Pharmacogenetics, pharmacogenomics and population admixture: implications for drug development and prescription

Sir

The implications of pharmacogenetics and geographical ancestry for improving global health, with an emphasis on its impact in developing countries, were the focus of a recent article in Nature Review Genetics¹. From my personal perspective, as a Brazilian pharmacogeneticist, an important issue relevant to this discussion, namely population admixture in many nations of the developing world, was not given the attention it duly deserves in this context². Admixture was addressed by Daar and Singer¹ with two points: first, when acknowledging the substantial (26%) Caucasian ancestry in African-Americans and second, when concluding that a “deeper understanding of the genotypes of local populations with little admixture may make it possible...to predict drug responses without the need to test each individual.” However, studies using informative genetic markers of continental ancestry indicate that admixture in the developing world is often not 'little'. For example, in Brazil, the world's fifth largest country, inhabited by an heterogeneous population of 184 million people, Native American, African and European markers are present in similar proportions in mitochondrial DNA (matrilineal ancestry) of self-identified white individuals. By contrast, their patrilineal ancestry (the Y chromosome) is >95% European³. The impact of population admixture is also evident in biparental, autosomal genetic markers⁴. Accordingly, the differences in allele frequencies between African and European ancestors with respect to genes of pharmacogenetic relevance, are attenuated (for example, CYP2A6^*1B and CYP2C9^*3 (Refs 5,6)) or absent (for example, TPMT^*3A and TPMT^*3C (Ref. 7)) between self-declared white versus black Brazilians. Furthermore, and not surprisingly, admixed populations deviate from the idea of 'race' or 'ethnic'-specific alleles of pharmacogenomic interest. As an example, the defective CYP2C9^*5 allele, previously found in African-Americans and sub-Saharan Africans, but not in Europeans and their descendants, has been recently detected in a white Brazilian man⁸. The relative contributions of European, African and Native American roots to his genetic pool were, in corresponding order, 92.0%, 7.5% and 0.5%. Significantly, mitochondrial DNA markers for matrilineal inheritance indicated the presence of the haplotype L3d, which is characteristic of West African populations, a major source of enslaved Africans who were trafficked into colonial Brazil. The CYP2C9^*5 allele was also detected in the proband's mother and in one of his brothers, which is consistent with inheritance through the matrilineal African ancestry⁹.

Collectively, the above examples, taken from a single developing nation, highlight the hazards of extrapolating pharmacogenomic data on the basis of race or ethnic categorization. The tri-hybrid roots of the Brazilian population are shared by other developing, as well as developed nations in the Americas, such as people who are categorized as Latino or Hispanic in the United States. The recent reporting of significant pharmacogenetic differences in β2–adrenergic receptor polymorphisms and bronchodilator responses to albuterol between the two largest Latino groups, namely Mexicans and Puerto Ricans⁹, is a striking example of the hazards that are associated with ignoring stratification within ethnic groups, as is often done in the pharmacogenetic literature. This concern extends to the momentous debate about the 'race-targeted' drug, BiDil (for treating heart failure in African-Americans). However, African-Americans do not share a common ancestry, having originated from a vast swathe of Africa¹⁰, and the European contribution to their genetic pool, although averaging 26%, varies widely, with some individuals having up to 70% of alleles associated with Caucasians. Because of the European contribution to their genomic pool, African-Americans carry alleles (for example, TPMT3A and CYP1A1^*2C) that have not been detected in sub-Saharan Africans. Intra-ethnic diversity, a corollary to the variable extent and dynamics of population admixture, adds complexity to the scientific appraisal, regulatory decisions and, eventually, prescription of race-targeted drugs.

Health disparities are not exclusive to the developing world, although they can be extreme and devastating, and pharmacogenetics or genomics has the potential to benefit people worldwide. Daar & Singer¹, writing from Canada, believe that pharmacogenetics “might present an opportunity [for pharmaceutical companies in developing countries], especially if they learn to harness our increasing knowledge of the link between population genomic variation and health” (italics added). The dichotomy that is implicit in this idea is inconsistent with the authors' statement that “Pharmacogenetics has so far had little impact on health care in general, or on the pharmaceutical industry in particular”, which is supposedly inclusive of the big pharmaceutical companies in developed countries. Different criteria have been proposed to distinguish pharmacogenetics from pharmacogenomics. I witness with great concern the perceived emerging trend¹ of relinquishing the idea of personalized drug therapy that is tailored to individual genetic characteristics — the original promise of pharmacogenetics — in favour of a (pharmacogenomic?) model of population-based drug development and prescription with all its potential pitfalls, especially when extended to admixed populations in developing or developed nations.