A case study of acenocoumarol sensitivity and genotype-phenotype discordancy explained by combinations of polymorphisms in VKORC1 and CYP2C9

Abstract

Aims

To determine the cause of a genotype-phenotype discordancy for acenocoumarol sensitivity.

Methods

A patient, highly sensitive to acenocoumarol, and previously determined to carry only a single CYP2C9 *3 allele, was genotyped for additional functionally defective alleles in the CYP2C9 and VKORC1 genes. Family members were also analyzed to trace the pedigree.

Results

The acenocoumarol-sensitive patient was found to possess, in addition to CYP2C9 *3 allele, a CYP2C9 *11 allele and the VKORC1 AA diplotype which were all traced back through the parental lines.

Conclusions

Acenocoumarol sensitivity in this subject is the consequence of inheritance of multiple functionally defective alleles in both the CYP2C9 and VKORC1 genes. The study provides additional data in support of diminished CYP2C9 activity due to the presence of the rare *11 allele.

Introduction

Single nucleotide polymorphisms (SNPs) in the CYP2C9 gene, especially the CYP2C9 *2 and CYP2C9 *3 variants which are common in European populations, are linked to coumarin anticoagulant sensitivity [1, 2]. Recently, one of us (HHT) reported a case of acenocoumarol sensitivity in a patient heterozygous at the CYP2C9 *3 locus [3]. This patient reached stable anticoagulation at an acenocoumarol dose of < 1 mg day⁻¹, and exhibited oral clearances of (S)-acenocoumarol and (R)-acenocoumarol that were 7% and 53%, respectively, that of a CYP2C9 *1/*1 subject. Moreover, formation clearances for the major (S)-7-hydroxy (S)-6-hydroxy and (R)-7-hydroxy acenocoumarol metabolites were barely 10% of a CYP2C9 *1 control population. In the absence of any evidence for interfering comedications, this overt ‘poor metabolizer’ phenotype is much more characteristic of a CYP2C9 *3 homozygote than one heterozygous for this variant allele.

Currently, some 25 coding region polymorphisms are listed for CYP2C9 on the Human P450 Allele Website (http://www.imm.ki.se/CYPalleles; accessed January 3, 2006), at least half of which are associated with diminished activity in vitro. In Caucasians, CYP2C9 *2 and CYP2C9 *3 occur at allele frequencies of 12% and 7%, respectively [4], but many of the newer SNPs have allele frequencies less than 1%, thereby making it difficult to conduct in vivo studies that unequivocally establish the clinical consequences of the mutation. Genotype-phenotype discordancies, however, provide a fertile ground for both new SNP discovery and functional evaluation of rare P450 polymorphisms, as exemplified by recent studies with both the CYP2D6 and CYP2C9 genes [5–7].

Recently, genetic variations in the VKORC1 gene have emerged as a dominant factor in oral anticoagulant response, contributing up to 30% of interindividual variability in dose requirements of warfarin and acenocoumarol [8–10]. Therefore, we re-analyzed DNA from our acenocoumarol-sensitive subject for additional functionally defective SNPs in CYP2C9 and also evaluated his VKORC1 genotype to attempt to gain insight into this genotype-phenotype discordancy.

Methods

Subjects

The acenocoumarol-sensitive patient and his father, mother and brother were recruited for the study. No family members other than the index patient underwent treatment with acenocoumarol. Written informed consent was obtained from all subjects and institutional approval was obtained from the Human Subjects Committees at both the University of Maastricht and the University of Washington.

Genotyping

Genomic DNA was isolated from peripheral blood of each subject and the CYP2C9 *2, *3, *4, *5 and *11 alleles genotyped by published procedures [2], as well as the VKORC1 SNPs, G2653C, G3673A, T5808G, C6009T, C6484T, G6853C and G9041A, that permit VKORC1 haplotype (H1-H9) and clade assignment (A or B) in Europeans [8, 11]. As described previously, haplotypes H1 and H2 belong to clade A, while haplotypes H7, H8 and H9 are grouped together as clade B [8].

Results

We confirmed the presence of a single CYP2C9 *3 allele in the acenocoumarol sensitive patient as reported previously [3]. Moreover, we also identified this subject as a carrier of the rarer CYP2C9 *11 allele (Figure 1, black box). The family study provides evidence that these two alleles are carried on different chromosomes, and so this individual’s CYP2C9 status is that of a compound CYP2C9 *3/*11 heterozygote (see Figure 1). Additionally, the patient was homozygous for the VKORC1 A clade (Figure 1), a diplotype that occurs in ∼15% of Europeans [8].

Figure 1.

Family study of warfarin susceptibility alleles. Genotyping revealed the index patient (black box) had inherited unusual combinations of CYP2C9*3, CYP2C9*11 and VKORC1 A clade haplotypes, all of which contribute to acenocoumarol sensitivity

Discussion

The genotype-phenotype discordancy for this acenocoumarol-sensitive patient is resolved by the finding of an additional functionally defective CYP2C9 *11 allele, compounded further by the VKORC1 AA diplotype known to associate with a low dose oral anticoagulant phenotype. Recent studies in multiple populations have repeatedly established the importance of VKORC1regulatory polymorphisms to oral anticoagulant outcomes [8–10, 12–14]. Indeed, we have found that European-American warfarin patients that are homozygous for the less common VKORC1 A clade require on average a 45% lower dose than the population mean of 5.5 mg day⁻¹ [8].

In contrast, the functional consequences of the CYP2C9 *11 allele are less well established. In vitro studies with recombinant CYP2C9.11 expressed in E. coli [15] and insect cells [11] are indicative of a defective enzyme. However, because CYP2C9 *11 occurs at only ∼2.0% allele frequency in subjects of African descent and ∼0.7% in Caucasians [11, 16, 17], and may even be absent in Japanese [17], in vivo studies have been limited. Allabi et al. reported reduced CYP2C9 activity in three *1/*11 Beninese subjects towards phenytoin [18], but not losartan [19]. King et al. reported a trend toward lower warfarin dose in three European *1/*11 patients, but this did not reach statistical significance [20]. Recently, Takahashi & coworkers [17] concluded that CYP2C9 *2/*3/*11 genotypes were an independent covariate contributing to interpatient variability in warfarin dose in Caucasian patients. In a separate study, we found that four *1/*11 European-American warfarin patients required a 33% lower warfarin dose than wild-type subjects, and suggested that the magnitude of the detrimental effect of the *11 allele on CYP2C9 activity towards warfarin lay between that of the well characterized *2 and *3 alleles [11]. The results of the present investigation are consistent with a similarly intermediate effect of the CYP2C9 *11 allele on acenocoumarol metabolism.

In conclusion, this family’s genetic profile, for both pharmacokinetic (CYP2C9) and pharmacodynamic (VKORC1) genes known to modulate patient response to oral anticoagulants, is unusual on two counts, both of which confer coumarin anticoagulant sensitivity. Firstly, there is a preponderance of VKORC1 haplotypes representing the A clade, the minor of the two clades in European Americans [8]. Secondly, the family carries two, discrete, functionally defective CYP2C9 alleles, one of which (CYP2C9 *11) is a rare allele in Europeans [11, 16, 17]. Therefore, multiple genetic liabilities in different susceptibility genes combine to explain the patient’s acenocoumarol sensitivity. This case study demonstrates the utility of a comprehensive pharmacogenomic approach targeting both VKORC1 and CYP2C9 for understanding sensitivity to coumarin anticoagulants, and suggests further that CYP2C9 *11 should be included in routine test panels for genotyping of oral anticoagulant patients.

This work was supported in part by NIH grants GM68769, NS053646 & ES07033.