CYP2D6 genotyping typically is performed by platforms that simultaneously detect multiple variants across the gene (Fig. 21.1). Several commercial assays are currently available, including the AmpliChip^® (Affymetrix/Roche [14]), Luminex [15], and AutoGenomics [16] assays, as well as other laboratory-developed tests (e.g., single-nucleotide extension assay) [17]. The AmpliChip^® CYP450 Test (Roche Diagnostics, Indianapolis, IN) is an oligonucleotide microarray hybridization method that tests variants in both CYP2D6 and CYP2C19. The Tag-It™ Luminex (Luminex Molecular Diagnostics, Toronto, Canada) platform is a bead array with oligonucleotides bound to microspheres and genotyping by allele-specific primer extension. The AutoGenomics (Carlsbad, CA) platform is a film-based microarray tested on the INFINITI^® PLUS Analyzer. These assays typically interrogate 15–20 important CYP2D6 variants including the deletion and duplication alleles.

CYP2D6 alleles are designated by the common star (*) allele nomenclature system, which often include multiple variants on the same haplotype. The CYP2D6*1 allele is the wild-type haplotype encoding normal enzyme activity; however, this is typically assigned in the absence of other detected variants. Consequently, when *1 is reported by targeted genotyping, a rare CYP2D6 star (*) allele not included in the genotyping panel would not be detected, which can only be identified by gene sequencing. Commonly interrogated CYP2D6 alleles are summarized in Table 21.3.

Two functional CYP2D6 alleles or one functional and one reduced-function CYP2D6 allele predict normal or extensive metabolizers. Two reduced-function alleles or one reduced-function and one nonfunctional allele predict intermediate metabolizers. Two nonfunctional alleles predict a poor metabolizer phenotype. Duplicated functional CYP2D6 alleles predict an ultrarapid metabolizer phenotype [8]. In addition to duplicated functional alleles, duplicated nonfunctional or reduced-function alleles also have been described. As such, determining which CYP2D6 allele is duplicated is important for proper interpretation when a gene duplication is identified in addition to a heterozygous genotype [18].

CYP2D6 genotyping is complicated by the fact that several variants occur on a number of important star (*) alleles and commercial assays cannot determine the phase of identified genotypes [19]. For example, the nonfunctional *4 allele is defined by several variants, including 100C > T and 1846G > A. The 1846G > A variant is the defining functional mutation for *4, which disrupts exon splicing and results in a frameshift and loss of enzyme activity. However, the *10 reduced-function allele also has the 100C > T variant but without 1846G > A. Of note, 1846G > A can also exist on a haplotype without 100C > T (*4M), and this allele is very rare. When one copy of each variant is detected, the most probable genotype is heterozygous *4 (one copy of a nonfunctional allele), since 100C > T and 1846G > A are presumed to be on the same haplotype. As genotyping platforms generally do not predict CYP2D6 haplotypes, individual laboratories are responsible for interpreting results and reporting CYP2D6 genotypes. Without haplotyping by more involved molecular assays, it is possible that one copy of both 100C > T and 1846G > A could be misinterpreted as a *4M/*10 compound heterozygote (one nonfunctional allele and one decreased-function allele). Although this scenario would be rare in the general population, the *4/*10 genotype is more appropriately defined by two copies of 100C > T and one copy of 1846G > A.

The CYP2D6 gene is highly homologous to a related pseudogene (CYP2D7P1), which complicates the design of PCR primers used for genotyping. As such, primers are designed to avoid co-amplification of pseudogenes by employing a specific long-range PCR of the entire CYP2D6 gene. Given the technical challenges with long-range PCR, suboptimal or degraded DNA may not perform well for many commercial CYP2D6 assays.

When a CYP2D6 platform is unable to identify if a duplication is a functional or nonfunctional allele in a heterozygous sample, laboratories may report an “indeterminant” result. As noted above, family studies can facilitate identifying which allele is duplicated in these cases. Although laboratory guidelines for CYP2D6 genotyping in relation to tamoxifen therapy have recently been reported [19], no current professional guidelines detail which alleles should be included in clinical CYP2D6 assays. Therefore, different laboratories may include different CYP2D6 alleles in their testing panels, which can result in conflicting CYP2D6 genotypes and predicted phenotypes between laboratories.

Cytochrome P450-2C19 (CYP2C19) is another important member of the cytochrome P450 superfamily that metabolizes commonly prescribed medications including anticonvulsants, antidepressants, antifungals, proton-pump inhibitors, antithrombotics, and chemotherapy, antimalarial, and antiulcer drugs [20]. Another drug metabolized by CYP2C19 is the antiplatelet agent clopidogrel, commonly prescribed for patients with acute coronary syndromes and those undergoing percutaneous coronary intervention (PCI). CYP2C19 converts clopidogrel in a two-step enzymatic reaction to an active metabolite, which binds irreversibly to platelet receptors and inhibits aggregation for the duration of the platelet life span. About 25 % of ACS/PCI patients have reduced platelet inhibition due in part to CYP2C19 loss-of-function alleles [21], which reduces the effectiveness of clopidogrel treatment.

Although over 30 CYP2C19 variant alleles have been identified, the effect of some rare alleles on enzyme function has not been established. Notable among the CYP2C19 alleles are *2 to *8, which impart loss of function, and *17, which has been reported as an increased-function allele (http://​www.​cypalleles.​ki.​se/​cyp2c19.​htm). In contrast to the role of the CYP2C19*2 to *8 alleles in reduced clopidogrel effectiveness, *17 may be associated with an enhanced response to clopidogrel and some antidepressants [22]. Variant CYP2C19 allele frequencies vary between racial and ethnic groups, with *2 being present in approximately 30 % of Asians and 15 % of Caucasians and African Americans, while *3 has a frequency of approximately 8 % in the Asian population, but is rare in other populations.

Warfarin is a commonly prescribed vitamin K antagonist for the prevention of thromboembolism among patients with atrial fibrillation, deep vein thrombosis, and other indications. However, the drug has a very narrow therapeutic index and a large interindividual variability in response, in part due to inherited genetic variability within genes involved in warfarin pharmacokinetics and pharmacodynamics. For example, cytochrome P450-2C9 (CYP2C9) is the principal enzyme involved in S-warfarin inactivation and elimination. Like other cytochrome P450 genes, CYP2C9 is highly polymorphic with several known variant alleles encoding reduced enzyme activity (http://​www.​cypalleles.​ki.​se/​cyp2c9.​htm). Importantly, in vitro studies have shown that the common *2 and *3 alleles are functionally defective, exhibiting only approximately 70 % and 5 % of wild-type activity toward S-warfarin, respectively [28]. Consequently, these CYP2C9 alleles result in reduced warfarin inactivation, higher warfarin blood levels with standard warfarin doses, and increased bleeding risks [29, 30]. The frequencies of *2 and *3 are approximately 15 % and 6 %, respectively, among Caucasians and 2–4 % among Asians and African Americans [31]. Of note, other variant alleles are more prevalent in individuals of African descent (e.g., *5, *6, and *8) [31], which may provide additional utility for genetically guided warfarin dosing in these populations.

Warfarin acts as a vitamin K antagonist by inhibiting the regeneration of reduced vitamin K, an essential cofactor for the clotting cascade. The target enzyme for warfarin is VKORC1, which catalyzes the rate-limiting step in the vitamin K cycle [32]. Importantly, common VKORC1 haplotypes that result in reduced gene expression have been reproducibly implicated as the major genetic determinant of warfarin dose variability [33–38]. The most commonly interrogated VKORC1 allele associated with warfarin sensitivity is the c.-1639G > A promoter polymorphism (Fig. 21.3) [38]. Like CYP2C9, the frequencies of VKORC1 alleles vary between racial and ethnic groups, with c.-1639G > A allele frequencies of approximately 10 %, 45 %, and 70–90 %, among African-American, Caucasian, and Asian individuals, respectively [31]. Of note, rare VKORC1 coding region mutations also have been identified that are strongly correlated with warfarin resistance.

Taken together, common CYP2C9 and VKORC1 variant alleles account for approximately 35 % of interindividual dose variability [36, 37], which can be used in conjunction with known clinical variables (e.g., age, race, body weight, gender) to predict individual therapeutic warfarin doses. These data prompted the US Food and Drug Administration (FDA) to modify the warfarin label noting the importance of pharmacogenetic testing with dosing recommendations based on CYP2C9 and VKORC1 genotypes (Table 21.6 ).