Abstract
Clopidogrel pharmacotherapy is associated with substantial interindividual variability in clinical response, which can translate into an increased risk of adverse outcomes. Clopidogrel, a recognized substrate of hepatic carboxylesterase 1 (CES1), undergoes extensive hydrolytic metabolism in the liver. Significant interindividual variability in the expression and activity of CES1 exists, which is attributed to both genetic and environmental factors. We determined whether CES1 inhibition and CES1 genetic polymorphisms would significantly influence the biotransformation of clopidogrel and alter the formation of the active metabolite. Coincubation of clopidogrel with the CES1 inhibitor bis(4-nitrophenyl) phosphate in human liver s9 fractions significantly increased the concentrations of clopidogrel, 2-oxo-clopidogrel, and clopidogrel active metabolite, while the concentrations of all formed carboxylate metabolites were significantly decreased. As anticipated, clopidogrel and 2-oxo-clopidogrel were efficiently hydrolyzed by the cell s9 fractions prepared from wild-type CES1 transfected cells. The enzymatic activity of the CES1 variants G143E and D260fs were completely impaired in terms of catalyzing the hydrolysis of clopidogrel and 2-oxo-clopidogrel. However, the natural variants G18V, S82L, and A269S failed to produce any significant effect on CES1-mediated hydrolysis of clopidogrel or 2-oxo-clopidogrel. In summary, deficient CES1 catalytic activity resulting from CES1 inhibition or CES1 genetic variation may be associated with higher plasma concentrations of clopidogrel-active metabolite, and hence may enhance antiplatelet activity. Additionally, CES1 genetic variants have the potential to serve as a biomarker to predict clopidogrel response and individualize clopidogrel dosing regimens in clinical practice.
Introduction
Clopidogrel is a second-generation thienopyridine derivative used extensively as an orally administered antiplatelet agent. The use of clopidogrel has become a standard treatment for patients with acute coronary syndromes and those undergoing percutaneous coronary interventions. Nevertheless, it has been consistently documented that a significant percentage (5–40%) of individuals treated with clopidogrel do not receive the anticipated therapeutic benefit, which in turn has been associated with an increased risk of adverse outcomes (Angiolillo et al., 2007; Karaźniewicz-Łada et al., 2012).
Clopidogrel is a prodrug that undergoes a complex metabolism. It is initially absorbed in its inactive prodrug (parent) form; after a multistep biotransformation sequence in the liver, it is ultimately metabolized to its active 5-thiol metabolite (Fig. 1). However, the majority of administered clopidogrel never enters this bioactivation cascade because ~85% of the absorbed prodrug is rapidly hydrolyzed to the inactive metabolite, clopidogrel carboxylic acid (Hagihara et al., 2009). This reaction is catalyzed by hepatic carboxylesterase 1 (CES1) (Tang et al., 2006). Accordingly, only ~15% of a clopidogrel dose is available to undergo further oxidative metabolism catalyzed by CYP1A2, 2B6, and 2C19, resulting in the formation of the thiolactone derivative 2-oxo-clopidogrel. A portion 2-oxo-clopidogrel is then hydrolyzed by CES1 to form 2-oxo-clopidogrel carboxylate, an inactive metabolite, while the balance of 2-oxo-clopidogrel is further hydrolyzed to the unstable but active 5-thiol metabolite. This final activation step is mediated by CYP2B6, CYP2C9, CYP2C19, and CYP3A4 (Kazui et al., 2010). The 5-thiol clopidogrel active metabolite (clopidogrel-AM) is a labile bioreactive compound that forms a disulfide bridge that binds irreversibly to P2Y12 receptors located on the platelet membrane, causing irreversible blockade. Finally, CES1 again plays a role in the further hydrolysis of the clopidogrel-AM, ultimately forming the 5-thiol carboxylic acid metabolite (Bouman et al., 2011).
With regard to formation of the clopidogrel-AM and its dependence on cytochrome P450 (P450) enzymes, numerous studies have focused on CYP2C19 and its functional variants, such as the loss of or reduced function alleles *2, *3, *4,*5, *6, *7, and *8 and the gain-of-function allele *17. Of these, CYP2C19*2 may have particular significance with regard to its influence on clopidogrel metabolism and the ensuing pharmacokinetic and pharmacodynamic response (Mega et al., 2009; Shuldiner et al., 2009). However, CYP2C19*2 explains only 5–12% of the variability in clopidogrel response (Karaźniewicz-Łada et al., 2012), and the majority of therapeutic variability in clopidogrel treatment remains unknown.
In humans, CES1 is the most predominant hydrolytic enzyme, catalyzing the hydrolysis of numerous ester- and amide-containing endogenous compounds, toxins, and medications to their respective free acids (Imai et al., 2006; Ross and Crow, 2007). CES1 contributes to 80–95% of total hydrolytic activity in the human liver. Significant interindividual variability in the expression and activity of CES1 has been consistently observed and reported in the biomedical literature. This variability is likely to be the result of both genetic and environmental factors (Hosokawa et al., 1995; Fukami et al., 2008; Yoshimura et al., 2008; Yang et al., 2009; Zhu et al., 2009a; Shi et al., 2011; Ross et al., 2012). We have reported that the CES1 single-nucleotide polymorphisms (SNPs) G143E (rs71647871) and D260fs (rs71647872) exhibit markedly decreased enzymatic activity toward the hydrolysis of the CES1 substrates methylphenidate, oseltamivir, and trandolapril (Zhu et al., 2008; Zhu et al., 2009b; Zhu and Markowitz, 2009). Furthermore, we and others have demonstrated that some therapeutic agents can significantly inhibit CES1 activity (Shi et al., 2006; Zhu et al., 2010; Rhoades et al., 2012). Thus, we hypothesized that variable CES1 function is an important contributing factor to interindividual variability of clopidogrel activation and antiplatelet activity. In the present study, we investigated the influence of CES1 inhibition on the activation of clopidogrel, and the effect of several CES1 nonsynonymous variants on the hydrolysis of clopidogrel and its intermediate metabolite 2-oxo-clopidogrel in vitro.
Materials and Methods
Materials.
S-(+)-clopidogrel, 2-oxo-clopidogrel, 2-bromo-3′-methoxy acetophenone (MPB) derivatized clopidogrel active metabolite (cis-clopidogrel thiol metabolite), clopidogrel carboxylate, and the internal standard d4-clopidogrel were purchased from Toronto Research Chemicals, Inc. (Toronto, Canada). The hydrolytic metabolites of 2-oxo-clopidogrel and clopidogrel-AM were obtained via incubation of the parent compounds (100 µM) with the cell supernatant 9000 (s9) fractions (1 mg protein/ml) prepared from the transfected cells stably expressing wild-type CES1. The hydrolytic reaction was completed in 90 minutes at 37°C. The completion of the bioconversion was confirmed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis. The derivatizing agent MPB and the CES inhibitor bis(4-nitrophenyl) phosphate (BNPP) were obtained from Sigma-Aldrich (St. Louis, MO). The Flp-In-293 cells, pcDNA5/FRT/V5-His TOPOTA Expression Kit, hygromycin B, and Lipofectamine 2000 were obtained from Invitrogen (Carlsbad, CA). Taq polymerase was purchased from Takara (Takara EX Taq HS, Shiga, Japan). Human CES1A1 cDNA cloned into a pCMV-SPORT6 vector was from the American Type Culture Collection (ATCC, Manassas, VA). Pooled human liver s9 fraction was obtained from BD Biosciences (Woburn, MA). All other chemicals and agents were of the highest analytical grade commercially available.
CES1 Inhibition Study in Human Liver s9 Fractions.
Pooled human liver s9 fractions were pre-incubated with an NADPH-generating system (0.1 mg/ml of yeast glucose-6-phosphate dehydrogenase, 3 mg/ml of NADP+, and 0.07 M glucose-6-phosphate) in the presence and absence of the CES1 inhibitor BNPP at 37°C for 5 minutes. The reaction was initiated by adding clopidogrel. The final concentrations of the liver s9 fractions, BNPP, and clopidogrel in the reaction system were 4 mg/ml, 10 µM, and 20 µM, respectively, and the total volume was 200 µl. Samples (20 µl) were collected at 0.25, 0.5, 1.0, 2.0, 4.0, 6.0, 8.0, and 22.0 hours after initiation of the reaction. The reaction was terminated by adding a 5-fold volume of acetonitrile containing the internal standard d4-clopidogrel (25 ng/ml) and the derivatizing reagent MPB (5 mM). The concentrations of clopidogrel and its five metabolites including 2-oxo-clopidogrel, clopidogrel-AM, clopidogrel carboxylate, 2-oxo-clopidogrel carboxylate, and clopidogrel-AM carboxylate were determined by a validated LC-MS/MS assay. The area under the time-concentration curve (AUC) of each compound was calculated using linear-trapezoidal rule by the software WinNonlin 5.2.1 (Pharsight, Mountain View, CA). The differences between BNPP treated and untreated groups were compared using the independent Student’s t test, and P < 0.05 was considered statistically significant.
Establishment of Cell Lines Stably Expressing CES1 Variants G18V, S82L, and A269S.
The Flp-In-293 cell lines stably expressing wild-type CES1 and its variants G143E and D260fs have been developed previously and applied to our ensuing CES1 pharmacogenetic studies (Zhu et al., 2008; Zhu et al., 2009b; Zhu and Markowitz, 2009). The minor allele frequency (MAF) of G143E is estimated to be 3.7%, 4.3%, 2.0%, and 0% in Caucasian, Black, Hispanic, and Asian populations, respectively, whereas D260fs appears to be a rare mutation in all racial and ethnic groups studied to date (Zhu et al., 2008). In the present study, we developed three additional cell lines expressing the CES1 variants G18V (rs3826190), S82L (rs62028647), and A269S (rs115629050) to study the effect of these nonsynonymous variants on CES1 enzymatic activity and clopidogrel hydrolysis. Briefly, the mutant CES1A1 plasmids were obtained via a site-directed mutagenesis assay using specific primers and probes. All plasmids were bidirectionally sequenced to confirm that the desired constructs were generated. The plasmids were cotransfected with the pOG44 plasmid at a ratio of 1:10 into Flp-In-293 cells using the transfection reagent Lipofectamine 2000 in serum-free RPMI 1640 medium. The cell lines were established after hygromycin B selection (100 µg/ml) for 3 weeks. The choice of these three CES1 variants was based on their relatively high MAFs, and all of these variants were predicted by the in silico programs SIFT (https://sift.jcvi.org) and Polyphen2 (https://genetics.bwh.harvard.edu/pph2/) to be possibly detrimental to enzyme function. The MAFs of G18V, S82L, and A269S range from 0.072 to 0.340, 0.026 to 0.368, and 0.014 to 0.057, respectively, among different populations (detailed information about the MAFs of these SNPs in different populations can be found in the Supplemental Table 1). Cells were cultured in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum. To prepare cell s9 fractions, the cells were rinsed and suspended in phosphate-buffered saline buffer solution (pH 7.4). The cells were then sonicated, and the s9 fractions were collected after centrifugation at 9000g for 30 minutes at 4°C. The protein concentration was determined using a Pierce BCA assay kit (Rockford, IL).
Hydrolysis of Clopidogrel and 2-Oxo-Clopidogrel by Wild-Type and Mutant CES1.
An in vitro incubation study was conducted to determine the catalytic activity of CES1 and its variants G18V, S82L, G143E, D260fs, and A269S on the hydrolysis of clopidogrel and 2-oxo-clopidogrel. Selected concentrations of clopidogrel (0.5, 1.5, 5.0, 15.0, 50.0, and 150.0 µM) and 2-oxo-clopidogrel (0.05, 0.15, 0.5, 1.5, and 5.0 µM) were incubated with cell s9 fractions at 37°C for 10 and 20 minutes, respectively. The final concentrations of the s9 fractions were 0.05 mg/ml and 0.2 mg/ml for clopidogrel and 2-oxo-clopidogrel, respectively. The reaction was terminated by the addition of a 5-fold volume of acetonitrile containing the internal standard d4-clopidogrel (25 ng/ml). After centrifugation at 16,000g at 4°C for 10 minutes, the supernatant was collected and analyzed for concentrations of the hydrolytic products (i.e., clopidogrel carboxylate and 2-oxo-clopidogrel carboxylate) by use of established validated LC-MS/MS assay. Data were fit to the Michaelis-Menten equation, and kinetic parameters Km and Vmax were calculated using nonlinear regression analysis (Graphpad Prism software version 4.0; Graphpad Software, Inc., San Diego, CA). Additionally, in an effort to capture any substrate specificity exhibited by any variant, a well-characterized CES1 selective substrate, methylphenidate, was included in the incubation study at a single concentration of 100 µM. The experimental procedures were identical to those previously published (Zhu et al., 2008). The catalytic activity of CES1 and its variants toward methylphenidate was evaluated by measuring the concentrations of the formed hydrolytic metabolite ritalinic acid.
High-Performance Liquid Chromatography–Tandem Mass Spectrometry Assay.
A high-performance liquid chromatography–tandem mass spectrometry assay was developed for the simultaneous quantification of clopidogrel and its five metabolites, including the intermediate metabolite 2-oxo-clodipgrel, the active metabolite clopidogrel-AM, and three hydrolytic metabolites (i.e., clopidogrel carboxylate, 2-oxo-clopidogrel carboxylate, and clopidogrel-AM carboxylate). This assay was a modification of a previously published method (Tuffal et al., 2011). For the human liver s9 inhibition study, the samples were prepared by mixing 20 µl of reaction mixtures with 100 µl of acetonitrile containing the internal standard d4-clopidogrel (25 ng/ml) and the derivatizing agent MPB (5 mM). Clopidogrel-AM is very unstable, so we carefully derivatized it with MPB to form the stable derivative clopidogrel-AM-MPB for analysis (Takahashi et al., 2008; Tuffal et al., 2011). The mixtures were left standing at room temperature for 10 minutes to allow the derivatization reaction to proceed to completion. No derivatizing agent was used for the samples from the clopidogrel and 2-oxo-clopidogrel cell s9 hydrolysis studies. All samples were centrifuged at 16,000g at 4°C for 10 minutes to remove proteins. The resulting supernatant was then collected for high-performance liquid chromatography–tandem mass spectrometry analysis.
A Shimadzu high-performance liquid chromatography system (Shimadzu, Kyoto, Japan) coupled to an Applied Biosystems API 3000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA) was employed. Ionization was achieved via electrospray ionization in the positive mode, and ions were monitored by multiple reaction monitoring. Clopidogrel, 2-oxo-clopidogrel, clopidogrel-AM, clopidogrel carboxylate, 2-oxo-clopidogrel carboxylate, clopidogrel-AM carboxylate, and the internal standard d4-clopidogrel were monitored via the m/z transition 322.1 > 194.0, 338.0 > 183.0, 504.0 > 354.0, 308.0 > 197.9, 324.0 > 169.0, 490.0 > 340.0, and 326.2 > 188.1, respectively. The compounds were separated on a C18 reverse phase column with the mobile phase consisting of 48% of acetonitrile, 2 mM ammonium acetate, and 0.2% formic acid, and delivered at a flow rate of 0.3 ml/min. The assay is highly sensitive, selective, and reliable. The lower limit of quantification of all analytes was estimated to be 1 nM. Accuracy and precision were within U.S. Food and Drug Administration guidelines [FDA (2001) http://www.fda.gov/downloads/Drugs/…/Guidances/ucm070107.pdf]. A representative chromatogram obtained after the incubation of 20 µM of clopidogrel with human liver s9 fractions for 30 minutes in the presence of a NADPH-generating system can be found in the associated Supplemental Fig. 1. As anticipated, two isomers of the thiol metabolites were observed (fourth panel from the top), which are Z compounds with reference to the C3–C16 double bond of 7S clopidogrel (Pereillo et al., 2002; Tuffal et al., 2011). These two isomers differ in terms of whether the C4 configuration is R or S. After being separated by a conventional reverse-phase column (e.g., C18), the first and second peaks are noted to correspond to H3 and H4 metabolites, respectively (Tuffal et al., 2011). It is essential that the analytical assay is enantioselective as both the H3 and H4 metabolites can be formed after hepatic metabolism, but only the H4 isomer is believed to be pharmacologically active. The identity of the active H4 metabolite was confirmed by performing comparisons with pure analytical reference standards.
Results
BNPP Affected Clopidogrel Metabolism and Activation.
Coincubation of BNPP with human liver s9 fractions significantly increased the concentrations of clopidogrel, 2-oxo-clopidogrel, and clopidogrel-AM while the concentrations of all formed carboxylate metabolites were significantly decreased (Fig. 2). Notably, the AUC0–22h of clopidogrel-AM increased by more than 100% in the presence of the CES1 inhibitor BNPP (Table 1). The ratios of clopidogrel to 2-oxo-clopidogrel and 2-oxo-clopidogrel to clopidogrel-AM were not increased following coadministration of BNPP, indicating that the increased formation of clopidogrel-AM was not due to the interaction between BNPP and P450 isoenzymes. The data suggest that CES1 inhibition led to enhanced formation of clopidogrel-AM as well as the increased concentrations of the intermediate metabolite 2-oxo-clopidogrel.
The CES1 Variants G143E and D260fs Are Loss-of-Function Variants for Clopidogrel and 2-Oxo-Clopidogrel Hydrolysis.
Consistent with previous reports, both clopidogrel and 2-oxo-clopidogrel were efficiently hydrolyzed by wild-type CES1 (Tang et al., 2006; Bouman et al., 2011). The Vmax and Km values for clopidogrel were determined to be 3558 ± 371 pmol/min/mg protein and 62.7 ± 15.4 µM, respectively, while the Vmax and Km values for 2-oxo-clopidogrel were 158.1 ± 16.2 pmol/min/mg protein and 2.4 ± 0.6 µM, respectively. The enzymatic activity of the CES1 variants G143E and D260fs were completely impaired in terms of catalyzing the hydrolysis of clopidogrel and 2-oxo-clopidogrel (Fig. 3; the D260fs data overlapped with G143E and are not shown in the figure). However, the variants G18V, S82L, and A269S produced no significant effect on CES1-mediated hydrolysis of clopidogrel or 2-oxo-clopidogrel. Similar to clopidogrel and 2-oxo-clopidogrel, no catalytic activity of G143E or D260fs was observed with regard to methylphenidate metabolism, while the activity of G18V, S82L, and A269S appeared to remain intact and was comparable to that of wild-type enzyme (Fig. 4). Thus, the study demonstrated that both 143E and D260fs are loss-of-function alleles, and the G18V, S82L, and A269S are nonfunctional variants for all three tested CES1 substrates. The results also suggest that the commonly used in silico programs SIFT and Polyphen2 may not be useful approaches for the prediction of the function of CES1 nonsynonymous variants.
Discussion
CES1 is encoded in humans by the CES1 gene, which consists of three isoforms: CES1A1, CES1A2, and CES1A3 (Hosokawa et al., 2007; Fukami et al., 2008). CES1A1 and CES1A3 are inversely located on chromosome 16 q13-q22.1 while CES1A2 is a variant of the CES1A3 gene (Fukami et al., 2008). Both CES1A1 and CES1A2 are functional whereas CES1A3 is a pseudogene due to a premature stop codon located in exon 3 (Fukami et al., 2008; Hosokawa et al., 2008; Zhu and Markowitz, 2012). CES1A1 and CES1A2 are identical with the exception of the exon 1 and promoter regions. In the liver, the majority of CES1 is the product of the CES1A1 gene because the transcription efficiency of the CES1A2 gene is only ~2% of that of CES1A1, probably due to additional Sp1 and CCAAT/enhancer binding protein binding sites in the promoter region of CES1A1 (Fukami et al., 2008; Hosokawa et al., 2008). Thus, only CES1A1 genetic variants are likely to produce a significant impact on CES1 expression and activity.
CES1 expression and activity vary markedly among individuals. Consolidated evidence supports the presence of both genetic and nongenetic factors as significant contributors to observed variability (Hosokawa et al., 1995; Fukami et al., 2008; Yoshimura et al., 2008; Yang et al., 2009; Zhu et al., 2009a; Ross et al., 2012). We have identified and characterized two novel CES1 nonsynonymous variants G143E and D260fs within the CES1A1 and CES1A2 genes, respectively, in a human subject who was participating in a normal volunteer pharmacokinetic study of dl-methylphenidate (Patrick et al., 2007; Zhu et al., 2008). Systemic blood concentrations of both d- and l-methylphenidate were grossly elevated after a single modest dose (0.3 mg/kg) of dl-methylphenidate relative to typical values found in the published literature as well as in his 19 study peers. Additionally, all hemodynamic measures (i.e., systolic blood pressure, diastolic blood pressure, mean arterial blood pressure, and heart rate) in this poor metabolizer differed significantly from his study peers (Zhu et al., 2008). Our in vitro functional studies have demonstrated that catalytic function of both G143E and D260fs are profoundly impaired in terms of hydrolyzing dl-methylphenidate and other CES1 substrates, including the prodrugs trandolapril and oseltamivir (Zhu et al., 2008; Zhu et al., 2009b; Zhu and Markowitz, 2009). Subsequently, a clinical study was conducted in patients with attention deficit-hyperactivity disorder treated with methylphenidate. The results showed that patients carrying the 143E allele required significantly lower doses of methylphenidate for symptom reduction relative to noncarriers (Nemoda et al., 2009). These findings are consistent with the expectation that impaired hydrolysis of methylphenidate would result in higher systemic and central nervous system concentrations and, accordingly, a decreased need for upward dosage titration. Recently, and consistent with our published in vitro findings (Zhu and Markowitz, 2009), a healthy volunteer pharmacokinetic study demonstrated that the G143E variant significantly impaired the activation of the prodrug oseltamivir (Tarkiainen et al., 2012). Notably, the AUC0–∞ of oseltamivir in a homozygous variant type (143EE) was found to be 360% greater than in the noncarrier peers.
Beyond the nonsynonymous SNPs, a number of SNPs within the promoter and 5′-untranslated region (5′-UTR) of CES1A1 and CES1A2/CES1A3 genes have been reported (Geshi et al., 2005; Yoshimura et al., 2008; Sai et al., 2010; Yamada et al., 2010). Among them, -816A>C was reported to be significantly associated with the efficacy of the angiotensin-converting enzyme inhibitor prodrug imidapril (Geshi et al., 2005). However, a later study demonstrated that this SNP resides in the nonfunctional pseudogene CES1A3 (Sai et al., 2010). Furthermore, it was not found to be associated with the activation of the partial CES1 substrate irinotecan, leaving its clinical significance unresolved (Sai et al., 2010). Additionally, the variant -75T>G within the 5′-untranslated region (5′-UTR) of CES1A1 gene was recently found to be associated with isoniazid-induced hepatotoxicity in patients with latent tuberculosis (Yamada et al., 2010). A subsequent clinical study revealed that the AUC ratios of irinotecan metabolites (SN-38+SN-38G) to irinotecan, an indicator of in vivo CES1 and CES2 activity, were significantly lower in cancer patients carrying the minor G allele relative to wild-type patients (Sai et al., 2010). Most recently, the -75G allele was associated with the worsening of appetite in attention deficit-hyperactivity disorder patients treated with methylphenidate (Bruxel et al., 2012). In addition to the promoter, 5′-UTR, and nonsynonymous variants, more than 900 other CES1 variants, such as synonymous and intronic variants, have been documented in several SNP databases (e.g., the National Center for Biotechnology Information dbSNP). However, the function of essentially none of these variants has been systematically evaluated to date.
Beyond the presence of functional CES1 variants, patient age and potential drug-drug interactions may contribute to interindividual variability in CES1 function as well. Data recently generated in our laboratory and by others demonstrate that the expression of CES1 is markedly lower in neonates and infants, and gradually increases in a developmental manner; full maturation in expression and function is observed by age 6 to 9 years (Yang et al., 2009; Zhu et al., 2009a; Shi et al., 2011). Additionally, a number of commonly used medications have been recently identified as CES1 inhibitors or inducers (Shi et al., 2006; Fukami et al., 2010; Zhu et al., 2010; Hatfield and Potter, 2011; Rhoades et al., 2012). However, the magnitude of the effects and the clinical significance of CES1 inhibitors/inducers and developmental age need to be validated by in vivo assessments, and both areas of study are in their relative infancy.
CES1 is responsible for the initial hydrolysis of ~85% of clopidogrel, converting it to its inactive carboxylate metabolite, leaving a balance of ~15% of clopidogrel for further hepatic metabolism (Fig. 1). Furthermore, 2-oxo-clopidogrel and clopidogrel-AM are also hydrolyzed by CES1, forming their respective inactive metabolites. In the present study, we have demonstrated that coincubation of clopidogrel with 10 μM BNPP resulted in significant inhibition of the hydrolysis of clopidogrel, 2-oxo-clopidogrel, and clopidogrel-AM. Based upon the AUCs generated for the three carboxylate metabolites assessed in the inhibition study (Table 1), the inhibitory rates of BNPP (10 μM) on the hydrolysis of clopidogrel, 2-oxo-clopidogrel, and clopidogrel-AM were determined to be 81.6, 60.6, and 44.0%, respectively.
Several P450 enzymes, including CYP1A2, 2B6, 2C9, 2C19, and 3A4, are involved in the two-step activation process of clopidogrel. To fully appreciate the effect of CES1 inhibition on the metabolism and activation of clopidogrel, it is critical that the CES1 inhibitor does not significantly interact with these P450 enzymes at the concentration(s) used in the study. In the present study, 10 μM BNPP significantly inhibited CES1 activity, resulting in decreased hydrolytic metabolism and increased formation of clopidogrel-AM. However, at this concentration, the inhibitor did not significantly alter the ratios of clopidogrel-to-2-oxo-clopidogel and 2-oxo-clopidogrel-to-clopidogrel-AM, indicating that 10 μM BNPP had no significant effect on the activity of those P450 enzymes under these experimental conditions. Thus, the concentration (10 μM) of BNPP appeared to inhibit CES1 as intended, without perturbing native P450 function.
Clopidogrel undergoes a fairly complex metabolism, and two sequential metabolic reactions lead to the formation of clopidogrel-AM. Accordingly, we employed a relatively lengthy incubation (22 hours) approach with multiple time-point sample collections. Under our experimental conditions, the formation of clopidogrel-AM and clopidogrel-AM carboxylate peaked at 2.5 and 4.5 hours, respectively, after the initiation of the incubation period. If only a brief incubation period were employed (i.e., ≤60 minutes), the study would have failed to detect the influence of CES1 inhibition on the production of clopidogrel-AM. The prolonged incubation time permitted a more thorough characterization of the effect of CES1 inhibition on clopidogrel metabolism and activation. The results showed CES1 inhibition can lead to significant increases in the metabolic production of both active and intermediate metabolites of clopidogrel. This is speculated to be the result of the accumulation of the parent compound coupled with the impairment of hydrolytic metabolism.
Beyond the demonstrated effects of chemical inhibition of CES1 and the ensuing influence on clopidogrel biotransformation, we further assessed the influence of CES1 genetic variants on clopidogrel metabolism by use of transfected cell lines. Functional CES1 SNPs such as G143E and D260fs exhibited null activity for the hydrolysis of both clopidogrel and 2-oxo-clopidogrel. However, catalytic activity of the variants G18V, S82L, and A269S on the hydrolysis of clopidogrel or 2-oxo-clopidogrel was comparable to the wild-type enzyme, indicating that these variants are nonfunctional SNPs for clopidogrel and 2-oxo-clopidogrel metabolism though all of these three variants were predicted to be potentially damaging to enzyme function by the in silico programs SIFT and Polyphen2.
Clopidogrel-AM is very unstable, and is currently not commercially available. Thus, we were not able to evaluate whether these variants exert similar effects on clopidogrel-AM hydrolysis. However, given the fact that clopidogrel-AM shares a high degree of similarity in its chemical structure with clopidogrel and 2-oxo-clopidogrel, we anticipate that G143E and D260fs are loss-of-function variants for clopidogrel-AM as well. Native expression of major drug metabolizing enzymes, including P450 enzymes and CES1, are undetectable in the parent human embryonic kidney 293 cells that were used to create the transfected cell lines (Bouman et al., 2011). Therefore, the cells serve as an excellent model to study the effect of genetic variation on the function of CES1. However, the oxidative intermediate and active metabolites of clopidogrel cannot be observed in the transfected human embryonic kidney 293 cells due to the lack of expression of P450 enzymes in the cells.
The clinical benefits of clopidogrel therapy have been demonstrated in numerous large-scale trials, which have established this agent as one of the mainstay treatments in patients with acute coronary syndromes and in those undergoing percutaneous coronary intervention, as also reflected in clinical practice guidelines (Levine et al., 2011; Wright et al., 2011). However, numerous investigations have revealed broad interindividual variability in the response profiles to clopidogrel therapy. Importantly, this has been shown to have prognostic implications (Angiolillo et al., 2007). In particular, patients with reduced clopidogrel-mediated antiplatelet effects who persist with high on-treatment platelet reactivity (also known as “poor responders”) have an increased risk of recurrent ischemic events, including stent thrombosis. Conversely, patients with enhanced clopidogrel-mediated antiplatelet effects who have low on-treatment platelet reactivity (also known as “hyperresponders”) have an increased risk of bleeding complications (Ferreiro et al., 2010).
Enormous efforts have been made to determine the causes of the marked variability in clinical outcomes. Genetic polymorphisms of drug metabolizing enzymes such as CYP2C19 are considered to be an important contributing factor to varied clopidogrel treatment outcomes. However, all biomarkers currently used for the prediction of clopidogrel response, including CYP2C19 genotypes, can only explain a very small portion of the observed variability. Given the very broad application in clinical practice of this pivotal antiplatelet agent, defining the determinants of variability in clopidogrel response is critical and may have important therapeutic implications. Our present study suggests that CES1 can markedly affect clopidogrel metabolism and activation, indicating that functional CES1 SNPs may be associated with higher plasma concentrations of clopidogrel-AM and enhanced antiplatelet activity. In fact, while this article was in preparation, Lewis et al. (2013) reported that the CES1A1 variant G143E discovered in our laboratory (Zhu et al., 2008) was associated with significantly increased plasma concentrations of clopidogrel-AM and greater clopidogrel response in participants in the Pharmacogenomics of Anti-Platelet Intervention study (n = 566) and in 350 patients with coronary heart disease. This clinical observation is fully in agreement with our in vitro data from the present study.
Beyond CES1 genetic polymorphisms, CES1-mediated drug-drug interactions have the potential to influence clopidogrel activation and the ensuing pharmacologic effect as well. Additional clinical investigations are needed to evaluate the effect of CES1-mediated drug-drug interactions on the therapeutic outcomes of clopidogrel and to examine whether CES1 genetic variants can be used as a biomarker to predict clopidogrel response and individualize clopidogrel doses in clinical practice.
Authorship Contributions
Participated in research design: Zhu, Wang, Gawronski, Markowitz.
Conducted experiments: Zhu, Wang, Gawronski, Brinda.
Performed data analysis: Zhu, Wang.
Wrote or contributed to the writing of the manuscript: Zhu, Angiollilo, Markowitz.
Footnotes
↵This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- AM
- active metabolite
- AUC
- area under the time-concentration curve
- BNPP
- bis(4-nitrophenyl) phosphate
- CES1
- carboxylesterase 1
- LC-MS/MS
- liquid chromatography-tandem mass spectrometry
- MAF
- minor allele frequency
- MPB
- 2-bromo-3′-methoxy acetophenone
- P450
- cytochrome P450
- SNP
- single-nucleotide polymorphism
- 5′-UTR
- 5′-untranslated region
- Received November 5, 2012.
- Accepted December 27, 2012.
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics