The effect of the cytochrome P450 CYP2C8 polymorphism on the disposition of (R)-ibuprofen enantiomer in healthy subjects

Abstract

Aims

To study the effect of CYP2C8*3, the most common CYP2C8 variant allele on the dis-position of (R)-ibuprofen and the association of CYP2C8*3 with variant CYP2C9 alleles.

Methods

Three hundred and fifty-five randomly selected Spanish Caucasians were screened for the common CYP2C8 and CYP2C9 mutations. The pharmacokinetics of (R)-ibuprofen were studied in 25 individuals grouped into different CYP2C8 genotypes.

Results

The allele frequency of CYP2C8*3 (0.17) was found to be higher than that reported for other Caucasian populations (P = 0.0001). The frequencies of CYP2C9*2 and CYP2C9*3 were 0.19 (0.16–0.21) and 0.10 (0.08–0.12), respectively. An association between CYP2C8*3 and CYP2C9*2 alleles was observed, occurring together at a frequency 2.4-fold higher than expected for a random association of alleles (P = 0.0001). The presence of the CYP2C8*3 allele was found to influence the pharmacokinetics of (R)-ibuprofen in a gene–dose effect manner. Thus, after administration of 400 mg ibuprofen, the plasma half-life (95% confidence intervals) for individuals with genotypes CYP2C8*1/*1, CYP2C8*1/*3 and CYP2C8*3/*3, was 2.0 h (1.8–2.2), 4.2 h (1.9–6.5; P < 0.05) and 9.0 h (7.8–10.2; P < 0.002), respectively. A statistically significant trend with respect to the number of variant CYP2C8*3 alleles was also observed for the area under the concentration-time curve (P < 0.025), and drug clearance (P < 0.03).

Conclusion

Polymorphism of the CYP2C8 gene was found to be common, with nearly 30% of the population studied carrying the variant CYP2C8*3 allele. The presence of the latter caused a significant effect on the disposition of (R)-ibuprofen. This suggests that a substantial proportion of Caucasian subjects may show alterations in the disposition of drugs that are CYP2C8 substrates.

Introduction

Interindividual variability in drug metabolism is a major cause of adverse drug effects. In many cases such variability is linked to polymorphisms in genes coding for drug-metabolizing enzymes. Among the enzyme families that compose the cytochrome P-450 multi-enzyme (CYP) system, the most relevant for drug metabolism are the CYP2 and CYP3 families. CYP2C enzymes account for about 20% of the total CYP in liver microsomes [1, 2] and are involved in the metabolism of about 20% of clinically used drugs [3]. All members of the CYP2C gene subfamily, namely CYP2C8, CYP2C9, CYP2C18 and CYP2C19, are polymorphic. An updated list of allelic variants is available at (http://www.imm.ki.se/CYPalleles/).

In contrast to other members of the CYP2C gene subfamily, little is known regarding CYP2C8 gene polymorphism, because mutations have only recently been described, and because most of them seem to occur at low frequencies. Four variant alleles causing amino acid changes have been reported, and are designated as CYP2C8*2 to *5, with the wild-type being CYP2C8*1[4–6]. Two of these variant alleles, CYP2C8*3 and CYP2C8*4, occur at frequencies of greater than 5% among Caucasians [5]. The CYP2C8*3 allelic variant encodes an enzyme containing two amino acid changes, namely R139K and K399R [4]. CYP2C8*2 and CYP2C8*5 have been detected in black and oriental individuals, respectively [4, 6], but do not occur among Caucasians. CYP2C8*4 does not seem to cause a significant effect on enzyme activity [5].

Heterologously expressed, CYP2C8*3 causes impaired paclitaxel metabolism [4], however no studies have been conducted in vivo to test the hypothesis that inheritance of CYP2C8*3 influences metabolism. Two factors have hampered the study of the effect of variant CYP2C8 alleles in vivo. One is the relatively low frequency of these variant alleles, which for CYP2C8*3 ranges from 0.095–0.155 in Caucasians [4, 5, 7, 8] (Table 1). This makes it difficult to identify subjects that are homozygous for CYP2C8*3, unless a large population sample is analysed. The second problem has been the identification of a suitable substrate of CYP2C8. Many are either nonspecific or too toxic to be used as probe drugs in vivo. CYP2C8 plays a significant role in the metabolism of the anticancer drug paclitaxel [5, 9, 10], the antimalarial drug amodiaquine [11], the hypoglycaemic agent troglitazone [12], as well as amiodarone [13], verapamil [14] and ibuprofen [15]. In addition, CYP2C8 plays a secondary role in the metabolism of fluvastatin [16], amitriptyline [17], perphenazine [18], diclofenac [19], gallopamil [20], omeprazole [21] and carbamazepine [22]. Out of these substrates we selected ibuprofen because it is widely used and relatively well tolerated [23]. The main phase I enzymes involved in ibuprofen metabolism are CYP2C8 and CYP2C9. Ibuprofen metabolism is stereoselective, and it has been shown that CYP2C8 plays a key role in the hydroxylation of the (R)-enantiomer [15, 24]. It may be possible, therefore, to use the disposition of (R)-ibuprofen as an index of CYP2C8 activity in vivo.

Table 1.

CYP2C8*3 allele frequencies in different Caucasian populations

Origin	Number of subjects	Allele frequency	Reference
Finnish	25	0.120	[7]
Swedish	1468	0.095	[8]
British	107	0.155	[5]
French	170	0.130	[4]
Spaniards (present study)	355	0.170	–

In this study we have analysed the CYP2C8 genotype in a large population to determine the frequency of CYP2C8^*3, and to identify individuals with nonmutated, heterozygous and homozygous mutated CYP2C8 genotypes. The effect of the different genotypes on (R)-ibuprofen pharmacokinetics was then investigated.

Methods

The protocol was approved by the Ethics Committee of the University Hospital Infanta Cristina, Badajoz, Spain. Written consent was obtained from participants, all of whom were Spanish nationals. Because in a previous study we detected significant differences in the frequency of CYP2C9 variant alleles between Spanish and other Caucasians [25], for the present study we selected a new group of 355 healthy subjects, of whom 182 were women with mean age ± SD of 45.4 ± 14.8 years, and 173 were men with mean age of 43.5 ± 14.3 years. The subjects underwent a medical examination to confirm good health. Over 95% of the subjects agreed to provide DNA samples for genotyping analyses. Blood samples were frozen immediately after collection and kept at −80 °C until analysed. Genomic DNA was prepared from leucocytes and dissolved in sterile 10 m m Tris-HCl, pH 8.0, 1 m m ethylenediaminetetraacetic acid at a final concentration of 400–600 µg ml⁻¹. The samples were stored at 4 °C in sterile plastic vials.

Genotyping analyses were carried out to detect CYP2C8*3, CYP2C9*2 and CYP2C9*3. According standard procedures, CYP2C8*1 and CYP2C9*1 variant alleles were assigned as those lacking the mutations analysed in the corresponding genes. In this regard it should be stated that other variant alleles are rare among Caucasian individuals, and therefore the risk for a misclassification of other variant alleles as CYP2C8*1 or CYP2C9*1 is negligible. Screening for CYP2C8*3 was carried out by amplification-restriction as described previously [4]. Briefly, primers 5′-CTTCCGTGCTAC ATGATGACG-3′ and 5′-CTGCTGAGAA AGGCATG AAG-3′ were used in a mismatch PCR-RFLP test. The 117 bp PCR product was purified and digested with the endonuclease XmnI. This produced two fragments of 92 and 25 bp in the CYP2C8*1 PCR product, but did not digest the CYP2C8*3 PCR product. Screening for the CYP2C9 variants was performed as described elsewhere [26].

(R)-ibuprofen pharmacokinetics

Subjects received 400 mg racemic ibuprofen (Dalsy®, Abbot Laboratories, S.A., Madrid, Spain) between 8.00 am and 10.00 am after an overnight fast. Plasma samples were collected from 0 to 12 h after administration, and immediately frozen until analysis. The determination of (R)-ibuprofen was carried by HPLC with ultraviolet detection at 220 nm by using a LiChroCart 250–4 Whelk01(S,S) column [27]. Plasma samples 0.5 ml were acidified with 0.2 ml 0.1 m HCl and vortexed. They were then extracted for 15 min with 3 ml n-hexane/diethyl ether (95/5 v/v) and centrifuged for 10 min at 3000 × g. Two millilitres of the supernatant was evaporated under a nitrogen stream and the residue was resuspended in 250 µl mobile phase composed of n-hexane/t-butyl methyl ether/acetic acid 70/30/0.1 (v/v). The limit of determination was 0.2 mg litre⁻¹. The interday and intraday coefficients of variation were less than 11% and less than 10%, respectively. Pharmacokinetic analyses were carried out using WinNonlin 1.1 software (Scientific Consulting Inc., Apex, North Carolina, USA).

Statistical analysis

The intergroup comparison values were performed using the statistical package SPSS 11.0.1. (SPSS Inc., Chigaco, Ill, USA). The Kruskal–Wallis test was used for comparison of pharmacokinetic parameters between individuals with different genotypes. The confidence intervals for the differences between means were calculated using the t-test for unpaired observations. The chi-squared (χ²) test was used for comparison of genotype frequencies and association of CYP2C8 and CYP2C9 variant alleles, unless the conditions for the application of this test were not adequate. In such cases, Fisher's exact test was used to calculate the P-value.

Results

The allele frequency (95% CI) of CYP2C8*3 in the population studied (n = 355 individuals) was 0.17 (0.14–0.20), which is significantly higher than that of the pooled frequencies reported for other Caucasian subjects (Table 1) (n = 1770, frequency (95% CI) 0.10 (0.09–0.11), 95% CI on the differences 0.040–0.099, chi-squared = 25.0, P = 0.001). Frequencies (95% CI) of 0.19 (0.16–0.21) and 0.10 (0.08–0.12) were observed for the CYP2C9*2 and CYP2C9*3 alleles, respectively, which were higher than those observed in a large study of Swedish subjects, 0.11 (0.10–0.12) and 0.07 (0.06–0.07) respectively. The differences (95% CI, P-value) between allele frequencies were 0.080 (0.049–0.111, P = 0.0001) for CYP2C9*2 and 0.031 (0.007–0.055, P = 0.001) for CYP2C9*3[8]. No unexpected bands were detected after PCR amplification of either CYP2C8 or CYP2C9 gene fragments. All variants alleles were at Hardy–Weinberg equilibrium in the population studied.

Most individuals that possessed the CYP2C8*3 allele were also carriers of CYP2C9*2 (Table 2). The proportion of individuals carrying combinations of both alleles (0.259) is higher than expected from the allele frequencies. The difference (95% CI, P-value) between the actual and the expected proportion of carriers of CYP2C8*3 and CYP2C9*2 combination was 0.149 (0.093–0.205, P = 0.0001). No association between CYP2C8*3 and the other CYP2C9 alleles was observed.

Table 2.

CYP2C8 and CYP2C9 genotypes in the Spanish Caucasian population studied

Genotype	CYP2C9*1/*1	CYP2C9*1/*2	CYP2C9*1/*3	CYP2C9*2/*2	CYP2C9*2/*3	CYP2C9*3/*3	Number of subjects	Allele frequency
CYP2C8*1/*1	158	26	52	0	3	3	242	0.682
CYP2C8*1/*3	16	73	3	6	6	1	105	0.296
CYP2C8*3/*3	1	1	0	5	1	0	8	0.023
Number of subjects	175	100	55	11	10	4	355	–
Allele frequency	0.493	0.282	0.155	0.031	0.028	0.011	–	1.000

Figure 1 shows the plasma concentration–time courses for (R)-ibuprofen in individuals with the different CYP2C8 genotypes. There were nine subjects homozygous for CYP2C8*1, 10 with the genotype CYP2C8*1/CYP2C8*3, and six who were homozygous for CYP2C8*3. None of these subjects had the CYP2C9*3 allele. To determine the effect of CYP2C9*2 allele alone, three individuals with genotype CYP2C8*3 but who lacked CYP2C9*2 were included in the heterozygous panel (Figure 1). No significant differences were observed between carriers and noncarriers of CYP2C9*2 in any of the parameters studied, and therefore the group composed of heterozygous subjects was analysed as a whole for comparison with other genotypes. Figure 1 shows that the presence of the CYP2C8*3 allele is a determinant of pharmacokinetics for (R)-ibuprofen. The pharmacokinetic parameters for individuals with no, one and two CYP2C8*3 alleles are shown in Table 3. There was a statistically significant gene–dose effect for AUC (P = 0.025), t_1/2 (P < 0.025) and drug clearance (P = 0.030).

Figure 1.

The time course of plasma (R)-ibuprofen concentrations in individuals with different CYP2C8 and CYP2C9 genotypes. Left panel: open circles correspond to nine individuals with genotypes CYP2C8*1/*1 plus CYP2C9*1/*1. Middle panel: open circles correspond to seven individuals with genotypes CYP2C8*1/*3 plus CYP2C9*1/*2, filled circles correspond to three individuals with genotypes CYP2C8*1/*3 plus CYP2C9*1/*1. Right panel: open circles correspond to five individuals with genotypes CYP2C8*3/*3 plus CYP2C9*2/*2, filled circles correspond to a subject with the genotype CYP2C8*3/*3 plus CYP2C9*1/*2

Table 3.

Pharmacokinetics of (R)-ibuprofen in individuals with different CYP2C8 genotypes

	CYP2C8*1/*1	CYP2C8*1/*3	CYP2C8*3/*3	Overall individuals
Number of subjects	9a	10b	6c	25
Cmax (mg litre−1)	20.6 (14.4–27.0)	25.3 (18.5–32.1)	15.2 (7.6–22.9)	21.2 (17.5–24.9)
Tmax (hours)	1.0 (0.6–1.3)	1.42 (0.91–1.94)	2.1 (1.6–2.6)[0.58–1.70, P < 0.001]	1.41 (1.11–1.71)
AUC (mg.h litre−1)	59.4 (35.8–83.0)	101.4 (80.1–122.7)[12.59–71.37, P < 0.008]	105.6 (65.0–146.3)[3.71–88.77, P < 0.036]	87.3 (71.8–102.8)
t1/2 (hours)	2.0 (1.8–2.2)	4.2 (1.9–6.5)[0.10–4.44, P < 0.05]	9.0 (7.8–10.2)[5.74–8.15, P < 0.001]	4.56 (3.16–5.97)
Clearance (litres h−1)	3.5 (2.5–4.4)	2.2 (1.6–2.8)[0.23–2.34, P < 0.02]	2.1 (1.3–2.9)[0.30–2.49, P < 0.03]	2.63 (2.15–3.11)

Values are mean (95% confidence intervals) data. Cmax, maximum concentration. Tmax, time required to reach peak concentration. AUC, area under the concentration—time curve. t1/2, plasma half-life.

^aAll had the CYP2C9*1/*1 genotype.

^bThree individuals had the CYP2C9/1/*1 genotype and seven individuals the CYP2C9*1/*2 genotype

^cOne subject had the CYP2C9*1/*2 genotype and five had the CYP2C9*2/*2 genotype. No subjects carrying CYP2C9*3 alleles were included. Pharmacokinetic parameters from individuals with variant genotypes were compared with those from the nine subjects lacking CYP2C8 and CYP2C9 mutations. Confidence intervals on the differences between means and P-values are shown in brackets.

Discussion

Several polymorphisms of CYP have been studied extensively, but little information about variability in the allele frequency, and the effect on enzyme activity of the CYP2C8 polymorphisms is available. In this study we observed the highest allele frequency for CYP2C8*3 described so far. Such finding is in agreement with growing evidences indicating that, even within caucasian individuals, major differences in CYP allele frequencies exist [28–32].

The frequency of CYP2C8*3 alleles in the present Spanish population is similar to that reported for British subjects, but greater than that observed in Swedish individuals (Table 1). Our data indicate that over 30% of the population analysed carry variant CYP2C8*3 alleles, making it a common polymorphism among Caucasians. The presence of CYP2C8*3 alleles generally occurred in combination with CYP2C9*2 alleles. Indeed, individuals carrying both alleles were at a 2.4-fold higher frequency than expected for a random association of alleles. In subjects homozygous for CYP2C8*3, the prevalence of CYP2C9*2 homozygotes was 20.2-fold higher than expected. This is in agreement with previous reports indicating an association between CYP2C8*3 and CYP2C9*2 alleles, although in the present population, both alleles occurred with a frequency of nearly twofold higher than that in a previous report [8].

Both CYP2C8 and CYP2C9 contribute equally to the metabolism of (R)-ibuprofen in vivo[15]. In order to minimise an effect of CYP2C9 genotype on (R)-ibuprofen pharmacokinetics, subjects carrying CYP2C9*3 alleles were excluded. However, those with CYP2C9*2 were not excluded, because in most cases CYP2C8*3 is co-inherited with CYP2C9*2. Although a role for the CYP2C9*2 enzyme in decreased R-ibuprofen metabolism cannot be excluded, our findings suggest that the effect on (R)-ibuprofen pharmacokinetics is attributable to the CYP2C8*3 allele, because there were no significant differences between the carriers of CYP2C8*3 who did or did not possess the CYP2C9*2 allele. Because of the high degree of association between CYP2C8*3 and CYP2C9*2, we could include only one subject that carried one CYP2C9*2 allele in the CYP2C8*3 homozygote group, which precluded a statistical comparison between carriers and non-carriers of CYP2C9*2. Nevertheless, the concentration versus time curve of the former subject is similar to those of subjects double homozygous for CYP2C8*3 and CYP2C9*2 alleles. Our findings are in agreement with data indicating that CYP2C9*2 does not contain mutations that affect substrate binding capacity [32], and with the observation that CYP2C9 genotype does not influence (R)-ibuprofen pharmacokinetics [27]. Nevertheless, further analyses involving sufficient numbers with different combinations of CYP2C8 and CYP2C9 genotypes are required to elucidate fully the contribution of CYP2C9*2 to decreased ibuprofen metabolism.

In summary our findings indicate that the CYP2C8 polymorphism is a determinant of the pharmacokinetics of (R)-ibuprofen. This study also suggests that the pharmacokinetics of drugs that are substrates, for example the highly toxic paclitaxel of the CYP2C8 enzyme, may be altered in nearly 20–30% of Caucasian subjects, depending on their ethnic origin. Further studies should focus on the impact of CYP2C8 polymorphism on the disposition and/or side effects caused by these drugs. Genotyping techniques, or phenotyping methods such as that described here, may be useful tools to predict impaired metabolism when using drugs that are substrates of CYP2C8 and that have reduced therapeutic range.