- Top of page
- Subjects and Methods
Abstract: The frequency distribution of SNPs and haplotypes in the ABCB1, SLCO1B1 and SLCO1B3 genes varies largely among continental populations. This variation can lead to biases in pharmacogenetic studies conducted in admixed populations such as those from Brazil and other Latin American countries. The aim of this study was to evaluate the influence of self-reported colour, geographical origin and genomic ancestry on distributions of the ABCB1, SLCO1B1 and SLCO1B3 polymorphisms and derived haplotypes in admixed Brazilian populations. A total of 1039 healthy adults from the north, north-east, south-east and south of Brazil were recruited for this investigation. The c.388A>G (rs2306283), c.463C>A (rs11045819) and c.521T>C (rs4149056) SNPs in the SLCO1B1 gene and c.334T>G (rs4149117) and c.699G>A (rs7311358) SNPs in the SLCO1B3 gene were determined by Taqman 5′-nuclease assays. The ABCB1 c.1236C>T (rs1128503), c.2677G>T/A (rs2032582) and c.3435C>T (rs1045642) polymorphisms were genotyped using a previously described single-base extension/termination method. The results showed that genotype and haplotype distributions are highly variable among populations of the same self-reported colour and geographical region. However, genomic ancestry showed that these associations are better explained by a continuous variable. The influence of ancestry on the distribution of alleles and haplotype frequencies was more evident in variants with large differences in allele frequencies between European and African populations. Design and interpretation of pharmacogenetic studies using these transporter genes should include genomic controls to avoid spurious conclusions based on improper matching of study cohorts from Brazilian populations and other highly admixed populations.
The efficacy of drug therapy results from complex interplay between multiple processes that govern drug disposition and response. Although passive diffusion accounts for cellular uptake of some drugs and metabolites, increased emphasis is being placed on the role of membrane transporters in absorption of oral medications across the gastrointestinal tract. Many current United States Food and Drug Administration (FDA)-approved drugs are substrates of these transporters .
The MDR1 multidrug transporter is one of the better-characterized members of the ATP-binding cassette (ABC) family of transporters. Although this transporter was initially identified in the context of multidrug resistance (MDR) against anticancer drugs, its range of known drugs and substrates has greatly expanded. In addition to chemotherapeutic drugs, the MDR1 transporter has been found to transport a wide variety of substrates representing nearly every category of clinically important drugs, including anti-arrhythmics, antidepressants, antipsychotics and antivirals [2–4].
More than 50 polymorphisms have been described in the ABCB1 gene . Single-nucleotide polymorphisms (SNPs) c.1236C>T (Gly411Gly) in exon 12, c.2677G>T/A (Ala893Ser/Thr) in exon 21 and c.3435C>T (Ile1145Ile) in exon 26 are three SNPS that have been most commonly investigated in pharmacogenetic studies [6,7]. The c.2677G>T/A SNP causes an amino acid change within a structurally important transmembrane domain of the translated protein. However, the effects of this polymorphism are controversial and drug-specific [6,8–11]. The c.3435C>T SNP is associated with a decrease in mRNA stability and lower expression levels . The organic anion-transporting polypeptides (OATPs) are sodium-independent transporters encoded by genes of the solute carrier family SLCO. These transporters are present in the basolateral membrane of hepatocytes and are major determinants of uptake of several drugs from the portal circulation into hepatocytes [1,13,14]. OATP1B1 and OATP1B3 are encoded by SLCO1B1 and SLCO1B3, respectively. Several SNPs and other sequence variations have been described in SLCO1B1 [15,16], some of which are associated with altered transport in vivo and in vitro. The c.388A>G (Asn130Asp), c.463C>A (Pro155Thr) and c.521T>C (Val174Ala) SNPs and their derived haplotypes were the focus of several investigations into drug transport, efficacy and tolerance. Genetic variants of the SLCO1B1 gene have been shown to reduce hepatic uptake and increase plasma concentrations of statins . In vivo studies suggested that the c.388A>G variant is associated with increased OATP1B1 activity, and the c.521T>C SNP is associated with reduced transport in vitro. A genome-wide study from the SEARCH collaborative group  showed that carriers of 521C allele taking 80 mg of simvastatin were found to be at a significantly increased risk of myopathy (odds ratio 4.5; 95% confidence interval 2.6–7.7).
The most common SNPs at the SLCO1B3 locus are c.699G>A (Met233Ile) and c.344T>G (Ser112Ala). These SNPs were associated with altered transport activity in vitro in COS-7 cells  but not in other cell lines [20,21]. The influence of the 699G>A/334T>G haplotype in digoxin pharmacokinetics was recently demonstrated by Tsujimoto et al. .
The Brazilian population was formed by successive migratory waves. Amerindian people occupied the Brazilian territory when the Portuguese arrived in 1500 and colonized the country. Then, between the 16th and 19th centuries, Africans were brought to Brazil as slaves. In addition to the Portuguese, other migratory waves occurred in the 19th and 20th centuries, mainly from Italy, Germany and Spain . All of these migratory events contributed to the formation of a multi-ethnic and highly admixed population. This heterogeneity was documented in several genetic studies that used either uniparental or autosomal markers to demonstrate a typical, although non-uniform, tri-ethnic (European, African and Amerindian) pattern for the Brazilian population. This admixture process occurred in different ways in the various geographical regions of the country. In north-eastern Brazil, the African contribution is high; in the north, the contribution of Native Americans is pronounced; and in the south, there are reduced Amerindian and African influences when compared with the other geographical regions [23–27].
The frequency distribution of SNPs and haplotypes in the ABCB1, SLCO1B1 and SLCO1B3 genes varies largely among continental populations [17,28–32]. This variation can lead to biases in pharmacogenetic studies conducted in admixed populations such as those from Brazil and other Latin American countries. Because the proportion of Amerindian, African and European ancestry varies in each individual and between Brazilian regions, the ‘racial’ classification officially used in Brazil based on self-perceived skin colour [33,34] can misrepresent real stratifications in pharmacogenetic association studies. In the present study, we evaluated the influence of ‘racial’ classifications, geographical origins and genetic ancestry in the distribution of ABCB1, SLCO1B1 and SLCO1B3 polymorphisms and haplotypes in a representative sample of the Brazilian population.
- Top of page
- Subjects and Methods
The present investigation reports the first large-scale study of ABCB1, SLCO1B1 and SLCO1B3 polymorphism diversity in a representative sample of the Brazilian population. Within a country the size of Brazil, population composition varies widely among regions. Therefore, it is not unexpected that population heterogeneity and diversity influenced the distribution of ABCB1, SLCO1B1 and SLCO1B3 genotypes and haplotypes within self-reported colour groups, across geographical regions, and according to the proportions of European, Amerindian and African genetic ancestry [23,34,39].
Individuals differ genetically in their susceptibility to particular diseases and their responses to various drugs. However, personalized treatments are difficult to develop because disease susceptibility and drug response generally have poorly characterized genetic determinants. It is therefore tempting to use the ethnicity of patients as a proxy for some of the variation in allele frequencies in genes that could underlie these traits . Transporter gene polymorphisms are known to occur at variable frequencies in different continental populations [14,17,29,30,32,40], but fine geographical distribution information is not available for most populations.
ABCB1 polymorphisms display a marked interethnic variation that is possibly a result of selection owing to specific environmental pressures. The observed trend in the Brazilian population for increased frequency of the ABCB1 c.1236T, c.2677nonG and c.3435T alleles and the T-NonG-T haplotype as the average proportion of European ancestry increases is compatible with the reported frequencies of these variants in Europe and sub-Saharan Africa [28,31,32,41–45]. The allele and haplotype distributions are similar to those reported in other Brazilian studies, but they differ from those described in African American or European-derived American populations. This is probably due to different levels of admixture in these populations [31,32,45]. These results did not differ when self-reported colour or genomic ancestry was considered.
The SLCO1B1 c.388A>G SNP was significantly associated with colour and geographical region and the interaction between these factors. Our study demonstrates that the SLCO1B1 c.388G variant was also strongly associated with decreases in European ancestry and increases in African ancestry. The c.521T>C SNP distribution was weakly associated with colour; however, this polymorphism was not associated with genetic ancestry. The c.463C>A SNP was not associated with colour, region or genetic ancestry. The haplotype frequencies of SLCO1B1 varied among regions and self-reported skin colour, but when genomic ancestry was considered, only the SCLO1B1*1b haplotype was associated with increased African ancestry. The global analyses reported by Pasanen et al.  showed that the c.521T>C SNP varied markedly between populations. The lowest c.521C frequencies were observed in sub-Saharan Africans and the highest in Native American populations. In contrast, the c.388G allele at the c.388A>G SNP had a higher prevalence in Africa. Pasanen et al.  have also shown that the SLCO1B1*1b haplotype was the most common haplotype in African and Native American populations, whereas haplotype *5 was only seen in Europe and the Middle East. In our admixed population, all haplotypes were identified in individuals self-classified as white or brown. However, not all haplotypes were identified in those self-classified as black, despite high levels of admixture . Although haplotype *1b occurs with a high frequency in Native Americans, we did not observe a significantly higher contribution of Amerindian ancestry in carriers of this haplotype in admixed Brazilians.
The SLCO1B3 gene has not been previously investigated to the extent of SCLO1B1 and ABCB1. This distinguishes between what has been previously described in this paper and the previous investigations into other groups. The frequencies of the SLCO1B3 c.344T>G and c.699G>A SNPs are similar in Caucasian and Asian populations but markedly lower in sub-Saharan Africans. The frequencies of the SLCO1B3 c.344T>G and c.699G>A SNPs were the major alleles in both Caucasian and Asian populations, whereas the major alleles in sub-Saharan Africans were c.344T and c.699G . In Brazilians, the SLCO1B3 haplotypes were associated with self-reported colour, geographical region, the interaction between colour and geographical region and genomic ancestry. The G-A haplotype was strongly associated with European ancestry, which was expected from worldwide population studies.
Geographical, social and cultural barriers have given rise to reproductively isolated human populations. Within these populations, random drift and/or natural selection has produced genetic differentiation. Historically, proxies such as skin colour, race and ethnicity have been used to make inferences regarding population structure, even in the absence of corroborative genetic data . As a result, there is a large body of literature comparing phenotypes between cohorts defined, for example, as ‘blacks’ and ‘whites’. It has also been shown that even when using proxies such as skin colour to match cases and controls, hidden admixture can still occur. This has been demonstrated by Shriver et al.  in European American populations and by the present study in Brazilian populations.
One significant consequence of population genetic structures is increased confounding in case–control association studies. In areas where people from different regions have mixed extensively, such as in Brazil, the connection between skin colour and ancestry has been substantially weakened .
One major goal of the present study was to quantify the correspondence between self-identified race/ethnicity and genetic ancestry in genes that are important for pharmacogenetic studies. In addition, because case and control subjects are sometimes recruited from different geographical regions and because pooled samples are sometimes used for genome-wide association studies’ (GWAS) purposes, it is important to evaluate the assumption that matching only at the level of self-identified race/ethnicity is sufficient. In the present investigation, we demonstrate that ancestry in Brazilians is better explained by a continuous ancestral variable. This variation could lead to stratification bias if genomic controls are not included in pharmacogenetic analyses. Moreover, the intrinsic heterogeneity of the Brazilian population must be acknowledged in the design and interpretation of pharmacogenetic studies using these transporter genes to avoid spurious conclusions based on improper matching of study cohorts.