OUP user menu

Establishment of a CYP2C19 Genotyping Assay for Clinical Use

Mark A. Cervinski PhD, Mary C. Schwab, Joel A. Lefferts PhD, Lionel D. Lewis MBBChMD, Kimberly A. Lebel MT, Allison M. Tyropolis MT, Solveig M.V. Pflueger MD, Gregory J. Tsongalis PhD
DOI: http://dx.doi.org/10.1309/AJCP9K2KDOCPCBSV 202-207 First published online: 1 February 2013


Conversion of clopidogrel (Plavix) to its active metabolite is catalyzed largely by the P450 enzyme 2C19 (CYP2C19). Numerous allelic variants of CYP2C19 exist. The *1 allele is considered wild type, whereas the *2 and *3 alleles have no in vivo enzymatic activity. Conversely, the *17 allele has increased expression, resulting in increased clopidogrel activation. Poor metabolizers (*2/*2 and *2/*3 genotypes) experience higher rates of therapeutic failure. For this reason, we have validated a CYP2C19 genotyping assay for the *1, *2, *3, and *17 alleles. Genomic DNA extracted from 30 deidentified EDTA whole-blood samples from patients was analyzed at 2 independent facilities using specific TaqMan realtime polymerase chain reaction primers and probes. Concordant genotypes were generated on all samples tested. Of the 30 samples, 15 were CYP2C19*1/*1, 8 were CYP2C19*1/*17, 5 were CYP2C19*1/*2, and 2 were CYP2C19*2/*17. There were no CYP2C19*3 alleles or *2/*2 homozygous genotypes detected. This CYP2C19 genotyping assay is appropriate for clinical testing, demonstrating excellent interlaboratory concordance, enabling the selection of the most effective clopidogrel treatment regimen for patients undergoing percutaneous coronary intervention.

Key Words
  • Clopidogrel
  • CYP2C19
  • TaqMan

Dual-antiplatelet therapy with aspirin and clopidogrel (Plavix, Sanofi, Bridgewater, NJ) has become the standard of care for patients with acute coronary syndrome. Clopidogrel, a prodrug, requires metabolic conversion by a number of cytochrome P4501 enzymes to an active metabolite (2-{1-[(1S)-1-(2-chlorophenyl)-2-methoxy-2-oxoethyl]-4-sulfonyl-3-piperidinyl-diene}acetic acid),2 which selectively and irreversibly inhibits the P2Y12 G-protein–coupled adenosine diphosphate (ADP) receptor on circulating platelets.3,4 Numerous studies have demonstrated that there is considerable interindividual variability in the clinical efficacy of clopidogrel therapy, resulting in some patients responding less well with subsequent recurrent cardiovascular morbidity and/or mortality. Many factors may affect the therapeutic efficacy of clopidogrel, including the concomitant use of other CYP450 enzyme inhibitor drugs; the patient’s age; comorbidities, which include renal failure and type II diabetes mellitus; and poor patient adherence to therapy.5 However, it is also evident that a significant portion of this variability in therapeutic efficacy is due to the individual patient’s genetic makeup of drug-metabolizing enzymes.

Exploration of the pharmacogenomic variables affecting clopidogrel’s therapeutic efficacy has identified many possible targets, including gene products, involved in intestinal absorption, such as the adenosine triphosphate–binding cassette transporter ABCB1, enzymatic conversion/activation (CYP2C9, 2C19, 2B6, 3A4, and 3A5), and response to clopidogrel’s active metabolite by the platelet P2Y12 ADP receptor.511 Although many studies have examined all of the above potential variables, the only gene product consistently correlated with in vitro inhibition of platelet adhesion and clinical outcome in patients is the CYP2C19 genotype.

The highly polymorphic CYP2C19 gene has at least 22 identified allelic variants,1,12 and although some of these allelic variants produce enzymes with decreased or absent function, the effect of many of these allelic variants on enzyme expression or function is as yet not fully elucidated. Of these allelic variants, only 4 (CYP2C19*1, *2, *3, and *17) are of significant allelic frequency generally identified in the vast majority of the human population Table 1. Although additional variants with diminished or absent enzymatic function may be identified via sequencing, the allelic frequency of these variants is typically less than 1%.13 The CYP2C19*1 or wild-type allele encodes an enzyme with normal activity and is the most commonly identified allele, with an allele frequency of approximately 70% in the white population.13 Patients homozygous for the *1 allele have 2 functional copies of the CYP2C19 gene; they are conventionally considered to yield active enzyme and are regarded as extensive metabolizers of CYP2C19 substrates, including clopidogrel. Both the CYP2C19*2 allele (rs-4244285), which contains a splice site defect, and the CYP2C19*3 allele (rs-4986893), which contains a premature stop codon, are not translated into a functional enzyme. As neither of these 2 alleles codes for a functional enzyme, they are considered loss-of-function alleles, and patients who carry either of these variants in combination with the *1 allele are considered to have an intermediate metabolizer phenotype. Individuals with 2 loss-of-function alleles do not express any functional CYP2C19 and are considered to be phenotypically poor metabolizers of CYP2C19 substrates. Conversely, the presence of the CYP2C19*17 (rs-12248560) allele is associated with increased transcription factor binding, leading to increased expression of the CYP2C19 protein. Patients homozygous for the *17 allele or who possess the *17 allele in combination with the *1 allele are generally considered to have an ultra-rapid metabolizer phenotype.10,11,14 The phenotypic characterization of the heterozygous pairing of the *17 allele with a deficiency allele such as the *2 allele is somewhat controversial. Although the phenotypic characterization of *2/*17 patients is beyond the scope of this article, there is evidence that patients carrying both the *2/*17 alleles demonstrate reduced platelet response to clopidogrel, as measured by light transmittance aggregometry.14

View this table:
Table 1

Because of this CYP2C19 genotype treatment failure relationship, the US Food and Drug Administration (FDA) required that the clopidogrel product label include a boxed warning.15 This warning cautions the physician that patients identified to be poor metabolizers of clopidogrel are at higher risk for subsequent thrombotic cardiovascular events following an initial episode of acute coronary syndrome or following percutaneous coronary intervention. Although the FDA did not require a definitive statement in the product label requiring that individual patients should have their CYP2C19 genotype determined prior to initiating clopidogrel therapy, it does state that testing is available and that alternative treatment strategies should be considered for those patients determined to have a genotype consistent with a poor metabolizer phenotype for conversion of clopidogrel to its active metabolite. Further guidance on when CYP2C19 testing should be considered and treatment strategies guided by the results of CYP2C19 genotype results are beyond the scope of this article and are well described elsewhere.12

The clopidogrel FDA-approved product insert does not specify which methodology should be used to determine a patient’s CYP2C19 genotype. If clopidogrel therapy is planned, a laboratory may consider numerous available technologies for this testing. This study reports the validation of a laboratory-developed test that uses TaqMan real-time poly-merase chain reaction (PCR) (Applied Biosystems, Foster City, CA) for the CYP2C19*2, *3, and *17 (rs-12248560) alleles and the relatively short turnaround time of this assay.

Materials and Methods

Specimens and Controls

Genomic DNA was extracted from 30 randomly selected, deidentified remnant EDTA-anticoagulated, patient-derived whole-blood samples with the EZ1 DNA blood kit on the BioRobot EZ1 workstation (Qiagen, Valencia, CA). The extracted DNA was then analyzed for purity and yield via absorbance spectroscopy (A260/A280) using the NanoDrop ND 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA) and split into 2 aliquots. One aliquot was then distributed to each of the 2 testing laboratories involved in this study (Dartmouth-Hitchcock Medical Center, Lebanon, NH, and Baystate Health, Springfield, MA). Control materials consisting of DNA with the known genotypes for CYP2C19, namely for the *1, *2, *3, and *17 alleles, were obtained from ParagonDx (Morrisville, NC). One additional sample with a *2/*3 CYP2C19 genotype was obtained from Coriell (cat. NA16688, Camden, NJ) and characterized as part of the Genetic Testing Reference Materials Coordination Program (GeT-RM) (Centers for Disease Control and Prevention, Atlanta, GA; www.cdc.gov/dls/genetics/rmma-terials/) for purposes of validating the *3 allelic discrimination assay.16


The CYP2C19 genotype of the 30 samples was determined via allelic discrimination plots using TaqMan realtime PCR primers and probes on the Applied Biosystems 7500 Fast Real-Time PCR system. The 30 patient samples (including appropriate controls) were analyzed in 3 allelic discrimination assays each for the presence or absence of CYP2C19*2, CYP2C19*3, and CYP2C19*17 variant alleles. All of the TaqMan probes used in this study were obtained as AB Assay on Demand reagents (Life Technologies, Carlsbad, CA) and are designed to specifically detect the CYP2C19*2, *3, and *17 alleles. The Assay on Demand reagents included both the necessary primers and fluores-cently labeled (FAM and VIC) TaqMan MGB probes to amplify and detect the presence or absence of the *2, *3, and *17 variant alleles.

The PCR cycling conditions were as follows: initial denaturation at 95°C for 10 minutes, followed by 40 cycles of denaturation at 92°C for 15 seconds and annealing/extension at 60°C for 60 seconds. Fluorescence signal intensity due to degradation of the TaqMan probe was quantified during the annealing/denaturation phase of each PCR cycle. Genotype analysis of the 30 deidentified patient DNA samples was performed in triplicate across 4 runs over 2 or 4 consecutive days in the Dartmouth-Hitchcock Medical Center and Baystate Health Molecular Pathology Laboratories, respectively. The presence of wild-type and variant alleles was defined by comparing the relative end-point fluorescence created by the degradation of each fluorescently labeled TaqMan probe (FAM and VIC). One additional sample for validation of the *3 allelic discrimination assay was assayed separately to validate the *3 allelic discrimination assay.


Analysis of control materials of known genotype via the allelic discrimination assay revealed distinct clusters on the allelic discrimination plots representing homozygous wildtype, heterozygous, and homozygous mutant (minor allele or variant) genotypes. The analysis of control materials demonstrated that the allelic discrimination assay was able to accurately genotype the patient samples with respect to the *2, *3, and *17 alleles. A dilution series of control material revealed that the cycle threshold increased with decreasing DNA concentration and demonstrated that the allelic discrimination assay was able to correctly assign genotype in dilute samples where the genomic DNA concentration was as low as 10 ng/ μL Figure 1.

Figure 1

Allelic discrimination plot of CYP2C19*17 patient samples. Homozygous wild-type (WT), heterozygous CYP2C19*17, and homozygous CYP2C19*17 patient samples were analyzed at differing total input DNA concentrations: A-C, 100, 50, and 10 ng/μL DNA homozygous at CYP2C19*17, respectively; D-F, 100, 50, and 10 ng/μL DNA heterozygous at CYP2C19*17, respectively; and G-J, 100, 50, 10, and 1 ng/μL DNA, homozygous WT, respectively. The no-template control sample is represented on this plot by the black square at the lower left corner.

Comparison of the CYP2C19 genotype assigned to the DNA samples by each laboratory revealed that there were no discordant results and that 15 patient samples were homozygous for the *1 allele, 5 were heterozygous for *1/*2, 8 were heterozygous for *1/*17, and 2 were heterozygous for *2/*17Table 2. There were no carriers of the *3 allele in this group of deidentified patients, nor were there any patients homozy-gous for the *2 allele. To validate the *3 allelic discrimination, we analyzed a separate, extracted DNA sample obtained via the GeT-RM (Centers for Disease Control and Prevention) program with a known *2/*3 CYP2C19 genotype following analysis of the 30 deidentified samples. The results from this allelic discrimination assay revealed a *2/*3 genotype, consistent with the genotype assigned by the supplier.

View this table:
Table 2


Here we describe a laboratory-developed test that uses real-time TaqMan PCR to define a CYP2C19 genotype that could be used for patients being considered for or already on antiplatelet treatment with clopidogrel. A number of prospective observational studies have suggested that individuals carrying either 1 or 2 copies of the CYP2C19 loss-of-function alleles (genotype *2 or *3) have lower plasma concentrations of clopidogrel’s active metabolite, as compared with those individuals not carrying a loss-of-function allele, and the individual patient’s CYP2C19 genotype may dramatically compromise treatment success.6,9 The lower concentration of clopidogrel’s active metabolite results in a demonstrably lower magnitude of inhibition of platelet aggregation in ex vivo studies.79,17 Patients carrying both CYP2C19 loss-of-function alleles experience greater rates of treatment failure, with increased rates of secondary myocardial infarction, stent thrombosis, and overall mortality than their extensive metabolizer and ultra-metabolizer counterparts.6,7,9,17,18

The samples used for this analysis were derived from a population that, based on New Hampshire census data, is 93.9% white, in whom the *2 allele was the predominant deficiency allele. Accordingly, we did not detect any patients homozygous for the *2 allele or any patients with *3 alleles in this sample comparison set, but this is not unexpected as the allelic frequency of these deficiency alleles varies between populations, and the presence of the *3 allele is rare in white populations.6 Analysis of a separate sample obtained via the GeT-RM program demonstrated a *2/*3 CYP2C19 genotype, consistent with this sample’s previously determined genotype. Predictably, in white populations, the CYP2C19*17 genotype (rs-12248560) was next most frequent to the *1 allele. The CYP2C19*17 genotype is characterized by a single-nucleotide polymorphism (c.-806C>T) with an additional linked polymorphism (c.-3402C>T) in the 5’ flanking region of the gene. These 2 polymorphisms promote increased transcription factor binding, which leads to increased gene transcription, ultimately increasing the amount of CYP2C19 protein produced and leading to an ultra-rapid metabolizer phenotype.10,11

Previously published observational studies show that CYP2C19 is the major pharmacogenomic variable in clopidogrel activation.6,9 As such, patients who are intermediate (heterozygous carriers) or poor metabolizers of clopidogrel (homozygous carriers of a loss-of-function allele) have lower concentrations of the active metabolite. Such individuals demonstrate a lower magnitude of ex vivo inhibition of platelet aggregation and experience greater rates of treatment failure, with increased rates of secondary myocardial infarction, stent thrombosis, and overall mortality than their extensive metabolizer and ultra-metabolizer counterparts.6,7,9,17,18

One of the limitations of our assay is that it is designed to detect only the *2, *3, and *17 alleles. Numerous known CYP2C19 variants are at very low allele frequencies that will not be detected by this allelic discrimination assay. Due to probe construction, our assay will also not likely differentiate the *2 from the neighboring *10 (rs-6413438) allelic variant. However, the allelic frequency of the *10 allele is less than 1% in the population for which this assay was designed.13 All allelic discrimination assays are subject to the same limitations, which can be overcome by the addition of separate primers and probes should the frequency of specified rare alleles be at a significant frequency in the patient population being served. Only sequencing of the CYP2C19 gene will reveal all possible allelic variants. However, this increased resolution would come at additional expense and increased turnaround time.

Although the patient’s CYP2C19 genotype is one of many factors underlying clopidogrel’s treatment success or failure, rapid determination of the CYP2C19 genotype may provide the clinician with an opportunity to select an alternative antiplatelet drug such as prasugrel or other antiplatelet medications that do not rely on CYP2C19 activation should the genotyping assay reveal that the patient is either an intermediate or poor metabolizer of clopidogrel. Many PCR-based technologies may be used to determine a patient’s CYP2C19 genotype. The decision of which technology any laboratory should pursue will likely depend on the technologies already available. However, given the choice between multiple technologies, our experience has shown that the TaqMan real-time PCR assay as used in this study retains many advantages, most notably the availability of characterized reagents from the AB on-demand menu as well as the advantage of a rapid turnaround time. In our laboratory, the total time required to set up and perform the TaqMan PCR reaction is approximately 2 hours, with the PCR cycling conditions outlined in our materials and methods accounting for approximately half of this time. Assuming that the laboratory has an efficient method to rapidly extract genomic DNA from anticoagulated whole-blood samples, multiple samples could be processed and results easily made available the same day.

The real-time PCR assay on the Spartan RX platform (Spartan, Ottawa, Canada) has been reported to have a similarly short total assay time, but the CYP2C19 assay only detects the presence of the *2 allele and is not available in the United States.19,20 Other technologies, such as the Hologic Invader (Hologic, Bedford, MA), GenMark eSensor (GenMark Diagnostics, Carlsbad, CA), and Luminex assay (Luminex, Austin, TX), have longer turnaround times. A typical Invader assay requires 3 hours of incubation,21 whereas GenMark eSensor and Luminex platforms have published total assay times of 4 and 8 hours.22,23 The GenMark and Luminex platforms also require post-PCR manipulation, potentially leading to contamination of the laboratory with amplicon. In-house designed PCR-restriction fragment length polymorphism assays would also result in a longer laboratory turnaround time as the PCR amplification, digestion, and analysis of the product via gel or capillary electrophoresis would likely result in an average analytical time spanning 2 days. It is now also possible to find reference laboratories that will fully sequence the CYP2C19 gene. Although whole-gene sequencing will identify rare allelic variants, it is important to note that testing for the *2 and *3 alleles will identify approximately 84% of white patients, 90% of African American patients, and almost 100% of Asian patients with the poor metabolizer phenotype.24

The limitation of our study is that our validation cohort consisted of only 30 patient samples and we did not verify the patients’ CYP2C19 genotype via sequencing. It is possible that a rare allelic variant may exist in this validation cohort; however, no other genotyping technique, with exception of whole-gene sequencing, which would come at increased expense and a greatly increased turnaround time, would routinely identify all variant alleles. It is also important to note that although the CYP2C19 genotype is a strong predictor of treatment effectiveness, it is not the only factor that will predict treatment efficacy.

The decision of which antiplatelet therapy to prescribe for a patient is often made soon after the initial clinical presentation, and an assay such as that described in this study using TaqMan probes would afford the clinician the opportunity to incorporate the patient’s CYP2C19 genotype into the decision-making process.


View Abstract