Introduction

Personalized medicine based on an individual’s genetic make-up has been becoming a reality as the need for pharmacogenomics has moved from the research setting into the clinical laboratory (Jannetto et al. 2004). However, the selection of appropriate technologies to perform personalized medicine is extremely critical, and requires careful consideration of several factors including prior knowledge of the polymorphisms, sensitivity, and specificity of the methods used, clinical sample requirements, clinical utilities, and the development cost.

Cytochrome P450 (P450) is responsible for most enzymatic oxidation reactions, and is often rate-limited for the fate of a drug in the body. Therefore, it is expected that dysfunction of P450 may lead to unexpected drug effects and toxicity (Nagata and Yamazoe 2002). Several key drug-metabolizing enzymes, such as P450 (CYP) 2D6, 2C9, 2C19, 3A4, and 3A5, are inherited mutations (or polymorphisms) that lead to different drug responses (Weinshilboum 2003; Evans and McLeod 2003). Polymorphisms of human drug metabolism are associated with specific phenotypes and genotypes. The phenotype is distinguished by the activities or contents of an enzyme. Individuals who have normal metabolic activity are called extensive metabolizers (EM) whereas individuals with defective metabolic activity are called poor metabolizers (PM; Meyer 1990). It has been reported that phenotypes associated with genetic variations often contribute to differences in drug metabolism and the dynamics of drug–drug interactions (Solus et al. 2004).

Although pharmacogenomic intervention cannot assume to be cost-effective for all clinical practice, some key aspects of cost-effectiveness of pharmacogenomics were described (Phillips and Van Bebber 2004). One aspect previously described is the prevalence of a specific genetic mutation in a population because testing was no longer cost-effective for lower mutation prevalence (Eckman et al. 2002). In this study, we screened the single nucleotide polymorphism associated with enzyme activity of CYP 2D6, 2C9, 2C19, 3A4, and 3A5, and identified a high prevalence of functional polymorphism in the Taiwanese population. The study was the first comprehensive polymorphism analysis of drug-metabolizing cytochrome genes to be performed in the Taiwanese population. The data would allow us to develop an inexpensive and simple genotyping method for detecting the high prevalence of PM and ultrarapid metabolizer (UM) alleles in the Taiwanese/Asian population.

Materials and methods

DNA preparation

Genomic DNA was extracted from 180 unrelated Han-Chinese volunteers living in Taiwan. Clinical samples were collected with written informed consent from all the participants recruited. DNA was isolated from blood samples using QIAamp DNA Blood kit (QIAGEN, Valencia, CA, USA) according to the manufacturer’s instructions. The quality of the isolated genomic DNA was checked using agarose gel electrophoresis analysis, the quantity was determined by spectrophotometer, and stored at −80°C until use.

SNP genotyping by sequencing

Fragments of DNA flanking the genomic region of the selected SNPs were amplified by ABI 9700 thermal cycler using two pairs of forward and reverse primers. The information regarding primers and genotyped SNPs is listed in Table 1. The fragments of polymer chain reaction (PCR) products were sequenced by an ABI 3700 automatic sequencer according the manufacturer’s protocol (Applied Biosystems, Foster City, CA, USA). The sequence data were analyzed using PolyPhred software to identify the potential candidate SNPs. The potential SNPs were manually checked to ensure the presence of a true SNP and the allele of each individual. Three independent manual confirmations were performed for all the sequence data and only those data that were confirmed were subjected to the subsequent statistical analysis. The identified SNPs and related primers are listed in Table 1.

Table 1 Single nucleotide polymorphisms (SNPs) of five cytochrome P-450 genes (newly identified SNPs) by polymer chain reaction (PCR)d-based sequencing methods
Full size table

Results

Six SNPs associated with the PM/UM phenotype had a high prevalence in the Taiwanese population

We investigated the frequencies of five P450 genes of PM/UM-associated alleles in 180 Han-Chinese volunteers in Taiwan. As Table 1 showed, 21 SNPs were detected by our screening in the Taiwanese population. Three SNPs, VGV 3030 in CYP2C19 exon 1, VGV1995 in CYP3A4 exon 11, and VGV2643 in CYP2D6 intron 5, were the newly identified SNPs in the current study. Six of the 21 SNPs in five genes associated with PM/UM phenotypes showed a frequency of more than 1% (Table 1). Compared with other ethnic populations, the allele frequencies of six SNPs associated with PM/UM in the Taiwanese population are similar to those of other Asian populations (Chinese–Asian, Japanese, Korean), but different from the Caucasian and African–American populations (Table 2). In Table 2, the total numbers of participants analyzed are less than 180 for some genes because those participants with missing genotyping or an ambiguous genotype calling were not analyzed.

Table 2 Allele frequencies in different ethnic populations
Full size table

High prevalence PM/UM genotype frequencies in the Taiwanese population

The high prevalence (frequencies more than 1%) of genotypes associated with abnormal metabolic rates are listed in Table 3. We did not find the homozygous CYP2C9*3 and CYP2D6*5 alleles in this study. We detected several participants who were heterozygous with both CYP2C19*2 and *3 loci. Nevertheless, we could not determine whether or not these two inactive alleles are on the same chromosome or on a different one. The participants who were heterozygous with both CYP2C19*2 and *3 loci should be defined as PM according to the haplotype analysis in the Japanese population because CYP2C19*2 and *3 were suggested to exist exclusively on the same chromosome in this previous report (Fukushima-Uesaka et al. 2005). In general, the PM, IM (intermediate metabolizer), EM and UM phenotypic categories correspond to two inactive alleles, one inactive allele, two normal active alleles, and two over-active alleles respectively. Because the CYP2D6*10 allele reduces, but does not abolish enzyme activity, being homozygous with CYP2D6*10 therefore represents the IM phenotype.

Table 3 Genotype frequencies of six SNPs identified in the present study
Full size table

Discussion

The likely response of an individual to a drug, such as the risk of a toxic event, is thus a complex equation involving multiple variables. However, inherited unique genetic polymorphisms (usually inactivating) are one of the major causes of variations in drug responsiveness (Evans and Johnson et al. 2001). It has been reported that the drug dosages used in clinical trials with east Asian participants are typically lower than those used in trials with western participants (Yu et al. 1996; Ross et al. 2001). Selecting polymorphisms with high frequencies of drug-metabolizing enzymes in different populations is a necessary and critical task. Therefore, an easy-to-use tool urgently needs to be developed by integrating the population’s specific DME polymorphisms for clinical practice in the near future.

CYP2D6*5 represents deletion of the CYP2D6 gene and its frequency is larger than 1% in the Taiwanese population. Theoretically, the genotype frequency of homozygous CYP2D6*5 is about 0.3% (0.0552) and represents the PM phenotype. However, we could not find homozygous CYP2D6*5 in this study. Similarly, CYP2D6*10 is another high frequency allele representing the IM phenotype in this study population. Most Taiwanese are homozygous CYP2D6*10 and belong to the IM phenotype. Both transheterozygous CYP2D6*5/*10 and homozygous CYP2D6*10/*10 represent the IM phenotype. Because of this, in the Taiwanese population it is likely that the IM transheterozygous genotyping with CYP2D6*5/*10 could be wrongly genotyped as CYP2D6*10/*10 due to having the same IM classification.

The CYP2C19*1/*2 and *1/*3 with the IM classification accounted for about 52% (0.449 + 0.067) of CYP2C19 genotypes. Over 50% of CYP2D6 genotyped individuals were related to the IM phenotype. The abnormal metabolic genotypes of the major drug-metabolizing cytochromes CYP2C9, CYP2C19, and CYP2D6, are usually linked to the IM phenotype. These observations may provide some indication as to why the drug dosages used in clinical trials with east Asian participants are usually lower than those used in trials with western participants.

The genotype of homozygous CYP3A5*3 is classified as PM and occurred in 47.2% Taiwanese. Previous studies on CYP3A5 in vitro and in humans have provided inconsistent information on whether CYP3A5 plays a significant role in the metabolism of CYP3A substrates in vivo (Williams et al. 2003). Further clarification is required to differentiate the relative contributions of CYP3A4 and CYP3A5 in vivo.