Analysis of CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms in a population from South-Eastern Europe

Abstract

The CYP2C9 enzyme metabolizes a wide range of relevant drugs, among which are oral anticoagulants. VKORC1 is the pharmacodynamic target of the oral anticoagulants. The genetic polymorphisms CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A are the major determinants of the inter-individual variability in the dosage requirements of oral anticoagulants. This study provides a first evaluation of these 3 polymorphisms in a Romanian population. A total of 332 Romanian individuals were genotyped for the CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms using the PCR-RFLP technique. Sixty-two individuals (18.7%) were heterozygous for CYP2C9*2, whereas 47 individuals (14.1%) were heterozygous for CYP2C9*3. Fourteen individuals (4.2%) had a CYP2C9*2 homozygous, CYP2C9*3 homozygous or CYP2C9*2/CYP2C9*3 compound heterozygous genotype. These individuals are predicted to have the lowest CYP2C9 enzymatic activity. The allele frequencies of the CYP2C9*2 and CYP2C9*3 polymorphisms were 11.3% and 9.3% respectively. For the VKORC1 -1639 G>A polymorphism, there were 170 heterozygotes (51.2%) and 55 (16.6%) homozygotes for the A allele. The frequency of the A allele was 42.2%. Overall, the distribution of the CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms observed in our cohort is in accordance with other Caucasian populations. A large number of Romanians are expected to harbour at least one CYP2C9 variant allele and/or one VKORC1 -1639 G>A allele. This frequency has major implications in the pharmacogenomics of oral anticoagulants in Romanians.

Introduction

The CYP2C subfamily of cytochrome P450 comprises 4 members: CYP2C8, CYP2C9, CYP2C18 and CYP2C19. Among them, CYP2C9 is the most abundant isoform in the liver, metabolizing a wide range of clinically relevant drugs, such as oral anticoagulants (warfarin, acenocoumarol and phenprocoumon), antidiabetic drugs (tolbutamide and glipizide), the anticonvulsant phenytoin, and a wide range of non-steroidal anti-inflammatory drugs (e.g. diclofenac, ibuprofen and piroxicam) 1,2. The gene encoding the CYP2C9 enzyme is polymorphic. Several variants of the CYP2C9 gene have been described, but the most prevalent and most studied of them are the CYP2C9*2 and CYP2C9*3 polymorphisms. The CYP2C9*2 allele is the result of a C>T transition in position 430 of the CYP2C9 gene, leading to an Arg-to-Cys substitution at residue 144 of the CYP2C9 molecule. The CYP2C9*3 allele is the result of an A>T transversion in position 1075 of the CYP2C9 gene, leading to an Ile-to-Leu substitution at residue 359 of the CYP2C9 molecule 3. Both alleles lead to a significant reduction in the enzymatic activity of the CYP2C9 molecule, representing the major causes of decreased CYP2C9 enzymatic activity, at least in Caucasians 3.

Vitamin K epoxide reductase catalyses the transformation of vitamin K 2,3-epoxide into vitamin K hydroxyquinone, which is essential for the synthesis of factors II, VII, IX and XI of the coagulation system, as well as proteins C and S. The enzyme vitamin K epoxide reductase, which is the pharmacodynamic target of oral anticoagulants, is encoded by the VKORC1 gene. While rare mutations of this gene produce clotting factor deficiencies or warfarin resistance, common functional polymorphisms are associated with variations in the dosage requirement of oral anticoagulants 4. Two common polymorphisms of the VKORC1 gene, namely -1639 G>A and 1173 C>T, which are in complete linkage disequilibrium, were shown to confer a high risk towards over-anticoagulation, predisposing the individuals carrying these polymorphisms to haemorrhagic incidents 5.

The distribution of the CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms has been extensively assessed in many populations; however, data for the Romanian population are lacking. Here, we analysed the frequency of these polymorphisms in a cohort of Romanians and compared it with similar data from other populations.

Materials and methods

Study participants

Between 2007 and 2010, a total of 332 individuals were enrolled in this study. All participants originated from the Transylvania region (North-Western and central parts of Romania). The study group comprised 155 women (46.7%) and 177 men (53.3%). The median age of the study participants was 69 years (range 34–83 years). Written and informed consent was obtained from all the participants prior to their enrolment. This study was reviewed and approved by the Ethics Committee of the “Iuliu Haţieganu” University of Medicine and Pharmacy of Cluj-Napoca. After admission to the study, 3 ml of blood was collected from each patient into an EDTA tube, and the genomic DNA was obtained using a commercially available procedure (Wizard Genomic DNA Purification Kit; Promega, Madison, WI, USA).

Genotyping for the CYP2C9*2 and CYP2C9*3 polymorphisms

Genotyping for the CYP2C9*2 and CYP2C9*3 polymorphisms was performed with PCR-RFLP (polymerase chain reaction–restriction fragment length polymorphism), essentially as previously described 6. Briefly, a 372-bp amplicon was obtained by PCR to study the CYP2C9*2 allele. The amplicon was digested overnight with the Sau96I restriction enzyme (Fermentas MBI, Vilnius, Lithuania), giving rise to 3 fragments with lengths of 179, 119 and 74 bp in the case of the wild-type allele. However, the CYP2C9*2 allele abolishes one restriction site of the Sau96I enzyme, giving rise to only 2 fragments, with lengths of 253 and 119 bp. To analyse the CYP2C9*3 variant, a 130-bp amplicon was obtained by PCR. This amplicon was digested overnight with the StyI restriction enzyme (Fermentas MBI), and the wild-type allele was resistant to the StyI digestion. However, the CYP2C9*3 allele creates a restriction site for StyI, cutting the 130-bp amplicon into 2 fragments, with lengths of 104 and 26 bp.

Genotyping for the VKORC1 -1639 G>A polymorphism

Genotyping for the VKORC1 -1639 G>A polymorphism was performed with PCR-RFLP, essentially as previously described 7. Briefly, a 290-bp amplicon was obtained by PCR. The amplicon was digested overnight with MspI restriction enzyme (Fermentas MBI). If the G allele was present, then the 290-bp amplicon was cut in 2 fragments, with lengths of 167 and 123 bp, whereas the presence of the A allele rendered the amplicon resistant to digestion with the MspI restriction enzyme.

Statistical analysis

The distribution of the CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms was compared between Romanians and other populations by Fisher's exact test, using the GraphPad InStat version 3.0 statistical software (GraphPad, San Diego, CA, USA). The level of statistical significance was set at 0.05.

Results

CYP2C9*2 and CYP2C9*3

Sixty-two individuals (18.7%) were heterozygous for CYP2C9*2, whereas 47 individuals (14.1%) were heterozygous for CYP2C9*3. Two individuals (0.6%) were homozygous for CYP2C9*2, whereas 3 individuals (0.9%) were homozygous for CYP2C9*3. A total number of 9 individuals (2.7%) were CYP2C9*2/CYP2C9*3 compound heterozygotes. Thus, 14 individuals (4.2%) from the cohort analysed are predicted to have the lowest CYP2C9 enzymatic activity, with respect to the CYP2C9*2 and CYP2C9*3. Overall, the CYP2C9*2 allele had a frequency of 11.3%, whereas the CYP2C9*3 allele had a frequency of 9.3%. The observed genotype frequencies for both CYP2C9 polymorphisms were consistent with the Hardy–Weinberg equilibrium (data not shown). The results regarding the distribution of CYP2C9*2 and CYP2C9*3 are shown in detail in Table1. The absence of the CYP2C9*2 or CYP2C9*3 polymorphisms was considered as a wild-type allele and was denoted as ‘CYP2C9*1’.

Table 1.

Genotype and alleles frequencies observed for the CYP2C9 polymorphisms in our study

Genotype distribution (N = 332 individuals)
Genotype	Frequency; n (%)
CYP2C9*1/CYP2C9*1	209 (63)
CYP2C9*1/CYP2C9*2	62 (18.7)
CYP2C9*1/CYP2C9*3	47 (14.1)
CYP2C9*2/CYP2C9*2	2 (0.6)
CYP2C9*2/CYP2C9*3	9 (2.7)
CYP2C9*3/CYP2C9*3	3 (0.9)
Alleles frequency (N = 664 alleles)
Allele	Frequency; n (%)
CYP2C9*1	527 (79.4)
CYP2C9*2	75 (11.3)
CYP2C9*3	62 (9.3)

VKORC1 -1639 G>A

A total number of 170 individuals (51.2%) were heterozygous for the -1639 G>A polymorphism, whereas 55 individuals (16.6%) had a homozygous AA genotype. Thus, the A allele had a frequency of 42.2% in our group. The observed genotype frequencies for VKORC1 -1639 G>A were consistent with the Hardy–Weinberg equilibrium (data not shown). The results regarding the distribution of the VKORC1 -1639 G>A polymorphism are presented in detail in Table2.

Table 2.

Genotype and alleles frequencies observed for the VKORC1 -1693 G>A polymorphism in our study

Genotype distribution (N = 332 individuals)
Genotype	Frequency; n (%)
GG	107 (32.2)
GA	170 (51.2)
AA	55 (16.6)
Alleles frequency (N = 664 alleles)
Allele	Frequency; n (%)
G	384 (57.8)
A	280 (42.2)

CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A in Romanians compared with other populations

Table3 shows the distribution of CYP2C9*2 and CYP2C9*3 across various populations and the differences in the allelic frequencies of these polymorphisms between these populations and the Romanians, as calculated by Fisher's exact test. Table4 shows the distribution of the VKORC1 -1639 G>A polymorphism across various populations and the differences in allelic frequencies (for the A allele) between Romanians and these populations, as calculated by Fisher's exact test.

Table 3.

Allelic frequencies of CYP2C9*2 and CYP2C9*3 polymorphisms in Romanians compared with different populations

Population	Sample size	CYP2C9*2 allele frequency (P-value*)	CYP2C9*3 allele frequency (P-value*)	Reference
Romanian	332	11.3	9.3	Present study
Faroese	312	8.8 (NS)	5.3 (0.008)	8
French	151	15 (0.01)	8 (NS)	9
Belgian	121	10 (NS)	7.4 (NS)	10
Spanish	157	14.3 (NS)	16.2 (<0.001)	11
Italian	360	12.5 (NS)	9.7 (NS)	12
Hungarian	535	12.5 (NS)	8.8 (NS)	13
Croatian	200	16.5 (0.001)	9.5 (NS)	14
Russian	290	10.5 (NS)	6.7 (0.04)	15
Greek	283	12.9 (NS)	8.13 (NS)	16
Turkish	499	10.6 (NS)	10 (NS)	6
Chinese	394	0.1 (<0.0001)	3.6 (<0.0001)	9
Japanese	140	0 (<0.0001)	1.8 (<0.0001)	17
Indian	346	4 (<0.0001)	8 (NS)	18
Korean	574	0 (<0.0001)	1.1 (<0.0001)	3
Egyptian	247	12 (NS)	6 (0.007)	19
Beninese	111	0 (<0.0001)	0 (<0.0001)	10
Brazilian	331	8.6 (NS)	6.5 (0.003)	20
Bolivian	778	4.8 (<0.0001)	3 (<0.0001)	21
Canadian (Native Indian)	153	3 (<0.0001)	6 (0.007)	22
Inuit	151	0 (<0.0001)	0 (<0.0001)	22

^*Fisher's exact test of Romanians versus other populations.

Table 4.

Allelic frequencies of the VKORC1 -1639 G>A polymorphism in Romanians compared with other populations

Population	Sample size	Frequency of the A allele (P-value*)	Reference
Romanian	332	42.2	Present study
Hungarian	510	39 (NS)	23
Greek	98	48.5 (NS)	24
French	222	42 (NS)	5
Italian	147	39.8 (NS)	25
Lebanese	161	52.8 (NS)	26
Iranian	126	55.56 (NS)	27
Japanese	341	91.8 (<0.0001)	28
Chinese	178	91.6 (<0.0001)	29
Indian (Northern)	102	14.22 (<0.0001)	30
South African	993	4 (<0.0001)	31

^*Fisher's exact test of Romanians versus other populations.

Discussions

CYP2C9*2 and CYP2C9*3, as demonstrated by our study, have a much higher frequency in Romanians than in Asian (Chinese, Japanese, Korean) or African populations (Ethiopians, Beninese), where these variants are very rare or sometimes absent. The two CYP2C9 polymorphisms were also significantly enriched in Romanians compared with South Americans (Bolivians, Brazilians), which have intermediate frequencies between Asians and Europeans. However, the allelic frequencies of CYP2C9*2 and CYP2C9*3 observed in our group are similar to those observed in most of the Caucasian populations living in Europe. In our group, 109 individuals (32.8%) were CYP2C9*2 or CYP2C9*3 heterozygotes, predicting a CYP2C9 intermediate enzymatic activity, whereas 14 individuals (4.2%) are predicted to have the lowest CYP2C9 enzymatic activity (those with CYP2C9*2 homozygous, CYP2C9*3 homozygous and CYP2C9*2/CYP2C9*3 compound heterozygous genotype). Assuming that CYP2C9*2 and CYP2C9*3 are homogeneously distributed in all Romanians and that other CYP2C9 defective alleles are rare in our population, we expect approximately 7,770,000 individuals harbouring at least one CYP2C9*2 or CYP2C9*3 allele, taking into account the population of Romania with approximately 21,000,000 inhabitants. Moreover, approximately 882,000 Romanians are expected to have a CYP2C9*2 homozygous, CYP2C9*3 homozygous or CYP2C9*2/CYP2C9*3 compound heterozygous genotype. These individuals would have the lowest CYP2C9 enzymatic activity, with respect to the CYP2C9*2 and CYP2C9*3 alleles. This frequency becomes important for the CYP2C9 substrates prescribed in Romania, but especially for the oral anticoagulants, in which the CYP2C9 status has major implications.

In Asians (e.g. Chinese, Japanese), the VKORC1 -1639 G>A polymorphism is extremely prevalent, with allelic frequencies for the A allele usually greater than 90%. In general, the A allele has an increasing frequency from Western Europe towards Eastern Asia, a finding that explains, at least in part, the inter-ethnical variations in the dosage requirements of the oral anticoagulants, with the Asians requiring lower doses for efficient anticoagulation. However, we found the frequency of the A allele in the Romanian group to be 42.2%, which is in good concordance with most of the Caucasian populations living in Europe, where the frequency of this allele is approximately 40%. We observed the GA and AA genotypes in 225 individuals (67.8%), which represent roughly two thirds of the group analysed. By extrapolation to the entire Romanian population, we may expect an approximate number of 14,000,000 inhabitants who would require lower doses of the oral anticoagulants. In our group, the AA homozygous genotype was observed in 55 individuals (16.6%), which represents roughly one sixth of the group analysed. Thus, we may expect approximately 3,500,000 Romanian inhabitants who would require the lowest dose of oral anticoagulants, based on the VKORC1 -1639 G>A status.

All the participants to this study were Romanians. The projections we made regarding the distribution of CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms assume a relatively homogeneous population. We made this assumption because Romania has indeed a relatively homogeneous population, around 90% of the inhabitants being Romanians. However, we admit that in certain sub-populations living in our country, we might have a different distribution of these polymorphisms, given that in general these sub-populations were not admixed with the Romanians over the centuries. Obviously, this would have important consequences regarding the oral anticoagulants metabolism in those sub-populations.

The last years have proven that the oral anticoagulants have a more complex metabolism. Along the CYP2C9 and VKORC1, new genes, such as CYP4F2, GGCX and CALU (the latter especially in African Americans) emerged as modifiers of the dose requirements 32. It would be helpful to evaluate the frequencies of the CYP4F2 rs2108622 polymorphism, which contributes to 1–7% of mean weekly warfarin dose variance and GGCX rs11676382 polymorphism, which contributes to a 6.1% reduction in warfarin dose requirement per G allele, to obtain a more complete phamacogenomic profile of the oral anticoagulants in our country.

Conclusions

This report is the first on the distribution of the CYP2C9 and VKORC1 -1639 G>A polymorphisms in a Romanian population. The distribution that we found for these polymorphisms is in good agreement with those for other European populations.

The genetic variation at the CYP2C9 and VKORC1 loci accounts for most of the inter-individual variability in the dosage requirements of oral anticoagulants. We expect a large number of Romanians harbouring at least one CYP2C9 detrimental variant or at least one A allele at the VKORC1 -1639 G>A position. The oral anticoagulants (especially acenocoumarol) are widely prescribed in our country; thus, a large number of these patients are at risk of over-anticoagulation, at least in the induction phase of the treatment. Genotyping for the CYP2C9*2, CYP2C9*3 and VKORC1 -1639 G>A polymorphisms should be performed whenever possible, prior to the initialization of the therapy, to identify the individuals at risk of over-anticoagulation. Moreover, new genetic polymorphisms such as CYP4F2 rs2108622 and GGCX rs11676382, as well as other CYP2C9 alleles, should be evaluated on the Romanian population to have a more encompassing profile of the oral anticoagulants pharmacogenomics.

Conflicts of interest

The authors confirm that there are no conflicts of interest.