Portal de Periódicos da CAPES

Evolutionary Genetics of Coronary Heart Disease

2009; Lippincott Williams & Wilkins; Volume: 119; Issue: 3 Linguagem: Inglês

10.1161/circulationaha.108.809970

1524-4539

Keyue Ding, Iftikhar J. Kullo,

Genetic Associations and Epidemiology

HomeCirculationVol. 119, No. 3Evolutionary Genetics of Coronary Heart Disease Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessResearch ArticlePDF/EPUBEvolutionary Genetics of Coronary Heart Disease Keyue Ding and Iftikhar J. Kullo Keyue DingKeyue Ding From the Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minn. and Iftikhar J. KulloIftikhar J. Kullo From the Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minn. Originally published27 Jan 2009https://doi.org/10.1161/CIRCULATIONAHA.108.809970Circulation. 2009;119:459–467Nothing in biology makes sense except in the light of evolution.1— —Theodosius Dobzhanksy, 1973No biological problem is solved until both the proximate and the evolutionary causation has been elucidated. Furthermore, the study of evolutionary causes is as legitimate a part of biology as is the study of the usually physico-chemical proximate causes.2— —Ernst W. Mayr, 1982Susceptibility to common diseases such as coronary heart disease (CHD) may in part reflect historical or evolutionary legacies,3,4 and interest in studying evolutionary biology to gain novel insights into human health and disease is increasing. The evolutionary history of the human species may provide valuable insights into the origin of common diseases beyond what is possible by investigating only the most immediate or "proximal" causes of disease. The potential role of evolutionary biology in explaining disease causation was highlighted by Williams and Nesse3 and is often referred to as darwinian medicine. Although the relevance of an evolutionary perspective may vary depending on the disease under study, a strong argument could be made for studying the evolutionary genetics of CHD, a leading cause of human morbidity and death.Within the last decade, several important advances have made it possible to study "modern" diseases from an evolutionary perspective. The Human Genome Project5 provided a reference human genome, and the subsequent International HapMap Project6,7 described genetic variations (mostly single nucleotide polymorphisms [SNPs]) among individuals and the patterns of variation across the genome. Both projects provide the raw material to study natural selection in the human genome, as reviewed in recent reports.8–10 In addition, sequencing of multiple other genomes, including those of primates, provides a framework for generating important insights into the origin and expression of human diseases.4 Such an avenue of investigation may help answer why, compared with their closest relative, the chimpanzee, humans are susceptible to CHD,11–13 why CHD has assumed epidemic proportions,14 and why substantial differences in susceptibility to CHD are present between ethnic groups.15 Genetic factors underlying both "complex" and mendelian diseases may be affected by natural selection16; conversely, regions of the human genome under natural selection17 are likely to harbor functional loci (gene coding and regulatory regions18) that may influence disease susceptibility. Inferences about natural selection may therefore help identify disease susceptibility loci in humans and facilitate disease mapping studies.19In this review, we discuss the current hypotheses and knowledge about the evolutionary genetics of several risk factors for CHD, including hypertension, dyslipidemia, and the metabolic syndrome. Several candidate genes in pathways of blood pressure, glucose and lipid metabolism, blood coagulation, and inflammation that may be under natural selection are enumerated. An evolutionary perspective might explain why contemporary humans are at high risk for CHD and also helps to better understand variation in disease susceptibility.Natural SelectionNatural selection is the inevitable consequence of inherited variation in fitness, which is defined as the relative ability of an organism to survive and transmit its genes to the next generation.20 Under natural selection, individuals with advantageous or "adaptive" traits tend to be more successful in producing offspring, and those with nonadvantageous traits will be selected against. Positive selection can increase the prevalence of adaptive traits with a genetic basis. CHD has not been subject to direct selective pressure because it often manifests in middle age and beyond and is unlikely to have an effect on reproductive success.21 However, evolutionary genetics provides a framework for understanding how susceptibility alleles (see the glossary in the online-only Data Supplement) for CHD may have been affected by natural selection.Nesse22 proposed several possible explanations why previously protective alleles may be maladaptive in the present environment. Selection will tend to maintain the frequency of genes that increase reproductive success even if these genes have other effects that increase disease susceptibility in older age.23 For example, genetically determined high cytokine response levels may be associated with adverse cardiovascular outcomes in older individuals but may increase reproductive success in young age by conferring resistance to fatal infectious diseases24; a time lag between environmental change and adaptation by natural selection may also increase disease susceptibility (eg, the proatherogenic environment is relatively recent in human history); natural selection may not be able to accomplish other potentially protective changes even given plenty of time; and finally, the process of selection is stochastic, and protective alleles may be lost and disease alleles retained by chance, resulting in increased disease susceptibility.25Methods for Detecting Signals of Natural SelectionNatural selection takes places at the levels of transmitted genotypes and occurs on varying time scales. Several approaches have been developed to detect signals of natural selection.26 Broadly, 2 types of selective forces have shaped the evolution of species: purifying selection, which favors conservation of existing phenotypes, and positive selection, which promotes the emergence of new phenotypes. Both purifying selection and positive selection leave distinctive signatures in the form of patterns of genetic variation that can be detected through statistical tests of genotype or sequence data. If a gene has an important role in organismal homeostasis, then changes such as nonsynonymous mutations will be selected against and eliminated both between species and within species and result in an increase in the proportion of low-frequency variants within species. Positive selection typically manifests as rapid divergence of functional sites between species and a reduction in polymorphic variation within species.In his landmark "neutral" theory of molecular evolution, Kimura27 proposed that most of the DNA sequence substitution observed both within and between species has no effect on the fitness of the individual organism and that most evolutionary change is the result of genetic drift acting on neutral alleles (ie, genes under no selective pressure). The neutral theory is widely used as a "null model" to identify signatures of natural selection in genes under selective pressure by comparison with the background distribution of genetic variation.28 Deviations from this null hypothesis are indicative of selection, and several tests have been designed to detect such deviation.29 Sabeti et al30 have reviewed in detail statistical tests to detect positive natural selection based on distinctive patterns of genetic variation in the genome.Broadly, these statistical tests can be classified into 4 categories discussed briefly below (more information about these tests can be obtained in several excellent reviews16,17,19,31,32). Comparison of DNA sequence between species, especially for genetic variants that alter protein function, can be used to detect positive selection that occurred several million years ago. More recent selection can be gleaned from human population genetic data as alterations in levels of nucleotide diversity (or the allele frequency spectrum), linkage disequilibrium (LD), and haplotype structure. Most methods applicable to population genetic data make certain assumptions about the demographic history of populations (a constant population size and no population structure),19 whereas methods comparing species are relatively robust to demographic factors.33 To tease out the demographic effects, one needs to carry out tests of neutrality that account for the demographic history of the population (see Demography and Recombination Versus Selection below). Examples of several methods that have been used to detect signatures of natural selection across the human leukocyte antigen locus are provided in Table I of the online-only Data Supplement.Tests Based on Divergence of SpeciesOver a prolonged period, positive selection can increase the fixation rate of beneficial function-altering mutations (eg, substitutions that change amino acids [ie, nonsynonymous substitutions]). A simple way of detecting selection from comparative genomic data is to calculate the ratio between the rate of nonsynonymous substitutions and the rate of synonymous substitutions (ie, dN/dS).34 This ratio provides a means of detecting selective pressure at the protein level: dN/dS=1 for no selection, dN/dS 1 for positive selection. Two caveats should be noted for dN/dS-based tests. First, purifying or positive selection may selectively influence synonymous or nonsynonymous residues within a single gene based on the functions of individual residues at the protein, RNA, or regulatory level. Second, this method is not easily applied to cis-regulatory regions, many of which are putative and not validated.35 However, several studies have shown that conserved noncoding regions in mammals bear the signature of natural selection.36–38Tests Based on Allele Frequency SpectrumThe allele frequency spectrum represents a summary of the allele frequencies of various mutations in a sample. Natural selection affects the distribution of alleles within populations, and several tests (eg, Tajima's D, Fu and Li's F, Fu and Li's D, and Fay and Wu's H) have been developed to detect selection based on the allele frequency spectrum.16,17,19 Selection on 1 allele changes the frequency of other alleles that are in LD with it (ie, when the frequency of an allele increases in a population, the frequency of the linked variants also increases). This effect, also called hitchhiking or a selective sweep, is the process by which a new advantageous mutation eliminates or reduces variation in linked neutral sites as it increases in frequency in the population (Figure 1). Download figureDownload PowerPointFigure 1. Depiction of a selective sweep. A sample of 8 haplotypes is shown for standing genetic variation on the background of which a beneficial mutation arises. As the beneficial allele increases in frequency, it drags along linked neutral polymorphisms.Tests Based on Population Differentiation (FST)Natural selection may increase the degree of differentiation among geographically separate populations that are subject to distinct environmental and/or cultural pressures. The signature can arise only when populations are at least partially isolated reproductively (eg, after the major human migrations out of Africa ≈50 000 to 70 000 years ago). This is best summarized by the statistic FST, which measures the variance in allele frequency among populations relative to total variation in the entire population (Figure 2).39 Local adaptation appears to have featured in recent human evolutionary history,40 and screening the whole genome for local adaptation may be a means of identifying susceptibility genes for diseases for which prevalence varies as a function of ethnicity.41 For example, we calculated FST for 15 559 common SNPs in 416 candidate genes in 6 causal pathways for CHD and found significant population differentiation for 9 genes.42Download figureDownload PowerPointFigure 2. An example of population differentiation. A, Three alleles exist for a marker locus in populations (Pop) 1 and 2, and significant difference can be seen in the frequency of alleles, resulting in a high FST (FST=0.30). B, The allele frequency is similar between populations 3 and 4, so the FST is low (FST=0.02). Different colors indicate different alleles at a given locus. C, An example of high FST between Europeans and Asians (Han Chinese and Japanese in Tokyo) for the IL4 gene based on the HapMap database is shown. The x axis indicates the chromosomal location (chromosome 5).Population differentiation leading to varying frequency of gene variants across populations has made admixture mapping of disease susceptibility genes possible.43 In an admixed group such as blacks in the United States, one can classify the genome into sections that come from their African or European ancestors. Genomic regions in which individuals with a particular complex disease tend to have an unusually high proportion of ancestry from either Europeans or Africans may harbor a disease risk variant that influences disease susceptibility.Tests Based on LDThese tests rely on the LD structure of a given locus in the genome. Levels of LD will increase in selected regions if the rise in frequency of the beneficial mutation occurs faster than the recombination rate. Haplotypes containing the beneficial allele are conserved over significantly longer distances than other nearby haplotypes with similar frequencies.44 Thus, positive selection manifests as a haplotype at high frequency with homozygosity that extends over large regions and can be detected with the extended haplotype homozygosity test.44Figure 3 illustrates extended haplotype homozygosity for the derived allele of the SNP rs2032582 in the ATP-binding cassette, subfamily B, member 1 gene (ABCB1, also known as the multidrug resistance gene) using haplotype bifurcation diagrams.45 This gene is associated with export of chemical substance from within cells to the extracellular space and may have been under selective pressure related to xenobiotic toxin exposure. The derived allele in Europeans and Chinese demonstrated clear long-range LD as a predominance of 1 thick branch. Download figureDownload PowerPointFigure 3. Haplotype bifurcation diagrams for SNP rs2032582 (triallelic in Europeans) in the ABCB1 gene. In each subplot, the root is an ancestral allele (left) and a derived allele (right). The derived allele in Europeans and Chinese demonstrated clear long-range LD as a predominance of 1 thick branch. A indicates Africans; B, Europeans; and C, Chinese. From Figure 5 of Wang et al.45 Copyright © 2007. Adapted by permission from Lippincott Williams & Wilkins.Demography and Recombination Versus SelectionIt should be noted that various demographic scenarios can mimic the effects of natural selection on genetic variation. The challenge for population genetics-based methods is to separate a signature of natural selection from the confounding effects of population demographic history (eg, population bottlenecks, expansions, and subdivision).30 For example, population expansion can mimic the effect of a selective sweep (increased proportion of low-frequency variants), whereas a population bottleneck can mirror the pattern seen under balancing selection (an excess of intermediate-frequency variants). Population isolation can lead to significantly less genetic diversity than humanity as a whole46 and extended haplotypes over several centimorgans.47 Thus, it is necessary to tease out the effect of population demographic history if significant deviations from evolutionary neutrality are observed. To do this, one needs to carry out a test of neutrality that accounts for the demographic history of the population. For example, a locus of interest can be compared with the pattern of genetic diversity across the genome.48 In contrast to demographic factors, the effects of selection will be locus specific. Tests of evolutionary neutrality also need to account for the recombination rate in the region under study because recombination is a determinant of haplotype structure given the negative correlation between recombination rates and extent of LD.49Evolutionary Models of the Genetic Architecture of CHDThe genetic architecture of CHD refers to the number of genetic polymorphisms that affect the risk of CHD, the distribution of their allele frequencies and effect sizes, and their genetic mode of action (additive, dominant and/or epistasis, and pleiotropy).50 Any evolutionary change such as mutation, genetic drift, and natural selection may affect genetic architecture. An evolutionary genetics perspective provides a better understanding of the number and frequency of susceptibility alleles for CHD and thereby helps determine optimal strategies to detect such alleles. Theoretical models for the allelic structure of rare mendelian disease and common complex diseases have been described.51,52Natural selection has contrasting effects on the genetic architecture of rare mendelian diseases and common complex diseases. The causal alleles for many dominant mendelian diseases are generally new mutations (ie, derived alleles) for which selective pressure exists against their propagation (ie, purifying selection). Examples include familial hypercholesterolemia (eg, mutations in the low-density lipoprotein receptor gene [LDLR]), long-QT syndrome (eg, the calcium channel, voltage-dependent, L type, α1C subunit gene [CACNA1C]), and hypertrophic cardiomyopathy (eg, myosin, heavy chain 7, cardiac muscle, β [MYH7]). However, natural selection may operate in different directions in the heterozygous and homozygous forms of recessive mendelian diseases. Two classic examples, sickle cell disease and cystic fibrosis, illustrate the selective advantage for the heterozygote against malaria and diarrhea illness, respectively, whereas the homozygotes with the disease generally die before they reach reproductive age.In contrast, the allelic spectrum of common complex diseases includes both common51,53 and rare variants.54 Recent genome-wide association studies55 have discovered many susceptibility variants of modest effect size that are prevalent in the general population, indicating that the variants were either neutral or advantageous in the past. Lifestyle changes such as culture and environment (eg, climate and diet) would have led to specific selective pressures as humans spread out of Africa. Present-day human beings live in a markedly different environment from their ancestors, so the common complex diseases might result from a mismatch between the ancestral alleles and current environments (ie, the ancestral-susceptibility model for complex diseases).56,57 In other words, ancestral alleles reflect adaptations to the lifestyle of ancient human populations but are deleterious in the "modern" environment (Figure 4). The rare variant–common disease hypothesis posits that much of the disease susceptibility is due to many rare variants at different loci, each conferring a moderate increase in risk.54 These alleles are not as deleterious for reproductive fitness as are rare mutations in mendelian disease and therefore are not eliminated by strong purifying selection. Detection of such rare alleles will require sequencing of the genome or candidate genes because conventional genome-wide association studies are not powered to detect such alleles.54Download figureDownload PowerPointFigure 4. The ancestral-susceptibility model for common, complex diseases. The ancestral allele is maladaptive in the modern environment and associated with increased disease susceptibility, whereas the derived allele is adaptive in the current environment and may be protective.Candidate Genes for CHD With Evidence for Natural SelectionTwo major challenges in studying the genetic basis and evolutionary genetics of CHD are its phenotypic complexity and the presence of multiple causal factors. Although considerable overlap exists among various CHD phenotypes, the underlying pathophysiology may be quite different. For example, myocardial infarction often results from rupture of a vulnerable atherosclerotic plaque, and risk factors and their interactions that determine plaque vulnerability may differ from those that influence coronary atherosclerotic burden as measured by the presence and quantity of coronary artery calcium.58 Epidemiological research over the last several decades has led to the discovery of several risk factors for CHD, including elevated lipid levels, diabetes mellitus, smoking, hypertension, inflammation, oxidative stress, and increased coagulability. In this section, we enumerate genes in causal pathways of atherosclerosis, including blood pressure regulation, lipoprotein and glucose metabolism, coagulation, and inflammation, that may be subject to various degrees of selective pressures resulting from climatic and dietary changes and host response to pathogens (the Table59–80). Candidate genes in these pathways may influence susceptibility to cardiovascular disease; we briefly discuss a few examples of such genes. It needs to be emphasized that for the loci identified as being under selection, further work is needed to address potential confounding caused by population history and structure and to establish definitive evidence of selection, mechanism of selection, and functional effects of the allelic variants under selection. Table. Candidate Genes for CHD for Which Evidence for Natural Selection Has Been ReportedGeneSignature of SelectionMode of SelectionMolecular Function and Implications for CHDHDL indicates high-density lipoprotein. ACE2 is the angiotensin I-converting enzyme 2 gene; AGT, angiotensinogen gene; CYP3A5, cytochrome P450, family 3, subfamily A, polypeptide 5 gene; GNB3, guanine nucleotide binding protein (G protein), β polypeptide 3 gene; GRK4, G protein–coupled receptor kinase 4 gene; SCG2, secretogranin II gene; APOE, apolipoprotein E gene; LDLR, low density lipoprotein receptor gene; PCSK9, proprotein convertase subtilisin/kexin type 9 gene; SCARB1, scavenger receptor class B, member 1 gene; ANGPTL4, angiopoietin-like 4 gene; CAPN10, calpain-10 gene; TCF7L2, transcription factor 7-like 2 gene; F7, coagulation factor VII gene; F9, coagulation factor IX gene; IL4, interleukin 4 gene; IL4R, interleukin 4 receptor α chain gene; IL13, interleukin 13 gene; IL1A, interleukin 1α gene; and MMP3, matrix metallopeptidase 3 gene.Blood pressure ACE2Increased nucleotide diversity59BalancingRegulates cardiovascular and renal function; an insertion/deletion polymorphism is associated with hypertension AGTIncreased derived allele frequency60PositiveRegulates salt metabolism; a regulatory polymorphism (A-6G) is associated with hypertension CYP3A5High FST; reduced nucleotide diversity61PositiveInvolved in cortisol and/or aldosterone metabolism; the CYP3A5*3 polymorphism is associated with hypertension GNB3Long haplotypes62PositiveMediates G protein–coupled signaling; the C825T polymorphism is associated with hypertension GRK4High FST (blacks vs Asians); long haplotypes63PositivePhosphorylates the activated forms of G protein–coupled receptors; a nonsynonymous variant (R65L) is associated with hypertension SCG2Reduced nucleotide diversity; increased derived allele frequency64PositiveNeuropeptide with potent angiogenic activity; a common regulatory variation has been implicated in blood pressure control and susceptibility to hypertensionLipid metabolism APOEIncreased frequency of haplotypes carrying ε3 allele65PositiveEssential for the normal catabolism of triglyceride-rich lipoproteins; the ancestral allele ε4 is associated with CHD LDLRA novel sterol response element has appeared in primates66PositiveBinds circulating LDL cholesterol; mutations are associated with familial hypercholesterolemiaLow FST; increased nucleotide diversity67Balancing PCSK9dN/dS>1 in several lineages in C-terminal domain68PositiveInduces LDLR degradation; increased activity is associated with higher circulating cholesterol levels; certain nonsynonymous variations are associated with hypercholesterolemia and CHDLong haplotypes; high FST69Positive SCARB1Low level of nucleotide diversity70PositivePlays a central role in HDL cholesterol metabolism; several nonsynonymous variants are associated with higher levels of HDL cholesterol and lower levels of LDL cholesterol ANGPTL4Excess of rare alleles71PurifyingRegulates the partitioning of fatty acids between sites of storage and sites of oxidation; loss-of-function alleles have been associated with lower plasma triglyceride and higher HDL cholesterol levelsEnergy metabolism CAPN10High FST (Africans vs non-Africans); long haplotypes72,73Positive; balancingCalcium-dependent, cysteine protease; an intronic polymorphism is associated with type II diabetes TCF7L2High FST (East Asians vs Europeans and Africans); long haplotypes in East Asians74positiveRegulates transcription activity of target genes; 2 polymorphisms (rs12255372 and rs7903146) are associated with type II diabetesBlood coagulation F7High FST (Africans vs Asians; Europeans vs Asians)75PositivePlasma levels are associated with the risk of myocardial infarction F9Reduced nucleotide diversity76Positive or backgroundDeficiency of F9 leads to hemophilia BInflammation IL4High FST (Africans vs Asians)77PositiveInflammatory cytokine that regulates Th1/Th2 cell development and differentiation IL4RA higher ratio of replacement to silent polymorphisms in humans compared with that between humans and orangutans78PurifyingInflammatory cytokine receptor that binds the α chain of IL4 IL13Increased derived allele frequency; high FST (Africans vs Europeans)79PositiveInflammatory cytokine that regulates Th1/Th2 cell differentiation IL1AIncreased nucleotide diversity59BalancingPleiotropic inflammatory cytokine MMP3Rapid evolution of T(n); high FST (Europeans vs Africans); low haplotype heterozygosity80PositiveInduces the breakdown of extracellular matrix; a regulatory variant is associated with CHDHypertensionWeder81 has summarized hypotheses that have been put forth to explain hypertension in the framework of evolution. Differential susceptibility to hypertension is likely in part due to our history of adaptation to climatic changes.82 For example, the sodium hypothesis83 posits that sodium-conserving mechanisms conferred a survival advantage among our ancestors in the hot and humid climate of Africa but may lead to hypertension in temperate climates. Thus, sodium-conserving genotypes in US blacks may predispose to excessive sodium retention and sodium-sensitive hypertension in their current milieu.84 The ancestral alleles (ie, the sodium-conserving alleles) show strong latitudinal gradients in allele frequency (clines) and are more prevalent in Africans than in populations from northern Europe, in whom signatures of positive selection (eg, high levels of LD and low haplotype diversity) are noted for the derived allele. The genetic patterns observed in candidate genes affecting sodium handling could have implications for the diagnosis and treatment of hypertension in geographically and ethnically defined populations around the world.60The geographic distribution of the A(-6)G variant in the promoter region of the human angiotensinogen (AGT) gene suggests that the G(-6) variant has been selectively advantageous outside Africa.60 The G(-6) variant, the derived allele, is present at higher frequency in Asians and Europeans than in Africans,60 and evidence exists of a selective sweep in the vicinity of the polymorphism because haplotypes carrying the derived G(-6) allele showed elevated levels of LD in non-African populations. Genetic drift is not a likely explanation because frequencies of other AGT alleles are not similarly affected.60 A population bottleneck followed by population growth can result in relatively high LD. However, the differences in LD are pronounced around the A(-6)G polymorphisms but not in the entire gene.60,85 Other examples of correlation between latitude and "heat-adapted" alleles include variants in CYP3A5,61GNB3,62ADRB2,62 and SCNN1A62 (online-only Data Supplement Table II). Among these alleles, the variant 825T in GNB3 may account for a significant portion of worldwide variation in blood pressure.62Lipid MetabolismHumans, in whom cholesterol metabolism has diverged in a variety of ways from that of many distant mammals such as rodents and dogs, are susceptible to hypercholesterolemia given the present-day "atherogenic" diet.86 Several selective pressures, including climatic and dietary changes, may have influenced lipoprotein metabolism. Higher serum cholesterol may have been advantageous during the rapid increase in human size during human evolution and for its role in steroid hormone synthesis. For these reasons, the ε4 allele of the apolipoprotein E (APOE) gene may have been advantageous in the past, but it is maladaptive in the setting of nutritional abundance typical of the developed world, leading to increased risk of CHD.87,88 Consequently, the derived ε3 allele is adaptive in the modern environment.65,89Two genes that play a key role in cholesterol homeostasis, LDLR and proprotein convertase subtilisin/kexin type 9 (PCSK9), a newly discovered regulator of LDLR, may have been under selection. LDLR has been shown to be differentially expressed among mammals,47,90 and an anthropoid primate-specific sequence element (a novel sterol regulatory element) that accounts for the strong activity characterizing human and other anthropoid primates has been identified.66 Species-specific re