As the 5-HTTLPR polymorphism was identified, about 20 years ago, it has been one of the most extensively investigated polymorphisms in neuropsychiatric disorders. Between only 2012 and 2014, 221 studies were published addressing the association between 5-HTTLPR polymorphism and a wide range of different phenotypes, including neuropsychiatric disorders.
The 5-HTTLPR polymorphism consists of a 44-bp insertion/deletion in the 5′-regulatory region of the serotonin transporter gene (5HTT).1 This polymorphism is associated with variations in transcriptional activity: the long variant (L; 44-bp insertion inside an imperfect repeat sequence region) has approximately three times the basal activity of the short promoter (S; 44-bp deletion).2 Functionally, the 5-HTTLPR-S leads to lower quantities of 5HTT, reducing the uptake of the neurotransmitter serotonin into the pre-synaptic cells of the brain. Other rare 5-HTTLPR alleles have been described,3 but most of the studies focused in the two common S- and L-alleles.
Several investigations conducted in different worldwide populations suggested that the S-allele may confer an increased genetic risk for affective and neuropsychiatric disorders. Note, however, that in case–controls studies significant allele frequency differences between the two sample sets may be unrelated to the causal variant of the particular phenotype under investigation. False-positive/negative results and/or lack of reproducibility in association studies have frequently been associated with population stratification due to admixture. Different patterns of nonrandom association of alleles (linkage disequilibrium, LD) between the causal variant and the marker(s) could also be cited as another important factor that leads to errors in association studies.4
There is no study evaluating in large scale the continental distributions of the frequency of the 5-HTTLPR polymorphism. This task can be made through an extensive compilation of the control samples of medical studies. We have performed this survey, and found that the average frequencies of the S-allele in Europe, Asia and Africa are 43%, 63% and 23%, respectively (Supplementary Table S1). Until recently, the amplitude of the 5-HTTLPR variation considering Native Americans was entirely unknown. Bisso-Machado et al.5 designed a study to fill this gap, genotyping 230 individuals from 17 Native American populations and found a surprising result: allele S presented the highest frequencies so far available, between 68 and 100% (average 90%). The observation of these different numbers across human populations has stimulated researchers to postulate adaptive hypotheses to explain them.6 Independently of the random and/or deterministic causes behind this human diversity pattern, an allele presenting wide variation in given ancestral groups can certainly represent strong confounding effects on genetic association studies performed, for instance, in Latin American admixed populations, as their gene pool were formed by relatively recent events involving three major distinct parental stocks (African, European and Native American).
In addition, heterogeneity within these populations and/or geographical group categories should be considered. According to the World Populations Prospects of the United Nations (http://esa.un.org/wpp/documentation/WPP%202010%20publications.htm), the estimated population of Latin America was over half a billion in 2011. However, commonly, the 'Latino' classification is used to categorize distinct Latin American populations, not considering the strong degree of genetic heterogeneity that exists within and among them. For instance, Ruiz-Linares et al.7 studied 7342 subjects from five countries, and found that the Native American average ancestry ranged from 9% in Brazil to 64% in Peru. Furthermore, wide variation in ancestry proportions within each country was also described.
On the basis of the LD analysis, Vinkhuyzen et al.8 suggested that two single-nucleotide polymorphisms (SNPs; rs2129785 and rs11867581), identified from the commercial Illumina genome-wide SNP platform, could be used as proxies for 5-HTTLPR, as the TA haplotype (frequency=0.412) was coupled with allele 5-HTTLPR-S. In other words, available databases that include subjects whose phenotypic traits of interest are known could be used to investigate their association with 5-HTTLPR alleles.
We decided to test the strength of these two putative tag SNPs in Native American populations (Table 1), owing to the important contribution of this parental group to the formation of contemporary Latin American populations. We genotyped both SNPs in 120 Native Americans from 14 populations whose data for the 5-HTTLPR polymorphism are now known.5 The Tagger option within the Haploview software (http://www.broadinstitute.org/scientific-community/science/programs/medical-and-population-genetics/haploview/downloads) was used for haplotype prediction. Interestingly, whereas Vinkhuyzen et al.8 showed that haplotype T (rs2129785) A (rs11867581) tags with allele 5-HTTLPR-S (r2=0.775), our results showed similar probabilities for TA (frequency=0.370) or TG (frequency=0.452) to be coupled with S, resulting in a low proxies for 5-HTTLPR (r2= 0.1132). In addition, data from a Paleo-Indian whose genome is known (named ‘Anzick’, an individual belonging to the classical North American hunter-gathering Clovis culture who lived around 13 000 calendar years ago)9 was also considered in present study. Anzick was homozygote for the 5-HTTLPR-L, rs2129785-T and rs11867581-G alleles, relatively rare in present-day Amerindians. These results suggest an even more heterogeneous population scenario, certainly connected with the complex evolutionary and demographic histories of Native Americans. For instance, they have experimented marked episodes of genetic drift, and successive bottleneck events. Furthermore, signals of positive natural selection associated with autochthonous environmental and cultural conditions have also been described. These phenomena can easily lead to distinct LD patterns in the Amerindian group when compared with the others who contributed to the formation of the modern Latin American populations. Note, also, that different admixture models (for example, hybrid isolation or continuous-gene flow) create distinct LD patterns.10
In conclusion, the two putative tag SNPs cannot be used as proxies for 5-HTTLPR, at least considering Native Americans and their descendents.
Heils A, Teufel A, Petri S, Stöber G, Riederer P, Bengel D et al J Neurochem 1996; 66: 2621–2624.
Lesch KP, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S et al Science 1996; 274: 1527–1531.
Ehli EA, Hu Y, Lengyel-Nelson T, Hudziak JJ, Davies GE . Mol Psychiatry 2012; 17: 185–192.
Cardon LR, Palmer LJ . Lancet 2003; 361: 598–604.
Bisso-Machado R, Ramallo V, Tarazona-Santos E, Salzano FM, Bortolini MC, Hünemeier T . Am J Phys Anthropol 2013; 151: 492–494.
Dobson SD, Brent LJ . Front Hum Neurosci 2013; 7: 588.
Ruiz-Linares A, Adhikari K, Acuña-Alonzo V, Quinto-Sanchez M, Jaramillo C, Arias W et al PLoS Genet 2014; 10: e1004572.
Vinkhuyzen AA, Dumenil T, Ryan L, Gordon SD, Henders AK, Madden PA et al Mol Psychiatry 2011; 16: 1073–1075.
Rasmussen M, Anzick SL, Waters MR, Skoglund P, DeGiorgio M, Stafford TW Jr et al Nature 2014; 506: 225–229.
Long JC . Genetics 1991; 127: 417–428.
We are grateful to Aline Brugnera Felkl for the technical assistance. This study was supported by CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil).
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on the Molecular Psychiatry website
About this article
Cite this article
Hünemeier, T., Bisso-Machado, R., Salzano, F. et al. Native American ancestry leads to complexity in 5-HTTLPR polymorphism association studies. Mol Psychiatry 20, 659–660 (2015). https://doi.org/10.1038/mp.2015.39