalyzed by a high K , isozyme (3), findings that

Rettie et al.

are consistent with the limited clinical consequences of these two interactions. The metabolite profiles presented in this study highlight a possible role for P-450 1A2 in the inhibition of (R)-warfarin metabolism caused by concurrent administration of warfarin and either enoxacin or diltiazem. Studies are currently underway to test the hypothesis that 1A2 represents the high K , (R)-6hydroxylase. In summary, it can be concluded that three forms of P-450, representing different gene families, are dominant in the oxidative metabolism of warfarin. The three forms, 1A2,2C9, and 3A4, are highly stereoselective and regioselective in the specific reactions they catalyze. Form 1A2 is highly stereoselectivefor (R)-warfarinand regioselective for the 6-position. Form 3A4 is highly regio- and stereoselective for the formation of (R)-10-hydroxywarfarin, a finding that is in agreement with a recent report concerning the metabolic profile of warfarin catalyzed by a closely related P-450 3A4 species expressed in yeast (31). Conversely, 2C9 is highly stereoselective for @)-warfarin and generates the (S)-7-hydroxy and (S)-6-hydroxy metabolites in a ratio of approximately 3.5 to 1. Since the preponderance of biological activity resides in (S)-warfarin, and since 2C9 displays the regioselectivity, stereoselectivity, and kinetic properties appropriate for the human liver microsomal (S)-7-hydroxylase, it may be concluded that 2C9 is a principal form of human liver P-450 that modulates the in vivo anticoagulant activity of the drug. Furthermore, it may be concluded that those drug interactions with warfarin that arise as a result of decreased clearance of the biologically more potent S-enantiomer have as their common basis the inhibition of P-450 2C9.

Acknowledgment. This work was supported in part by National Institutes of Health Program Project Grant GM 32165. We thank Dr. John Miners, Flinders Medical Centre, Adelaide, Australia, for his kind gift of sulfaphenazole.

References (1) Barry, M., and Feely, J. (1990) Enzyme induction and inhibition. Pharmucol. Ther. 48, 71-94. (2) Toon, S., Low, L. K., Gibaldi, M., Trager, W. F., OReilly, R. A,, Motley, C. H., and Goulart, K. A. (1986) The warfarin-sulfinpyrazone interaction: Stereochemical considerations. Clin. Pharmacol. Ther. 39, 15-24. (3) Rettie, A. E., Eddy, A. C., Heimark, L. D., Gibaldi, M., and Trager, W. F. (1989) Characteristics of warfarin hydroxylation catalyzed by human liver microsomes. Drug Metab. Dispos. 17, 265-270. (4) Fasco, M. J., Vatsis, K. P., Kaminsky, L. S., and Coon, M. J. (1979) Regioselective and stereoselective hydroxylation of R and S warfarin by different forms of purified cytochrome P-450 from rabbit liver. J. Biol. Chem. 253,7813-7820. (5) Guengerich, F. P., Dannan, G. A., Wright, S. T., Martin, M. V., and Kaminsky, L. S. (1982) Purification and characterization of liver microsomal cytochromes P-450: electrophoretic, spectral, catalytic and immunochemical properties and inducibility of eight isozymes isolated from rata treated with phenobarbital of p-naphthoflavone. Biochemistry 21, 6019-6030. (6) Wang, P. P., Beaune, P., Kaminsky, L. S.,Dannan, G. A., Kadlubar, F. F., Larrey, D., and Guengerich, F. P. (1983) Purification and characterization of six cytochrome P-450 isozymes from human liver microsomes. Biochemistry 22, 5375-5382. (7) Abernethy, D. R., Kaminsky, L. S.,and Dickinson, T. H. (1991) Selective inhibition of warfarin metabolism by diltiazem in humans. J. Pharmacol. Exp. Ther. 257, 411-415. (8) Boobis, A. R., Sesardic, D., and Gooderham, N. J. (1990) Methods in drug metabolism. In ComprehensiveMedicinal Chemistry (Hansch, C., Ed.) Vol. 5, pp 443-479, Pergamon Press, Oxford, U.K. (9) Aoyama, T., Gonzalez, F. J., and Gelboin, H. V. (1989) Human cDNA-expressed cytochrome P450 IA2: Mutagen activation and substrate specificity. Mol. Carcinog. 2, 192-198.

Human P-450’s: Warfarin-Drug Interactions (10) Crespi, C. L., Steimel, D. T., Aoyama, T., Gelboin, H. V., and Gonzalez, F. J. (1990) Stable expression of human cytochrome P450IA2 cDNA in a human lymphoblastoid cell line: role of the enzyme in the metabolic activation of aflatoxin B1. Mol. Carcinog. 3, 5-8. (11) Nhamburo, P. T., Kimura, S., McBride, 0. W., Kozak, C. A., Gelboin, H. V., and Gonzalez, F. J. (1990) The human CYP2F gene subfamily: Identification of a cDNA encoding a new cytochrome P450, cDNA-directed expression, and chromosome mapping. Biochemistry 29, 5491-5499. (12) Yamano, S., Tatsuno, J., and Gonzalez, F. J. (1990) The CYP2A3 gene product catalyzes coumarin 7-hydroxylation in human liver microsomes. Biochemistry 29,1322-1329. (13) Relling, M. V., Aoyama, T., Gonzalez, F. J., and Meyer, U. A. (1990) Tolbutamide and mephenytoin hydroxylation by human cytochrome P45Os in the CYP2C subfamily. J. Pharmacol. Exp. Ther. 252,442-447. (14) Bush, E. D., Low, L. K., and Trager, W. F. (1983) A sensitive and specific stable isotope assay for warfarin and its metabolites. Biomed. Mass Spectrom. 10, 395-398. (15) Lawrence, R. F., Rettie, A. E., Eddy, A. C., and Trager, W. F. (1990) Chemical synthesis, absolute configuration, and stereochemistry of formation of 10-hydroxywarfarin: A major oxidative metabolite of (+)-@)-warfarin from hepatic microsomal preparations. Chirality 2, 96-105. (16) Wilkinson, L. (1987) SYSTAT The system for statistics. SYSTAT Inc., Evanston, IL. (17) Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265-272. (18) Estabrook, R. W., Peterson, J., Baron, J., and Hildebrandt, A. (1972) The spectrophotometric measurement of cytochromes aasociated with drug metabolism. Methods Pharmacol. 2,303-350. (19) Back, D. J., Tjia, J. F., Karbwang, J., and Colbert, J. (1988) In vitro inhibition studies of tolbutamide hydroxylase activity of human liver microsomes by azoles, sulphonamides and quinolines. Br. J. Clin. Pharmacol. 26, 23-29. (20) Brian, W. R., Srivastava, P. K., Umbenhauer, D. R., Lloyd, R. S., and Guengerich, F. P. (1989) Expression of a human liver cytochrome P-450 protein with tolbutamide hydroxylase activity in Saccharomyces cereuisiae. Biochemistry 28, 4993-4999. (21) Miners, J. O., Smith, K. J., Robson, R. A., McManus, M. E., Veronese, M. E., and Birkett, D. J. (1988) Tolbutamide hydroxy-

Chem. Res. Toxicol., Vol. 5, No. l , 1992 59 lation by human liver microsomes. Kinetic characterization and relationship to other cytochrome P-450 dependent xenobiotic oxidations. Biochem. Pharmacol. 37, 1137-1144. (22) Kunze, K. L., Eddy, A. C., Gibaldi, M., and Trager, W. F. (1991) Metabolic enantiomeric interactions: The inhibition of human (S)-warfarin-7-hydroxylaseby @)-warfarin. Chirality 3, 24-29. (23) Miners, J. O., Foenander, T., Wanwimolruk, S., Gallus, A. S., and Birkett, D. J. (1982) The effect of sulphinpyrazone on oxidative drug metabolism in man: Inhibition of tolbutamide elimination. Eur. J. Clin. Pharmacol. 22, 321-326. (24) Pond, S. M., Birkett, D. J., and Wade, D. N. (1977) Mechanisms of inhibition of tolbutamide metabolism: phenylbutazone, oxyphenbutazone, sulfaphenazole. Clin. Pharmacol. Ther. 22, 573-579. (25) Srivastava, P. K., Yun, C-H., Beaune, P. H., Ged, C., and Guengerich, F. P. (1991) Separation of human liver microsomal tolbutamide hydroxylase and (SI-mephenytoin 4’-hydroxylase cytochrome P-450 enzymes. Mol. Pharmacol. 40, 69-79. (26) Yasumori, T., Kawano, S., Nagata, K., Shimada, M., Yamazoe, Y., and Kato, R. (1987) Nucleotide sequence of a human liver cytochrome P-450 related to the rate male specific form. J. Biochem. 102, 1075-1082. (27) Romkes, M., Faletto, M. B., Blaisdell, J. A., Raucy, J. L., and Goldstein, J. A. (1991) Cloning and expression of complementary DNA’s for multiple members of the human cytochrome P-450 IIC subfamily. Biochemistry 30, 3247-3255. (28) Veronese, M. E., Mackenzie, P. I., Doecke, C. J., McManus, M. E., Miners, J. O., and Birkett, D. J. (1991) Tolbutamide and phenytoin hydroxylations by cDNA-expressed human liver cytochrome P-4502C9. Biochem. Biophys. Res. Commun. 175, 1112-1118. (29) Kimura, S., Pastewka, J., Gelboin, H. V., and Gonzalez, F. J. (1987) cDNA and amino acid sequences of two members of the human P450IIC gene subfamily. Nucleic Acids Res. 15, 10053-10054. (30) Toon, S., Hopkins, K. J., Garstang, F. M., Aarons, L., Sedman, A., and Rowland, M. (1987) Enoxacin-warfarin interaction: Pharmacokinetic and stereochemical aspects. Clin. Pharmacol. Ther. 42, 33-41. (31) Brian, W. R., Sari, M.-A., Iwasaki, M., Shimada, T., Kaminsky, L. S., and Guengerich, F. P. (1990) Catalytic activities of human liver cytochrome P-450 IIIA4 expressed in Saccharomyces cereuisiae. Biochemistry 29, 11280-11292.