Introduction

CYP2C19 is a polymorphic enzyme involved in the metabolism of around 15% of the drugs prescribed, including anxiolytics, antidepressants, antineoplastics, proton pump inhibitors and antiplatelet agents, among others (Rosemary and Adithan 2007). Differences for these drugs have been observed in pharmacologic responses and adverse drug reactions due to several factors, such as the presence of single-nucleotide polymorphisms (SNPs) in the CYP2C19 gene that generate several alleles, influencing the enzymatic function (Li-Wan-Po et al. 2009). These alleles associated with the drug’s metabolic ability define the CYP2C19 genotype, which can be related to the following four metabolizer phenotypes: poor metabolizer (PM), homozygous or compound heterozygous genotypes for the loss-of-function (LOF) CYP2C19 alleles; intermediate metabolizer (IM), heterozygous genotypes for the LOF CYP2C19 alleles; extensive metabolizer (EM), carrying two functional or wild-type alleles; and ultrarapid metabolizer (UM), both heterozygous and homozygous for the CYP2C19*17 that increase gene expression and thereby CYP2C19 activity (Sim et al. 2006; Santos et al. 2011). However, it has been claimed that only homozygous CYP2C19*17/*17 displays the UM phenotype (Li-Wan-Po et al. 2009). In Latin America, population studies have shown differences in the genetic diversity of CYP2C19, mainly regarding LOF alleles related to the PM phenotype, such as CYP2C19*2, CYP2C19*3, CYP2C19*4 and CYP2C19*5, principally (Bravo-Villalta et al. 2005; Luo et al. 2006; Isaza et al. 2007; Linden et al. 2009; Hoyo-Vadillo et al. 2010; Salazar-Flores et al. 2012). However, scarce population data exists in Latin American for CYP2C19*17, the unique allele associated to the UM phenotype. This allele shows relatively high frequencies in African American (21%), European (range 18–27%), and Brazilian populations (range 15.8–26.3%) (Santos et al. 2011); whereas low frequencies have been found in Hispanic American (12%) and Asian populations (0–4%) (Kearns et al. 2010).

In Mexico, in addition to the wild-type allele CYP2C19*1, the CYP2C19*2 allele has almost exclusively been detected in Mestizos (range 6.9–10.3%) and in Amerindian groups (range 3.6–31%) (Hoyo-Vadillo et al. 2010; Salazar-Flores et al. 2012). Interestingly, one previous pharmacokinetic study in western Mexico, the same population studied here, estimated the frequency of PM (6%), EM (90%), and UM (4%) phenotypes using Omeprazole as probe drug (González et al. 2003). Therefore, the primary purpose of this study was to detail the genetic variability of CYP2C19 in western Mexicans, and to check its correspondence with: i) previous genetic population studies of CYP2C19 throughout Latin America; and ii) the metabolizer phenotype frequency previously described in the same Mexican population.

Material and methods

Population sample

The total population sample consisted of 238 healthy unrelated Mexican Mestizos from the state of Jalisco (western Mexico). Mestizos are the result of admixture mostly between Spaniards, Amerindians and Africans after the European contact, and presently they represent most of the Mexican population (∼90%) (Salazar-Flores et al. 2012). All volunteers for the study signed a written informed consent form according to the ethical guidelines of the Helsinki Declaration. The protocol was approved by the Committee of Ethics and Research of the CUCiénega (UdeG).

Methods

Genomic DNA was extracted from peripheral blood using the standard phenol–chloroform method (Sambrook et al. 1989), and it was quantified with a Nanodrop 2000 TM instrument (Thermo Scientific, Wilmington, USA). To analyse the genetic variability of CY2C19, the alleles CYP2C19*2, CYP2C19*3, CYP2C19*4 and CYP2C19*5 were genotyped by SNaPshot TM following the same conditions described in a previous report (Salazar-Flores et al. 2012). Detection of CYP2C19*17 (rs12248560, c. −806 C_T) was performed by TaqMan qPCR assay according to the manufacturer’s instructions (Applied Biosystems, Foster City, USA); the Step One TM Real-Time PCR system was employed for this purpose (Applied Biosystems). Moreover, a subset of samples was regenotyped for CYP2C19*2 (rs4244285, c.19154 G_A) by TaqMan qPCR assay to ensure accuracy. The presence of wild-type allele CYP2C19*1 was deduced in each individual by the absence of the SNPs studied here. The phenotype was inferred theoretically from the genotype, as mentioned in the Introduction section. Allele and genotype frequencies were determined by the gene counting method. For the SNPs detected, Hardy–Weinberg equilibrium (HWE) and pairwise comparisons with previous reports were carried out by chi-square tests (http://www.quantpsy.org/chisq/chisq.htm).

Results

We describe the genetic variation of CY2C19 in Mexican-Mestizos. Although this issue has been addressed previously (Hoyo-Vadillo et al. 2010; Salazar-Flores et al. 2012), this report includes for the first time the analysis of CYP2C19*17 in a Mexican population sample. We only detected the polymorphisms CYP2C19*2 and CYP2C19*17 with frequencies of 8.6% and 14.3%, respectively; thus, the wild-type allele CYP2C19*1 exhibited the highest frequency with 77.1% (table 1). Genotype frequencies involving alleles CYP2C19*1, CYP2C19*2 and CYP2C19*17 were distributed according to the HWE (P = 1.000). Dual genotyping of CYP2C19*2 (TaqManqPCR vs. SNaPshot) allowed confirmation of dubious SNaPshot results, specifically for some heterozygous CYP2C19*1/*2, which could be confused with homozygous CYP2C19*1/*1 due to the low signal for the CYP2C19*2 allele. Finally, all CYP2C19 genotypes were established without doubt.

Table 1 Genetic variation, predicted and observed phenotypes for CYP2C19 in a western Mexican Mestizo population.
Full size table

We compared the allele frequencies estimated here for CYP2C19*2 andCYP2C19*17 with respect to previously studied Latin American populations (table 2), which are characterized by the absence or negligible frequency (≤0.1%) of CYP2C19*3, CYP2C19*4 and CYP2C19*5 alleles. The allele frequency estimated here for CYP2C19*2 was similar to previous reports from Latin American and Native American populations (χ 2=8.071; P = 0.4265), except for Caucasian Brazilians (P<0.00002), probably explained by higher European ancestry than the rest of the populations. Similarly, the CYP2C19*17 allele frequency of western Mexicans was comparable with Hispanic Americans and Native Americans (table 2). Finally, the global differentiation test (excluding Caucasian Brazilians) demonstrated homogeneity among these populations for CYP2C19*2 and CYP2C19*17 (P>0.05).

Table 2 Allelic frequency distribution of CYP2C19*2, CYP2C19*3 and CYP2C19*17 in Latin and Native American populations.
Full size table

Discussion

Among alleles related to the PM phenotype, the nearly exclusive presence of CYP2C19*2 (8.6%) in the studied population is in agreement with previous reports on Mexican, Latin American and Native American populations (table 2). As could be expected by the allele frequencies estimated in Mexican Mestizos, the main genotype is the wild-type CYP2C19*1/*1 (60.1%), followed by genotypes related to the UM phenotype CYP2C19*1/*17 (21%), and CYP2C19*17/*17 (2.1%). Conversely, unique genotypes related to IM and PM phenotypes observed here were CYP2C19*1/*2 (13%) and CYP2C19*2/*2 (0.4%), respectively (table 1).

The prevalence of homozygous CYP2C19*2/*2 estimated in this work (0.4%), and those previously estimated in Mexico (≤1.4%), suggest a low frequency of the PM phenotype in this country (Hoyo-Vadillo et al. 2010; Salazar-Flores et al. 2012). However, the predicted PM frequency based on the genotype CYP2C19*2/*2 was in disagreement with the PMs frequency observed in the pharmacokinetic study of González et al. (2003) (0.4 vs. 6.3%; χ 2=9.58, P = 0.00196), who used omeprazole as probe-drug to estimate the prevalence of CYP2C19 phenotypes in Mexican population (table 1). Although both the expression of the gene CYP3A4 and environmental factors might influence the PM phenotype for CYP2C19, this finding could also imply additional CYP2C19 alleles (Andersson et al. 1994; González et al. 2003; Linden et al. 2009; Kearns et al. 2010), similar to the Amerindian mutation CYP2D6*82 recently described in Mexican population (Contreras et al. 2011). The plausible presence of novel alleles is consistent with the elevated genetic variability described for CYP2C19, which includes at least 28 polymorphic alleles and encourages deeper studies to explore new functional mutations (Rosemary and Adithan 2007). Therefore, to confirm this hypothesis for CYP2C19, DNA sequencing would be required in Mexican individuals with PM phenotype but discordant CYP2C19 genotype (e.g. CYP2C19*1/*1), which by now is out of the purposes of this study.

CYP2C19 genotype frequencies estimated in the Mexican population analysed are ostensibly correlated with the frequency of altered metabolizer phenotypes (table 1). In this case, UM would be the most prevalent altered phenotype based on the frequency of the CYP2C19*17 allele. However, there are two criteria for predicting the prevalence of UMs: some authors describe CYP2C19*1/*17 individuals as UMs (Wang et al. 2009), whereas others consider them as EMs, within the same group of the wild-type homozygous CYP2C19*1/*1 (Sugimoto et al. 2008). A previous study based on a MEDLINE search shed light to this topic: researchers concluded that only homozygous CYP2C19*17/*17 could be considered UMs and that they are likely at an increased risk to suffer therapeutic ineffectiveness, principally for drugs with a very narrow therapeutic window, such as Clopidogrel (Li-Wan-Po et al. 2009). We were able to contribute to this topic by comparing our results with the previously mentioned pharmacokinetic study on CYP2C19 in Mexican population González et al. (2003) (table 1). Interestingly, the 4% prevalence for the UM phenotype is in agreement with the 2.1% prevalence of homozygous CYP2C19*17/*17 observed here (χ 2=1.048; P = 0.306), assuming that this would be the unique genotype resulting in UM (table 1). Therefore, our results support the hypothesis of the recessive inheritance of CYP2C19*17 which by definition comprises the allele that causes the phenotype only in homozygous, and never in heterozygous. Eventually, this inheritance pattern must be confirmed by comparison between the CYP2C19*17 genotype and the pharmacokinetic profile in a population sample.

To our knowledge, CYP2C19*17 is the unique allele related to the UM phenotype (Li-Wan-Po et al. 2009; Kearns et al. 2010). Thus, under a dominant inheritance pattern of CYP2C19*17, the observed genotypes including this allele, such as CYP2C19*17/*17 and CYP2C19*1/*17, should correspond to the prevalence of UMs. We tested this hypothesis, comparing the predicted frequency of UMs based on the genotypes CYP2C19*17/*17 plus CYP2C19*1/*17 with regard to the prevalence of UMs previously observed in the pharmacokinetic study (23.1 vs. 4%; χ 2=22.16; P = 0.0000025) (González et al. 2003) (table 1). This difference allows discarding the dominant inheritance pattern of CYP2C19*17 for expressing the UM phenotype at least for those obtained by the administration of omeprazol as the probe drug.

The importance of discarding the dominant inheritance pattern of CYP2C19*17 is based on the significant reduction of the predicted frequency of UMs in Mexican population based on CYP2C19*17 (dominant 23.1% vs. recessive 2.1%). This fact would impact the therapeutic activity of various clinically relevant drugs, such as voriconazole, tamoxifen, and proton pump inhibitors among others. The frequency of the CYP2C19*17 allele in Mexico, and the knowledge of its inheritance pattern permit us to estimate a relatively low health-care impact based on the predicted prevalence of UMs (CYP2C19*17/*17 = 2.1%), as well as PMs (CYP2C19*2/*2 = 0.4%). Interestingly, these conclusions are probably valid for Native American and Latin American populations as suggested by the homogeneous distribution demonstrated for CYP2C19 at the continental level.

Conclusion

In brief, we describe the genetic variability of CYP2C19 in a Mexican population which was similar to Latin American and Native American populations. Comparison with a previous pharmacokinetic report on the same Mexican population allowed the following: i) to suggest the probable presence of novel CYP2C19 alleles related to the PM phenotype in our population; and ii) to discard the dominant inheritance pattern of allele CYP2C19*17 for resulting in the UM phenotype.