Pharmacogenetic relevance of CYP4F2 V433M polymorphism on acenocoumarol therapy

The initiation of oral anticoagulation therapy is associated with one of the highest adverse event rates for any single drug.¹  Up to one-half of patients with atrial fibrillation and no contraindication to warfarin therapy, who are at high risk of stroke, are currently not receiving anticoagulant therapy because physicians are often reluctant to initiate it in elderly patients or patients with risk of bleeding.²  In addition, the initiation of this therapy is always critical and the stabilization is difficult during the first weeks and the risk higher.³  Moreover, the broad variability in patient response also modulates the initial risk. It is now well known that common genetic polymorphisms have a strong influence on the interindividual coumarinics response. In extensively studied white populations, CYP2C9 genotype predicts approximately 10% of the interpatient variability in warfarin dose. Similarly, in white and Asian populations, VKORC1 genotype predicts 25% of such variability, with similar results for acenocoumarol.⁴  The importance of these strong genetic effects was recognized by recent relabeling of warfarin by the Food and Drug Administration to raise awareness in the clinical community.⁵  However, it is important to note that patient demographics, clinical factors, and (known) genetic variants combined only explain 45% to 55% of the total dose variant.⁶  From the recently published studies using genome-wide scan strategy for common genetic variants with influence of warfarin response, it could be concluded that none of the other candidate genes (besides VKORC1 and CYP2C9) will exert such a deep influence in response.^6,7  These data do not discard, however, the fact that single nucleotide polymorphisms (SNPs) in new genes could also be involved in intermediate phenotypes, so explaining the modest percentages in interpersonal variability and hence be involved in the risk of therapy. In this vein, Caldwell et al⁸  used the Affymetrix (Santa Clara, CA) drug-metabolizing enzymes and transporters panel recently to describe that a variant in cytochrome P450 4F2 (CYP4F2) (rs2108622, V433M) influences warfarin requirements in long-term treatment patients. We have evaluated the role of CYP4F2 V433M polymorphism on acenocoumarol by analyzing the earliest response in a well-defined sample of patients starting therapy

We have analyzed the CYP4F2 V433M SNP on a previously selected sample⁹  of 100 white men younger than 75 years (62.2 ± 6.5 years; ± SD), of similar weight (body surface area [BSA], 1.81 ± 0.15 m²; ± SD), with nonvalvular atrial fibrillation, who were not taking any medication known to interfere with acenocoumarol and who started anticoagulation therapy with 3 mg acenocoumarol for 3 consecutive days. We evaluated the initial response over 3 consecutive days by analyzing the reduction in vitamin K-dependent hemostatic proteins and by considering its overall anticoagulant effect (international normalized ratio [INR]) according to V433M genotype. The dose needed to achieve a steady anticoagulation (INR = 2-3) recorded after 3 months was also evaluated. Plasma levels of clotting factors were determined using human immunodepleted plasma, based on the prothrombin time (factors II, VII, and X [FII, FVII, and FX, respectively]) or on the partial activated thromboplastin time (factor IX [FIX]; ACL 3000 with HemosIL reagents [Instrumentation Laboratory, Lexington, MA]).

Genotyping of other candidate SNPs has recently been published.^9,10  Genotyping of V433M was performed by a validated TaqMan Drug Metabolism Genotyping Assay C_16179493_40 (Applied Biosystems, Foster City, CA). The frequency observed for V433M SNP (Table 1) was similar to that described for other white populations^8,11  and to that previously described in the HapMap.

Comparisons between 2 groups were performed by the analysis of variance test and Kruskal-Wallis test. A multivariate stepwise linear regression model was performed to evaluate the potential contribution of VKORC1, CYP450, and CYP4F2 polymorphisms to interindividual variability in therapeutic acenocoumarol doses and INR value at third day. All analyses were carried out using SPSS version 15.0 software (SPSS, Chicago, IL). Approval for this study was acquired from the Ethic Committee of Morales Meseguer Hospital (Murcia, Spain), and patient informed consent was obtained in accordance with the Declaration of Helsinki.

We evaluated the influence of V433M genotype in the initial response to 3 mg acenocoumarol over 3 consecutive days. To define the weight of CYP4F2 polymorphism more accurately, our experimental design took into account the influence of factors, such as age, BSA, sex, disease, and other drugs.⁹ 

This is the first study to show that V433M polymorphism correlates with the decrease of FII, FVII, FIX, or FX in the earliest response to acenocoumarol, and these effects are gene dosage-dependent. Thus, whereas carriers of the V433V genotype experienced the biggest reduction in all vitamin K–dependent proteins analyzed, heterozygous subjects gave an intermediate response (Table 1). FVII reached the lower levels (12 IU/mL in V433V vs 38 IU/mL in M433M; P = .001), according to the shorter half-life of this factor, which also makes it the most sensitive to anticoagulants.¹²  The effect of this polymorphism on FX levels showed the same tendency, although differences did not reach statistical significance (Table 1). Such decreases were not related to CYP2C9 genotype.⁹  Linear regression analysis, including the effects of CYP2C9, VKORC1, and CYP4F2 genotypes, confirmed that only CYP4F2 V433M exerts an independent effect on clotting factors (Table 1).

We consistently observed a deep impact of the V433M genotype on INR after 3 days, which also had a gene dose-dependent effect (Table 1). As shown, V433V patients displayed the highest INR values. Worthy of note was that 6 of 9 patients with INR more than 3.5 after 3 days had V433V genotype.

The dose required to achieve a steady INR was also influenced by CYP4F2 genotype (Table 1), similar to data reported for warfarin therapy⁸ : 433M carriers needed approximately 4 mg/week more (∼26% increase in dose) than V433V subjects (P = .043; Table 1). However, we did not observe a clear gene dosage-dependent effect in this parameter, as described by Caldwell et al.⁸  This minor discrepancy may be explained by differences between warfarin and acenocoumarol or the size of our sample. Finally, combined genotypes of VKORC1 and CYP4F2 exacerbated the response. Thus, double homozygous VKORC1 1173T CYP4F2 433V (n = 5) showed the highest INR (3.1; 2.0-4.7) and required the lowest dose (11 ± 3 mg/week).

Results from multivariate linear regression models discarded the overall influence of CYP450 2C9 genotypes (P = .201), BSA (P = .145), and age (P = .071) in our study. The predictor contributions of VKORC1 and CYP4F2 to variability of both early INR and dose are shown in Table 2. Interestingly, the addition of V433M polymorphism to VKORC1 genotype raised the R² from 8% to 14% for INR and from 12% to 19% for dose requirements.

CYP4F2 participates in the inactivation pathway of vitamin E.¹³  The V433M polymorphism has an effect on the cytochrome function because the 433M variant has decreased activity.¹⁴  The role of CYP4F2 in γ-carboxylation or vitamin K cycle is unknown, but it might have a functional effect on warfarin⁸  and acenocoumarol requirements. Relying on the similarity of the vitamins E and K, CYP4F2 might hydroxylate the vitamin K phytyl side chain, hence interfering in the vitamin K recycling.⁸  An alternative hypothesis is that CYP4F2 could be involved in the metabolism of acenocoumarol because the V433M polymorphism is associated with significantly different levels of FII, FVII, FIX, and FX only after acenocoumarol, not before therapy.

In conclusion, our data provide new information about the pharmacogenetics of acenocoumarol, as we confirm that the CYP4F2 V433M polymorphism plays a relevant role in the earliest response to acenocoumarol and the required dose. CYP4F2 V433M and VKORC1 genotyping may help to recommend safe doses in patients with a genetic profile associated with poor outcomes when treated by traditional trial-and-error dosing. The answer to the recent controversial question of whether genotyping before dosing would be clinically useful^15,–17  will require further randomized prospective and large trials that evaluate adverse effect of empiric versus genotype adjustment of doses, but obviously any new genetic element involved in adjustment of therapeutic dose will increase the reliability of pharmacogenetic tests.

This work was supported by Ministerio de Ciencia y Tecnología (Madrid, Spain), Fondo Europeode Desarrollo Regional (European Union), Red Temática de Investigación Cooperativa en Enfermadades Cardiovasculares (Madrid), Instituto de Salud Carlos III (Madrid) Grant SAF 2006-06212, and Fundación Séneca (Murcia, Spain) Grant 05759/PI/07.

Contribution: V.P.-A. collected patients, was responsible for their management, and analyzed clotting factors; V.R. collected patients, was responsible for their management, analyzed clotting factors, contributed to the design of the research, and performed the statistical analysis; A.I.A. and N.G.-B. were responsible for genetic analysis; J.C. and V.V. contributed to the design of the research; and R.G.-C. contributed to the design of the research and was responsible for genetic analysis. All authors contributed to the writing of the paper. Fundación Séneca (Murcia, Spain) Grant 05759/PI/07.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Rocio González-Conejero, Centro Regional de Hemodonación, Ronda de Garay s/n, 30003, Murcia, Spain; e-mail: rocio.gonzalez@carm.es.