Pharmacogenetics of Warfarin: Current Status and Future Challenges

Pharmacogenetics of warfarin: current status and future challenges

M Wadelius1 and Warfarin is an anticoagulant that is difficult to use because of the wide 2 variation in dose required to achieve a therapeutic effect, and the risk of M Pirmohamed serious bleeding. Warfarin acts by interfering with the recycling of vitamin K 1Department of Medical Sciences, Clinical in the liver, which leads to reduced activation of several clotting factors. Pharmacology, Uppsala University Hospital, Thirty genes that may be involved in the biotransformation and mode of Uppsala, Sweden and 2Department of action of warfarin are discussed in this review. The most important genes Pharmacology and Therapeutics, University of affecting the pharmacokinetic and pharmacodynamic parameters of warfarin Liverpool, Liverpool, UK are CYP2C9 (cytochrome P450 2C9) and VKORC1 (vitamin K epoxide Correspondence: reductase complex subunit 1). These two genes, together with environ- Dr M Wadelius, Department of Medical mental factors, partly explain the interindividual variation in warfarin dose Sciences, Clinical Pharmacology, Uppsala requirements. Large ongoing studies of genes involved in the actions of University Hospital, Entrance 61 3rd floor, warfarin, together with prospective assessment of environmental factors, SE-751 85 Uppsala, Sweden. E-mail: [email protected] will undoubtedly increase the capacity to accurately predict warfarin dose. Implementation of pre-prescription genotyping and individualized warfarin therapy represents an opportunity to minimize the risk of haemorrhage without compromising effectiveness. The Pharmacogenomics Journal (2007) 7, 99–111. doi:10.1038/sj.tpj.6500417; published online 19 September 2006

Keywords: warfarin; vitamin K epoxide reductase complex subunit 1; VKORC1; cytochrome P450 enzyme; CYP2C9; vitamin K-dependent protein

Introduction

Warfarin is a widely used coumarin anticoagulant prescribed for patients with venous thrombosis and pulmonary embolism, chronic atrial fibrillation and prosthetic heart valves. Interindividual differences in drug response, a narrow therapeutic range and the risk of bleeding, all make warfarin a difficult drug to use clinically. Warfarin dose requirements, the stability of anticoagulation and risk of bleeding are influenced by environmental factors such as the intake of vitamin K, illness, age, gender, concurrent medication and body surface area, and by genetic variation.1–8 To be able to improve the benefit–harm profile associated with warfarin therapy, all these factors need to be taken into account. There is increasing interest in whether pharmacogenetics can accurately predict warfarin dose. There have been some recent advances in this area, but much more work needs to be done: at least 30 genes may be involved in the mode of action of warfarin (Table1 and Figure 1). Within these genes, there are thousands of publicly available single nucleotide polymorphisms (SNPs) with unknown function. There are also over 100 genetic variants that are known to change protein function that can be identified through a systematic search of the Received 24 April 2006; revised 13 July 2006; accepted 24 July 2006; published online 19 literature. Well-known examples are polymorphisms that change cytochrome September 2006 P450 enzyme activity, apolipoprotein E (APOE) variants, factor V Leiden and Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 100

Table 1 Genes involved in the mechanism of action of warfarin

Protein name Gene Location Exons Transcript Protein Function of protein (bp) (aa)

Biotransformation of warfarin Transport Alpha-1-acid glycoprotein 1, ORM1 Chr 9: 114 083 890– 6 802 201 A plasma glycoprotein that Orosomucoid 1 114 087 309 bp functions as a carrier of warfarin in the blood16,17 Alpha-1-acid glycoprotein 2, ORM2 Chr 9: 114 171 703– 6 760 201 A plasma glycoprotein that Orosomucoid 2 114 175 086 bp functions as a carrier of warfarin in the blood16,17 P-glycoprotein, Multidrug ABCB1 Chr 7: 85 668 428– 29 4643 1279 A cellular efflux pump for resistance protein 1 (MDR1) 85 877 818 bp xenobiotics.20 Warfarin is a weak inhibitor and maybe a substrate18 Metabolism Cytochrome P450 2C9 CYP2C9 Chr 10: 96 688 405– 9 1847 490 Polymorphic hepatic drug 96 739 137 bp metabolizing enzyme. Metabolism of S-warfarin26,28 Cytochrome P450 1A1 CYP1A1 Chr 15: 72 798 943– 7 2601 512 Extrahepatic oxidation, inducible. 72 804 930 bp Metabolism of R-warfarin28,43,45 Cytochrome P450 1A2 CYP1A2 Chr 15: 72 828 257– 7 1618 516 Hepatic oxidation, inducible. 72 834 505 bp Metabolism of R-warfarin28,43 Cytochrome P450 2A6 CYP2A6 Chr 19: 46 041 284– 9 1751 494 Polymorphic hepatic drug 46 048 180 bp metabolizing enzyme. Metabolism of S-warfarin?35 Cytochrome P450 2C8 CYP2C8 Chr 10: 96 786 520– 9 1923 490 Polymorphic hepatic drug 96 819 244 bp metabolizing enzyme. Minor pathway for R- and S-warfarin26,28 Cytochrome P450 2C18 CYP2C18 Chr 10: 96 432 700– 9 2418 490 Found in the liver and lung. Minor 96 485 937 bp pathway for R- and S-warfarin28,44 Cytochrome P450 2C19 CYP2C19 Chr 10: 96 437 901– 9 1901 490 Polymorphic hepatic drug 96 603 007 bp metabolizing enzyme. Minor pathway for R- and S-warfarin28,44 Cytochrome P450 3A4 CYP3A4 Chr 7: 97 889 181– 13 2768 503 Hepatic oxidation, inducible. 97 916 385 bp Metabolism of R-warfarin28 Cytochrome P450 3A5 CYP3A5 Chr 7: 97 780 394– 13 1707 502 Polymorphic hepatic and 97 812 183 bp extrahepatic oxidation. Metabolism of R-warfarin?46 Cytochrome P450 inducibility Pregnane X receptor (PXR) NR1I2 Chr 3: 120 982 021– 9 2753 473 Mediates induction of CYP2C9, 121 020 021 bp CYP3A4, other CYP enzymes and ABCB151–53 Constitutive androstane NR1I3 Chr 1: 158 012 528– 9 1337 348 Transcriptional regulation of a receptor (CAR) 158 021 028 bp number of genes including CYP2C9 and CYP3A454

Biotransformation of vitamin K Transport Apolipoprotein E APOE Chr 19: 50 100 879– 4 1179 317 Apolipoprotein E serves as a ligand 50 104 489 bp for receptors that mediate the uptake of vitamin K61–64 Vitamin K cycle Vitamin K epoxide reductase VKORC1 Chr 16: 31 009 677– 3 997 163 A hepatic epoxide hydrolase that 31 013 777 bp catalyses the reduction of vitamin K. The target of warfarin56,69,70 Epoxide hydrolase 1, EPHX1 Chr 1: 222 304 587– 9 1605 455 A hepatic epoxide hydrolase in the microsomal 222 339 995 bp endoplasmic reticulum that may be complexed with VKOR86,87,89 NAD(P)H dehydrogenase, NQO1 Chr 16: 68 300 807– 6 2448 274 A detoxifying enzyme that has the quinone 1 68 317 893 bp potential to reduce the quinine form of vitamin K62,90,91

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 101

Table 1 Continued

Protein name Gene Location Exons Transcript Protein Function of protein (bp) (aa)

Calumenin CALU Chr 7: 127 973 368– 7 3316 315 Binds to the vitamin K epoxide 128 005 478 bp reductase complex and inhibits the effect of warfarin98,99 Gamma-glutamyl carboxylase GGCX Chr 2: 85 687 865– 15 3155 758 Carboxylates vitamin K-dependent 85 700 237 bp coagulation factors and proteins in the vitamin K cycle92–94 Vitamin K-dependent proteins Coagulation factor II, F2 Chr 11: 46 697 331– 14 1997 622 Converts fibrinogen to fibrin, prothrombin 46 717 631 bp activates FV, FVIII, FXI, FXIII, protein C93,108 Coagulation factor VII F7 Chr 13: 112 808 124– 9 2459 466 Is converted to FVIIa and then 112 822 348 bp converts FIX to FIXa and FX to FXa93,108 Coagulation factor IX F9 Chr X: 138 340 437– 8 2780 461 Makes a complex with FVIIIa and 138 373 137 bp then converts FX to its active form93,108 Coagulation factor X F10 Chr 13: 112 825 128– 8 1524 488 Converts FII to FIIa in the presence 112 851 846 bp of factor Va93,108 Protein C PROC Chr 2: 127 892 246– 9 1756 461 Activated protein C counteracts 127 903 048 bp coagulation together with protein S by inactivating FVa and VIIIa93,108 Protein S PROS1 Chr 3: 95 074 647– 15 3275 676 Cofactor to protein C that 95 175 395 bp degrades coagulation factors Va and VIIIa93,108 Protein Z PROZ Chr 13: 112 860 971– 8 1488 400 Is together with protein Z- 112 874 700 bp dependent protease inhibitor, a cofactor for the inactivation of FXa93,100 Growth-arrest-specific GAS6 Chr 13: 113 546 903– 15 2499 678 Participates in many processes, for protein 6 113 590 421 bp example, potentiation of agonist- induced platelet aggregation62 Other coagulation proteins Anti-thrombin III SERPINC1 Chr 1: 170 604 596– 7 1559 464 Inhibits FIIa, FIXa, Xa, XIa and XIIa. 170 618 130 bp Anti-thrombin deficiency increases risk of thrombosis108 Coagulation factor V F5 Chr 1: 166 215 067– 25 6914 2224 A cofactor that activates FII 166 287 379 bp together with FXa. An F5 mutation leads to risk of thrombosis108

Protein and gene names, location in NCBI build 35, number of exons, size of transcript and protein and function of the proteins are included in the table.

other mutations in the coagulation system that cause either analyses.9,10 The efficacy and safety is, however, contingent a bleeding tendency or increase the risk of thrombosis. In on maintaining the anticoagulation within a clinically this review, we will analyse the different pathways involved acceptable ‘therapeutic range’. This may be easier to achieve in warfarin’s action and critically evaluate the likelihood of within the confines of a randomized controlled trial than whether genetic variation in these pathways may truly during everyday real-world clinical practice. impact on the safety and ease of use of warfarin in clinical Warfarin has a narrow therapeutic index and thus the practice. dose required to achieve therapeutic anticoagulation is very close to the dose that leads to over-anticoagulation. Furthermore, the maintenance dose varies between different Variability in the effect of warfarin individuals, and ranges from 0.5 mg/day to more than 10 mg/day. This unpredictability leads to difficulties in The efficacy of warfarin and other vitamin K antagonists in maintaining patients within a therapeutic anticoagulation preventing and treating thrombosis has been well demon- range, which usually is an international normalized ratio strated in numerous randomized controlled trials and meta- (INR) of 2.0–3.0. A recent analysis of 2223 patients showed

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 102

Figure 1 An overview of warfarin interactive pathways. This figure illustrates the genes thought to be involved in the action and biotransformation of warfarin and vitamin K.

that patients were outside the INR target range one-third of that the average cost per patient of a bleeding episode was the time, with 15.4% of INR values above 3.0 and 16.7% of $15 988 (range $2707–$64 446) with a mean length of stay of INR values below 2.0.11 There was higher mortality, 6 days.13 increased risk of stroke and increase in the rate of hospitalization when patients were outside the anticoagula- Warfarin interactive pathways tion range. The most feared adverse effect associated with antic- At least 30 genes may be involved in the mechanism by which oagulation is bleeding. Major and fatal bleeding events warfarin exerts its anticoagulant effect (Table 1 and Figure 1). occur at a rate of 7.2 and 1.3/100 patient years, respectively, The most important gene in the pharmacokinetics of warfarin according to a meta-analysis of 33 studies.12 Bleeding rates is CYP2C9 (cytochrome P 2C9 gene), whereas the central may be lower in specialized anticoagulation clinics,13 and 450 gene in the pharmacodynamics of warfarin is VKORC1 when monitoring is more frequent.11 Bleeding events are (vitamin K epoxide reductase complex subunit 1 gene). most likely to occur within the first 90 days of therapy, but the incidence never falls to zero. The risk of bleeding is higher when INR is over 3, but bleeding can also occur when Transportation of warfarin the INR is within the therapeutic range.13 Apart from the mortality and morbidity associated with warfarin-related The molecular basis of the pharmacokinetics of warfarin has bleeds, there is also a cost element: a recent analysis showed been extensively studied.14 Warfarin is rapidly absorbed

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 103

from the stomach and the upper gastrointestinal tract, both Europeans and Afro-Americans) all lead to a reduction in with a bioavailability of 100%.15 In the circulating blood, warfarin dose requirement.39,40 warfarin is 99% protein bound largely to albumin and alpha- S-warfarin may also be metabolized by CYP2C8, CYP2C18 1-acid glycoproteins. The latter are encoded by ORM1 and CYP2C19 to form 4-hydroxywarfarin, although these (orosomucoid 1 gene or alpha-1-acid glycoprotein 1 gene) are minor pathways.28 The genes encoding these P450 and ORM2 (orosomucoid 2 gene or alpha-1-acid glycopro- isoforms contain many functional polymorphisms. Two tein 2 gene).16,17 It has been shown that warfarin binds studies have so far found no effect of the CYP2C19*2 variant preferentially to certain genetic variants of alpha-1-acid allele on warfarin therapy.2,41 The role of other CYP2C glycoproteins that can be separated by chromatography.17 isoforms has not been adequately evaluated, but would be Whether this has any effect on warfarin dose requirement predicted to be small. Furthermore, the coumarin hydro- seems rather unlikely given the binding to albumin. xylase variant CYP2A6*2 has been suspected to cause Based on an inhibition assay, there is some evidence that warfarin sensitivity.35,42 However, these reports from one the transport of warfarin across plasma membranes of cells, laboratory have not been replicated and there is no good for example in the liver, may be mediated by P-glycoprotein evidence that CYP2A6 actually metabolizes warfarin. (multidrug resistance protein 1), which is encoded by R-warfarin, which is the less active enantiomer, is mainly 18 ABCB1 (MDR1; ATP-binding cassette transporter B1 gene). metabolized by cytochrome P450 enzymes CYP1A2 (to 6- and However, the evidence is scant and seems less probable 8-hydroxywarfarin) and CYP3A4 (to 10-hydroxywarfar- given that warfarin has a very good bioavailability. Poly- in).26,28,43 In addition, CYP1A1, CYP2C8, CYP2C18, morphisms in ABCB1 have been linked to changes in mRNA CYP2C19 and CYP3A5 may be involved in the metabolism and protein expression, and to the pharmacokinetic profiles of R-warfarin.26,28,43–47 There are as yet no published studies of various drugs.19 The widely studied synonymous exon 26 indicating that polymorphisms in these enzymes influence C3435T variant has been the subject of numerous studies warfarin dosing.1,48 with conflicting results.19–22 Interestingly, it has been shown Many of the P450 isoforms involved in the metabolism of that a haplotype containing the exon 26 C3435T variant warfarin are inducible; indeed, this is the mechanism of the (which could be expected to reduce drug efflux) was well-known interactions that occur when warfarin is co- overrepresented among patients requiring a low dose of prescribed with drugs, for instance the aromatic anti- warfarin to maintain therapeutic anticoagulation.1 This convulsants and herbal medicines such as St John’s needs to be replicated in another cohort, but is unlikely to Wort.49,50 The mechanism of induction of the P450 isoforms be of major importance. is dependent on the nuclear hormone receptors pregnane X receptor (PXR) and constitutive androstane receptor (CAR), but nothing has yet been published on whether warfarin Biotransformation of warfarin dose requirement is affected by variation in the genes encoding these receptors, NR1I2 (pregnane X receptor gene) The influence of genetic variation on warfarin pharmacoki- and NR1I3 (constitutive androstane receptor gene).50–54 netics has been the focus of several review articles.4,14,23–25 Warfarin is administered as a racemate comprising R- and S-enantiomers: the S-form being 3–5 times more active than Distribution and hepatic uptake of vitamin K the R-form.26,27 Once warfarin has entered the liver,

S-warfarin is metabolized by cytochrome P450 2C9 (CYP2C9) Warfarin targets the vitamin K epoxide reductase complex in to 7-hydroxywarfarin.26,28 Many different polymorphisms the liver, thereby interfering with the recycling of vitamin K in CYP2C9 that vary according to ethnicity and in terms (Figure 1).25,55–59 A high intake of fat-soluble vitamin K can of their functional effects have been described (http:// reverse the action of warfarin, and a low or erratic intake of www.imm.ki.se/CYPalleles/cyp2c9.htm). Most of the war- dietary vitamin K may be partly responsible for unstable farin studies have so far concentrated on CYP2C9*2 and control of anticoagulation in warfarin patients.60 Vitamin

CYP2C9*3 variants. Compared with extensive metabolizers, K1 is absorbed from the small intestine along with dietary who are homozygous for the wild-type *1 allele, homo- fat, transported by chylomicrons in the blood and subse- zygosity for *2 reduces CYP2C9 enzyme activity to 12% quently cleared by the liver through an APOE receptor- whereas homozygosity for *3 reduces enzyme activity to specific uptake.61–63 Uptake of chylomicrons and thus 29–31 5%. In accordance with this, many studies have vitamin K1 into the liver varies between different APOE shown that patients with the CYP2C9*2 and CYP2C9*3 alleles, the rank order being *E44*E34*E2.61,64 Consistent variant alleles require lower mean daily warfarin doses with this, patients with the APOE*E2 allele, who allegedly 1–3,5–8,32–38 (Table 2). A systematic review and meta-analysis have the least efficient uptake of vitamin K1, have an of nine studies has established that the CYP2C9*2 and increased risk of warfarin-associated intracerebral haemor- CYP2C9*3 alleles lead to 17 and 37% reduction in the daily rhage.65 In a Swedish cohort, it was shown that CYP2C9*1/ warfarin dose, respectively.24 In all studies, the overall variance *1 individuals (extensive metabolizers) who were homo- in warfarin dose accounted for by CYP2C9*2 and CYP2C9*3 zygous for APOE*E4 were given significantly higher warfarin was below 20%. Moreover, CYP2C9 alleles *4 (identified in the doses than other CYP2C9 extensive metabolizers.66 In Japanese), *5 and *6 (found in Afro-Americans) and *11 (rare in agreement with this, a Dutch study showed that APOE*E4

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 104

Table 2 A selection of studies on CYP2C9 polymorphisms in warfarin-treated patients

Reference Patient populations n Warfarin dose requirement Risk of adverse event

Furuya et al. (1995)6 British Caucasian 94 CYP2C9*2 associated with low dose Not studied Steward et al. (1997)7 American Caucasian 1 CYP2C9*3 associated with low dose High initial INR Aithal et al. (1999)114 British Caucasian 88 CYP2C9*2 and *3 associated with Low-dose requirement associated with low dose raised INR during induction and risk of major bleeding Ogg et al. (1999)8 British 233 CYP2C9*3 associated with low dose CYP2C9*3 associated with risk of early bleeding Taube et al. (2000)33 British 561 CYP2C9*2 and *3 associated with No increased risk of severe over- low dose anticoagulation or bleeding Margaglione et al. Italian Caucasian 180 CYP2C9*2 and *3 associated with Variant alleles associated with risk of (2000)34 low dose bleeding Freeman et al. American Caucasian 78% 38 CYP2C9*2 and *3 associated with Not studied (2000)35 Afro-American 22% low dose Loebstien et al. Israeli 156 CYP2C9*2 and *3 associated with Not studied (2001)3 low dose Leung et al. (2001)127 Hong Kong Chinese 89 CYP2C9 208Val carriers require a Not studied lower dose Tabrizi et al. (2002)36 American Caucasian 81% 153 CYP2C9*2 and *3 associated with Not studied Afro-American 19% low dose Higashi et al. (2002)5 American Caucasian 91% 185 CYP2C9*2 and *3 associated with Variant alleles associated with high Asian 4% low dose initial INR, longer time to stable INR Afro-American 3% and risk of serious bleeding Hispanic 2% Scordo et al. (2002)2 Italian Caucasian 93 CYP2C9*2 and *3 associated with Not studied low dose and S-warfarin-clearance Wadelius et al. Swedish Caucasian 225 CYP2C9*2 and *3 associated with Variant alleles not associated with (2004)1 low dose bleeding Peyvandi et al. Italian 125 CYP2C9*2 and *3 associated with Variant alleles associated with INR43 (2004)37 low dose during induction Joffe et al. (2004)115 American Caucasian 73 CYP2C9*2 and *3 associated with A tendency to an association between low dose variant alleles and INR46, but no association with bleeding Lindh et al. (2005)38 Swedish 219 CYP2C9*2 and *3 associated with Variant alleles associated with INR43 low dose during induction

INR, international normalized ratio.

carriers required slightly higher maintenance doses of the vitamin K epoxide reductase (Figure 1).57 In 2004, the gene anticoagulant phenprocoumon, but surprisingly carriers of encoding this enzyme was identified as VKORC1.69,70 Rare APOE*E4 required lower maintenance doses of acenocou- mutations in the human VKORC1 gene that convey marol.67 In Italian patients, where the E4 allele is rare, no resistance to warfarin have been identified.69,71 Further- association was found between warfarin dose requirements more, a number of studies have shown that common SNPs and APOE genotype.68 The contradictory results of these in VKORC1 are strongly associated with warfarin mainte- candidate gene association studies reflect the lack of a clear nance dose in several populations (Table 3).72–83 Addition- description of the exact role of different APOE genotypes in ally, a similar relationship has been demonstrated with the vitamin K uptake and intracellular handling. other vitamin K antagonists acenocoumarol and phenprocoumon,84,85 with one study suggesting an association with coumarin-related bleeding.85 The associated VKORC1 SNPs The vitamin K cycle are within a region of strong linkage disequilibrium, and a combination of several SNPs does not contribute greater Genes involved in the vitamin K cycle have recently been information than one individual SNP.73,74,81 The molecular shown to be crucial determinants of the response to mechanism by which VKORC1 polymorphisms lead to warfarin. Warfarin and other vitamin K antagonists exert variation in response to warfarin has not been resolved. It their anticoagulant effects by preventing the regeneration of has been suggested that VKORC1 is regulated at a transcrip- vitamin K from vitamin K epoxide by inhibiting the enzyme tional level, as at least one of five correlated SNPs is a

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 105

promoter polymorphism (À1639G4A, rs9923231) and is be complexed with microsomal epoxide hydrolase (encoded associated with low mRNA levels in liver specimens.74 In by EPHX1), to produce a multiprotein complex that is addition, a study has shown that a VKORC1 promoter responsible for vitamin K epoxide reduction.86,87 Micro- cloned into a human hepatoma cell line was 44% more somal epoxide hydrolase by itself does not possess vitamin K active if it contained the wild-type À1639G than the A epoxide reductase activity.88 Interestingly, a recent study in allelic variant.75 This is biologically plausible given that the an Israeli population has shown an association between SNP resides within an E-box site, which can be important in high doses of warfarin and a coding EPHX1 polymorphism determining tissue-specific transcription.75 However, no (rs1051740) in CYP2C9 extensive metabolizers.89 However, study has yet demonstrated that the change in mRNA levels the authors’ contention that this EPHX1 polymorphism is associated with a change in protein levels or indeed in leads to high-dose requirements beyond the effect of functional activity in vitro. The VKORC1 gene is located in a CYP2C9 needs to be replicated in another population. large haplotype block where there are numerous SNPs that Moreover, it has been suggested that the antioxidant potentially could mediate the functional effect, and even enzyme nicotine adenine dinucleotide phosphate though the evidence suggests transcriptional regulation of (NAD(P)H) dehydrogenase, also called flavoprotein DT- VKORC1, no single causative SNP or set of SNPs has been diaphorase, has the potential to reduce dietary vitamin identified. In addition, there is no definite proof that K.62,90,91 Its gene, NQO1 (NAD(P)H dehydrogenase, quinone transcriptional regulation is involved, as another study 1 gene), has not been studied with respect to warfarin utilizing transfected HepG2 cells failed to show that a treatment. construct containing the G variant had activity that was Reduced vitamin K is an essential cofactor for the different from that containing the A variant.84 activation of vitamin K-dependent proteins by gamma- Although the gene for VKORC1 has been identified, the glutamyl carboxylase.92–94 Gamma-glutamyl carboxylase is mechanism by which it functions as a reductase is unclear. an integral endoplasmic reticulum protein that localizes in The protein resides in the endoplasmic reticulum, and may close proximity to the vitamin K epoxide reductase

Table 3 A selection of studies on VKORC1 polymorphisms in warfarin-treated patients

Reference Patient populations n Warfarin dose requirement Risk of adverse event

Rost et al. (2004)69 Warfarin resistant 6 Coding polymorphisms increase dose Not studied requirement in warfarin resistant D’Andrea et al. (2005)72 Italian Caucasian 147 1173C4T associated with dose Not studied Harrington et al. (2005)71 Warfarin resistant 4 Coding polymorphisms increase dose Not studied requirement in warfarin resistant Wadelius et al. (2005)73 Swedish Caucasian 201 À1639G4A, 1173C4T and 2255C4T Not studied associated with dose Rieder et al. (2005)74 American Caucasian 554 Haplotypes defined by À4451C4A, 497T4G, Not studied 1542G4C and 3730G4A associated with dose Yuan et al. (2005)75 Taiwan Chinese 120 À1639G4A associated with dose Not studied Sconce et al. (2005)76 British Caucasian 335 À1639G4A associated with dose Not studied Veenstra et al. (2005)77 Hong Kong Chinese 69 Haplotypes defined by À4451C4A, 497T4G, Not studied 1542G4C and 3730G4A associated with dose Geisen et al. (2005)78 European warfarin resistant 12 À1639G4A and 1173C4T associated with Not studied dose in warfarin resistant Vecsler et al. (2006)79 Israeli 100 1542G4C associated with dose Not studied Mushiroda et al. (2006)80 Japanese 828 À1639G4A, 1173C4T, 1542G4C, 2255C4T Not studied and 3730G4A associated with dose Takahashi et al. (2006)81 American Caucasian 47% 243 1173C4T associated with dose Not studied Japanese 26% Afro-American 26% Lee et al. (2006)82 Singapore Chinese 53% 275 Haplotypes defined by À4931T4C, Not studied Malay 31% À1639G4A, 1173C4T, 1542G4C and Indian 16% 2255C4T associated with dose Aquilante et al. (2006)83 American Caucasian 91% 350 À1639G4A is associated with dose Not studied Afro-American 7% Hispanic 1% Asian 0.3%

Translation into rs numbers: À4451C4A ¼ rs17880887, À1639G4A ¼ rs9923231, 497T4G ¼ rs2884737, 1173C4T ¼ rs9934438, 1542G4C ¼ rs8050894, 2255C4T ¼ rs2359612, 3730G4A ¼ rs7294.

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 106

complex. A very rare autosomal recessive bleeding disorder exaggerated in patients with a hereditary deficiency of due to combined deficiency of the vitamin K-dependent protein C or S, which leads to a relative hypercoagulable coagulation factors II , VII, IX and X, and proteins C, S and Z state at the start of warfarin treatment.106,107 Rare genetic is caused by mutations in the gamma-glutamyl carboxylase variants of PROC and PROS1 did not, however, affect gene (GGCX).92,95 SNPs as well as microsatellite markers that warfarin dose requirement in a Japanese population.96 Two might affect warfarin dosing have recently been identified in vitamin K-dependent proteins, encoded by PROZ and GAS6, the GGCX gene. For example, an intronic polymorphism have not been studied with respect to warfarin sensitivity. that increases warfarin dose requirements was identified in a Two non-vitamin K-dependent clotting proteins of interest Swedish population.73 A microsatellite in intron 6 has been for warfarin pharmacogenetics are anti-thrombin III and associated with warfarin dose in the Japanese;96 a similar factor V. Anti-thrombin III inhibits factors II, IX, X, XI and analysis in a Swedish population showed that warfarin dose XII, and anti-thrombin III deficiency, both the congenital requirements increase with the number of microsatellite form caused by mutations in SERPINC1 (anti-thrombin III repeats.97 On the other hand, a coding polymorphism gene) and the acquired form, may create a hypercoagulable (rs699664) that leads to a change from arginine to state during warfarin induction.107,108 A point mutation glutamine at residue 325 is not associated with warfarin in the factor V gene (Arg506Gln or FV Leiden), which sensitivity or resistance.73,89 Taken as a whole, the effect of commonly causes thromboembolism and warfarin treat- GGCX seems to be rather modest.96,97 ment, is not known to affect dose requirement.109 The endoplasmic reticulum chaperone protein calumenin, encoded by CALU (calumenin gene), can bind to the vitamin K cycle and inhibit its activity.98,99 It has been Alternative approaches shown that silencing of the CALU gene with small interfering RNA results in a fivefold increase in gamma- Although we understand a lot about the pharmacokinetics carboxylase.99 Furthermore, overexpression of calumenin and dynamics of warfarin (Figure 1), it is possible, and in the liver produces warfarin resistance in rats by protecting indeed likely, that other genes are involved in the outcome vitamin K epoxide reductase from inhibition by warfarin.98 of treatment. Such genes may act in trans (e.g. transcription Whether this is a mechanism of warfarin resistance in man factors) and may therefore not be identified by the is unknown at present, particularly as calumenin is candidate gene approach. Owing to recent advances in expressed at low levels in the human liver. Only one coding genotyping technologies, it is now feasible to find these polymorphism in the human CALU gene (rs2290228) has so other genes through genome-wide association studies. far been related to warfarin dose requirements.79 Compared with targeted analysis of candidate genes based on the known actions and metabolism of warfarin, a genome-wide approach is advantageous because (a) it has a Vitamin K-dependent proteins better chance of identifying previously unknown genes that influence warfarin therapy and (b) the cost and effort per Many vitamin K-dependent proteins have been implicated genotype produced is significantly lower than for the in warfarin sensitivity. The main vitamin K-dependent analysis of a limited number of candidate genes. However, proteins are clotting factors II (prothrombin), VII, IX and there is a need for large sample sizes to ensure adequate X, proteins C, S and Z and growth-arrest-specific protein 6, statistical power (which in effect renders these studies encoded by F2, F7, F9, F10, PROC, PROS1, PROZ and expensive), and better statistical approaches need to be GAS6.62,93,100 Two independent studies have shown that a developed. polymorphism in F2 causing a change from threonine to methionine at residue 165 leads to increased sensitivity to warfarin,96,101 whereas a third study did not show this.83 It Future challenges for clinical practice has also been shown that promoter polymorphisms in F7 have an effect on warfarin sensitivity.83,96,101 Mutations in The studies discussed above clearly show that genetic the propeptide of F9, causing a change from alanine to variation, especially in CYP2C9 and VKORC1, is extremely valine or threonine at residue À10, lead to a rapid drop in important for the variability in the response to warfarin. factor IX during warfarin treatment and are the reason for Polymorphisms in the VKORC1 and CYP2C9 genes and a bleeding in rare cases.102,103 Promoter polymorphisms and a limited subset of environmental determinants account for synonymous coding polymorphism in exon 7 of F10 have around 50–60% of the variance in warfarin dose require- also been studied, but no effect on warfarin sensitivity was ment.73,76,77,79,81–83 In six studies, the relative contribution seen.83,96 of VKORC1 is greater than that of CYP2C9,73,74,78,81–83 in Unlike other vitamin K-dependent factors, protein C and S two, CYP2C9 has a greater quantitative contribution,72,76 work as natural anticoagulants. After administration of whereas VKORC1 and CYP2C9 contribute equally in one warfarin, protein C and S decline more rapidly than other study.79 Sconce et al.76 have gone on to develop a dosing vitamin K-dependent proteins, and this may contribute to table based on a regression equation combining age, height the poor antithrombotic efficacy during the first day of and CYP2C9*2 and CYP2C9*3, and the VKORC1 SNP anticoagulant therapy.104,105 The temporary imbalance is À1639G4A.

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 107

The data from various pharmacogenetic studies worldwide Fourth, genetic individualization of warfarin therapy have been considered by the FDA in an open hearing needs to be shown to be cost-effective. If it greatly adds to (http://www.fda.gov/ohrms/dockets/ac/05/slides/2005-4194S1_ the cost of treating patients, and given the huge usage of Slide-Index.htm).73,74,110,111 The interesting questions are warfarin in the general population, it may be difficult to whether this will lead to a recommendation for genotyping persuade health-care organizations to fund genetic testing. in the label for warfarin, and if this would change clinical Hopefully, the rapid development of genetic technology will practice and, more importantly, improve the use and safety lead to more sophisticated assays at a lower cost, and this is of warfarin. Before these questions can be answered several likely to facilitate incorporation of genetic analyses in important issues need to be considered. clinical practice. A small retrospective study has already First, the estimates for the variance in warfarin dosing suggested that CYP2C9 genotyping is potentially effective in have been derived from retrospective studies in homoge- preventing bleeding with a marginal cost.120 However, this neous populations. Thus, it is unclear how a combined needs to be performed in a larger study, and is currently variance of 55–60% will translate into predictive values in being assessed as part of the UK prospective study (see diverse populations. Furthermore, the retrospective nature below). of the studies undertaken so far is likely to underestimate Fifth, it has been stated that regulation is likely to the environmental contribution and overestimate the be the key factor that will drive the implementation of genetic contribution. This is a consistent feature of genetic pharmacogenetics into clinical practice. This is true association studies, which is perhaps best exemplified by the to an extent. The possibility that pharmacogenetic informa- association of ACE gene polymorphisms and risk of tion is going to be incorporated into the warfarin label myocardial infarction.112 is an important development. A similar example is Second, it could be argued that the maintenance dose of azathioprine, which for a long time has been known to be warfarin could be achieved rapidly by more intensive metabolized by thiopurine methyltransferase (TPMT), monitoring particularly in specialist anticoagulant clinics. which is polymorphically expressed, with low expressers However, this has not been studied in comparison to genetic being at higher risk of leukopenia. Although the poly- individualization. The crucial issue to assess here is how morphic metabolism is mentioned in the label for azathio- closely the induction dose predicts the maintenance dose. prine, there is no mandatory statement regarding dose Encouragingly, a randomized study of 5 mg of warfarin as a individualization according to genotype or phenotype. starting dose versus an initial dose calculated on the basis A Europe-wide survey has shown that TPMT testing before of weight, age, serum albumin and presence of malignancy azathioprine use occurs in only about 12% of cases,121 resulted in the latter regimen leading to a slightly but and in Australasia, pharmacogenetic testing for drug significantly quicker time to onset of anticoagulation (5 metabolizing enzymes are performed rarely in clinical versus 4.2 days).113 Genetic individualization of dose might practice.122 Pharmacogenetic labelling is in these cases for further speed up this process. information only and not mandatory, and the absence of Third, the ultimate aim of individualizing warfarin dosing clear guidelines may lessen the probability that the test is is not only to improve the stability of anticoagulation used. Many factors are needed for regulators to change the control, but also to reduce the risk of bleeding with warfarin. nature of the warnings in the product label, the most Some,5,8,34,114 but not all,1,33,115 studies have shown an important of which are a strong research base and good association between bleeding and genetic factors such as evidence of clinical relevance.122 Other factors may also be CYP2C9 polymorphisms (Table 2). Prospective and retro- important including overcoming financial and perception spective data have shown that the intensity of anticoagula- barriers, education regarding pharmacogenetics and ade- tion and deviation in anticoagulation control are the quate information on the benefits of pre-prescription strongest predictors for the risk of bleeding.116 It is likely testing.121 Such a multi-pronged approach is going to be to be more difficult to consistently show an association needed to incorporate pharmacogenetics into the prescrib- between genetic factors and warfarin-related bleeding ing of warfarin. because (a) it is relatively uncommon and therefore most Finally, various other strategies have been suggested of the studies are under-powered with respect to bleeding to improve the safety of anticoagulation therapy including as an end point, (b) there are differences between studies in computer decision support systems,123 the use of the definitions used for the severity of bleeding, (c) some patient self-monitoring devices124 and the use of drugs that patients bleed at normal INR values,13 and in these patients inhibit other targets in the anticoagulation pathway, for in particular, there may be underlying causes such as example the oral thrombin inhibitors.125 Whether we tumours,117,118 and (d) there may be other co-incidental should use pharmacogenetic-based warfarin therapy in genetic polymorphisms that contribute to the risk of competition or in conjunction with these other approaches bleeding, for example those involved in platelet aggrega- is not clear. tion.119 Nevertheless, a meta-analysis of CYP2C9 genetic To address these issues, as clinicians, we feel that it will be polymorphisms showed that the relative bleeding risk for necessary to undertake large prospective studies of variation CYP2C9*2 was 1.91 (95% CI 1.16–3.17) and for CYP2C9*3 in response to warfarin therapy. It has clearly been shown 1.77 (95% CI 1.07–2.91).24 For either variant, the relative that prospective randomized controlled trials based on risk was 2.26 (95% CI 1.36–3.75). CYP2C9 genotyping are feasible.111,126 It should therefore

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 108

be possible to conduct randomized controlled trials based warfarin, and translation of this knowledge into clinical on CYP2C9 and VKORC1 polymorphisms. Concerning other guidelines, is likely to have a major impact on the safety of genes outlined in this review, it is important to note that warfarin. two studies in Swedish patients (one in 200 patients and another in 1500 patients) are examining polymorphisms in all these pathways in collaboration with the Sanger Abbreviations Institute, UK, and are due to report soon. These studies are ABCB1 ATP-binding cassette transporter B1 gene, retrospective and will not be able to determine the relative P-glycoprotein gene or MDR1 contributions of genetic and environmental factors and the APOE apolipoprotein E gene interaction between them. These questions are, however, CALU calumenin gene likely to be answered in a prospective study of up to 2000 CAR constitutive androstane receptor CYP1A1 cytochrome P450 1A1 gene patients that is currently ongoing in the UK (http:// CYP1A2 cytochrome P450 1A2 gene www.genres.org.uk/prp/projectsliverpool2.htm). This study CYP2A6 cytochrome P450 2A6 gene is not only looking at all the genes mentioned here, but it is CYP2C18 cytochrome P450 2C18 gene assessing environmental factors including the clinical (age, CYP2C19 cytochrome P450 2C19 gene CYP2C8 cytochrome P450 2C8 gene gender, ethnicity, disease, concurrent medication, adher- CYP2C9 cytochrome P450 2C9 gene ence to treatment), pharmacological (R- and S-warfarin CYP3A4 cytochrome P450 3A4 gene levels), biochemical (vitamin K and epoxide levels) and CYP3A5 cytochrome P450 3A5 gene haematological (clotting factor levels) phenotypes. The EPHX1 epoxide hydrolase 1, microsomal gene F2 coagulation factor II gene or prothrombin gene study will be able to assess the cost-effectiveness of pre- F5 coagulation factor V gene prescription genotyping, and provide values for positive and F7 coagulation factor VII gene negative prediction, and numbers needed to screen. These F9 coagulation factor IX gene developments will provide the necessary framework to F10 coagulation factor X gene FII coagulation factor II or prothrombin undertake prospective randomized controlled trials to assess FIIa coagulation factor II activated or thrombin the clinical utility of pre-prescription genotyping for FIX coagulation factor IX warfarin. FIXa coagulation factor IX activated FV coagulation factor V FVII coagulation factor VII FVIIa coagulation factor VII activated FX coagulation factor X Conclusion and summary FXa coagulation factor X activated GAS6 growth-arrest-specific 6 gene GGCX gamma-glutamyl carboxylase gene Despite the fact that warfarin is an old drug, there is MDR1 multidrug resistance gene 1, P-glycoprotein gene or ABCB1 currently unprecedented interest in the pharmacology and NQO1 NAD(P)H dehydrogenase, quinone 1 gene effectiveness of warfarin, which is partly due to a general NR1I2 pregnane X receptor gene interest in whether pharmacogenetics can improve the use NR1I3 constitutive androstane receptor gene ORM1 orosomucoid 1 gene or alpha-1-acid glycoprotein 1 gene of common medicines. This research has led to the ORM2 orosomucoid 2 gene or alpha-1-acid glycoprotein 2 gene identification of striking genetic predisposing factors in PROC protein C gene two genes, CYP2C9 and VKORC1, explaining a large part of PROS1 protein S gene the interindividual variation in warfarin dose requirement. PROZ protein Z gene PT INR prothrombin time international normalized ratio To what extent variability in other genes in the warfarin PXR pregnane X receptor interactive pathways influences warfarin therapy remains to SERPINC1 anti-thrombin III gene be resolved. Most studies to date have had an inadequate SNP single nucleotide polymorphism sample size to be able to detect small genetic effects in these VKORC1 vitamin K epoxide reductase complex subunit 1 gene other genes, and the findings highlighted in some of the genetic association studies may thus be due to pure chance. For the intraindividual variation in warfarin dose, environ- Acknowledgments mental factors such as the intake of vitamin K and interacting medications will be more important than The support of the UK Department of Health, which is funding the genetic factors. Whether the identified genetic and environ- prospective UK warfarin pharmacogenetics study, is gratefully acknowledged. The Uppsala warfarin study is supported by the mental factors will improve the use and safety of warfarin in Swedish Society of Medicine, Foundation for Strategic Research, clinical practice is unclear, but is likely to be resolved in the Heart and Lung Foundation and the Clinical Research Support (ALF) next couple of years with ongoing and newly planned at Uppsala University. The support of David Bentley and the studies of all known warfarin interactive pathways. Irrespec- Wellcome Trust Sanger Institute is acknowledged. The sponsors tive of whether pre-prescription genotyping impacts directly had no role in the writing of this review. on the use of warfarin, we are learning much more about the pharmacology of warfarin because of the current interest in Duality of Interest warfarin pharmacogenetics. An indirect benefit of this will be an increase in the knowledge of how to prescribe None.

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 109

References netics and MDR1 gene expression. Br J Clin Pharmacol 2005; 60: 159– 171. 1 Wadelius M, So¨rlin K, Wallerman O, Karlsson J, Yue QY, Magnusson PK 23 Kirchheiner J, Brockmoller J. Clinical consequences of cytochrome et al. Warfarin sensitivity related to CYP2C9, CYP3A5, ABCB1 (MDR1) P450 2C9 polymorphisms. Clin Pharmacol Ther 2005; 77: 1–16. and other factors. Pharmacogenomics J 2004; 4: 40–48. 24 Sanderson S, Emery J, Higgins J. CYP2C9 gene variants, drug dose, and 2 Scordo MG, Pengo V, Spina E, Dahl ML, Gusella M, Padrini R. Influence bleeding risk in warfarin-treated patients: a HuGEnet systematic review of CYP2C9 and CYP2C19 genetic polymorphisms on warfarin and meta-analysis. Genet Med 2005; 7: 97–104. maintenance dose and metabolic clearance. Clin Pharmacol Ther 25 Linder MW. Genetic mechanisms for hypersensitivity and resistance to 2002; 72: 702–710. the anticoagulant Warfarin. Clin Chim Acta 2001; 308: 9–15. 3 Loebstein R, Yonath H, Peleg D, Almog S, Rotenberg M, Lubetsky A et 26 Rettie AE, Korzekwa KR, Kunze KL, Lawrence RF, Eddy AC, Aoyama T et al. Interindividual variability in sensitivity to warfarin – nature or al. Hydroxylation of warfarin by human cDNA-expressed cytochrome nurture? Clin Pharmacol Ther 2001; 70: 159–164. P-450: a role for P-4502C9 in the etiology of (S)-warfarin–drug 4 Takahashi H, Echizen H. Pharmacogenetics of CYP2C9 and inter- interactions. Chem Res Toxicol 1992; 5: 54–59. individual variability in anticoagulant response to warfarin. Pharmaco- 27 Takahashi H, Echizen H. Pharmacogenetics of warfarin elimination and genomics J 2003; 3: 202–214. its clinical implications. Clin Pharmacokinet 2001; 40: 587–603. 5 Higashi M, Veenstra D, Kondo L, Wittkowsky A, Srinouanprachanh S, 28 Kaminsky L, Zhang Z. Human P450 metabolism of warfarin. Pharmacol Farin F et al. Association between CYP 2C9 genetic variants and Ther 1997; 73: 67–74. anticoagulation-related outcomes during warfarin treatment. JAMA 29 Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. Impaired 2002; 287: 1690–1698. (S)-warfarin metabolism catalysed by the R144C allelic variant of 6 Furuya H, Fernandez-Salguero P, Gregory W, Taber H, Steward A, CYP2C9. Pharmacogenetics 1994; 4: 39–42. Gonzalez FJ et al. Genetic polymorphism of CYP2C9 and its effect on 30 Haining RL, Hunter AP, Veronese ME, Trager WF, Rettie AE. Allelic warfarin maintenance dose requirement in patients undergoing variants of human cytochrome P450 2C9: baculovirus-mediated anticoagulation therapy. Pharmacogenetics 1995; 5: 389–392. expression, purification, structural characterization, substrate stereo- 7 Steward DJ, Haining RL, Henne KR, Davis G, Rushmore TH, Trager WF selectivity, and prochiral selectivity of the wild-type and I359L mutant et al. Genetic association between sensitivity to warfarin and forms. Arch Biochem Biophys 1996; 333: 447–458. expression of CYP2C9*3. Pharmacogenetics 1997; 7: 361–367. 31 Crespi CL, Miller VP. The R144C change in the CYP2C9*2 allele alters 8 Ogg MS, Brennan P, Meade T, Humphries SE. CYP2C9*3 allelic variant interaction of the cytochrome P450 with NADPH:cytochrome P450 and bleeding complications. Lancet 1999; 354: 1124. oxidoreductase. Pharmacogenetics 1997; 7: 203–210. 9 Horton JD, Bushwick BM. Warfarin therapy: evolving strategies in 32 Aithal GP, Day CP, Kesteven PJ, Daly AK. Association of polymorphisms anticoagulation. Am Fam Physician 1999; 3: 635–646. in the cytochrome P450 CYP2C9 with warfarin dose requirement and 10 Aguilar MI, Hart R. Oral anticoagulants for preventing stroke in risk of bleeding complications. Lancet 1999; 353: 717–719. patients with non-valvular atrial fibrillation and no previous history of 33 Taube J, Halsall D, Baglin T. Influence of cytochrome P-450 CYP2C9 stroke or transient ischemic attacks. Cochrane Database Syst Rev 2005; polymorphisms on warfarin sensitivity and risk of over-anticoagulation 3: CD001927. in patients on long-term treatment. Blood 2000; 96: 1816–1819. 11 Jones M, McEwan P, Morgan CL, Peters JR, Goodfellow J, Currie CJ. 34 Margaglione M, Colaizzo D, D’Andrea G, Brancaccio V, Ciampa A, Evaluation of the pattern of treatment, level of anticoagulation control, Grandone E et al. Genetic modulation of oral anticoagulation with and outcome of treatment with warfarin in patients with non-valvar warfarin. Thromb Haemost 2000; 84: 775–778. atrial fibrillation: a record linkage study in a large British population. 35 Freeman BD, Zehnbauer BA, McGrath S, Borecki I, Buchman TG. Heart 2005; 91: 472–477. Cytochrome P450 polymorphisms are associated with reduced 12 Linkins LA, Choi PT, Douketis JD. Clinical impact of bleeding in patients warfarin dose. Surgery 2000; 128: 281–285. taking oral anticoagulant therapy for venous thromboembolism: a 36 Tabrizi AR, Zehnbauer BA, Borecki IB, McGrath SD, Buchman TG, meta-analysis. Ann Intern Med 2003; 139: 893–900. Freeman BD. The frequency and effects of cytochrome P450 (CYP) 13 Fanikos J, Grasso-Correnti N, Shah R, Kucher N, Goldhaber SZ. Major 2C9 polymorphisms in patients receiving warfarin. J Am Coll Surg 2002; bleeding complications in a specialized anticoagulation service. Am J 194: 267–273. Cardiol 2005; 96: 595–598. 37 Peyvandi F, Spreafico M, Siboni SM, Moia M, Mannucci PM. CYP2C9 14 Daly AK, King BP. Pharmacogenetics of oral anticoagulants. Pharma- genotypes and dose requirements during the induction phase of oral cogenetics 2003; 13: 247–252. anticoagulant therapy. Clin Pharmacol Ther 2004; 75: 198–203. 15 Palareti G, Legnani C. Warfarin withdrawal. Pharmacokinetic– 38 Lindh JD, Lundgren S, Holm L, Alfredsson L, Rane A. Several-fold pharmacodynamic considerations. Clin Pharmacokinet 1996; 30: increase in risk of overanticoagulation by CYP2C9 mutations. Clin 300–313. Pharmacol Ther 2005; 78: 540–550. 16 Otagiri M, Maruyama T, Imai T, Suenaga A, Imamura Y. A comparative 39 Schwarz UI. Clinical relevance of genetic polymorphisms in the human study of the interaction of warfarin with human alpha 1-acid glycoprotein CYP2C9 gene. Eur J Clin Invest 2003; 33(Suppl 2): 23–30. and human albumin. JPharmPharmacol1987; 39: 416–420. 40 Tai G, Farin F, Rieder MJ, Dreisbach AW, Veenstra DL, Verlinde CL et al. 17 Nakagawa T, Kishino S, Itoh S, Sugawara M, Miyazaki K. Differential In-vitro and in-vivo effects of the CYP2C9*11 polymorphism on binding of disopyramide and warfarin enantiomers to human warfarin metabolism and dose. Pharmacogenet Genomics 2005; 15: alpha(1)-acid glycoprotein variants. Br J Clin Pharmacol 2003; 56: 475–481. 664–669. 41 Takahashi H, Kashima T, Nomizo Y, Muramoto N, Shimizu T, Nasu K et 18 Sussman N, Waltershied M, Butler T, Cali J, Riss T, Kelly J. The al. Metabolism of warfarin enantiomers in Japanese patients with heart predictice nature of high throughput toxicity screening using a human disease having different CYP2C9 and CYP2C19 genotypes. Clin hepatocyte cell line. Cell Notes 2002; 3: 7–10. Pharmacol Ther 1998; 63: 519–528. 19 Ishikawa T, Hirano H, Onishi Y, Sakurai A, Tarui S. Functional 42 Tabrizi AR, McGrath SD, Blinder MA, Buchman TG, Zehnbauer BA, evaluation of ABCB1 (P-glycoprotein) polymorphisms: high-speed Freeman BD. Extreme warfarin sensitivity in siblings associated with screening and structure–activity relationship analyses. Drug Metab multiple cytochrome P450 polymorphisms. Am J Hematol 2001; 67: Pharmacokinet 2004; 19: 1–14. 144–146. 20 Kroetz DL, Pauli-Magnus C, Hodges LM, Huang CC, Kawamoto M, 43 Zhang Z, Fasco MJ, Huang Z, Guengerich FP, Kaminsky LS. Human Johns SJ et al. Sequence diversity and haplotype structure in the cytochromes P4501A1 and P4501A2: R-warfarin metabolism as a human ABCB1 (MDR1, multidrug resistance transporter) gene. probe. Drug Metab Dispos 1995; 23: 1339–1346. Pharmacogenetics 2003; 13: 481–494. 44 Kaminsky LS, de Morais SM, Faletto MB, Dunbar DA, Goldstein JA. 21 Fischer V, Einolf HJ, Cohen D. Efflux transporters and their clinical Correlation of human cytochrome P4502C substrate specificities with relevance. Mini Rev Med Chem 2005; 5: 183–195. primary structure: warfarin as a probe. Mol Pharmacol 1993; 43: 234–239. 22 Chowbay B, Li H, David M, Cheung YB, Lee EJ. Meta-analysis of the 45 Grossman SJ, Herold EG, Drey JM, Alberts DW, Umbenhauer DR, influence of MDR1 C3435T polymorphism on digoxin pharmacoki- Patrick DH et al. CYP1A1 specificity of Verlukast epoxidation in mice,

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 110

rats, rhesus monkeys, and humans. Drug Metab Dispos 1993; 21: 69 Rost S, Fregin A, Ivaskevicius V, Conzelmann E, Hortnagel K, Pelz HJ 1029–1036. et al. Mutations in VKORC1 cause warfarin resistance and 46 Huang W, Lin YS, McConn II DJ, Calamia JC, Totah RA, Isoherranen N multiple coagulation factor deficiency type 2. Nature 2004; 427: et al. Evidence of significant contribution from CYP3A5 to hepatic drug 537–541. metabolism. Drug Metab Dispos 2004; 32: 1434–1445. 70 Li T, Chang CY, Jin DY, Lin PJ, Khvorova A, Stafford DW. Identification 47 Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J et al. Sequence of the gene for vitamin K epoxide reductase. Nature 2004; 427: diversity in CYP 3A promoters and characterization of the genetic basis 541–544. of polymorphic CYP 3A5 expression. Nat Genet 2001; 27: 383–391. 71 Harrington DJ, Underwood S, Morse C, Shearer MJ, Tuddenham EG, 48 Daly AK, Aithal GP. Genetic regulation of warfarin metabolism and Mumford AD. Pharmacodynamic resistance to warfarin associated response. Semin Vasc Med 2003; 3: 231–238. with a Val66Met substitution in vitamin K epoxide reductase complex 49 Wang Z, Gorski JC, Hamman MA, Huang SM, Lesko LJ, Hall SD. The subunit 1. Thromb Haemost 2005; 93: 23–26. effects of St John’s wort (Hypericum perforatum) on human 72 D’Andrea G, D’Ambrosio RL, Di Perna P, Chetta M, Santacroce R, cytochrome P450 activity. Clin Pharmacol Ther 2001; 70: 317–326. Brancaccio V et al. A polymorphism in the VKORC1 gene is associated 50 Moore LB, Goodwin B, Jones SA, Wisely GB, Serabjit-Singh CJ, Willson with an interindividual variability in the dose-anticoagulant effect of TM et al. St John’s wort induces hepatic drug metabolism through warfarin. Blood 2005; 105: 645–649. activation of the pregnane X receptor. Proc Natl Acad Sci USA 2000; 97: 73 Wadelius M, Chen LY, Downes K, Ghori J, Hunt S, Eriksson N et al. 7500–7502. Common VKORC1 and GGCX polymorphisms associated with 51 Lehmann JM, McKee DD, Watson MA, Willson TM, Moore JT, Kliewer warfarin dose. Pharmacogenomics J 2005; 5: 262–270. SA. The human orphan nuclear receptor PXR is activated by 74 Rieder MJ, Reiner AP, Gage BF, Nickerson DA, Eby CS, McLeod HL et al. compounds that regulate CYP3A4 gene expression and cause drug Effect of VKORC1 haplotypes on transcriptional regulation and interactions. J Clin Invest 1998; 102: 1016–1023. warfarin dose. N Engl J Med 2005; 352: 2285–2293. 52 Chen Y, Ferguson SS, Negishi M, Goldstein JA. Induction of 75 Yuan HY, Chen JJ, Lee MT, Wung JC, Chen YF, Charng MJ et al. A novel human CYP2C9 by rifampicin, hyperforin, and phenobarbital is functional VKORC1 promoter polymorphism is associated with inter- mediated by the pregnane X receptor. J Pharmacol Exp Ther 2004; individual and inter-ethnic differences in warfarin sensitivity. Hum Mol 308: 495–501. Genet 2005; 14: 1745–1751. 53 Geick A, Eichelbaum M, Burk O. Nuclear receptor response elements 76 Sconce EA, Khan TI, Wynne HA, Avery P, Monkhouse L, King BP et al. mediate induction of intestinal MDR1 by rifampin. J Biol Chem 2001; The impact of CYP2C9 and VKORC1 genetic polymorphism and 276: 14581–14587. patient characteristics upon warfarin dose requirements: proposal for a 54 Assenat E, Gerbal-Chaloin S, Larrey D, Saric J, Fabre JM, Maurel P et al. new dosing regimen. Blood 2005; 106: 2329–2333. Interleukin 1beta inhibits CAR-induced expression of hepatic genes 77 Veenstra DL, You JH, Rieder MJ, Farin FM, Wilkerson HW, Blough DK et involved in drug and bilirubin clearance. Hepatology 2004; 40: 951– al. Association of vitamin K epoxide reductase complex 1 (VKORC1) 960. variants with warfarin dose in a Hong Kong Chinese patient 55 Bell RG. Metabolism of vitamin K and prothrombin synthesis: antic- population. Pharmacogenet Genomics 2005; 15: 687–691. oagulants and the vitamin K–epoxide cycle. Fed Proc 1978; 37: 2599– 78 Geisen C, Watzka M, Sittinger K, Steffens M, Daugela L, Seifried E et al. 2604. VKORC1 haplotypes and their impact on the inter-individual and inter- 56 Bell RG, Sadowski JA, Matschiner JT. Mechanism of action of warfarin. ethnical variability of oral anticoagulation. Thromb Haemost 2005; 94: Warfarin and metabolism of vitamin K 1. Biochemistry 1972; 11: 1959– 773–779. 1961. 79 Vecsler M, Loebstein R, Almog S, Kurnik D, Goldman B, Halkin H et al. 57 Sadler JE. Medicine: K is for koagulation. Nature 2004; 427: 493–494. Combined genetic profiles of components and regulators of the 58 Shetty HG, Woods F, Routledge PA. The pharmacology of oral vitamin K-dependent gamma-carboxylation system affect individual anticoagulants: implications for therapy. J Heart Valve Dis 1993; 2: sensitivity to warfarin. Thromb Haemost 2006; 95: 205–211. 53–62. 80 Mushiroda T, Ohnishi Y, Saito S, Takahashi A, Kikuchi Y, Shimomura H 59 Hirsh J, Dalen J, Anderson D, Poller L, Bussey H, Ansell J et al. Oral et al. Association of VKORC1 and CYP2C9 polymorphisms with anticoagulants: mechanism of action, clinical effectiveness, and warfarin dose requirements in Japanese patients. J Hum Genet 2006; optimal therapeutic range. Chest 1998; 114: 445S–469S. 51: 249–253. 60 Sconce E, Khan T, Mason J, Noble F, Wynne H, Kamali F. Patients with 81 Takahashi H, Wilkinson GR, Nutescu EA, Morita T, Ritchie MD, Scordo unstable control have a poorer dietary intake of vitamin K compared to MG et al. Different contributions of polymorphisms in VKORC1 and patients with stable control of anticoagulation. Thromb Haemost 2005; CYP2C9 to intra- and inter-population differences in maintenance 93: 872–875. dose of warfarin in Japanese, Caucasians and African-Americans. 61 Kohlmeier M, Salomon A, Saupe J, Shearer MJ. Transport of vitamin K Pharmacogenet Genomics 2006; 16: 101–110. to bone in humans. J Nutr 1996; 126: 1192S–1196S. 82 Lee SC, Ng SS, Oldenburg J, Chong PY, Rost S, Guo JY et al. Interethnic 62 Berkner KL, Runge KW. The physiology of vitamin K nutriture and variability of warfarin maintenance requirement is explained by vitamin K-dependent protein function in atherosclerosis. J Thromb VKORC1 genotype in an Asian population. Clin Pharmacol Ther 2006; Haemost 2004; 2: 2118–2132. 79: 197–205. 63 Lamon-Fava S, Sadowski JA, Davidson KW, O’Brien ME, McNamara JR, 83 Aquilante CL, Langaee TY, Lopez LM, Yarandi HN, Tromberg JS, Schaefer EJ. Plasma lipoproteins as carriers of phylloquinone (vitamin Mohuczy D et al. Influence of coagulation factor, vitamin K epoxide K1) in humans. Am J Clin Nutr 1998; 67: 1226–1231. reductase complex subunit 1, and cytochrome P450 2C9 gene 64 Saupe J, Shearer MJ, Kohlmeier M. Phylloquinone transport and its polymorphisms on warfarin dose requirements. Clin Pharmacol Ther influence on gamma-carboxyglutamate residues of osteocalcin in patients 2006; 79: 291–302. on maintenance hemodialysis. Am J Clin Nutr 1993; 58: 204–208. 84 Bodin L, Verstuyft C, Tregouet DA, Robert A, Dubert L, Funck-Brentano 65 Rosand J, Hylek EM, O’Donnell HC, Greenberg SM. Warfarin- C et al. Cytochrome P450 2C9 (CYP2C9) and vitamin K epoxide associated hemorrhage and cerebral amyloid angiopathy: a genetic reductase (VKORC1) genotypes as determinants of acenocoumarol and pathologic study. Neurology 2000; 55: 947–951. sensitivity. Blood 2005; 106: 135–140. 66 Kohnke H, So¨rlin K, Granath G, Wadelius M. Warfarin dose related to 85 Reitsma PH, Heijden JF, Groot AP, Rosendaal FR, Buller HR. A C1173T apolipoprotein E (APOE) genotype. Eur J Clin Pharmacol 2005; 61: dimorphism in the VKORC1 gene determines coumarin sensitivity and 381–388. bleeding risk. PLoS Med 2005; 2: e312. 67 Visser LE, Trienekens PH, De Smet PA, Vulto AG, Hofman A, van Duijn 86 Cain D, Hutson SM, Wallin R. Assembly of the warfarin-sensitive CM et al. Patients with an ApoE epsilon4 allele require lower doses of vitamin K 2,3-epoxide reductase enzyme complex in the endoplasmic coumarin anticoagulants. Pharmacogenet Genomics 2005; 15: 69–74. reticulum membrane. J Biol Chem 1997; 272: 29068–29075. 68 Kohnke H, Scordo MG, Pengo V, Padrini R, Wadelius M. Apolipopro- 87 Morisseau C, Hammock BD. Epoxide hydrolases: mechanisms, tein E (APOE) and warfarin dosing in an Italian population. Eur J Clin inhibitor designs, and biological roles. Annu Rev Pharmacol Toxicol Pharmacol 2005; 61: 781–783. 2005; 45: 311–333.

The Pharmacogenomics Journal Pharmacogenetics of warfarin M Wadelius and M Pirmohamed 111

88 Guenthner TM, Cai D, Wallin R. Co-purification of microsomal epoxide 108 Dahlback B. Blood coagulation and its regulation by anticoagulant hydrolase with the warfarin-sensitive vitamin K1 oxide reductase of the pathways: genetic pathogenesis of bleeding and thrombotic diseases. vitamin K cycle. Biochem Pharmacol 1998; 55: 169–175. J Intern Med 2005; 257: 209–223. 89 Loebstein R, Vecsler M, Kurnik D, Austerweil N, Gak E, Halkin H et al. 109 Larsen TB, Lassen JF, Dahler-Eriksen BS, Petersen PH, Brandslund I. Common genetic variants of microsomal epoxide hydrolase affect Effect of anticoagulant therapy on the hypercoagulable state in warfarin dose requirements beyond the effect of cytochrome P450 patients carrying the factor V Arg506Gln mutation. Thromb Res 1998; 2C9. Clin Pharmacol Ther 2005; 77: 365–372. 92: 157–162. 90 Ross D, Siegel D. NAD(P)H:quinone oxidoreductase 1 (NQO1, DT- 110 Andersson T, Flockhart DA, Goldstein DB, Huang SM, Kroetz DL, Milos diaphorase), functions and pharmacogenetics. Methods Enzymol 2004; PM et al. Drug-metabolizing enzymes: evidence for clinical utility of 382: 115–144. pharmacogenomic tests. Clin Pharmacol Ther 2005; 78: 559–581. 91 Wallin R, Hutson S. Vitamin K-dependent carboxylation. Evidence that 111 Hillman MA, Wilke RA, Yale SH, Vidaillet HJ, Caldwell MD, Glurich I et at least two microsomal dehydrogenases reduce vitamin K1 to support al. A prospective, randomized pilot trial of model-based warfarin dose carboxylation. J Biol Chem 1982; 257: 1583–1586. initiation using CYP2C9 genotype and clinical data. Clin Med Res 2005; 92 Rost S, Fregin A, Koch D, Compes M, Muller CR, Oldenburg J. 3: 137–145. Compound heterozygous mutations in the gamma-glutamyl carbox- 112 Keavney B, McKenzie C, Parish S, Palmer A, Clark S, Youngman L ylase gene cause combined deficiency of all vitamin K-dependent et al. Large-scale test of hypothesised associations between the blood coagulation factors. Br J Haematol 2004; 126: 546–549. angiotensin-converting-enzyme insertion/deletion polymorphism 93 Berkner KL. The vitamin K-dependent carboxylase. J Nutr 2000; 130: and myocardial infarction in about 5000 cases and 6000 controls. 1877–1880. International Studies of Infarct Survival (ISIS) Collaborators. Lancet 94 Wu SM, Stafford DW, Frazier LD, Fu YY, High KA, Chu K et al. Genomic 2000; 355: 434–442. sequence and transcription start site for the human gamma-glutamyl 113 Shine D, Patel J, Kumar J, Malik A, Jaeger J, Maida M et al. A carboxylase. Blood 1997; 89: 4058–4062. randomized trial of initial warfarin dosing based on simple clinical 95 Brenner B, Sanchez-Vega B, Wu SM, Lanir N, Stafford DW, Solera J. A criteria. Thromb Haemost 2003; 89: 297–304. missense mutation in gamma-glutamyl carboxylase gene causes 114 Aithal G, Day C, Kesteven P, Daly A. Association of polymorphisms in combined deficiency of all vitamin K-dependent blood coagulation the cytochrome P450 CYP2C9 with warfarin dose requirement and factors. Blood 1998; 92: 4554–4559. risk of bleeding complications. Lancet 1999; 353: 717–719. 96 Shikata E, Ieiri I, Ishiguro S, Aono H, Inoue K, Koide T et al. Association 115 Joffe HV, Xu R, Johnson FB, Longtine J, Kucher N, Goldhaber SZ. of pharmacokinetic (CYP2C9) and pharmacodynamic (factors II, Warfarin dosing and cytochrome P450 2C9 polymorphisms. Thromb VII, IX, and X; proteins S and C; and gamma-glutamyl Haemost 2004; 91: 1123–1128. carboxylase) gene variants with warfarin sensitivity. Blood 2004; 103: 116 Fihn S, Callahan C, Martin D, McDonell M, Henikoff J, White R. The risk 2630–2635. for and severity of bleeding complications in elderly patients treated 97 Chen LY, Eriksson N, Gwilliam R, Bentley D, Deloukas P, Wadelius M. with warfarin. Ann Intern Med 1996; 124: 970–979. Gamma-glutamyl carboxylase (GGCX) microsatellite and warfarin 117 Ripley TL, Havrda DE, Blevins S, Culkin D. Early evaluation of hematuria dosing. Blood 2005; 106: 3673–3674. in a patient receiving anticoagulant therapy and detection of 98 Wallin R, Hutson SM, Cain D, Sweatt A, Sane DC. A molecular malignancy. Pharmacotherapy 2004; 24: 1638–1640. mechanism for genetic warfarin resistance in the rat. FASEB J 2001; 15: 118 Muhlau M, Schlegel J, Von Einsiedel HG, Conrad B, Sander D. Multiple 2542–2544. progressive intracerebral hemorrhages due to an angiosarcoma: a case 99 Wajih N, Sane DC, Hutson SM, Wallin R. The inhibitory effect of report. Eur J Neurol 2003; 10: 741–742. calumenin on the vitamin K-dependent gamma-carboxylation system. 119 Nurden AT, Nurden P. Inherited disorders of platelets: an update. Curr Characterization of the system in normal and warfarin-resistant rats. Opin Hematol 2006; 13: 157–162. J Biol Chem 2004; 279: 25276–25283. 120 You JH, Chan FW, Wong RS, Cheng G. The potential clinical and 100 Broze Jr GJ. Protein Z-dependent regulation of coagulation. Thromb economic outcomes of pharmacogenetics-oriented management of Haemost 2001; 86: 8–13. warfarin therapy – a decision analysis. Thromb Haemost 2004; 92: 590– 101 D’Ambrosio RL, D’Andrea G, Cappucci F, Chetta M, Di Perna P, 597. Brancaccio V et al. Polymorphisms in factor II and factor VII genes 121 Woelderink A, Ibarreta D, Hopkins MM, Rodriguez-Cerezo E. The modulate oral anticoagulation with warfarin. Haematologica 2004; 89: current clinical practice of pharmacogenetic testing in Europe: TPMT 1510–1516. and HER2 as case studies. Pharmacogenomics J 2006; 6: 3–7. 102 Kristensen SR. Warfarin treatment of a patient with coagulation factor 122 Gardiner SJ, Begg EJ. Pharmacogenetic testing for drug metabolizing IX propeptide mutation causing warfarin hypersensitivity. Blood 2002; enzymes: is it happening in practice? Pharmacogenet Genomics 2005; 100: 2676–2677. 15: 365–369. 103 van der Heijden JF, Rekke B, Hutten BA, van der Meer FJ, Remkes MG, 123 Fitzmaurice DA, Hobbs FD, Delaney BC, Wilson S, McManus R. Review Vermeulen M et al. Non-fatal major bleeding during treatment with of computerized decision support systems for oral anticoagulation vitamin K antagonists: influence of soluble thrombomodulin and management. Br J Haematol 1998; 102: 907–909. mutations in the propeptide of coagulation factor IX. J Thromb 124 Yang DT, Robetorye RS, Rodgers GM. Home prothrombin time Haemost 2004; 2: 1104–1109. monitoring: a literature analysis. Am J Hematol 2004; 77: 177–186. 104 Weiss P, Soff GA, Halkin H, Seligsohn U. Decline of proteins C and S 125 Albers GW, Diener HC, Frison L, Grind M, Nevinson M, Partridge S et and factors II, VII, IX and X during the initiation of warfarin therapy. al. Ximelagatran vs warfarin for stroke prevention in patients with Thromb Res 1987; 45: 783–790. nonvalvular atrial fibrillation: a randomized trial. JAMA 2005; 293: 105 Vigano S, Mannucci PM, Solinas S, Bottasso B, Mariani G. Decrease in 690–698. protein C antigen and formation of an abnormal protein soon after 126 Voora D, Eby C, Linder MW, Milligan PE, Bukaveckas BL, McLeod HL et starting oral anticoagulant therapy. Br J Haematol 1984; 57: 213–220. al. Prospective dosing of warfarin based on cytochrome P-450 2C9 106 McGehee WG, Klotz TA, Epstein DJ, Rapaport SI. Coumarin necrosis genotype. Thromb Haemost 2005; 93: 700–705. associated with hereditary protein C deficiency. Ann Intern Med 1984; 127 Leung AY, Chow HC, Kwong YL, Lie AK, Fung AT, Chow WH et al. 101: 59–60. Genetic polymorphism in exon 4 of cytochrome P450 CYP2C9 may be 107 Chan YC, Valenti D, Mansfield AO, Stansby G. Warfarin induced skin associated with warfarin sensitivity in Chinese patients. Blood 2001; necrosis. Br J Surg 2000; 87: 266–272. 98: 2584–2587.

The Pharmacogenomics Journal