PharmGKB summary: very important pharmacogene information for cytochrome P450, family 2, subfamily C, polypeptide 8

Introduction

Cytochrome P450, family 2, subfamily C, polypeptide 8 (CYP2C8) is a phase I metabolizing enzyme that plays an integral role in the biotransformation of structurally diverse xenobiotics and endogenous compounds [1]. CYP2C8 accounts for 7% of the CYP content in the liver and is expressed to a lesser extent in the kidney, adrenal gland, mammary gland, brain, ovary, uterus, and duodenum [2–5]. Over the last decade, CYP2C8 has garnered increased attention following the elucidation of its crystal structure, identification of clinically relevant substrates and inhibitors, and characterization of functional CYP2C8 single nucleotide polymorphisms (SNPs). This PharmGKB summary discusses CYP2C8 and its pharmacogenomic importance. A fully interactive version of this short review, with links to individual paper annotations can be found at http://www.pharmgkb.org/gene/PA125#tabview=tab3&subtab=33.

Substrates, inhibitors, and inducers

CYP2C8 is responsible for the biotransformation of 5% of currently used drugs that undergo phase I hepatic metabolism [4]. The enzyme's substrate-binding cavity can accommodate large and structurally unrelated compounds (e.g. paclitaxel and amiodarone) [6,7]. In addition, the CYP2C8 active site is similar in size, but different in shape, from that of CYP3A4 [6,8]. This likely explains why CYP2C8 and CYP3A4 often have overlapping substrates, but yield different metabolite profiles [4,9].

Table 1 provides a list of drugs for which CYP2C8 is a major contributor to metabolism. CYP2C8 also plays an intermediate or minor role in the oxidation of myriad other xenobiotics and endogenous compounds such as NSAIDs (e.g. ibuprofen and diclofenac) [24,25], statins (e.g. fluvastatin and simvastatin acid) [26,27], calcium channel blockers (e.g. verapamil) [28,29], opioids (e.g. morphine and methadone) [30,31], tyrosine kinase inhibitors (e.g. imatinib) [32,33], arachidonic acid [34,35], retinoids [36–39], and others. Further information on CYP2C8 substrates is provided at http://medicine.iupui.edu/clinpharm/ ddis/table.aspx and in comprehensive reviews [4,23,40].

Table 1.

Examples of substrates metabolized to a major extent by CYP2C8^a

Drug name	Drug class	Indication	References
Amiodarone	Class III antiarrhythmic	Supraventricular and ventricular arrhythmias	Ohyama et al. [10]
Amodiaquine, Chloroquine	4-aminoquinolines	Malaria	Li and colleagues [11–13]
Cerivastatinb	Statin	Hyperlipidemia	Wang et al. [14]
Montelukast	Leukotriene receptor antagonist	Asthma	Filppula et al. [15]
Paclitaxel	Taxane	Solid tumors	Rahman and colleagues [16, 17]
Pioglitazone, Rosiglitazone, Troglitazoneb	Thiazolidinedione	Type 2 diabetes	Baldwin and colleagues [18–20]
Repaglinide	Meglitinide	Type 2 diabetes	Bidstrup et al. [21]
Tazarotenic acid	Retinoid derivative	Acne, psoriasis	Attar et al. [22]

^aTable adapted from Totah and colleagues [4,23].

^bWithdrawn from the market.

In vitro, many compounds have been shown to inhibit CYP2C8 including gemfibrozil, trimethoprim, ketoconazole, montelukast, quercetin, and others [14,16,41–46]. In vivo, gemfibrozil is the most potent CYP2C8 inhibitor, primarily because of rapid, mechanism-based inactivation of CYP2C8 by its 1-O-β glucuronide metabolite [47–51]. In clinical studies, gemfibrozil has been shown to increase the plasma exposure of CYP2C8 substrates such as rosiglitazone [52], pioglitazone [53,54], repaglinide [55–57], cerivastatin [58], loperamide [59], R-ibuprofen [60], and montelukast [61]. In addition, gemfibrozil reduced imatinib metabolite formation in healthy volunteers [62].

CYP2C8 has been described as the most inducible member of the CYP2C subfamily [4,40]. Transcriptional activation of CYP2C8 is mediated by the pregnane X receptor (NR1I2), the constitutive androstane receptor (NR1I3), and the glucocorticoid receptor (NR3C1) [63,64]. Along these lines, a constitutive androstane receptor/pregnane X receptor-binding sequence in the distal promoter (– 8806 bp) is thought to play a key role in CYP2C8 induction [63]. In vitro, CYP2C8 is upregulated by the inducers rifampin (rifampicin), dexamethasone, and phenobarbital [65–68]. In clinical drug–drug interaction studies, rifampin has been shown to decrease the plasma exposure of major CYP2C8 substrates such as rosiglitazone [69,70], pioglitazone [71], and repaglinide [72,73].

CYP2C8 gene and common variants

CYP2C8 is located on chromosome 10q24 in a CYP2C gene cluster (centromere–CYP2C18–CYP2C19–CYP2C9– CYP2C8–telomere), of which CYP2C8 is the smallest gene (31 kb, nine exons) [5,74,75]. Given the close proximity of CYP2C8 and CYP2C9, some linkage disequilibrium exists between these genes [76]. Substantial interindividual variability exists in CYP2C8 protein expression and catalytic activity [18,77,78]. This variability is due, in part, to genetic polymorphisms. Over 450 CYP2C8 SNPs have been identified to date [40]. Some of these SNPs, particularly those in the coding region, are associated with variability in CYP2C8-mediated metabolism and altered drug disposition and response. In general, polymorphic CYP2C8 alleles have not been assigned an activity level or phenotype classification (e.g. poor metabolizer). This is primarily because of the relatively limited in vitro data at present, with conflicting results; substrate-dependent functional consequences; and discrepancies between in vitro and in vivo findings. CYP2C8*1 (or *1A) refers to the wild-type or reference allele [5]. In general, most CYP2C8 studies have only evaluated individual SNPs, usually referred to by ‘*’ alleles. Important variant alleles are summarized in Table 2, and relevant in vitro findings are described below.

Table 2.

Variant CYP2C8 alleles

CYP2C8 allele	dbSNP numbera	Nucleotide change	Location	Effect on protein	Enzyme activity (in vitro)	Approximate allele frequencies in major race groups	References
*1 (also known as *1A)	–	–	–	Wild type	Normal	–	Klose et al. [5]
*1B	rs7909236	–271C>A	Promoter	Increased promoter activity		CEU, 23%; HCB, 10%; JPT, 9%; YRI, 0%b	Bahadur and colleagues [79,80]
*1C	rs17110453	–370T>G	Promoter	Unknown	-	CEU, 12%; HCB, 28%; JPT, 34%; YRI, <1%b	Bahadur et al. [79]
*2	rs11572103	c.805A>T	Exon 5	p.I269F	Decreased	CEU, 0%; HCB, 0%; JPT, 0%; YRI, 19%b	Dai and colleagues [81–84]
*3	rs11572080	c.416G>A	Exon 3	p.R139K	Increased, decreased, or unchanged	CEU, 11–14%; HCB, 0%; JPT, 0%; YRI, 0%b	Bahadur and colleagues [79–92]
	rs10509681	c.1196A>G	Exon 8	p.K399R
*4	rs1058930	c.792C>G	Exon 5	p.I264M	Decreased or unchanged	CEU, 7%; HCB, 0%; JPT, 0%; YRI, <1%b	Bahadur and colleagues [79,80,83,84,86,88,93,94]
*5	rs72558196	c.475delA	Exon 3	p.T159 frameshift	None	Japanese, <1%	Soyama and colleagues [92,95–97]
*6	rs142886225	c.511G>A	Exon 4	p.G171S	Unchanged	Japanese, <1%	Saito and colleagues [96,98]
*7	rs72558195	c.556C>T	Exon 4	p.R186X	None	Japanese, <1%	Saito and colleagues [96,98]
*8	rs72558195	c.556C>G	Exon 4	p.R186G	Decreased	Japanese, <1%	Saito and colleagues [96,98]
*9	–	c.740A>G	Exon 5	p.K247R	Unchanged	Japanese, <1%	Saito and colleagues [96,98]
*10	–	c.1149G >T	Exon 7	p.K383N	Unchanged	Japanese, <1%	Saito and colleagues [96,98]
*11	rs78637576	c.820G>T	Exon 6	p.E274X	None	White, 0%; Asian, ≤ 1%; African-American, 0%	Yeo et al. [99]
*12	rs3832694	c.1382_1384delTTG	Exon 9	p.461delV	Unknown	Japanese, <1%	Saito et al. [96]
*13	–	c.669T>G	Exon 5	p.I223M	Unchanged	Japanese, <1%	Saito and colleagues [96,100–102]
*14	rs188934928	c.712G >C	Exon 5	p.A238P	Decreased	Japanese, <1%	Saito and colleagues [96,100–102]
Not assigned	rs66501115	c.1210C>G	Exon 8	p.P404A	Decreased	Japanese, <1%; African, 2%	Soyama and colleagues [85,94,103]

CEU, Utah residents with Northern and Western European ancestry from the CEPH collection; dbSNP, single nucleotide polymorphism database; HCB, Han Chinese in Beijing, China; JPT, Japanese in Tokyo, Japan; YRI, Yoruba in Ibadan, Nigeria.

^aRefSNP accession number in dbSNP (http://www.ncbi.nlm.nih.gov/snp/).

^bFrequencies based on HapMap data available in dbSNP.

CYP2C8*1B and *1C, rs7909236 (g. – 271C > A; CYP2C8* 1B) and rs17110453 (g. – 370G > T; CYP2C8*1C)

The – 271C > A SNP is designated as CYP2C8*1B and is present in about 23% of Whites and 10% of Asians; it is absent in Africans [79] (http://www.ncbi.nlm.nih.gov/projects/SNP/). The variant allele results in the creation of a C/EBPα transcription factor consensus sequence and has been associated with increased transcription factor binding and promoter activities, but not differences in protein expression, as compared with the wild-type allele [79,80]. The – 370G > T SNP is designated as CYP2C8*1C and is present in about 12% of Whites, 28–34% of Asians, and is rare in Africans [79] (http://www.ncbi.nlm.nih.gov/projects/SNP/). The functional consequences of the CYP2C8*1C allele have not been fully elucidated.

CYP2C8*2, rs11572103 (c.805A > T; p.I269F)

The c.805A > T variant is designated as CYP2C8*2 and is located in exon 5. CYP2C8*2 is common in Africans (19%) but is rare in Whites and Asians [81] (http://www.ncbi.nlm. nih.gov/projects/SNP/). In vitro, CYP2C8*2 has been associated with decreased enzyme activity and lower intrinsic clearance of paclitaxel, amodiaquine, and repaglinide compared with the wild-type gene [81–84].

CYP2C8*3, rs11572080 (c.416G > A; p.R139K) and rs10509681 (c.1196A > G; p.K399R)

CYP2C8*3 denotes two highly linked variants, rs11572080 and rs10509681, in exons 3 and 8, respectively [81]. The CYP2C8*3 allele is common in Whites (11–14%) but is rare in Africans and Asians [4,79,81,95,104] (http://www.ncbi.nlm.nih.gov/projects/SNP/). CYP2C8*3 is in strong partial linkage disequilibrium with CYP2C9*2 [76,105,106]. In vitro, conflicting data exist with regard to the effect of CYP2C8*3 on substrate metabolism. Compared with the wild-type enzyme, CYP2C8*3 has been associated with decreased enzyme activity and metabolism of paclitaxel, arachidonic acid, and amodiaquine [79,81–86,103]. Other studies have found no influence of CYP2C8*3 on paclitaxel metabolism [80,87,88]. For some substrates, such as pioglitazone, repaglinide, and cerivastatin, CYP2C8*3 has been associated with increased metabolism [84,89,90]. It has also been shown that CYP2C8*3 exhibits higher overall activity than CYP2C8*1 in the presence of the redox partners, cytochrome b5 and cytochrome P450 reductase [91]. The reasons for differential effects of CYP2C8*3 on substrate metabolism are not entirely clear. One plausible explanation is that the large binding site and multiple substrate recognition sites of CYP2C8 may allow for substrate-dependent interactions with CYP2C8*3 [84]. In addition, the effects of cytochrome b5 and cytochrome P450 reductase may mediate substrate-dependent interactions, as described above [84,91]. Another possible explanation is linkage disequilibrium with CYP2C9*2 [76,105,106].

CYP2C8*4, rs1058930 (c.792C > G; p.I264M)

The c.792C > G variant is designated as CYP2C8*4 and is located in exon 5. CYP2C8*4 is present in about 7% of Whites, but is rare or absent in Africans and Asians (http://www.ncbi.nlm.nih.gov/projects/SNP/). In vitro, CYP2C8*4 has been associated with reduced enzyme activity in some, but not all, studies and lower paclitaxel, repaglinide, ibuprofen, and arachidonic acid metabolism compared with the wild-type gene [79,80,83,84,86,93,94].

Rare variants

CYP2C8*5 through CYP2C8*14 are rare variants that are typically found in less than 1% of the population, mainly Asians (http://www.cypalleles.ki.se/cyp2c8.htm) [98]. Some of these variants have demonstrated functional consequences in vitro. CYP2C8*5 (rs72558196, c.475delA, exon 3) causes a frame shift, which results in a premature stop codon at position 177 [92]. rs72558195 is a triallelic SNP in exon 4, which results in a premature stop codon (CYP2C8*7, c.556C > T, p.R186X) or an Arg to Gly change at codon 186 (CYP2C8*8, c.556C > G) [98]. Decreased rosiglitazone hydroxylation has been reported for the CYP2C8*11 loss-of-function variant (c.820G > T, p.E274X, exon 6) [99]. CYP2C8*14 (c.712G > C, p.A238P, exon 5) results in decreased paclitaxel binding affinity and decreased intrinsic clearance of amiodarone [96,100,101]. The unassigned p.P404A SNP (c.1210C > G, exon 8) has been associated with reduced protein expression and less efficient metabolism of paclitaxel and amiodarone [85,94,103]. The contribution of CYP2C8*5 through CYP2C8*14, as well as other rare CYP2C8 variants, to variability in clinical drug response or rare adverse drug reactions is not known.

CYP2C8 haplotype blocks

Although most CYP2C8 studies have focused on individual SNPs, some work has been done to characterize CYP2C8 haplotypes and their impact on substrate disposition. One study used HapMap data to identify CYP2C8 tag SNPs in Whites. The authors found that CYP2C8 was contained in one haplotype block (40 kb), and six tag SNPs revealed seven common haplotypes (i.e. A, B, C1, C2, C3, D, and E) with a frequency greater than 2% [80]. Haplotype B (which contains g. – 271C > A) was associated with increased paclitaxel metabolism, whereas haplotype C (which combines C1, C2, and C3) was associated with decreased paclitaxel metabolism in vitro [80]. Carriers of haplotype B or haplotype D (which contains CYP2C8*3) had lower repaglinide plasma exposure, whereas carriers of haplotype C had higher repaglinide plasma exposure, as compared with non-carriers [80]. The CYP2C8 haplotype structure has also been characterized in other populations (e.g. Japanese) [96,107]. Given the close proximity of CYP2C8 to other CYP2C genes, some groups have characterized haplotypes containing variants across several genes in the CYP2C cluster [105,106,108–111].

Clinical associations between CYP2C8 variant alleles and drug disposition, response, and toxicity

Antidiabetic agents

Thiazolidinediones, rosiglitazone, and pioglitazone are peroxisome proliferator-activated receptor-γ agonists that are used in the treatment of type 2 diabetes. Most healthy volunteer studies have found CYP2C8*3 to be associated with higher oral clearance and lower plasma exposure of rosiglitazone and pioglitazone than in wild-type homozygotes [112–116]. Little is known about the impact of CYP2C8*2 on thiazolidinedione disposition, although a recent study has suggested that CYP2C8*2 influences the ratio of metabolite to pioglitazone plasma exposure in healthy volunteers [117]. In terms of clinical outcomes, a recent study has shown that CYP2C8*3 carriers have lower rosiglitazone trough concentrations, reduced therapeutic response, and lower risk of developing edema as compared with carriers of the CYP2C8*1/*1 genotype [118].

Repaglinide, a nonsulfonylurea insulin secretagogue, is used to lower postprandial glucose levels in patients with type 2 diabetes. Some clinical studies have reported higher oral clearance and lower plasma exposure of subclinical doses of repaglinide in CYP2C8*3 carriers versus wild-type homozygotes [56,119,120]. However, others have shown no association between repaglinide pharmacokinetics or pharmacokinetics and CYP2C8*3 at clinically relevant doses [121,122].

Paclitaxel

Paclitaxel is a chemotherapeutic agent that is used to treat breast, lung, and ovarian malignancies. Most, but not all, clinical reports suggest that CYP2C8*3 is not a major determinant of paclitaxel pharmacokinetics [123–127]. The discrepancy between in vitro and in vivo findings is likely a result of the contribution of drug transporters to paclitaxel disposition in humans [128]. Peripheral neuropathy is a troubling toxicity associated with paclitaxel therapy and is correlated with drug exposure [126,129]. Some studies have reported an association between CYP2C8*3 and an increased risk for paclitaxel neurotoxicity [126,130–132]. Additional work is needed to elucidate the clinical utility of CYP2C8 variants as predictors of neurotoxicity, as well as other toxicities (e.g. myelosuppression), in paclitaxel-treated patients.

Statins

Cerivastatin, an HMG-CoA reductase inhibitor and CYP2C8 substrate, was withdrawn from the market in 2001 because of a high incidence of rhabdomyolysis [133]. It is possible that rare, loss-of-function CYP2C8 variants (e.g. *5, *7, *11) may have predisposed some individuals to this adverse effect [90,97]. For example, CYP2C8*5 was identified in a Japanese individual who had rhabdomyolysis following cerivastatin therapy [97]. For other statins, no relationship has been observed between CYP2C8 polymorphisms and fluvastatin pharmacokinetics [134] or simvastatin-induced myotoxicity [135].

Antimalarial agents

Amodiaquine and chloroquine are used in the treatment of malaria, particularly in Africa. Although CYP2C8*2 has been associated with decreased amodiaquine metabolism in vitro, it was not a predictor of the efficacy of amodiaquine or major toxicities in African patients [82]. More recently, other data have suggested that host CYP2C8 variants (e.g. *2 or *3) may influence the risk of amodiaquine-resistant or chloroquine-resistant malaria parasites [136,137]. The potential role of CYP2C8 genetics in host–pathogen interactions and resistance merits further investigation.

Nonsteroidal anti-inflammatory drugs

Conflicting data exist with regards to the relationship between CYP2C8*3 and interindividual variability in R-ibuprofen and S-ibuprofen pharmacokinetics [138–141]. This is likely due to ibuprofen being a CYP2C8/2C9 substrate and the linkage disequilibrium that exists between these genes. In terms of adverse effects, some data suggest that the combined presence of CYP2C8*3 and CYP2C9*2 is a determinant of NSAID-induced gastrointestinal bleeding [142]. Additional studies are needed to delineate the relative contributions of CYP2C8 and CYP2C9 to the metabolism of various NSAIDS and the ability of common CYP2C8/2C9 haplotypes to predict NSAID efficacy and/or toxicity [143].

Other clinical associations

Bisphosphonates (e.g. zoledronic acid) are commonly used in the treatment of benign and malignant bone diseases. However, these agents are associated with the rare, but serious, adverse effect of jawbone necrosis. A genome-wide association study found an intronic CYP2C8 SNP (rs1934951) to be significantly associated with osteonecrosis of the jaw in multiple myeloma patients treated with bisphosphonate therapy [144]. Bisphosphonates are not metabolized by cytochrome P450 enzymes; therefore, the mechanism underlying this association is unclear, but it may be due to the influence of CYP2C8 on vascular tone, angiogenesis, and/or inflammation. However, other studies in different patient populations have not been able to replicate the findings of the genome-wide association study [145–147]. As such, additional clinical and mechanistic work is needed to determine whether CYP2C8 SNPs are predictors of bisphosphonate-related osteonecrosis of the jaw.

Calcineurin inhibitors (i.e. cyclosporine and tacrolimus) are key agents used to prevent allograft rejection in solid organ transplantation; however, they are associated with a high incidence of renal dysfunction. Data suggest that CYP2C8 polymorphisms may influence the risk for renal dysfunction in liver and kidney transplant patients treated with these agents. One study has reported that CYP2C8*3 is a predictor of renal toxicity in liver transplant patients treated with calcineurin inhibitors, particularly tacrolimus [86]. CYP2C8 is responsible for the metabolism of endogenous arachidonic acid to vasodilatory epoxyeicosatrienoic acid (EET) metabolites, which are thought to have protective functions in the kidney. It is hypothesized that diminished 14,15-EETand 11,12-EET production, as a result of CYP2C8*3, may predispose individuals to calcineurin inhibitor nephrotoxicity [86]. Similar findings have recently been observed in renal transplant recipients [148]. In terms of other immunosuppressant agents, studies have reported that the CYP2C8 SNPs, rs11572103 (*2) and rs11572076, are significantly associated with mycopheno-late-related anemia following kidney transplantation [149,150].

Conclusion

CYP2C8 plays a major role in the metabolism of many commonly used drugs, and several CYP2C8 SNPs have functional consequence in vivo. As a result, CYP2C8 has emerged as a significant pharmacogene. The CYP2C8 genotype may be important in determining the dosage of drugs and/or in selecting drugs to optimize efficacy and reduce adverse drug reactions. Additional clinical studies are needed to further elucidate the impact and clinical significance of key CYP2C8 variants and haplotypes on heterogeneity in drug disposition and response pheno-types in humans.