dmd 39:2020–2033, 2011 printed in u.s.a. the proton pump...

The Proton Pump Inhibitor, Omeprazole, but Not Lansoprazole orPantoprazole, Is a Metabolism-Dependent Inhibitor of CYP2C19:

Implications for Coadministration with Clopidogrel□S

Brian W. Ogilvie, Phyllis Yerino, Faraz Kazmi, David B. Buckley, Amin Rostami-Hodjegan,Brandy L. Paris, Paul Toren, and Andrew Parkinson

XenoTech, LLC, Lenexa, Kansas (B.W.O., P.Y., F.K., D.B.B., B.L.P., P.T., A.P.); University of Manchester, Manchester, UnitedKingdom (A.R.H.); and Simcyp Limited, Sheffield, United Kingdom (A.R.H.)

Received June 22, 2011; accepted July 27, 2011

ABSTRACT:

As a direct-acting inhibitor of CYP2C19 in vitro, lansoprazole is morepotent than omeprazole and other proton pump inhibitors (PPIs), butlansoprazole does not cause clinically significant inhibition ofCYP2C19 whereas omeprazole does. To investigate this apparentparadox, we evaluated omeprazole, esomeprazole, R-omeprazole,lansoprazole, and pantoprazole for their ability to function as direct-acting and metabolism-dependent inhibitors (MDIs) of CYP2C19 inpooled human liver microsomes (HLM) as well as in cryopreservedhepatocytes and recombinant CYP2C19. In HLM, all PPIs were foundto be direct-acting inhibitors of CYP2C19 with IC50 values varyingfrom 1.2 �M [lansoprazole; maximum plasma concentration (Cmax) �

2.2 �M] to 93 �M (pantoprazole; Cmax � 6.5 �M). In addition, weidentified omeprazole, esomeprazole, R-omeprazole, and omepra-zole sulfone as MDIs of CYP2C19 (they caused IC50 shifts after a30-min preincubation with NADPH-fortified HLM of 4.2-, 10-, 2.5-, and

3.2-fold, respectively), whereas lansoprazole and pantoprazole werenot MDIs (IC50 shifts < 1.5-fold). The metabolism-dependent inhibi-tion of CYP2C19 by omeprazole and esomeprazole was not reversedby ultracentrifugation, suggesting that the inhibition was irreversible(or quasi-irreversible), whereas ultracentrifugation largely reversedsuch effects of R-omeprazole. Under various conditions, omeprazoleinactivated CYP2C19 with KI (inhibitor concentration that supportshalf the maximal rate of inactivation) values of 1.7 to 9.1 �M and kinact

(maximal rate of enzyme inactivation) values of 0.041 to 0.046 min�1.This study identified omeprazole, and esomeprazole, but not R-omeprazole, lansoprazole, or pantoprazole, as irreversible (or quasi-irreversible) MDIs of CYP2C19. These results have important implica-tions for the mechanism of the clinical interaction reported betweenomeprazole and clopidogrel, as well as other CYP2C19 substrates.

Introduction

Omeprazole and other proton pump inhibitors (PPIs, i.e., esome-prazole, lansoprazole, dexlansoprazole, rabeprazole, and pantopra-zole) are well known for a relatively low incidence of adverseevents and pharmacokinetic drug-drug interactions (DDIs). Nev-ertheless, PPIs are the perpetrators of interactions with cyclospor-ine (omeprazole and rabeprazole), diazepam (esomeprazole andomeprazole), and warfarin (esomeprazole, lansoprazole, omepra-zole, and rabeprazole) (Wallace and Sharkey, 2011). In addition,all drugs that increase gastric pH can affect the bioavailability ofdrugs such as ampicillin esters, iron salts, and ketoconazole,among others (Wallace and Sharkey, 2011). Omeprazole, the firstapproved PPI, decreases the clearance of drugs such as diazepam,moclobemide, escitalopram, carbamazepine, saquinavir, sibut-ramine, proguanil, etravirine, disulfiram, phenytoin, voriconazole,and clopidogrel, and, by inducing CYP1A2, it increases the clear-ance of several antipsychotic drugs, such as imipramine, theoph-

This work was previously presented in part as follows: Paris BL, Yerino P,Ogilvie BW, and Parkinson A (2008) The proton pump inhibitors (PPIs) omeprazoleand rabeprazole, but not lansoprazole, are in vitro time-dependent inhibitors ofCYP2C19. 10th European Regional ISSX Meeting; 2008 May 18–21; Vienna,Austria. International Society for the Study of Xenobiotics, Washington, DC. DrugMetab Rev 40 (Suppl 1): 89–90; Parkinson A, Kazmi F, Buckley DB, Yerino P,Ogilvie BW, and Paris BL (2010) System-dependent outcomes during the evalu-ation of drug candidates as inhibitors of cytochrome P450 (CYP) and uridinediphosphate glucuronosyltransferase (UGT) enzymes: human hepatocytes versusliver microsomes versus recombinant enzymes. Drug Metab Pharmacokinet 25:16–27; Parkinson A (2011) Irreversible inhibition of CYP2C19 by some but not allproton pump inhibitors and its relevance to the anti-platelet effect of clopidogrel(Plavix). North Jersey ACS Drug Metabolism Discussion Group (http://www.njacs.org/drugmet.html); 2011 Apr 13; Somerset, NJ.

Article, publication date, and citation information can be found athttp://dmd.aspetjournals.org.

doi:10.1124/dmd.111.041293.□S The online version of this article (available at http://dmd.aspetjournals.org)

contains supplemental material.

ABBREVIATIONS: PPI, proton pump inhibitor; AUC, area under the plasma concentration versus time curve; CLint, in vitro intrinsic clearance;Cmax, maximum plasma concentration; P450, cytochrome P450; DDI, drug-drug interaction; EM, extensive metabolizer; HLM, human livermicrosomes; Ki, inhibition constant; KI, inhibitor concentration that supports half the maximal rate of inactivation; kinact, maximal rate of enzymeinactivation; LC/MS-MS, liquid chromatography/tandem mass spectrometry; MTDI, University of Washington Metabolism and Transport DrugInteraction database; MDI, metabolism-dependent inhibitor; FDA, U.S. Food and Drug Administration; DMSO, dimethyl sulfoxide; PM, poormetabolizer; MSM, mechanistic state model; MDM, mechanistic dynamic model; amu, atomic mass units.

0090-9556/11/3911-2020–2033$25.00DRUG METABOLISM AND DISPOSITION Vol. 39, No. 11Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics 41293/3721195DMD 39:2020–2033, 2011 Printed in U.S.A.

2020

http://dmd.aspetjournals.org/content/suppl/2011/07/27/dmd.111.041293.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on M

ay 21, 2018dm

d.aspetjournals.orgD

ownloaded from

http://dmd.aspetjournals.org/content/suppl/2011/07/27/dmd.111.041293.DC1

http://dmd.aspetjournals.org/

ylline, and tacrine (Angiolillo et al., 2011; Wallace and Sharkey,2011; MTDI database, http://www.druginteractioninfo.org).

The inhibitory DDIs listed above have been attributed, at least inpart, to inhibition of CYP2C19 by omeprazole. However, clinicallyrelevant DDIs with omeprazole are generally of low magnitude[�120% (2.2-fold) increase in plasma AUC of CYP2C19 sub-strates; (MTDI database)]. By way of comparison, there is up to a14.6-fold increase in the area under the curve (AUC) of omepra-zole when CYP2C19 is completely absent as occurs in geneticallydetermined CYP2C19 poor metabolizers (Furuta et al., 1999).Nearly 40 in vitro studies examining PPIs as P450 inhibitors havebeen published to date (MTDI database). Only omeprazole andlansoprazole (and their S-enantiomers) have been shown to inhibitCYP2C19, with IC50 or Ki values �1 �M. Of particular interest isthat none of these in vitro studies examined the PPIs as metabo-lism-dependent inhibitors (MDIs; also known as time-dependentinhibitors) of cytochrome P450 (P450) enzymes. On the basis ofthe lowest reported Ki values for CYP2C19 inhibition in the MTDIdatabase (i.e., 0.45 �M for lansoprazole and 1 �M for omeprazole)and the reported plasma Cmax for lansoprazole and omeprazole inCYP2C19-extensive metabolizers (2.2 and 3.9 �M, respectively;Li et al., 2004; Hassan-Alin et al., 2005), the [I]total/Ki values forlansoprazole and omeprazole would be 4.9 and 3.9, respectively[calculated as recommended in the Draft U.S. Food and DrugAdministration (FDA) Guidance for Industry, 2006, http://www.fda.gov/Drugs/DevelopmentApprovalProcess/DevelopmentResources/DrugInteractionsLabeling/ucm080499.htm], which means thatboth drugs would be expected to cause clinically relevant directinhibition of CYP2C19, but that lansoprazole would have a higherlikelihood of doing so than would omeprazole. However, lanso-prazole has been reported to cause no interaction with theCYP2C19 substrates diazepam and phenytoin (MTDI database,Wallace and Sharkey, 2011). The lack of clinically relevant directinhibition of CYP2C19 by lansoprazole is likely explained by itsrelatively short half-life (1.1 h; Li et al., 2004) and high plasmaprotein binding (�97%; Prevacid prescribing information, 2010,http://www.accessdata.fda.gov/drugsatfda_docs/label/2010/020406s074,021428s021lbl.pdf). However, omeprazole also has ashort half-life (0.7 h; Hassan-Alin et al., 2005) and high plasmaprotein binding (�95%; Prilosec prescribing information, 2011,http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/019810s092,022056s008lbl.pdf), and yet, as described above,omeprazole does cause clinically significant inhibition ofCYP2C19. This apparent discrepancy could be explained ifomeprazole, but not lansoprazole, were an irreversible inhibitor ofCYP2C19.

In support of this possibility, we presented preliminary in vitroevidence for metabolism-dependent inhibition of CYP2C19 byomeprazole (Paris et al., 2008). More recently, a large clinical studyin 282 healthy subjects (Angiolillo et al., 2011) demonstrated thatomeprazole inhibited the CYP2C19-dependent activation of clopi-dogrel as evidenced by a 40 to 47% decrease in the formation of H4,the purported active antiplatelet metabolite of clopidogrel. Inhibitionof clopidogrel activation was observed even when the two drugs wereadministered 12 h apart. In the same subjects, no such interaction wasobserved between pantoprazole and clopidogrel.

The interaction between clopidogrel and PPIs and the effect of theCYP2C19 poor metabolizer phenotype have prompted warnings from theFDA and European Medicines Energy (2010, http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm203888.htm; http://www.ema.europa.eu/docs/en_GB/document_library/Public_statement/2010/03/WC500076346.pdf), although the FDA

specifically warns against coadministration of clopidogrel and omeprazoleand further specifically suggests that pantoprazole may be a safer alternative(2010, http://www.fda.gov/Drugs/DrugSafety/ucm231161.htm).

The identity of the enzyme(s) responsible for the multistep activa-tion of clopidogrel is controversial (Savi et al., 1994; Clarke andWaskell, 2003; Abraham et al., 2010; Kazui et al., 2010; Zahno et al.,2010). Much of the pharmacogenomic data strongly implicateCYP2C19 as the most important enzyme for clopidogrel activation onthe basis of the poor response to clopidogrel in carriers of reduced-function CYP2C19 alleles (e.g., the *2 and *3 alleles) and increasedbleeding in carriers of the increased-function CYP2C19*17 allele(Sibbing et al., 2010). The DDI between PPIs and clopidogrel is ofparticular importance because PPIs are coprescribed in up to approx-imately two thirds of patients after discharge from the hospital be-cause PPIs lessen the severity of the gastrointestinal hemorrhageassociated with clopidogrel treatment (Ho et al., 2009).

In the present study, we examined the in vitro inhibitory potentialof omeprazole, its individual enantiomers and selected metabolites, aswell as lansoprazole and pantoprazole (Fig. 1) in pooled human livermicrosomes (HLM), pooled cryopreserved hepatocytes, and recombi-nant CYP2C19, with a special emphasis on the potential for thesedrugs to cause metabolism-dependent inhibition of CYP2C19. Theimplications of our results were explored by dynamic simulationsassessing the level of active CYP2C19 under multiple doses ofomeprazole. This physiologically based pharmacokinetic modelingand simulation provided additional insight into the ongoing debatesurrounding the interaction between PPIs and clopidogrel.

Materials and Methods

Chemicals and Reagents. Omeprazole and 4�-hydroxymephenytoin werepurchased from Sigma-Aldrich (St. Louis, MO). Esomeprazole, 5�-hydroxyo-meprazole, S-mephenytoin, lansoprazole, omeprazole sulfide, and pantopra-zole were purchased from Toronto Research Chemicals (North York, ON,Canada). R-Omeprazole and omeprazole sulfone were purchased from SynFineResearch (Richmond Hill, ON, Canada) or Toronto Research Chemicals.Human liver microsomes (pooled from 16 donors) and human hepatocyteswere prepared from nontransplantable livers and characterized at XenoTech,LLC (Lenexa, KS) as described previously (Parkinson et al., 2011), andrecombinant CYP2C19 (Bactosomes) was obtained from Cypex (Dundee,Scotland). All other reagents were obtained from commercial sources, asdetailed elsewhere (Ogilvie et al., 2006; Nassar et al., 2009; Parkinson et al.,2011).

CYP2C19 Inhibition: IC50 Determinations. CYP2C19 activity in HLMwas determined according to previously published procedures (Ogilvie et al.,2006; Nassar et al., 2009; Parkinson et al., 2011). In brief, incubations wereconducted in 200-�l incubation mixtures (pH 7.4) containing high-puritywater, potassium phosphate buffer (50 mM), MgCl2 (3 mM), EDTA (1 mM),NADP (1 mM), glucose 6-phosphate (5 mM), glucose-6-phosphate dehydro-genase (1 unit/ml), S-mephenytoin (approximately equal to Km, i.e., 40 �M,final), and HLM (0.1 mg protein/ml, except where otherwise noted). Allincubations were conducted in duplicate at 37°C for 5 min (except whereotherwise noted) and were terminated by the addition of 200 �l of acetonitrilecontaining an internal standard (d3-4�-hydroxymephenytoin). Aliquots of thestock and/or working solutions of the inhibitors (i.e., omeprazole, esomepra-zole, R-omeprazole, lansoprazole, or pantoprazole; final concentrations rang-ing from 0.1 to 100 �M in methanol) were added to buffer mixtures containingthe components described above but before addition of the NADPH-generatingsystem. To evaluate the potential for metabolism-dependent inhibition, theinhibitors (at the same concentrations used to evaluate direct inhibition) wereincubated at 37°C with NADPH-fortified HLM for approximately 30 min.After the 30-min preincubation, S-mephenytoin (40 �M, final) was added, andthe incubation was continued for 5 min to measure residual P450 activity.Precipitated protein was pelleted by centrifugation (920g for 10 min at 10°C).Calibration and quality control standards (4�-hydroxymephenytoin) were pre-pared in zero-time incubations. IC50 determinations with recombinant human

2021OMEPRAZOLE IS A METABOLISM-DEPENDENT INHIBITOR OF CYP2C19

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


CYP2C19 (15 pmol/ml) were conducted in the same manner except that theincubation time was only 2 min to prevent overmetabolism of omeprazole.Metabolite formation (4�-hydroxymephenytoin) was analyzed by liquid chro-matography-tandem mass spectrometry (LC-MS/MS) as described previously(Ogilvie et al., 2006; Parkinson et al., 2011).

Incubations with cryopreserved human hepatocytes (pooled, n � 3, 106

cells/ml) were conducted in 200-�l incubation mixtures at approximately 37°Cin Krebs Henseleit buffer in triplicate. In all cases, the solvent or omeprazolewas allowed to equilibrate for 10 min with hepatocytes before incubations. Forsamples with no preincubation with inhibitor, reactions were started by theaddition of hepatocytes to prewarmed Krebs Henseleit buffer containing in-hibitor and S-mephenytoin (40 �M, final concentration). For reactions with apreincubation, hepatocytes and inhibitor were incubated at 37°C for 30 min,and reactions were started by addition of S-mephenytoin (40 �M, final). In allcases, marker substrate reactions were conducted for 60 min and terminated bythe addition of an equal volume of acetonitrile and internal standard (d3-4�-hydroxymephenytoin). Metabolite formation (4�-hydroxymephenytoin) wasanalyzed by LC-MS/MS as described previously (Ogilvie et al., 2006; Parkin-son et al., 2011).

Metabolic Stability of Omeprazole. The metabolic stability of omeprazole(10 �M) was determined at three concentrations of HLM (0.1, 1, and 2.5mg/ml) under conditions similar to those described under CYP2C19 Inhibition:IC50 Determinations for CYP2C19 inhibition experiments. Omeprazole wasincubated for 0, 5, 10, 20, 30, 45, and 60 min in triplicate. Reactions wereterminated by the addition of an equal volume of acetonitrile and internalstandard (pantoprazole). Precipitated protein was removed by centrifugation(920g for 10 min at 10°C). Omeprazole disappearance was monitored byLC-MS/MS. Calibration standards were prepared in zero-time incubations.

Microsomal Binding of Omeprazole. The binding of omeprazole to mi-crosomal protein was determined by ultrafiltration with Millipore AmiconCentriplus centrifugal filter devices (15 ml, 30-kDa membrane) obtained fromThermo Fisher Scientific (Waltham, MA). Omeprazole (2 and 10 �M), wasincubated with pooled HLM (0, 0.1, 1, or 2.5 mg/ml) as described underMetabolic Stability of Omeprazole, but in the absence of an NADPH-gener-ating system, at 37°C for 10 min. Aliquots (1.1 ml) were then removed andadded to the ultrafiltration devices and centrifuged at 1900g in a Sorvall(Asheville, NC) RC 5C centrifuge with a Sorvall SS-34 rotor at room temper-ature for 5 min. Aliquots of the ultrafiltrate (100 �l) were transferred to glass

tubes, and an equal volume of acetonitrile was added and vortexed. Precipi-tated protein was removed by centrifugation (920g for 10 min at 10°C). Aftercentrifugation, an aliquot (100 �l) was transferred to an equal volume ofacetonitrile (with pantoprazole as the internal standard) and analyzed foromeprazole concentration by LC-MS/MS.

KI and kinact Determinations. To determine the KI and kinact values for theinactivation of CYP2C19, various concentrations of omeprazole (1–60 �M)were incubated for 2.5 to 15 min with pooled human liver microsomes (0.1 and2.5 mg/ml) at 37°C. For experiments conducted at 0.1 mg/ml HLM, after thepreincubations, S-mephenytoin (40 or 400 �M final concentration) was addedand residual CYP2C19 activity was determined as described under CYP2C19Inhibition: IC50 Determinations. For experiments conducted at 2.5 mg/ml,after the preincubation, an aliquot (8 �l) was transferred to another incubationtube (final volume 200 �l) containing S-mephenytoin (400 �M) and anNADPH-generating system to measure residual CYP2C19 activity as de-scribed under CYP2C19 Inhibition: IC50 Determinations. This procedure di-luted the microsomes to 0.1 mg/ml and diluted omeprazole to 1/25th of itsoriginal concentration. Reactions were performed in triplicate.

CYP2C19 Activity in Cultured Human Hepatocytes. Cultured humanhepatocytes were prepared and treated, and microsomes (0.02 mg/ml) wereisolated as described previously (Nassar et al., 2009). In brief, cultured humanhepatocytes were treated for 72 h with 0.1% (v/v) dimethyl sulfoxide (DMSO)(control) or 100 �M omeprazole. Isolated microsomes were washed andincubated for 30 min with marker substrate (40 �M S-mephenytoin) todetermine CYP2C19 activity as described under CYP2C19 Inhibition: IC50

Determinations.Assessment of MDI Reversibility by Ultracentrifugation. Omeprazole

(100 �M), R-omeprazole (100 �M), and esomeprazole (100 �M) were incu-bated in triplicate with NADPH-fortified pooled HLM (0.1 mg/ml) at 37°C for30 min in potassium phosphate buffer (50 mM, pH 7.4), MgCl2 (3 mM), EDTA(1 mM, pH 7.4), and chemically reduced NADPH (1 mM) as describedpreviously (Parkinson et al., 2011). Incubations (n � 3) with solvent alone (1%methanol v/v, final) served as controls. After the 30-min incubation, HLMwere 1) assayed directly for residual CYP2C19 activity; 2) re-isolated byultracentrifugation and then assayed for residual CYP2C19 activity; or 3)treated with potassium ferricyanide, re-isolated by ultracentrifugation, and thenassayed for residual CYP2C19 activity. For samples in groups 2 and 3,microsomal protein was re-isolated by ultracentrifugation (100,000g for 30

O OOmeprazole Esomeprazole

N

HN

O

NS

O N

HN

O

NS

O

N O

O

S

O N O

O

S

O

R-Omeprazole Omeprazole sulfone

HNNS

HNNS

O

O OOmeprazole sulfide 5´-Hydroxyomeprazole

N

HN

O

NS

N

HN

O

NS

OHO

N

O

O

CF3

N O

O

OO

F

Lansoprazole Pantoprazole

HNNS

HNNS

F

FIG. 1. Structures of the PPIs and omeprazole metabolites examined.

2022 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


min at 4°C in a Beckman ultracentrifuge with a 70 Ti rotor; Beckman Coulter,Fullerton, CA). The supernatant fraction was discarded, and the resultantmicrosomal pellets were rinsed three times with wash buffer (150 mM potas-sium chloride and 10 mM EDTA, pH 7.4) to remove residual omeprazoleand/or any reversible inhibitory metabolites. Microsomal pellets were resus-pended in 250 mM sucrose, and the microsomal protein concentration wasdetermined by the Pierce BCA protein assay (Thermo Fisher Scientific). Forsamples in group 3, HLM were incubated with potassium ferricyanide (2 mM)for 10 min at 37°C before re-isolation of microsomal protein by ultracentrif-ugation to disrupt nitrogen-based metabolite inhibitory complexes (Franklin,1977). Residual CYP2C19 activity was assessed at a final concentration of 0.1mg/ml HLM (supplemented with an NADPH-regenerating system) with S-mephenytoin (400 �M, i.e., 10 � Km) to reduce the inhibitory effects of anyresidual competitive inhibition.

Analytical Methods. LC-MS/MS methods were performed as describedpreviously (Ogilvie et al., 2006; Parkinson et al., 2011). Omeprazole analysiswas performed with an Applied Biosystems/Sciex API2000 high-performanceliquid chromatography-tandem mass spectroscopy system equipped with anelectrospray (TurboIonSpray) ionization source (Applied Biosystems, FosterCity, CA), two LC-10ADvp pumps with an SCL-10Advp controller, a SIL-HTA autosampler, and a DGU-14 solvent degasser (Shimadzu, Kyoto, Japan).The high-performance liquid chromatography column used was an AtlantisdC18, 5 �m, 100- by 2-mm column (Waters, Milford, MA), which waspreceded by a direct connection guard column with a C8, 4- by 2-mm cartridge(Waters). Masses were monitored in the multiple reaction monitoring mode:omeprazole 345.93 197.9 amu and internal standard, pantoprazole, 383.93199.9 amu. Mobile phases were buffer A (0.2% formic acid in water) andbuffer B (0.2% formic acid in methanol). Omeprazole and pantoprazole wereeluted with a linear gradient (25% B to 75% B) over 2.5 min.

Data Analyses and Simulations. All IC50, half-life, KI, and kinact values weredetermined by nonlinear regression with XLfit3 (version 3.0.5; ID BusinessSolutions Ltd., Guildford, Surrey, UK), which is an add-in for the computerprogram Microsoft Excel (Office 2003; Microsoft, Redmond, WA) or with GraFit(version 4.0.21; Erithacus Software, Horley, Surrey, UK), as detailed previously(Ogilvie et al., 2006; Nassar et al., 2009; Parkinson et al., 2011).

In vitro-to-in vivo extrapolation of CYP2C19 inactivation was performedusing a mechanistic static model (MSM) (Grimm et al., 2009) and a mecha-nistic dynamic model (MDM) (Rowland Yeo et al., 2010). The MSM wasimplemented based on the following equation:

AUCi

AUC�

1

�fm,CYP

1 � � kinact � �I�

kdeg � �KI � �I�� 1 � fm,CYP�

(1)

where fm,CYP represents the fraction of a hypothetical concomitantly admin-istered drug metabolized by a given P450 enzyme, [I] is the inactivatorconcentration, kinact and KI are the in vitro inactivation parameters, and kdeg isthe rate constant for enzyme degradation [which has not been determinedexperimentally in vivo for CYP2C19, but for which the average value wasreported to be 0.000445 min1 on the basis of in vitro data (Renwick et al.,2000; Yang et al., 2008)]. Given that the unbound plasma Cmax was previouslyfound to be more predictive than the total Cmax for MDIs (Obach et al., 2007;Grimm et al., 2009), this value was used in the MSM (eq. 1) and was based onthe total omeprazole plasma Cmax of 3.87 �M after 5 days of dosing (40 mgdaily) in CYP2C19-extensive metabolizers (EMs) (Hassan-Alin et al., 2005),coupled with its plasma protein binding of 95%, for an unbound plasma Cmax

of 0.19 �M.The MDM simulations were performed using the Simcyp Population-Based

Simulator (version 10.2; Simcyp Limited, Sheffield, UK). The differentialequations that form the basis of these simulations are described in detailelsewhere (Rowland Yeo et al., 2010). Input parameters can be found in thesupplemental data.

The purpose of the simulations was to assess the time-varying effect ofrepeated omeprazole administration (40 mg every 12 h) on the level of activeCYP2C19 in liver. This included the self-inhibition effect because the deac-tivation of CYP2C19 led to lower clearance and higher concentrations ofomeprazole itself. The simulations were performed twice using the lower (1.7

�M) and upper (9.1 �M) boundaries of observed KI values and taking intoaccount an unbound fraction of 0.75 (i.e., the free fraction of omeprazole inHLM at 1–2.5 mg/ml). kinact was assumed to be 0.045 min1, and turnover ofCYP2C19 was the same as that assumed for the MSM.

Although the main purpose of the simulations was to assess the level of activeenzyme, S-mephenytoin was used as a substrate on day 14 after administration ofomeprazole, and the effect on AUC was simulated on day 7 after administration ofomeprazole. The omeprazole dose was continued until day 14, and the AUC forS-mephenytoin was assessed from day 7 until the end of the study (day 14) andcompared with AUC in the absence of an inactivator.

Results

Inhibitory Effect of Omeprazole on S-Mephenytoin 4�-Hy-droxylation in Multiple Test Systems: IC50 Determinations, Met-abolic Stability, and Microsomal Protein Binding. Omeprazole wasevaluated as a direct-acting and MDI of CYP2C19 activity (S-mephe-nytoin 4�-hydroxylation) in pooled human liver microsomes (n � 16),recombinant CYP2C19 (Bactosomes), and pooled, cryopreserved hu-man hepatocytes (n � 3) at a substrate concentration approximatelyequal to Km (40 �M). The results are summarized in Fig. 2, a to d, andTable 1. The results show that omeprazole caused metabolism-depen-dent inhibition of CYP2C19 as evidenced by a left shift in the IC50

curves after a 30-min preincubation with NADPH-fortified HLM,recombinant CYP2C19, and human hepatocytes (IC50 shifts of 4.1-,6.9-, and 3.6-fold, respectively). A left shift of �1.5-fold is consid-ered indicative of metabolism-dependent inhibition (Parkinson et al.,2011). With no preincubation, omeprazole inhibited CYP2C19 in allthree systems with similar IC50 values ranging from 4.7 to 9.6 �M,which decreased to approximately 1.5 �M in all three test systemsafter preincubation with NADPH. The experiment presented in Fig. 2aalso included an IC50 determination with a 30-min preincubation stepwithout NADPH to confirm that the IC50 shift was dependent onNADPH (Table 1; for clarity, these data are not presented in Fig. 2a).Omeprazole was also examined as an inhibitor of CYP1A2, CYP2B6,CYP2C8, CYP2C9, CYP2D6, CYP3A4, and CYP2J2 and was foundto be a weak direct inhibitor of these enzymes (IC50 approximately100 �M), with no NADPH-dependent IC50 shift 1.5-fold (seesupplemental data).

The data presented in Fig. 2a were obtained under “low microsomalprotein-short incubation time” conditions (i.e., 0.1 mg/ml HLM for 5min). However, S-mephenytoin is a low-turnover substrate in HLM,for which reason S-mephenytoin 4�-hydroxylation activity is fre-quently assessed by incubating S-mephenytoin with high concentra-tions of HLM [i.e., �0.2 mg/ml) for 20 to 40 min (e.g., Walsky andObach, 2004; Nishiya et al., 2009; Parkinson et al., 2011)]. Becausemetabolism-dependent inhibition of CYP2C19 by omeprazole had notbeen previously described, we also examined the ability of omepra-zole to inhibit CYP2C19 under “high microsomal protein-long incu-bation time” conditions (i.e., 1 mg/ml HLM for 30 min). Under suchconditions we detected no metabolism-dependent inhibition ofCYP2C19 by omeprazole (i.e., IC50 shift �1.5; Fig. 2b).

Because of the difference in results between low microsomal pro-tein-short incubation time and high microsomal protein-long incuba-tion time conditions, we examined the metabolic stability and non-specific binding of omeprazole in HLM. In NADPH-fortified HLM at0.1 mg/ml, omeprazole was relatively stable with a half-life of 98 min(Fig. 2e). However, at 1 and 2.5 mg/ml HLM (i.e., the concentrationused in one of the kinact determinations, see Inhibitory Effects of theMajor Omeprazole Metabolites and Enantiomers on S-Mephenytoin4�-Hydroxylation in Human Liver Microsomes: IC50 Determinations)omeprazole rapidly disappeared from incubations, with half-life val-ues of only 14 and 5.7 min, respectively. Nonspecific binding ofomeprazole to HLM was also examined (by ultrafiltration) as a


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


possible cause of the discrepancy in results between low microsomalprotein-short incubation time and high microsomal protein-long incu-bation time conditions. However, as shown in Fig. 2f, 75% ofomeprazole (2 and 10 �M) remained free in the incubation from 0.1to 2.5 mg/ml.

Inhibitory Effects of the Major Omeprazole Metabolites andEnantiomers on S-Mephenytoin 4�-Hydroxylation in HumanLiver Microsomes: IC50 Determinations. Esomeprazole, R-omepra-zole, omeprazole sulfide, omeprazole sulfone, and 5�-hydroxyome-prazole were evaluated as direct-acting and MDIs of CYP2C19 ac-tivity (S-mephenytoin 4�-hydroxylation) in pooled human livermicrosomes (n � 16, 0.1 mg/ml) at a substrate concentration approx-imately equal to the Km (40 �M). The results are summarized in Table

1 and Fig. 3. The results show that esomeprazole (Fig. 3a), R-omeprazole (Fig. 3b), and omeprazole sulfone (Fig. 3d) caused me-tabolism-dependent inhibition of CYP2C19 as evidenced by a leftshift in the IC50 curves (1.5-fold) after a 30-min preincubation withNADPH-fortified HLM (10-, 2.5-, and 3.2-fold IC50 shifts, respec-tively), with similar “shifted” IC50 values ranging from 1.5 to 5.6 �M.The IC50 values after a 30-min preincubation in the absence ofNADPH were higher than those in the presence of NADPH, suggest-ing that the time-dependent inhibition of CYP2C19 by these com-pounds was in fact dependent on metabolism (Table 1). The IC50

values for omeprazole sulfide after a 30-min preincubation with orwithout NADPH remained similar to the value without preincubation(IC50 shift � 1.5-fold), suggesting that omeprazole sulfide is only a

xyla

tion

) 120

140

160

xyla

tion

) 120

140

160

0.1 mg/mL HLM, 5 min incubation noitabucni nim 03 ,MLH Lm/gm 0.1a b

Zero-min IC50 = 6.9 ± 0.7 µM

30-min IC = 1 7 ± 0 0 µM

Zero-min IC50 = 8.3 ± 0.3 µM

30-min IC = 6 3 ± 0 3 µM

ephe

nyto

in 4

´-hyd

ro(P

erce

nt o

f con

trol

20

40

60

80

100

120

ephe

nyto

in 4

´-hyd

ro(P

erce

nt o

f con

trol

20

40

60

80

100

12030-min IC50 = 1.7 ± 0.0 µM 30-min IC50 = 6.3 ± 0.3 µM

[Omeprazole](µM)

0.01 0.1 1 10 100 1000S-M

e

0

[Omeprazole](µM)

0.01 0.1 1 10 100 1000S-M

e

0

Recombinant CYP2C19 setycotapeh namuh devreserpoyrCc d4´

-hyd

roxy

latio

nof

con

trol)

80

100

120

140

160

4´-h

ydro

xyla

tion

of c

ontro

l)

80

100

120

140

160Zero-min IC50 = 9.6 ± 2.1 µM30-min IC50 = 1.4 ± 0.2 µM

Zero-min IC50 = 4.7 ± 0.3 µM

30-min IC50 = 1.3 ± 0.2 µM

[Omeprazole](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyto

in

(Per

cent

o

0

20

40

60

[Omeprazole](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyto

in

(Per

cent

0

20

40

60

(µM)(µM)

Omeprazole metabolic stability (HLM) e Omeprazole free fraction in HLM f

mai

ning

ontro

l)

100

120

140 0.1 mg/mL t1/2 = 98 ± 15 min1.0 mg/mL t1/2 = 14 ± 1 min2.5 mg/mL t1/2 = 5.7 ± 0.1 min

Zero protein control

2 µM Omeprazole

10 µM Omeprazole

Om

epra

zole

rem

(Per

cent

of c

o

0

20

40

60

80

0 20 40 60 80 100 120

p

0.1 mg/mL HLM

1 mg/mL HLM

2.5 mg/mL HLM

Incubation time(min)

0 10 20 30 40 50 60 70 Omeprazole(Percent unbound)

FIG. 2. Evaluation of omeprazole as a direct-acting and MDI of CYP2C19. Each symbolrepresents the average of duplicate determina-tions unless otherwise indicated. a, omepra-zole inhibited CYP2C19 in pooled HLMwith IC50 values as shown (S-mephenytoin4�-hydroxylation control rates � 97.3 and95 pmol � mg1 � min1 with and withoutpreincubation, respectively). b, omeprazoleinhibited CYP2C19 in pooled HLM withIC50 values as shown (S-mephenytoin 4�-hydroxylation control rates � 97.3 and74.3 pmol � mg1 � min1 with and withoutpreincubation, respectively). c, omeprazoleinhibited recombinant human CYP2C19with IC50 values as shown (S-mephenytoin4�-hydroxylation control rates � 0.690 and0.828 min1 with and without preincuba-tion, respectively). d, omeprazole inhibitedCYP2C19 cryopreserved human hepato-cytes with IC50 values as shown (S-mephenytoin 4�-hydroxylation controlrates � 24.7 and 24.8 pmol/million cells/min with and without preincubation, re-spectively). Each symbol represents the av-erage of triplicate determinations, and errorbars represent the S.D. e, the metabolicstability of omeprazole (10 �M) was eval-uated in HLM at the protein concentrationsused in this study (0.1, 1, and 2.5 mg/ml),as described under Materials and Methods.Half-life values for the disappearance ofomeprazole are as indicated. Each symbolrepresents the average of triplicate determi-nations, and error bars represent the S.D. f,the microsomal binding of omeprazole (2and 10 �M) was evaluated as describedunder Materials and Methods.

2024 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


direct-acting inhibitor of CYP2C19 (Fig. 3c; Table 1). 5�-Hydroxyo-meprazole, one of the major metabolites formed by CYP2C19, was aweak inhibitor of CYP2C19 (IC50 100 �M) (Fig. 3e). The exper-iments presented in Fig. 3 also included an IC50 determination with a30-min preincubation step without NADPH to confirm that the IC50

shift was dependent on NADPH (Table 1; for clarity, these data arenot presented in Fig. 3).

Inhibitory Effects of Lansoprazole and Pantoprazole on S-Mephenytoin 4�-Hydroxylation in Human Liver Microsomes andCryopreserved Human Hepatocytes: IC50 Determinations. Lanso-prazole and pantoprazole were evaluated as direct-acting and MDIs ofCYP2C19 activity (S-mephenytoin 4�-hydroxylation) in pooled hu-man liver microsomes (n � 16) and pooled, cryopreserved humanhepatocytes (n � 3) at a substrate concentration approximately equalto the Km (40 �M). The results are summarized in Table 1 and Fig. 4.Neither lansoprazole nor pantoprazole caused metabolism-dependentinhibition of CYP2C19 in either HLM or cryopreserved human hepa-tocytes (i.e., IC50 shifts �1.5-fold). Lansoprazole was a relativelypotent (i.e., IC50 � 1 �M), direct-acting inhibitor of CYP2C19 inHLM and cryopreserved human hepatocytes (Fig. 4, a and b), whereaspantoprazole was a relatively weak inhibitor (Fig. 4, c and d). Theexperiments presented in Fig. 4, a and c, also included an IC50

determination with a 30-min preincubation step without NADPH toconfirm that the IC50 shift was dependent on NADPH (Table 1; forclarity, these data are not presented in Fig. 4).

Inactivation of CYP2C19 by Omeprazole: KI and kinact Deter-minations. The results in Fig. 2 suggest that omeprazole is an MDI ofCYP2C19. Accordingly, experiments were performed to determine KI

(the concentration of omeprazole supporting half-maximal rate ofCYP2C19 inactivation) and kinact (the first-order rate constant forCYP2C19 inactivation). The results from three different experimentsare summarized in Fig. 5. The three experimental conditions were1) low HLM concentration with [S] � Km with no dilution step (Fig.5, a and b), 2) low HLM concentration with [S] � 10 � Km (Fig. 5,c and d), and 3) high [HLM] with [S] � 10 � Km with a 25-folddilution step (Fig. 5, e and f). Under all three conditions, the inacti-vation of CYP2C19 was dependent on the concentration of omepra-zole (over the full ranges examined), and the time course conformedto a first-order inactivation process (as indicated by the linearity ofplots of the log of the residual enzyme activity against time; Fig. 5, a,c, and e). The KI values were different depending on whether omepra-zole was preincubated with human liver microsomes at 2.5 mg/ml(and subsequently diluted 25-fold to determine residual CYP2C19activity with substrate � 10 Km; Fig. 5, e and f) or 0.1 mg/ml (and not

diluted to determine residual CYP2C19 activity, with substrate � 10Km or Km; Fig. 5, a–d). The KI value was approximately 9 �M in theformer case and approximately 2 �M in the latter cases. In all threecases, the kinact values were approximately 0.04 min1, which meansthat, in the presence of saturating concentrations of omeprazole, 4% ofCYP2C19 was inactivated every minute. The efficiency of CYP2C19inactivation (kinact/KI) decreased by a factor of approximately 5 whenthe concentration of HLM was increased and a dilution step was used.Taken together, these data suggest that omeprazole is an irreversibleinactivator of CYP2C19.

Effects of Omeprazole on CYP2C19 Activity in MicrosomesPrepared from Hepatocytes after 3 Days of Treatment. CYP2C19activity (S-mephenytoin 4�-hydroxylation) was measured in micro-somes isolated from fresh primary cultures of human hepatocytestreated with omeprazole (100 �M; 3 days). It should be noted that thisconcentration is approximately 19 times greater than the Cmax in poormetabolizers (PMs) [5.3 �M (Chen et al., 2009); who can be neitherinduced nor inhibited with respect to CYP2C19] and 26 times greaterthan the Cmax in EMs (Hassan-Alin et al., 2005) after a 40-mg oraldose of omeprazole. Omeprazole treatment decreased microsomalCYP2C19 activity in hepatocytes that initially expressed high levelsof CYP2C19, but it increased microsomal CYP2C19 activity in hepa-tocytes that initially expressed low levels of CYP2C19 (Fig. 6). Thisdual effect likely reflects the overall effect of two opposing actions ofomeprazole; in hepatocytes with high CYP2C19 activity, the predom-inant effect of omeprazole was irreversible inactivation, but in hepa-tocytes with low CYP2C19 activity, the predominant effect wasinduction via pregnane X receptor activation (Yueh et al., 2005).These data were obtained from many studies conducted to examinethe potential for drug candidates to induce P450 enzymes, in whichprimary cultures of human hepatocytes were treated with 100 �Momeprazole, a well known aryl hydrocarbon receptor activator andtherefore a positive control for CYP1A2 induction. Because the mi-crosomes are obtained from these studies in a way that washes outresidual inhibitor, these data provided additional evidence thatomeprazole is an irreversible inactivator of CYP2C19.

Irreversible or Quasi-Irreversible Inhibition of CYP2C19 byOmeprazole and Esomeprazole, but Not R-omeprazole: Ultracen-trifugation. Because esomeprazole was approximately 4-fold moreeffective as a MDI of CYP2C19 than R-omeprazole (Table 1; Fig. 3,a and b) with a 10-fold shift in IC50 value for the former and a 2.5-foldshift for the latter, omeprazole and its individual enantiomers werefurther examined to determine whether the inactivation of CYP2C19involved irreversible or quasi-irreversible inhibition on the basis of anultracentrifugation method described herein and previously (Parkin-son et al., 2011). A concentration of inhibitor that caused nearlycomplete (i.e., �95%) inhibition after 30-min preincubation withNADPH, but incomplete (��85%) inhibition after 30-min preincu-bation in the absence of NADPH was chosen (i.e., 100 �M for all).Pooled HLM were treated with inhibitor or solvent (methanol, 0.1%v/v) for 30 min in the presence of NADPH and in the presence orabsence of potassium ferricyanide (to reverse the formation of me-tabolite inhibitory complex associated with quasi-irreversible inacti-vation) (Fig. 7). After the 30-min incubation, pooled HLM sampleswere either 1) assayed directly for residual P450 activity (“prespin”);2) re-isolated by ultracentrifugation and then assayed for residualP450 activity (“postspin”); or 3) treated with potassium ferricyanide,re-isolated by ultracentrifugation, and then assayed for residual P450activity (“postspin K3[Fe(CN)6]”). As expected, substantial inhibi-tion was observed after treatment with omeprazole, esomeprazole, andR-omeprazole (Fig. 7, prespin) in a rank-order consistent with the IC50

shifts: esomeprazole � omeprazole R-omeprazole. It was surpris-

TABLE 1

Inhibition of CYP2C19 in human liver microsomes by omeprazole, itsenantiomers, or its major metabolites and lansoprazole and pantoprazole

Compound

IC50 (�M)a

IC50 Shift(fold)b

NoPreincubation

30-MinPreincubation(NADPH)

30-MinPreincubation( NADPH)

Omeprazole 6.9 � 0.7 8.7 � 0.6 1.7 � 0.0 4.2Esomeprazole 15 � 1 16 � 1 1.5 � 0.1 10R-Omeprazole 8.1 � 1.2 12 � 1 3.3 � 0.4 2.5Omeprazole sulfide 9.7 � 0.5 8.4 � 2.7 9.6 � 1.0 1.0Omeprazole sulfone 18 � 2 12 � 3 5.6 � 0.5 3.25-Hydroxyomeprazole 100 100 100 NALansoprazole 1.2 � 0.1 1.2 � 0.1 1.2 � 0.0 1.0Pantoprazole 93 � 7 100 65 � 4 1.4

NA, not applicable (i.e., IC50 values greater than the highest concentration examined).a Values are displayed to two significant figures � S.E. of the measurement.b Calculated from full precision values as (IC50 with no preincubation)/(IC50 with 30-min

preincubation NADPH) and rounded to two significant figures. IC50 shifts 1.5-fold appearin bold.


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


ing to note that centrifugation and re-isolation of HLM after preincu-bation largely reversed the inhibition caused by R-omeprazole (Fig. 7,postspin and postspin K3[Fe(CN)6]). In contrast, centrifugation didnot fully restore CYP2C19 activity after treatment with omeprazole oresomeprazole, which suggests that esomeprazole is the major contrib-utor to the inactivation of CYP2C19 by omeprazole. Potassium fer-ricyanide treatment of the samples caused some additional restorationof activity after treatment with omeprazole and esomeprazole, but itdid not fully restore CYP2C19 activity. Taken together, these resultssuggest that the metabolism-dependent inhibition of CYP2C19 observedwith racemic omeprazole is largely irreversible (because of covalentbinding to the apo protein, heme moiety, or both). However, given thepartial restoration of CYP2C19 activity with potassium ferricyanide,

contribution of a quasi-irreversible mechanism (because of metaboliteinhibitory complex formation) cannot be completely ruled out. Theresults of a previous experiment that was based on overnight dialysis arealso consistent with an irreversible or quasi-irreversible mechanism ofthe metabolism-dependent inhibition of CYP2C19 with racemic omepra-zole (Paris et al., 2008). However, the results presented in Fig. 7 showthat the metabolism-dependent inhibition of CYP2C19 caused by R-omeprazole is largely reversible after re-isolation of HLM, suggestingthat one or more metabolites of this enantiomer is simply a more potentinhibitor of CYP2C19 than the parent, whereas metabolism of esome-prazole largely leads to irreversible inactivation of CYP2C19.

Simulation of Time-Dependent Changes in Active CYP2C19.Active levels of hepatic CYP2C19 in the presence of omeprazole (40

xyla

tion

) 120

140

160

oxyl

atio

nl) 120

140

160

belozarpemosE a R-Omeprazole

Zero-min IC50 = 15 ± 1 µM

30 min IC = 1 5 ± 0 1 µM

Zero-min IC50 = 8.1 ± 1.2 µM

30 min IC = 3 3 ± 0 4 µM

ephe

nyto

in 4

´-hyd

ro(P

erce

nt o

f con

trol

20

40

60

80

100

120

ephe

nyto

in 4

´-hyd

ro(P

erce

nt o

f con

tro

20

40

60

80

100

120 30-min IC50 = 1.5 ± 0.1 µM 30-min IC50 = 3.3 ± 0.4 µM

nn

[R-Omeprazole](µM)

0.01 0.1 1 10 100 1000S-M

e

0

[Esomeprazole](µM)

0.01 0.1 1 10 100 1000S-M

e

0

c Omeprazole Sulfide d Omeprazole Sulfoneto

in 4

´-hyd

roxy

latio

nce

nt o

f con

trol)

60

80

100

120

140

160

oin

4´-h

ydro

xyla

tion

ent o

f con

trol)

60

80

100

120

140

160Zero-min IC50 = 9.7 ± 0.5 µM

30-min IC50 = 9.6 ± 1.0 µM


30-min IC50 = 5.6 ± 0.5 µM

[Omeprazole sulfone](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyt

(Per

c

0

20

40

[Omeprazole sulfide](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyto

(Per

c

0

20

40

60

ydro

xyla

tion

ntro

l)

100

120

140

160e 5´-Hydroxyomeprazole

Zero-min IC50 > 100 µM

30-min IC50 > 100 µM

0 01 0 1 1 10 100 1000S-M

ephe

nyto

in 4

´-hy

(Per

cent

of c

o n

0

20

40

60

80

100

[5´-Hydroxyomeprazole](µM)

0.01 0.1 1 10 100 1000S

FIG. 3. Evaluation of esomeprazole, R-omepra-zole, omeprazole sulfide, omeprazole sulfone,and 5�-hydroxyomeprazole as direct-actingand MDIs of CYP2C19. Each symbol repre-sents the average of duplicate determinationsunless otherwise indicated. a, esomeprazoleinhibited CYP2C19 in pooled HLM with IC50

values as shown (S-mephenytoin 4�-hydroxy-lation control rates � 80.9 and 88.1 pmol �mg1 � min1 with and without preincuba-tion, respectively). b, R-omeprazole in-hibited CYP2C19 in pooled HLM withIC50 values as shown (S-mephenytoin 4�-hydroxylation control rates � 106 and 108pmol � mg1 � min1 with and without preincu-bation, respectively). c, omeprazole sulfide in-hibited CYP2C19 in pooled HLM with IC50

values as shown (S-mephenytoin 4�-hydroxyla-tion control rates � 89.8 and 82.2 pmol � mg1 �min1 with and without preincubation, respec-tively). d, omeprazole sulfone inhibitedCYP2C19 in pooled HLM with IC50 val-ues as shown (S-mephenytoin 4�-hy-droxylation control rates � 85.4 and 84.7pmol � mg1 � min1 with and withoutpreincubation, respectively). e, 5�-hydroxyome-prazole weakly inhibited CYP2C19 in HLM(S-mephenytoin 4�-hydroxylation controlrates � 108 and 97 pmol � mg1 � min1 withand without preincubation, respectively).

2026 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


mg twice daily for 14 days) were simulated with Simcyp version 10.2as described under Materials and Methods. Figure 8 shows the activeCYP2C19 levels with time. Depending on the assumption regardingKI (1.7 or 9.1 �M), the level of active enzyme could decrease toapproximately 10 to 60% of the baseline, which would be associatedwith a 1.4- to 10-fold increase in the AUC of compounds mainlymetabolized by CYP2C19. However, the simulated effect of omepra-zole on the S-mephenytoin AUC showed only a 1.45-fold increasewhen the KI value was 9.1 �M and a 5.46-fold increase when the KI

value was 1.7 �M, rather than a 10-fold increase, partly because of thecontribution of other enzymes in its metabolism (e.g., CYP2B6) andpossibly due to the inhibitory contribution of metabolites (e.g.,omeprazole sulfone), not accounted for in the simulation.

Discussion

Given the relatively few drugs that are metabolized extensivelyby CYP2C19, it is not surprising that few DDIs have been attrib-uted to clinically relevant inhibition of CYP2C19. As noted in theIntroduction, there is evidence of clinically relevant inhibition ofCYP2C19 by omeprazole, but direct competitive inhibition ofCYP2C19 is an unlikely cause; if it were, lansoprazole would beexpected to inhibit CYP2C19 more than omeprazole, whereas theconverse is observed clinically. It should be noted that the dose oflansoprazole versus omeprazole (which depends on indication)could partly explain this difference. For the lowest dose indications(i.e., gastroesophageal reflux disease or maintenance of healing oferosive esophagitis), it is 20 mg daily for omeprazole (Prilosec

prescribing information, 2011, http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/019810s092,022056s008lbl.pdf) and15 mg daily for lansoprazole (Prevacid prescribing information,2010, http://www.accessdata.fda.gov/drugsatfda_docs/label/2010/020406s074,021428s021lbl.pdf). For the highest dose indication(i.e., Zollinger-Ellison or other hypersecretory syndromes), thedoses are equal (60 mg daily). On the basis of this comparisonalone, it seems unlikely that the slight difference in the low doseexplains the differences in clinical DDIs. This supposition is borneout by clinical data in the MTDI: negative interactions were reportedbetween lansoprazole (60 mg daily for 10 days) and the CYP2C19substrates diazepam and phenytoin (lansoprazole dose 60 mg dailyfor 9 days with the latter drug). No positive DDIs with CYP2C19substrates have been reported in the MTDI with lansoprazole atany dose. On the other hand, with a low dose of omeprazole (20 mgdaily) in various studies (8 –23 days), the exposure (AUC) of theCYP2C19 substrates diazepam and escitalopram increased from 26to 91%.

To our knowledge, with the exception of our preliminary findings(Paris et al., 2008; Parkinson et al., 2010), in vitro evidence formetabolism-dependent inhibition of CYP2C19 has not been previ-ously described for any PPI. In the current study, omeprazole, esome-prazole, R-omeprazole, and omeprazole sulfone were identified asMDIs of CYP2C19 (IC50 shifts after a 30-min preincubation withNADPH of 4.2, 10, 2.5, and 3.2, respectively), whereas lansoprazoleand pantoprazole were not MDIs (IC50 shifts �1.5). Furthermore, themetabolism-dependent inhibition of CYP2C19 by omeprazole and

oxyl

atio

nol

)

120

140

160

xyla

tion

)

140

160 a Lansoprazole

Cryopreserved human hepatocytesHuman Liver Microsomes

b Lansoprazole

Zero-min IC50 = 1.2 ± 0.1 µM

30 min IC 1 2 ± 0 0 µM

Zero-min IC50 = 1.0 ± 0.1 µM

30 min IC = 1 4 ± 0 1 µM

ephe

nyto

in 4

´-hyd

r o(P

erce

nt o

f con

tro

20

40

60

80

100

120

ephe

nyto

in 4

´-hyd

ro(P

erce

nt o

f con

trol

20

40

60

80

100

120 30-min IC50 = 1.2 ± 0.0 µM 30-min IC50 = 1.4 ± 0.1 µM

[Lansoprazole](µM)

0.01 0.1 1 10 100 1000

S-M 0

[Lansoprazole](µM)

0.01 0.1 1 10 100 1000S-M

e

0

c d

Cryopreserved human hepatocytesHuman Liver Microsomesin

4´-h

ydro

xyla

tion

ent o

f con

trol)

60

80100

120

140160

in 4

´-hyd

roxy

latio

nen

t of c

ontro

l)

60

80

100

120

140

160Pantoprazole Pantoprazole


30-min IC50 = 65 ± 4 µM


30-min IC50 = 23 ± 2 µM

[Pantoprazole](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyto

(Per

ce

0

2040

60

[Pantoprazole](µM)

0.01 0.1 1 10 100 1000S-M

ephe

nyto

(Per

ce

0

20

40

60

FIG. 4. Evaluation of lansoprazole andpantoprazole as direct-acting and MDIsof CYP2C19. Each symbol represents theaverage of duplicate determinations un-less otherwise indicated. a, lansoprazoleinhibited CYP2C19 in HLM with IC50

values as shown (S-mephenytoin 4�-hy-droxylation control rates � 83.8 and 79.6pmol � mg1 � min1 with and withoutpreincubation, respectively). b, lansopra-zole inhibited CYP2C19 in cryopreservedhuman hepatocytes with IC50 values asshown (S-mephenytoin 4�-hydroxylationcontrol rates � 30.4 and 26.6 pmol/mil-lion cells/min with and without preincu-bation, respectively). Each symbol repre-sents the average of triplicate determinations,and error bars represent the S.D. c, pantoprazoleinhibited CYP2C19 in pooled HLM with IC50

values as shown (S-mephenytoin 4�-hydroxyla-tion control rates � 61.7 and 71 pmol � mg1 �min1 with and without preincubation, respec-tively). d, pantoprazole inhibited CYP2C19 incryopreserved human hepatocytes with IC50

values as shown (S-mephenytoin 4�-hydroxyla-tion control rates � 24.5 and 14.2 pmol/millioncells/min with and without preincubation, re-spectively). Each symbol represents the averageof triplicate determinations, and error bars rep-resent the S.D.


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


esomeprazole was not reversed by ultracentrifugation, suggesting thatthe inhibition was irreversible, whereas ultracentrifugation largelyreversed such effects of R-omeprazole. Under various conditions,omeprazole inactivated CYP2C19 with KI values of 1.7 to 9.1 �M andkinact values (maximal rate of inactivation) of 0.041 to 0.046 min1

[corresponding to kinact/KI values ranging from 5.1 to 24 min1 �

mM1 depending on the experimental conditions used (Fig. 5)]. Thevariation in KI values, but not kinact values, is generally consistent withprevious reports investigating the impact of dilution (Van et al., 2006;

Parkinson et al., 2011). We propose that the quasi-irreversible or irre-versible metabolism-dependent inhibition of CYP2C19 by omeprazole(rather than reversible inhibition) explains, at least in part, the observedclinical interactions between omeprazole (and esomeprazole) andCYP2C19 substrates, including clopidogrel, notwithstanding the consid-erable debate surrounding the role of CYP2C19 in the activation of thelatter drug (Bouman et al., 2011; Sibbing et al., 2011).

Why Did Previous In Vitro Studies Miss the Metabolism-Dependent Inhibition of CYP2C19 by Omeprazole? The IC50 or Ki

ba

atio

nvi

ty)

4.5

No dilution, [S] = Km

0 04

ph

enyt

oin

4´-

hyd

roxy

lp

erce

nt

rem

ain

ing

act

i

2 5

3

3.5

4

0.01

0.02

0.03

0.04

= 2 4 ± 0 3 µM

k obs

(min

-1)

dNo dilution, [S] = 10Km

Pre-incubation time(min)

0 5 10 15

S-M

eLn

(p

2

2.5

[Omeprazole] (µM)0 10 20 30

0

KI= 2.4 ± 0.3 µM

kinact = 0.044 ± 0.002 min-1

n 4

´-h

ydro

xyla

tion

emai

nin

g a

ctiv

ity)

4

4.4

4.8

0 µM Omeprazole

0.03

0.04

c

(min

-1)

Pre-incubation time( i )

0 5 10 15

S-M

eph

enyt

oiLn

(p

erce

nt

re

3.2

3.61 µM Omeprazole

3 µM Omeprazole

10 µM Omeprazole

30 µM Omeprazole


0

0.01

0.02

KI = 1.7 ± 0.3 µM

kinact = 0.041 ± 0.003 min-1

k obs

(min)

25-fold dilution, [S] = 10Kmfe

oxyl

atio

nac

tivity

)

4 4

4.8

0.04

-Mep

hen

ytoi

n 4

´-h

ydro

Ln (

per

cen

t re

mai

nin

g

3.6

4

4.4

0 µM Omeprazole

5 µM Omeprazole

10 µM Omeprazole

15 µM Omeprazole

30 µM Omeprazole0.01

0.02

0.03

KI = 9.1 ± 1.7 µM

= 0 046 ± 0 002 min-1

k obs

(min

-1)

Pre-incubation time(min)

0 2 4 6 8 10

S- L 60 µM Omeprazole


0kinact

= 0.046 ± 0.002 min

FIG. 5. Determination of KI and kinact forthe metabolism-dependent inhibition ofCYP2C19 by omeprazole. Individual pointsrepresent the average of triplicate determi-nations �S.D. unless otherwise noted. Forgraphs in a and c, omeprazole was preincu-bated (at concentrations indicated in thelegend of c), and residual CYP2C19 activ-ity was determined as described under Ma-terials and Methods. For the graph ine, omeprazole was preincubated (at concen-trations indicated), and residual CYP2C19activity was determined as described underMaterials and Methods (after a 25-fold di-lution). The graphs in b, d, and f representthe direct plots of the initial rates of inac-tivation of CYP2C19. Values are the slopesof the initial rates of inactivation (kobs) at eachconcentration of omeprazole (shown �S.E.).

2028 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


values for inhibition of CYP2C19 in HLM by omeprazole reported inthe literature range from 150 to 1 �M in nearly 40 studies (MTDIdatabase). The lowest values (1 and 4 �M) were reported with someof the longest substrate incubation periods (60–120 min; presumablyused because of the low turnover of S-mephenytoin), and few (if any)studies specifically included a preincubation with NADPH-fortifiedHLM to examine the possibility of metabolism-dependent inhibitionof CYP2C19. On the basis of our data (Fig. 5), 60 to 120 min wouldprovide ample time for inactivation of CYP2C19 by omeprazoleduring the substrate incubation, therefore leading to “unintentional”metabolism-dependent inhibition and artificially low IC50 or Ki val-ues, as recently discussed (Parkinson et al., 2011). Our data also showthat the IC50 shift diminishes as higher protein concentrations andlonger substrate incubation times are used (Fig. 2b), probably becauseof the combination of inhibitor depletion (Fig. 2e) and the uninten-tional metabolism-dependent inhibition during the long substrate in-cubation (Parkinson et al., 2011).

Would We Predict Clinically Relevant CYP2C19 Inhibition?As noted in the Introduction, we would not predict clinically signif-

icant inhibition of CYP2C19 by omeprazole on the basis of compet-itive inhibition alone. A thorough prediction of the effect of omepra-zole on CYP2C19 substrates on the basis of the experimentallydetermined KI and kinact values necessitated the use of physiologicallybased pharmacokinetic modeling to allow for dynamic simulation ofchanges to omeprazole (including self-inhibition), substrate concen-trations, and enzyme turnover (i.e., a so-called MDM), with the MSM(eq. 1) used as a comparator.

In the MDM under conditions in which in vitro inhibitor depletionand microsomal protein binding of omeprazole were minimal (Fig. 2,e and f; i.e., KI � 1.7 �M, Fig. 5d), the level of active CYP2C19 ispredicted to decrease to �10% of baseline after approximately 7 daysof omeprazole administration (Fig. 8a), which could cause up to a10-fold increase in the AUC of compounds predominantly metabo-lized by CYP2C19. When the higher estimate of KI is used in theMDM (obtained with a 25-fold dilution under conditions in whichsignificant inhibitor depletion occurs; Fig. 2e), the level of activeCYP2C19 is predicted to decrease to �60% of baseline after approx-imately 12 days of omeprazole administration (Fig. 8b). Although

es

Control CYP2C19 activity in microsomes from cultured human hepatocytes

ba

Inhibition InductionEffect of omeprazole treatment (3 days)

YP

2C19

Rat

en

Con

trol C

Yrd

er B

ased

oR

ank

Or

10 150 525 20S-mephenytoin 4´-hydroxyation

(pmol/min/mg protein)

75 50 25 0 100 200 300 400100S-mephenytoin 4´-hydroxyation

(Percent of control activity)(p g p ) ( y)

FIG. 6. CYP2C19 activity (S-mephenytoin4�-hydroxylation) in microsomes isolatedfrom fresh-plated hepatocytes treated withDMSO (control) or omeprazole. Primarycultures of human hepatocytes weretreated for 3 consecutive days withDMSO or omeprazole (100 �M), and mi-crosomes were prepared from the hepato-cytes 24 h after the last treatment.CYP2C19 activity was determined as de-scribed under Materials and Methods. Ac-tivities were sorted in rank order from high-est to lowest control rates (a), and thecorresponding omeprazole-treated samplesare displayed in b in terms of percentage ofcontrol activity.

140

120

ion

Post-spin

P t i K [F (CN) ]

Pre-spin

80

100

´-hyd

roxy

lat

g/m

in)

Post-spin + K3[Fe(CN)6]

60

phen

ytoi

n 4´

(pm

ol/m

g

20

40

S-M

ep

0Esomeprazole

(100 µM)Omeprazole

(100 µM)R-Omeprazole

(100 µM)Solvent

(1% Methanol)

FIG. 7. Reversibility assessment of the metabolism-dependent in-hibition of CYP2C19 by omeprazole and its enantiomers with theultracentrifugation method. The potential reversibility of the metab-olism-dependent inhibition of CYP2C19 NADPH-fortified HLM(0.1 mg/ml) by omeprazole and its individual enantiomers (100�M) was evaluated with the ultracentrifugation method as describedunder Materials and Methods. Incubations labeled “prespin” wereconducted similarly to analogous samples in the IC50 determinations(conduced in triplicate and displayed as the average rates of S-mephenytoin 4�-hydroxylation � S.D.). For incubations labeledpostspin (conducted in triplicate and analyzed in triplicate), micro-somal protein was isolated by ultracentrifugation after 30-min in-cubations with the inhibitor and NADPH and analyzed in triplicatefor residual CYP2C19 activity (S-mephenytoin 4�-hydroxylation)displayed as the average rates � S.E. as described under Materialsand Methods. Half of the incubations from the postspin samplesincluded potassium ferricyanide (2 mM, final concentration) asindicated. Samples treated with methanol (1% v/v, final) served ascontrols. All rates were normalized to final microsomal proteinconcentrations as described under Materials and Methods.


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


S-mephenytoin is a model CYP2C19 probe substrate, only a 1.45- to5.46-fold increase in its AUC (depending on KI) is predicted by theMDM after 14 days of omeprazole administration (40 mg twice daily),partly because of the contribution of other enzymes to its metabolism(e.g., CYP2B6) and possibly because of less-than-complete inac-tivation of CYP2C19, and/or possibly due to the inhibitory contribu-tion of metabolites (e.g., omeprazole sulfone), not accounted for in thesimulation.

In the MSM, when the lowest value of KI is used, omeprazole ispredicted to cause a 2.85-fold increase in the AUC of a drug such asmoclobemide, which has an fmCYP2C19 of 0.72 (Yu et al., 2001). Thepredicted AUC increase with moclobemide falls to 1.96-fold with theKI value of 9.1 �M. If the fmCYP2C19 were equal to 1 and the lowestKI value was used, then omeprazole is predicted to cause up to a10.4-fold increase in the AUC of such a hypothetical “perfect”CYP2C19 substrate. However, because there is up to a 14.6-foldincrease in the AUC of omeprazole in CYP2C19 PMs relative to EMs(Furuta et al., 1999), this prediction suggests that omeprazole does notcompletely inactivate CYP2C19, which is consistent with either sce-nario in the MDM after several days of administration of omeprazole(Fig. 8).

The clinical example of moclobemide (fmCYP2C19 � 0.72) is gen-erally in agreement with these predictions, with AUC� increasesaveraging 2.21-fold (1.03- to 3.39-fold) in CYP2C19 EMs withomeprazole coadministration (Yu et al., 2001). The AUC increasespredicted with the lowest KI in the MSM value fall within thisobserved range. In addition, to achieve the recently reported 1.39-foldincrease in the AUC of clopidogrel with concomitant dosing ofomeprazole (Angiolillo et al., 2011), the fmCYP2C19 for clopidogrelwould need only to be approximately 0.31 according to the MSM.

Given the large difference in the predicted active CYP2C19 afterseveral days of omeprazole administration in the MDM depending onwhich KI value is used (Fig. 8), the importance of determining the KI

value under conditions at which microsomal protein binding andinhibitor depletion are minimized is underscored, similar to the casewith IC50 values (Parkinson et al., 2011).

Is There Clinical Evidence of CYP2C19 Inhibition by Omepra-zole? In vivo evidence for the metabolism-dependent inhibition ofCYP2C19 by omeprazole has already been reported. For instance,Klotz, 2006 reported that the healing rates of gastroesophageal refluxdisease after 4 weeks of therapy with esomeprazole were not depen-dent on CYP2C19 status, as they are for lansoprazole. On the basis ofthe metabolite ratios of 5�-hydroxyomeprazole (formed by CYP2C19)

to omeprazole sulfone (formed by CYP3A4), it was concluded thatCYP3A4 plays the major role in the metabolism of esomeprazole aftermultiple dosing, consistent with autoinhibition of CYP2C19 and con-version of CYP2C19 EMs to intermediate metabolizers or PMs afterrepeat dosing (Klotz, 2006). The changes in metabolite ratio suggestthat the effect of multiple dosing on omeprazole reflects the autoin-hibition of CYP2C19, rather than increased stability due to highergastric pH as originally suspected, especially because such time-dependent changes do not occur with other PPIs. On the other hand,the prescribing information for Nexium states “at repeated once-dailydosing with 40 mg, the systemic bioavailability is approximately 90%compared to 64% after a single dose of 40 mg” (Nexium prescribinginformation, 2010, http://www.accessdata.fda.gov/drugsatfda_docs/label/2010/021153s036,021957s009,022101s006lbl.pdf). In addition,Andersson and Weidolf (2008) reported that when 15 mg of omepra-zole (i.e., the racemate), esomeprazole, or R-omeprazole was admin-istered orally for 7 days, exposure to esomeprazole (plasma AUC)increased by approximately 2-fold over 7 days, whereas exposure toomeprazole increased by only 52%, and exposure to R-omeprazoleactually decreased by 9%. These results are consistent with our invitro results showing that the inhibition of CYP2C19 by R-omepra-zole appears to be largely reversible, whereas that of the racemate andesomeprazole are largely irreversible (Fig. 7).

In vivo evidence for metabolism-dependent inhibition of CYP2C19by omeprazole also comes from clinical DDI studies with omeprazoleas the perpetrator. For instance, in CYP2C19-extensive metabolizers(but not in poor metabolizers), the AUC of moclobemide increased byapproximately 31% after a single 40-mg dose of omeprazole, but itincreased by 121% after 8 days of dosing with 40 mg of omeprazole(Yu et al., 2001). Such an apparent increase in the exposure of avictim drug with repeated dosing of the perpetrator drug is oftenapparent with MDIs. In addition, omeprazole (but not lansoprazole orpantoprazole) has long been known to inhibit the metabolism ofdiazepam in vivo, and this inhibition occurs in CYP2C19 EMs but notPMs, further suggesting that the mechanism involves CY2C19 inhi-bition by omeprazole (Andersson et al., 1990; Lefebvre et al., 1992;Gugler et al., 1996).

Does Metabolism-Dependent Inhibition of CYP2C19 by OmeprazoleExplain the PPI-Clopidogrel Interaction? Our data and the predictionsdetailed above may explain, at least in part, the interaction betweenomeprazole (or esomeprazole) and clopidogrel. As noted in the Intro-duction, the FDA specifically warns against coadministration of clopi-dogrel and omeprazole (2010, http://www.fda.gov/Drugs/DrugSafety/

100

120

100

120 a KI bMµ 7.1 = KI = 9.1 µM

60

80

100

60

80

100

zym

e (%

)

40

60

40

60

Ac�

ve E

nz

0

20

0 2.5 5 7.5 10 12.5

0

20

0 2.5 5 7.5 10 12.5Time(days)

Time(days)

FIG. 8. Simulation of time-dependent changesin active CYP2C19. Active levels of hepaticCYP2C19 in the presence of omeprazole (40mg twice daily for 14 days) were simulated withthe Simcyp Simulator version 10.2, as describedunder Materials and Methods. a is based on aKI value of 1.7 �M; b is based on a KI

value of 9.1 �M.

2030 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


ucm231161.htm). Given that the in vivo half-life of omeprazole (andother PPIs) is short, and plasma protein binding is high, it is remarkablethat many recent publications attribute the clopidogrel-omeprazole inter-action to competitive (reversible) inhibition of CYP2C19 by omeprazole,with some suggestion that separation of dosing can prevent the interaction(Abraham et al., 2010; Tran et al., 2010; Bates et al., 2011). However, itshould be noted that for clopidogrel, genetic differences in the metabo-lism or transport of the drug and in the therapeutic target (the P2Y12

receptor on platelets), as well as environmental factors (e.g., diet, disease,coadministered drugs), have been implicated in the variation in its clinicaleffect (for recent reviews see Abraham et al., 2010; Society for Cardio-vascular Angiography and Interventions et al., 2010; Tran et al., 2010;Bates et al., 2011).

In addition, the interaction between omeprazole (or esomeprazole)and clopidogrel is particularly complex, as noted by Zhang et al.(2009). These authors note that clopidogrel itself is an MDI ofCYP2C19, increasing the ratio of 5�-hydroxyomeprazole to omepra-zole by �75% in CYP2C19 EMs (Chen et al., 2009; Zhang et al.,2009), and that clopidogrel and its 2-oxo metabolite (the precursor tothe active metabolite) also directly inhibit CYP2C19 with IC50 values�1 �M. The authors suggested that the “stronger” effect of omepra-zole on CYP2C19 may be due to the “time-dependent” inhibition wereported in our preliminary work (Paris et al., 2008; Zhang et al.,2009). Our finding of irreversible or quasi-irreversible inactivation ofCYP2C19 by omeprazole, with up to a predicted 90% decrement inactive CYP2C19 after approximately 7 days of dosing (Fig. 8a), isconsistent with this hypothesis. The fact that the FDA warning appliesonly to omeprazole (http://www.fda.gov/Drugs/DrugSafety/ucm231161.htm) and not the other PPIs is consistent with a lack ofmetabolism-dependent inhibition by the other PPIs examined in thisstudy.

Given the minor inhibitory effects of omeprazole on other P450enzymes (see supplemental data), the metabolism-dependent inhibi-tion of CYP2C19 by omeprazole and the reports of clinical interac-tions are consistent with a significant role for CYP2C19 in themetabolism of clopidogrel. In addition, as described in the Introduc-tion, direct inhibition of CYP2C19 by other PPIs is not likely to be thecause of a clinically significant interaction with clopidogrel, which isconsistent with the lack of in vitro metabolism-dependent inhibition ofCYP2C19 by lansoprazole and pantoprazole reported here and a lackof clinically significant pharmacokinetic interactions between eitherlansoprazole or pantoprazole and clopidogrel (MTDI database).

Potential Mechanisms of Inactivation of CYP2C19 by Omepra-zole. Follow-up studies will be needed to further elucidate the mech-

anism of inactivation of CYP2C19 by omeprazole (or esomeprazole).However, a few possibilities for the mechanism of CYP2C19 inacti-vation by omeprazole (or esomeprazole) are suggested by the reportedmetabolism of the individual enantiomers of omeprazole (Fig. 9,adapted from Abelo et al., 2000 and Andersson and Weidolf, 2008).Methylhydroxylation of omeprazole to form 5�-hydroxyomeprazolecould involve the intermediacy of a benzylic radical and heme alky-lation, analogous to the inactivation of CYP2C8 by gemfibrozil gluc-uronide (Ogilvie et al., 2006; Baer et al., 2009). The formation of5-O-desmethylomeprazole (a para-aminophenol) could lead to a re-active quinoneimine that could inactivate CYP2C19 if formed in itsactive site. The latter possibility may be unlikely given that the5-hydroxylation of lansoprazole also leads to para-aminophenol for-mation (Fig. 10; admittedly a tautomer of the analogous 5-O-des-methylomeprazole para-aminophenol), and yet lansoprazole is not anMDI of CYP2C19 (Fig. 4a).

However, the data presented in Fig. 7 suggest that esomeprazolehas a greater metabolism-dependent inhibition effect on CYP2C19than does R-omeprazole. These data are in fact consistent with clinicalobservations in which the plasma Cmax of esomeprazole (40 mg dailyof an oral solution) increases from 3.07 to 4.86 �M from day 1 to day5, and that of omeprazole (40 mg daily of an oral solution) increasesfrom 2.32 to 3.87 �M, whereas that of R-omeprazole (40 mg daily ofan oral solution) only increases from 1.62 to 1.98 �M (Hassan-Alin etal., 2005). In addition, published in vitro results in HLM suggest thatthe CLint for esomeprazole sulfoxidation (catalyzed by CYP3A4) is4.6-fold greater than that for R-omeprazole (Abelo et al., 2000),although the total CLint for R-omeprazole is approximately 3-foldhigher than for esomeprazole. This finding is even more apparent inthe clinical data, which show an approximately 14-fold higher AUCfor the sulfone when esomeprazole is administered for 1 day (20 or 40mg) than for R-omeprazole; this ratio increases to nearly 40-fold after5 days of dosing (Hassan-Alin et al., 2005) because CYP3A4 plays amore important role in esomeprazole metabolism after multiple dos-ing (Klotz, 2006). Because we found that omeprazole sulfone is alsoan MDI of CYP2C19, it is possible that the combination of effectsfrom esomeprazole and its sulfone explain the much greater inactiva-tion of CYP2C19 by esomeprazole than R-omeprazole, and this pos-sibility needs to be followed up.

At the same time, the CLint for the 5�-hydroxylation of R-omeprazole is approximately 10-fold higher than for esomeprazole,which, along with our results (Fig. 7), suggest that unless there is alsoa difference in the ultimate fate of R- versus S-5�-hydroxyomeprazole(e.g., partition ratio, benzylic radical formation, and oxygen rebound

94% 27%N O

N

O

S

OHO

HNN

5´-Hydroxyomeprazole

O

R-OmeprazoleCLint = 42.5 µL/min/mg

4%2C19

3A4EsomeprazoleCLint = 14.6 µL/min/mg

46% 2C19

3A4N

HN

OH

NS

O

5-O-Desmethylomeprazole

2% 27%N O

O

O

HNNS

OOmeprazole sulfone

FIG. 9. Metabolic scheme for omeprazoleenantiomers. The conversion of each enan-tiomer of omeprazole to the major metabo-lites and enzymes responsible for each isshown. The scheme is adapted from thatpublished by Andersson and Weidolf, 2008,which is in turn based on in vitro datapublished by Abelo et al., 2000, which isthe source of the CLint values.


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


or other inactivating pathways), this pathway may not explain theformation of a reactive metabolite that inactivates CYP2C19. Enan-tiomer-enantiomer interactions at the active site of CYP2C19 whenthe racemate is used (as in some experiments in this study) couldcomplicate the interpretation of results with the single enantiomers aspreviously described (Li et al., 2005), especially considering that thepresence of R-omeprazole acts as an alternative substrate and offerssubstrate-protection for the irreversible inactivation by esomeprazoleor its sulfone.

In conclusion, we have shown that omeprazole (but not pantoprazoleor lansoprazole) is an MDI of CYP2C19 in HLM, cryopreserved humanhepatocytes, and recombinant human CYP2C19. On the basis of the KI

and kinact values for the metabolism-dependent inhibition of CYP2C19 byomeprazole in HLM, we predict that this inactivation is clinically signif-icant. Furthermore, we provided evidence that esomeprazole is morelikely to irreversibly inactivate CYP2C19 than is R-omeprazole. Thesefindings have implications for the ongoing debate surrounding the inter-action between clopidogrel (as well as other CYP2C19 substrates) andomeprazole and, in particular, esomeprazole.

Acknowledgments

We thank the Hepatocellular Products and Services Department of Xeno-Tech, LLC (Lenexa, KS) for providing historical data on CYP2C19 activityand the effect of 3-day omeprazole treatment on hepatocytes.

Authorship Contributions

Participated in research design: Parkinson, Kazmi, Buckley, Paris, andOgilvie.

Conducted experiments: Yerino, Ogilvie, and Kazmi.Contributed new reagents or analytic tools: Toren and Rostami-Hodjegan.Performed data analysis: Yerino, Ogilvie, Kazmi, Toren, Buckley, and

Rostami-Hodjegan.Wrote or contributed to the writing of the manuscript: Ogilvie, Parkinson,

Kazmi, Yerino, Rostami-Hodjegan, Toren, and Buckley.

References

Abelo A, Andersson TB, Antonsson M, Naudot AK, Skånberg I, and Weidolf L (2000)Stereoselective metabolism of omeprazole by human cytochrome P450 enzymes. Drug MetabDispos 28:966–972.

Abraham NS, Hlatky MA, Antman EM, Bhatt DL, Bjorkman DJ, Clark CB, Furberg CD,Johnson DA, Kahi CJ, Laine L, et al. (2010) ACCF/ACG/AHA 2010 expert consensusdocument on the concomitant use of proton pump inhibitors and thienopyridines: a focusedupdate of the ACCF/ACG/AHA 2008 expert consensus document on reducing the gastroin-testinal risks of antiplatelet therapy and NSAID use. A Report of the American College ofCardiology Foundation Task Force on Expert Consensus Documents. J Am Coll Cardiol56:2051–2066.

Andersson T, Cederberg C, Edvardsson G, Heggelund A, and Lundborg P (1990) Effect ofomeprazole treatment on diazepam plasma levels in slow versus normal rapid metabolizers ofomeprazole. Clin Pharmacol Ther 47:79–85.

Andersson T and Weidolf L (2008) Stereoselective disposition of proton pump inhibitors. ClinDrug Investig 28:263–279.

Angiolillo DJ, Gibson CM, Cheng S, Ollier C, Nicolas O, Bergougnan L, Perrin L, LaCreta FP,Hurbin F, and Dubar M (2011) Differential effects of omeprazole and pantoprazole on thepharmacodynamics and pharmacokinetics of clopidogrel in healthy subjects: randomized,placebo-controlled, crossover comparison studies. Clin Pharmacol Ther 89:65–74.

Baer BR, DeLisle RK, and Allen A (2009) Benzylic oxidation of gemfibrozil-1-O-beta-glucuronide by P450 2C8 leads to heme alkylation and irreversible inhibition. Chem ResToxicol 22:1298–1309.

Bates ER, Lau WC, and Angiolillo DJ (2011) Clopidogrel-drug interactions. J Am Coll Cardiol57:1251–1263.

Bouman HJ, Schomig E, van Werkum JW, Velder J, Hackeng CM, Hirschhauser C, WaldmannC, Schmalz HG, ten Berg JM, and Taubert D (2011) Paraoxonase-1 is a major determinant ofclopidogrel efficacy. Nat Med 17:110–116.

Chen BL, Chen Y, Tu JH, Li YL, Zhang W, Li Q, Fan L, Tan ZR, Hu DL, Wang D, et al. (2009)Clopidogrel inhibits CYP2C19-dependent hydroxylation of omeprazole related to CYP2C19genetic polymorphisms. J Clin Pharmacol 49:574–581.

Clarke TA and Waskell LA (2003) The metabolism of clopidogrel is catalyzed by humancytochrome P450 3A and is inhibited by atorvastatin. Drug Metab Dispos 31:53–59.

Franklin MR (1977) Inhibition of mixed-function oxidations by substrates forming reducedcytochrome P-450 metabolic-intermediate complexes. Pharmacol & Thera A Chemother,Toxicol Metabol Inhib 2:227–245.

Furuta T, Ohashi K, Kobayashi K, Iida I, Yoshida H, Shirai N, Takashima M, Kosuge K, HanaiH, Chiba K, et al. (1999) Effects of clarithromycin on the metabolism of omeprazole in relationto CYP2C19 genotype status in humans. Clin Pharmacol Ther 66:265–274.

Grimm SW, Einolf HJ, Hall SD, He K, Lim HK, Ling KH, Lu C, Nomeir AA, Seibert E, SkordosKW, et al. (2009) The conduct of in vitro studies to address time-dependent inhibition ofdrug-metabolizing enzymes: a perspective of the pharmaceutical research and manufacturersof America. Drug Metab Dispos 37:1355–1370.

Gugler R, Hartmann M, Rudi J, Brod I, Huber R, Steinijans VW, Bliesath H, Wurst W, and KlotzU (1996) Lack of pharmacokinetic interaction of pantoprazole with diazepam in man. Br J ClinPharmacol 42:249–252.

Hassan-Alin M, Andersson T, Niazi M, and Rohss K (2005) A pharmacokinetic study comparingsingle and repeated oral doses of 20 mg and 40 mg omeprazole and its two optical isomers,S-omeprazole (esomeprazole) and R-omeprazole, in healthy subjects. Eur J Clin Pharmacol60:779–784.

Ho PM, Maddox TM, Wang L, Fihn SD, Jesse RL, Peterson ED, and Rumsfeld JS (2009) Riskof adverse outcomes associated with concomitant use of clopidogrel and proton pumpinhibitors following acute coronary syndrome. JAMA 301:937–944.

Kazui M, Nishiya Y, Ishizuka T, Hagihara K, Farid NA, Okazaki O, Ikeda T, and Kurihara A(2010) Identification of the human cytochrome P450 enzymes involved in the two oxidativesteps in the bioactivation of clopidogrel to its pharmacologically active metabolite. DrugMetab Dispos 38:92–99.

Klotz U (2006) Clinical impact of CYP2C19 polymorphism on the action of proton pumpinhibitors: a review of a special problem. Int J Clin Pharmacol Ther 44:297–302.

Lefebvre RA, Flouvat B, Karolac-Tamisier S, Moerman E, and Van Ganse E (1992) Influenceof lansoprazole treatment on diazepam plasma concentrations. Clin Pharmacol Ther 52:458–463.

Li XQ, Andersson TB, Ahlstrom M, and Weidolf L (2004) Comparison of inhibitory effects ofthe proton pump-inhibiting drugs omeprazole, esomeprazole, lansoprazole, pantoprazole, andrabeprazole on human cytochrome P450 activities. Drug Metab Dispos 32:821–827.

Li XQ, Weidolf L, Simonsson R, and Andersson TB (2005) Enantiomer/enantiomer interactionsbetween the S- and R-isomers of omeprazole in human cytochrome P450 enzymes: major roleof CYP2C19 and CYP3A4. J Pharmacol Exp Ther 315:777–787.

Nassar AE, King I, Paris BL, Haupt L, Ndikum-Moffor F, Campbell R, Usuki E, Skibbe J, BrobstD, Ogilvie BW, et al. (2009) An in vitro evaluation of the victim and perpetrator potential ofthe anticancer agent laromustine (VNP40101M), based on reaction phenotyping and inhibitionand induction of cytochrome P450 enzymes. Drug Metab Dispos 37:1922–1930.

Nishiya Y, Hagihara K, Kurihara A, Okudaira N, Farid NA, Okazaki O, and Ikeda T (2009)Comparison of mechanism-based inhibition of human cytochrome P450 2C19 by ticlopidine,clopidogrel, and prasugrel. Xenobiotica 39:836–843.

Obach RS, Walsky RL, and Venkatakrishnan K (2007) Mechanism-based inactivation of humancytochrome P450 enzymes and the prediction of drug-drug interactions. Drug Metab Dispos35:246–255.

Ogilvie BW, Zhang D, Li W, Rodrigues AD, Gipson AE, Holsapple J, Toren P, and ParkinsonA (2006) Glucuronidation converts gemfibrozil to a potent, metabolism-dependent inhibitor ofCYP2C8: implications for drug-drug interactions. Drug Metab Dispos 34:191–197.

Paris BL, Yerino P, Ogilvie BW, and Parkinson A (2008) The proton pump inhibitors (PPIs)omeprazole and rabeprazole, but not lansoprazole, are in vitro time-dependent inhibitors ofCYP2C19. Drug Metab Rev 40 (Suppl 1):89–90.

Parkinson A, Kazmi F, Buckley DB, Yerino P, Ogilvie BW, and Paris BL (2010) System-dependent outcomes during the evaluation of drug candidates as inhibitors of cytochrome P450(CYP) and uridine diphosphate glucuronosyltransferase (UGT) enzymes: human hepatocytesversus liver microsomes versus recombinant enzymes. Drug Metab Pharmacokinet 25:16–27.

Parkinson A, Kazmi F, Buckley DB, Yerino P, Paris BL, Holsapple J, Toren P, Otradovec SM,and Ogilvie BW (2011) An evaluation of the dilution method for identifying metabolism-dependent inhibitors (MDIs) of cytochrome P450 (CYP) enzymes. Drug Metab Dispos39:1370–1387.

Renwick AB, Watts PS, Edwards RJ, Barton PT, Guyonnet I, Price RJ, Tredger JM, PelkonenO, Boobis AR, and Lake BG (2000) Differential maintenance of cytochrome P450 enzymes incultured precision-cut human liver slices. Drug Metab Dispos 28:1202–1209.

Rowland Yeo K, Jamei M, Yang J, Tucker GT, and Rostami-Hodjegan A (2010) Physiologicallybased mechanistic modelling to predict complex drug-drug interactions involving simultane-ous competitive and time-dependent enzyme inhibition by parent compound and its metabolitein both liver and gut—the effect of diltiazem on the time-course of exposure to triazolam. EurJ Pharm Sci 39:298–309.

Savi P, Combalbert J, Gaich C, Rouchon MC, Maffrand JP, Berger Y, and Herbert JM (1994)The antiaggregating activity of clopidogrel is due to a metabolic activation by the hepaticcytochrome P450–1A. Thromb Haemost 72:313–317.

Sibbing D, Koch W, Gebhard D, Schuster T, Braun S, Stegherr J, Morath T, Schomig A, vonBeckerath N, and Kastrati A (2010) Cytochrome 2C19*17 allelic variant, platelet aggregation,bleeding events, and stent thrombosis in clopidogrel-treated patients with coronary stentplacement. Circulation 121:512–518.

Sibbing D, Koch W, Massberg S, Byrne RA, Mehilli J, Schulz S, Mayer K, Bernlochner I,Schomig A, and Kastrati A (2011) No association of paraoxonase-1 Q192R genotypes withplatelet response to clopidogrel and risk of stent thrombosis after coronary stenting. EurHeart J 32:1605–1613.

Society for Cardiovascular Angiography and Interventions, Society of Thoracic Surgeons,Writing Committee Members, Holmes DR Jr, Dehmer GJ, Kaul S, Leifer D, O’Gara PT, and

N OH

O

O N

O

O

CF3

O-Desmethylomeprazole

N

HNNS

5-Hydroxylansoprazole

HNNS

OH

FIG. 10. Para-aminophenol formation from omeprazole and lansoprazole. Themajor metabolites of omeprazole (O-desmethylomeprazole) and lansoprazole (5-hydroxylansoprazole) are shown. The highlighted area indicates the part of themolecules that are potentially reactive para-aminophenols.

2032 OGILVIE ET AL.

at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from


Stein CM (2010) ACCF/AHA Clopidogrel clinical alert: approaches to the FDA “boxedwarning”: a report of the American College of Cardiology Foundation Task Force on ClinicalExpert Consensus Documents and the American Heart Association. Circulation 122:537–557.

Tran M, Tafreshi J, and Pai RG (2010) Review article: combination of clopidogrel and protonpump inhibitors: implications for clinicians. J Cardiovasc Pharmacol Ther 15:326–337.

Van LM, Heydari A, Yang J, Hargreaves J, Rowland-Yeo K, Lennard MS, Tucker GT, andRostami-Hodjegan A (2006) The impact of experimental design on assessing mechanism-based inactivation of CYP2D6 by MDMA (Ecstasy). J Psychopharmacol 20:834–841.

Wallace JL and Sharkey KA (2011) Pharmacotherapy of gastric acidity, peptic ulcers, andgastroesophageal reflux disease, in Goodman and Gilman’s The Pharmacological Basis ofTherapeutics (Brunton L, Chabner B, and Knollman B eds), pp 1001–1019, McGraw-Hill,New York.

Walsky RL and Obach RS (2004) Validated assays for human cytochrome P450 activities. DrugMetab Dispos 32:647–660.

Yang J, Liao M, Shou M, Jamei M, Yeo KR, Tucker GT, and Rostami-Hodjegan A (2008)Cytochrome p450 turnover: regulation of synthesis and degradation, methods for determiningrates, and implications for the prediction of drug interactions. Curr Drug Metab 9:384–394.

Yu KS, Yim DS, Cho JY, Park SS, Park JY, Lee KH, Jang IJ, Yi SY, Bae KS, and Shin SG(2001) Effect of omeprazole on the pharmacokinetics of moclobemide according to the geneticpolymorphism of CYP2C19. Clin Pharmacol Ther 69:266–273.

Yueh MF, Kawahara M, and Raucy J (2005) High volume bioassays to assess CYP3A4-mediateddrug interactions: induction and inhibition in a single cell line. Drug Metab Dispos 33:38–48.

Zahno A, Brecht K, Bodmer M, Bur D, Tsakiris DA, and Krahenbuhl S (2010) Effects of druginteractions on biotransformation and antiplatelet effect of clopidogrel in vitro. Br J Pharma-col 161:393–404.

Zhang H, Ragueneau-Majlessi I, and Levy RH (2009) Interaction between clopidogrel and protonpump inhibitors: hypothesis to explain multifactorial CYP2C19 inhibition. Drug Metab Lett3:287–289.

Address correspondence to: Dr. Andrew Parkinson, XenoTech, LLC, 16825West 116th Street, Lenexa, KS 66219. E-mail: [email protected]


at ASPE

T Journals on M

ay 21, 2018dm


ownloaded from
