Farmacogenética populacional e ancestralidade na América Latina

Abstract

Polymorphism frequencies in genes coding for drug metabolizing enzymes vary among distinct ethnic groups. In admixed populations, such as Latin Americans, the knowledge of pharmacogenetic characteristics for distinct Latin American ethnic groups may be relevant for the implementation of therapeutic strategies adapted to them. The goal of this dissertation is to verify the ancestry and the frequency of pharmacogenetic variants in Latin American populations, aiming to clarify frequency patterns in Latin America. In two admixture studies using Brazilian populations comprising EPIGEN project we concluded that Southeast/South cohorts are predominantly Europeans (from North Europe/Middle East) and less Africans (from Center-East Africa), whereas Northeast cohort is prevalently African (from Center-West Africa) and less European (from Iberic Peninsula). Further, we found a correlation between genomic and self-reported ancestry in Brazilians. To characterize frequencies of the main variants from worldwide populations, we presented three systematic reviews considering CYP2D6, CYP2C9 and CYP2C19 genes, concluding that there are worldwide patterns in allele frequencies distributions. In a Latin American context, we showed in a systematic review using 121 pharmacogenetic biomarkers that patterns in this population are complicated due to its complex demographic history. Moreover, we presented an article using original data from the Iberoamerican Pharmacogenetics Network, particularly in Costa Rica, showing a significant correlation between genomic ancestry and CYPs in three populations from this country. Furthermore, from an admixture mapping study which associated Native American ancestry with increased risk of acute lymphoblastic leukemia relapse due to two SNPs, we verified that this genomic region is differentiated in Latin American populations, especially Native Americans. So, in a population and epidemiologic context, self-reported and genomic ancestry are informative to pharmacogenetic inferences. However, pharmacogenetic clinical implementation in Latin Americans must be performed by individual genotyping because drug response cannot be predicted by ancestry.