personalized medicine and human genetic...

Personalized Medicine and Human GeneticDiversity

Yi-Fan Lu1, David B. Goldstein1, Misha Angrist2, and Gianpiero Cavalleri3

1Center for Human Genome Variation, Duke University, Durham, North Carolina 277082Institute for Genome Sciences and Policy, Duke University, Durham, North Carolina 277083Molecular and Cellular Therapeutics, Royal College of Surgeons, Dublin 4, Ireland

Correspondence: [email protected]

Human genetic diversity has long been studied both to understand how genetic variationinfluences risk of disease and infer aspects of human evolutionary history. In this article, wereview historical and contemporary views of human genetic diversity, the rare and commonmutations implicated in human disease susceptibility, and the relevance of genetic diversityto personalized medicine. First, we describe the development of thought about diversitythrough the 20th century and through more modern studies including genome-wide asso-ciation studies (GWAS) and next-generation sequencing. We introduce several examples,such as sickle cell anemia and Tay–Sachs disease that are caused by rare mutations and aremore frequent in certain geographical populations, and common treatment responses thatare caused by common variants, such as hepatitis C infection. We conclude with commentsabout the continued relevance of human genetic diversity in medical genetics and person-alized medicine more generally.

We all differ at the level of our DNA se-quence, and geneticists obsess over trying

to understand the significance of this geneticdiversity. This is an important goal, as by un-derstanding human genetic diversity we canlearn about the evolutionary history of our spe-cies, where we have come from, and perhapswhere we are headed. More practically, under-standing human genetic diversity is essential tounderstanding the biology of our diseases ofvarious kinds, from the genetically more simpleto more complex, and how we respond to treat-ment at both the population and individual lev-

els (Torkamani et al. 2012). Indeed, improvingour knowledge of human disease biology is theprimary driver behind the largest and most sys-tematic studies of human genetic diversity to-day. These studies, and the population- and dis-ease-specific investigations made possible bythem, are essential for reducing health dispari-ties and improving health outcomes for the spe-cies as a whole. Unfortunately, largely because ofwhich DNA samples are most easily accessible,most genomics research programs have concen-trated their discovery efforts in populations ofEuropean ancestry (Need and Goldstein 2009;

Editor: Aravinda Chakravarti

Additional Perspectives on Human Variation available at www.perspectivesinmedicine.org

Copyright # 2014 Cold Spring Harbor Laboratory Press; all rights reserved

Advanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a008581

1

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org

on July 8, 2018 - Published by Cold Spring Harbor Laboratory Press http://perspectivesinmedicine.cshlp.org/Downloaded from

http://perspectivesinmedicine.cshlp.org/

Bustamante et al. 2011). As we discuss in thisessay, this approach is myopic and carries with ituntoward consequences for both the scientificand public health enterprises.

The successful completion of the HumanGenome Project in 2003 was the first in a seriesof large multinational public efforts that beganto move the field of medical genetics away frompurely descriptive documentation of patients’physical features coupled with laborious one-by-one examination of a small subset of theirgenes for potentially pathogenic changes. Forexample, the International HapMap Project’scollection of millions of genotypes from fourglobal populations was indispensable to thepursuit of hereditary changes in genes that con-tribute to disease by providing the platform forso called “genome-wide association studies”(GWAS). GWAS gave us the ability to efficientlyand comprehensively assay genetic variants thatare common in a population and identify thosethat appear more commonly in patients with agiven disease than they do in controls withoutthe disease. Such variants can sometimes pro-vide clues to the genetic basis of human disease(Manolio et al. 2009).

In parallel, researchers have capitalized onour improved understanding of population his-tory to identify disease-causing genes. Popula-tion-specific studies of disease, from myocardialinfarction in Icelanders to prostate cancer inAfrican-Americans, have cleverly exploited theenrichment of specific disease-susceptibility al-leles in more genetically homogeneous popula-tions (Torkamani et al. 2012).

Along the way, advances in our understand-ing of patterns of human genetic variation havealso informed our view of the history of mod-ern human populations. Our interpretation ofthe scientific data, however, has been influencedby the constantly evolving sociopolitical milieu.During the early part of the 20th century, twoschools of thought emerged on how natural se-lection influenced the frequency and distribu-tion of genetic variation. The “classical” schoolbelieved that most genetic variation was rareand that variants present in the population arealmost always deleterious. In the very occa-sional cases in which new mutations are advan-

tageous, they quickly became “fixed” (cases inwhich the new advantageous allele replacedthe ancestral one). The “balanced” school, onthe other hand, believed that genetic variationwas quite common and often actively main-tained by selection favoring multiple forms ofa gene in the population. This might be becauseof so-called overdominance, in which selec-tion favored the heterozygote or other formsof selection-maintaining diversity. In fact, eitherof these perspectives could readily be, and were,marshaled in support of eugenic perspectivesthat were common before and after the SecondWorld War. For example, under the classicalschool, it was easy to postulate a genetic under-class that carried a greater-than-average load ofdeleterious mutations. Because natural selec-tion pressures can be assumed to have differedamong diverse human populations, it was pos-sible to imagine that populations from somegeographic regions would have a “superior”complement of variants in terms of key pheno-typic characteristics as compared with othergeographic regions.

Early “modern” approaches to quantifyingbiological differences were based on physicalmeasurements that were heavily biased in theways they were deployed. The distinguishingcharacteristics used to construct racial classifi-cation were those to which human perceptionis most finely tuned (skin color, eye shape andcolor, hair color and texture, etc.). Direct, ob-jective methods of quantifying “genetic” varia-tion (as opposed to “physical” characteristics)simply did not exist. Further, physical measure-ments were typically prone to environmental(nongenetic) influences, blurring the relation-ship between the measurement and geneticmakeup of the individual.

The population characterization of the ABOblood group system by the Hirszfelds in the ear-ly 1900s, therefore, was seminal. It provided abiochemical marker that was closely aligned tounderlying genetic variation, and, in so doing,provided the first major system for exploringpatterns of human genetic diversity in an unbi-ased manner. Indeed, when tested in soldiers inarmies of World War I, the pattern of A and Bblood types showed frequency gradients that

Y.-F. Lu et al.

2 Advanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a008581

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



correlated with the geographic origin of the sol-diers (Hirszfeld and Hirszfeld 1919). Soldiersfrom Western Europe (English and French)had a lower frequency of the “B” blood group,which appeared to gradually increase as onemoved east toward Eastern European (Greeks,Turks, Russians) and Asian groups, suggestingthat gene frequencies changed gradually acrossgeographically defined populations.

A decisive break with the tradition of as-suming sharp genetic divisions among ethnicgroups came with the work of Richard Lewontinin 1972. Using multiple different polymorphicgenetic markers, comprising blood group sys-tems and serum protein markers that were as-certained in an unbiased manner, Lewontingenerated data from more than 100 populationssampled across seven socially constructed “ra-cial” groups (Caucasians, Africans, East Asians,South Asians, Amerindians, Oceanians, andAustralians). Lewontin showed that the vast ma-jority of human genetic diversity (�85%) iscaused by individual differences that are sharedacross “all” populations and races. Only a smallpercentage (�15%) was because of differences“between” populations and a smaller percent-age, again (�6%), was caused by differencesbetween “racial” groups (Lewontin 1972). Lew-ontin’s data suggested that although there issubstantial genetic variation within the humanpopulation, such variation has accumulatedover time; most of this variation appeared be-fore the expansion of Homo sapiens out of Af-rica and the resulting isolation of populationswithin continents. Put another way, the con-siderable genetic differences we see between in-dividuals has very little to do with so-called“racial” boundaries. Rather, it is merely the var-iation that was present in the original humanpopulation that seeded all the current humanpopulations. Therefore, the same polymorphicalleles (genetic variants) are found in most pop-ulations, although their frequencies may differsubstantially. Broadly speaking, there has beentoo little time for the accumulation of substan-tive divergence in a young species such as ours.The fact that the classical model predicted ex-tensive genetic differentiation between popula-tions was explained by the molecular evolution

pioneer Motoo Kimura, who hypothesized thatmost variation was selectively neutral (neitherenhancing nor retarding human survival) and,therefore, largely free from the influence of Dar-winian selective forces.

At the time of publication, Lewontin’s find-ings were controversial, but consensus graduallyemerged that genetic differences among pop-ulations are modest (Nei and Roychoudhury1972; Cavalli-Sforza et al. 1994). Before Lewon-tin, the general consensus was that genetic di-versity would be structured according to raciallabels and, thus, the labels were scientificallyjustified. The observation that patterns in hu-man genetic variation were largely gradual ac-cording to geographic boundaries and not sub-ject to sudden population-specific changes thatfollowed preconceived racial notions removedthe biological argument for race (or, we wouldargue, it should have).

Lewontin illustrated that genetic variationwas extensive and largely shared across popula-tions. But, it was not until the sequencing of thefirst human genome (actually, a consensus ofseveral human genomes) in 2003 that we appre-ciated just how extensive genetic variation reallywas in the human genome. Any two randomlyselected individuals of European descent willdiffer at �3 million points in their genome, or�0.1% of their .3 billion bases of DNA. Thefact that most of this variation is, in effect, se-lectively neutral presents an enormous chal-lenge for characterizing those alleles that con-tribute to our common diseases in substantiveways. In other words, the challenge is to identifythe few trait-altering variants that lie in an oceanof irrelevant ones.

A major breakthrough in this challenge wasthe development of GWAS. The basic frame-work used in these studies is to select key vari-ants that inform about virtually all commonvariations in the human genome. These spe-cially selected variants are often called taggingsingle-nucleotide polymorphisms (SNPs) be-cause they are near perfect surrogates for vari-ants not directly assayed. These variants couldbe tested easily by newly developed technologiesusing specially designed genotype chips. GWASchips are also relatively inexpensive; one can

Personalized Medicine and Human Genetic Diversity

Advanced Online Article. Cite this article as Cold Spring Harb Perspect Med doi: 10.1101/cshperspect.a008581 3

ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



now genotype a million variants for ,$50 asample. Applied to large studies involving thou-sands of disease (case) and nondisease (control)individuals, the GWAS approach provided theframework to associate specific genetic variantsand their cognate genomic regions with dis-eases, even if the study design was not well suit-ed to identifying the actual genetically causalvariants. The GWAS approach was successfulin that it provided much needed momentumin the push to identify disease genes. Neverthe-less, in most cases, even when applied to studiesinvolving hundreds of thousands of partici-pants, the approach failed to explain the major-ity of the presumed genetic component of anygiven trait (Manolio et al. 2009).

One explanation for this problem of “miss-ing heritability” lies in the fact that the GWASapproach only tests for genetic variants that arecommon in a population, that is to say, thosethat Lewontin first observed as shared acrossindividuals and populations. The reason forthis is that the research community (throughthe International HapMap Project) had a goodunderstanding of the nature and extent of com-mon variation; it was, after all, “common” andtherefore easy to find and test in large popu-lations. Thus, it was a logical starting point forgenome-wide studies. Further, it was not untilthe development of novel DNA-sequencingtechniques in the last few years that the studyof rare variants became logistically and finan-cially feasible (Cirulli and Goldstein 2010). Asa result, geneticists using GWAS in the late2000s were akin to the drunken man who wouldonly look for his lost keys under the street-lamp; he looked there because that is wherethe light was.

THE PATTERN OF GEOGRAPHIC VARIATIONFOR COMMON VARIATION MAY BE QUITEDIFFERENT FROM THAT OF VARIANTSINFLUENCING DISEASE RISK AND DRUGRESPONSE

Although most common variants are indeedcommon among most human populations, ithas long been known that rarer gene variantscan show markedly different patterns across hu-

man groups. Perhaps the best evidence of thiscomes from the successor to the HapMap Pro-ject, the 1000 Genomes Project (The 1000 Ge-nomes Project Consortium 2010), whose goalis to sequence the genomes of a large numberof humans to provide a comprehensive survey ofhuman genetic variation (Via et al. 2010). Inves-tigators in the 1000 Genomes Project discoveredthat 63% of novel variants (that is, those thathave never before been observed in humans) arefound in African ancestry populations as com-pared with 33% with European ancestry.

In the same study, several hundred thou-sand SNPs with large allele-frequency differ-ences were found across geographically distinctpopulations. Within these variants, there wasenrichment for so-called “nonsynonymous”variants, which are characterized by importantchanges in the DNA sequence that lead to struc-tural and functional changes in the proteinsproduced by these genes. This observation sug-gests that local populations adapted to theirspecific environments and the genetic changesthat allowed this to happen were selected forby evolution (The 1000 Genomes Project Con-sortium 2010). These results also illustrate thefact that Lewontin’s assessment related specifi-cally to common variants because those are theones most important to overall variation pre-sent in an individual. If you look at one individ-ual, most of the variants that individual has arecommon variants, and those are the ones thatfollow Lewontin’s pattern; they are mostly de-rived from the common human ancestral pop-ulation. But, if the variants that are most impor-tant to phenotype variation are more rare, thenthis assessment that Lewontin provides does notapply to those most responsible for phenotypicvariation.

MENDELIAN MUTATIONS ARE HIGHLYPOPULATION SPECIFIC FOR A NUMBEROF REASONS

Because the Moravian monk Gregor Mendelwas the first one to work out the basic laws ofheredity, we refer to diseases within a family that“obviously” follow the rules of inheritance de-scribed by Mendel as Mendelian diseases. Typ-

Y.-F. Lu et al.


ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



ically, these mutations have a major effect ondisease risk (and gene function) and relativelyfew genes can carry mutations that cause a givendisease and still allow the organism to survive.Some of the mutations responsible for Mende-lian diseases have long been known to show ahigh degree of population specificity. In someexceptional cases, this is clearly because of pos-itive, as opposed to negative, natural selection.The autosomal recessive disease sickle cell ane-mia (that is caused by two defective copies ofthe b hemoglobin gene and, thus, producing ahemoglobin protein with reduced function), forexample, is largely restricted to African, Medi-terranean, and South Asian ancestry popula-tions. In African-Americans, the allele frequen-cy of the sickle hemoglobin (Hb S) mutation is�4% (Ashley-Koch et al. 2000). Why? Becausealthough carrying two mutant Hb S alleles caus-es the devastating condition sickle cell anemia,carrying a single copy of Hb S does not usuallycause health problems. However, it does protectagainst malarial infection. For this reason, it hasbeen selected for in regions of the world wheremalaria has been endemic: Africa, the Mediter-ranean, and South Asia. Its frequency is signifi-cantly higher among populations that originatefrom these regions. In other words, carriers ofone defective copy of the hemoglobin gene areat an evolutionary advantage in regions of theworld where malaria is common and, therefore,this version of the hemoglobin gene has be-come more common in those areas (Aidoo etal. 2002).

Carrier advantage is not the principal rea-son why many Mendelian mutations can bethought of as more or less population specific.Most fundamentally, mutations that have a ma-jor impact on risk are rare because of naturalselection against them. That means that theyhave been relatively recently introduced intothe population by mutation, and the specificmutations are, therefore, usually geographicallyquite restricted. This, by itself, would mean thatthe mutations tend to be very different in dif-ferent geographic regions, but not the total bur-den of diseases they cause. In fact, the collectivefrequency of disease-causing mutations in spe-cific populations at specific genes can be quite

different from global averages because smallpopulation size and demographic history canalso be important.

Consider the Ashkenazi Jews, who are stat-istically more likely to carry mutations thatcause autosomal recessive Tay–Sachs diseasein which affected children die at an early agebecause their mutations deprive them of a par-ticular enzyme. In all likelihood, mutationscausing Tay–Sachs increased in frequency dur-ing a time when the Ashkenazi Jewish popu-lation was small. Perhaps when the Ashkenaz-im were beginning to establish themselves inEurope during the early Middle Ages, one ormore Tay–Sachs mutations arose by chanceand the small breeding population led to a“founder effect,” that is, persistence of particu-lar alleles because those alleles were overrepre-sented when the population in question firstemerged. Similarly, perhaps Tay–Sachs muta-tions were overrepresented after the AshkenaziJewish population underwent a “populationbottleneck,” that is, experienced a sharp contrac-tion. For example, the European Jewish popu-lation declined precipitously following perse-cution of Jews during the First Crusade in thelate 11th century and the subsequent spreadof Black Death in the mid-14th century. It maywell be that, by chance, Tay–Sachs mutationswere present in surviving members of the Ash-kenazim following these events and those mu-tations were, therefore, preferentially transmit-ted to subsequent generations (Slatkin 2004).Ashkenazi Jews’ historical propensity to prefer-entially choose mates within their group has alsoserved to keep Tay–Sachs alleles within theircommunity at a relatively high frequency. Today,the carrier frequency of Tay–Sachs disease ison the order of 1 in 30 in self-identified Ashke-nazi Jews, 10 times higher than in other popu-lations. Before widespread population carrierscreening of this disorder, 1 in 3600 childrenborn to Ashkenazi parents had Tay–Sachs dis-ease (Fernandes Filho and Shapiro 2004; Brayet al. 2010). Screening has since reduced Tay–Sachs births among Ashkenazim by some 90%(Ostrer and Skorecki 2013).

In other cases in which specific genetic dis-eases appear to be more common in a certain



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



population, it is not clear whether the high fre-quency of rare disease-causing mutations iscaused by chance, selective mating among car-riers within the population, carrier selectionadvantage, or some combination of these fac-tors. Cystic fibrosis, for example, is most com-mon in European ancestry populations. InCaucasians, the frequencies of cystic fibrosismutations in the cystic fibrosis transmembraneconductance regulator (CFTR) gene are sig-nificantly higher than in other populationsand cause the autosomal recessive disease in 1of 2500 newborns (Ratjen and Doring 2003).Over the years, geneticists have speculated asto why this is the case, often focusing on Dar-winian selection as an explanation: perhaps car-riers of CFTR mutations were more resistantto cholera and other dehydrating intestinal dis-eases (Bertranpetit and Calafell 1996). Or per-haps they were more resistant to contractingtuberculosis (Poolman and Galvani 2007). An-other hypothesis suggested that carrier frequen-cies rose in Europe after farmers on the con-tinent began raising dairy cattle, which led tothe transmission of various pathogens fromlivestock to humans, perhaps via cow’s milk(Alfonso-Sanchez et al. 2010). Although theseideas are intriguing, none have been proven tothe extent of the implication of malaria as theselective force accounting for the rise of the HbS allele. A particularly provocative hypothesiswas promulgated by Harpending and Cochranthat some of the mutations causing Tay–Sachsand other lysosomal storage diseases, several ofwhich also occur at increased frequency in Ash-kenazi Jews, were the result of positive selection;the idea was that somehow being a carrier forthese diseases was associated with greater in-telligence (Cochran et al. 2006). However, thereremains no real evidence to support this spec-ulation.

Whatever the reason for the emergence ofdisease-causing alleles at relatively high fre-quencies in specific populations, their existencesuggests the possibility that rarer variants arealso important in common diseases, and theremay be more population specificity than antic-ipated by Lewontin’s analysis of common vari-ation. Consequently, we would do well to pay

attention to the population frequencies of var-ious human diseases and traits to better under-stand their genetic underpinnings.

COMMON VARIATION INFLUENCINGDISEASE RISK AND DRUG RESPONSE

Even among common variants, some show rel-atively greater differentiation (frequency dif-ferences) among population groups because ofgenetic drift or selection, with clinically impor-tant consequences even at the level of the pop-ulation average. One of the most well-knowndiseases for which common genetic variationaffects both the spontaneous clearance of aninfectious agent and treatment response is hep-atitis C virus (HCV) infection. Treatment re-sponse refers to medical treatment with thecombination of peginterferon-a (PegIFN-a)and antiviral therapies to induce viral clearance,whereas spontaneous clearance is the auto-matic viral clearance without exogenous drugadministration. It was already well known thatAfrican ancestry individuals respond morepoorly to HCV drug treatment than Caucasianand Asian individuals. In 2009, GWAS discov-ered a SNP (also known as rs12979860) in theIL28B locus (abbreviated as IL28B polymor-phism below) that is highly associated with pa-tient drug responses to medicines designed totreat HCV (Ge et al. 2009). Allele frequenciesof IL28B polymorphism were found to differlargely among these ethnic populations, and ex-plain the differences of treatment success rateamong those populations. IL28B encodes inter-feron-l-3, which is an important cytokine forinnate immunity and one of the first respondersto the invasion of foreign pathogens. Some be-lieve the allele frequency of IL28B has been se-lected among different populations by one ormore pathogen and, thus, evolved at differentstages of human history. However, the exact nat-ural selection pressure that causes the distinctpattern of allele frequency is unclear. Overall,the discovery of IL28B polymorphism illustratesthat the frequency distribution of certain riskalleles is sufficient to affect the disease progres-sion and drug responses. Below, we will discussin detail how this common variant was discov-

Y.-F. Lu et al.


ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



ered and its impact on both treatment-inducedand spontaneous HCV clearance.

IL28B DISCOVERY FOR HCV TREATMENTRESPONSES

HCV is a positive-strand RNAvirus belonging tothe family Flaviviridae. HCV transmission ismainly through blood-to-blood contact andchronic infection usually results in fibrosis, cir-rhosis, livercarcinoma, and even liver failure. It isestimated that 170 million people are chronicallyinfected by HCV worldwide, and it is the majorcause for liver transplants in the United States.Because HCV has been a serious public healthproblem in the United States and worldwide,there have been efforts to develop treatmentsfor chronic HCV infection. However, the treat-ment success rate has been unsatisfactory.PegINF-a combined with ribavirin (RBV) thera-py has been widely used to treat chronically in-fected HCV patients since 2002. The treatmentsuccess rate is moderate (from 20% to 70%) andis dependent on a patient’s ancestry. Treatmentsuccess is defined as reaching sustained viro-logical response (SVR), when the blood viralload is suppressed below the detectable level for

24 wk after 48 wk of combination treatment(Ghany et al. 2009). In East Asian populations,the PegIFN-a plus RBV treatment for chroni-cally infected HCV patients has been shown toreach 76% of the overall SVR rate, which is dra-matically higher than the 56% SVR rate of Euro-pean-Americans and 24% of African-Ameri-cans (Liu et al. 2008; Ge et al. 2009). Before thegenetic discovery of IL28B, the reason for thedifferences observed among major ethnic groupswas unclear, and race had been used as a profil-ing feature to predict HCV treatment response.

The GWAS performed by Ge and colleagues,as well as studies performed by two othergroups, identified a SNP (rs12979860) on theIL28B locus associated with the response ofPegIFN-a plus RBV therapy. This genetic vari-ant (rs12979860) is a C-to-T substitution withC being the major allele in Europeans and EastAsians. The relative risk for SVR (chance toreach treatment success) is around threefoldhigher in C/C than non-C/C patients (includ-ing C/T and T/T), and is statistically highlysignificant (Fig. 1). Similar results were alsofound in several other studies; patients withthe homozygous C/C genotype at IL28B gener-ally have a two to three times higher treatment

100

n = 336

T/T T/C

European-Americans

African-Americans Hispanics Combined

SVR (%) Non-SVR (%)rs12979860

C/C T/T T/C C/C T/T T/C C/C T/T T/C C/C

n = 35

n = 559

n = 186n = 14n = 91

n = 70

n = 433

n = 102

SV

R (

%)

P = 1.06 × 10–25 P = 2.06 × 10–3 P = 4.39 × 10–3 P = 1.37 × 10–28

n = 30 n = 26 n = 392

80

60

40

20

0

Figure 1. The SVR% as a function of ancestry and IL28B genotype. Figure based on data from Ge et al. (2009).



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



success rate than patients with C/T or T/T ge-notypes (Ge et al. 2009; Suppiah et al. 2009;Tanaka et al. 2009). European-American pa-tients with C/C genotypes under different treat-ment regimens show �80% SVR, comparedwith 30% and 40% SVR rates of C/T and T/Tgenotypes, respectively. In African-Americans,patients with the C/C genotype show �50%of the SVR rate compared with ,20% of theSVR rate for C/T and T/T patients (Fig. 1).The overall effect of the IL28B polymorphismis, therefore, substantial in predicting HCVtreatment response. In general, regardless of eth-nicity, the C/C genotype has higher SVR ratethan non-C/C genotypes (twofold higher inEuropean-Americans and Hispanics, and three-fold in African-Americans). This result suggeststhat C/C universally favors treatment successversus non-C/C, although in African-Ameri-cans, the same C/C genotype shows a lowerSVR rate than in European-Americans (50%in African-Americans vs. 80% in European-Americans). The factors that cause this successrate difference in C/C genotype among individ-uals of different ethnicities are still unclear.

IL28B has received a great deal of attentionsince the GWAS discovery for its ability to pre-dict the pretreatment drug response outcome,and the potential for its biological antiviral ac-tivities. Before the GWAS, the reason behind thelink between ethnicity and drug responses waselusive, but we now clearly know that the IL28Ballele frequencies show very different distribu-tions across populations. Using random con-trols with unknown hepatitis C status, 90% ofthe East Asian population carried the IL28B Callele versus 70% in European-Americans. How-ever, in the African-American population, the Callele has become the minor allele (smaller allelefrequency) at 40%. Strikingly, according to thestudy performed by Ge and colleagues (2009),the C allele frequency showed linear correlationwith the SVR rate in four distinct populations(Table 1). This concordance strongly suggeststhat the difference observed in HCV treatmentresponse can be mostly explained by the allelefrequency distribution among populations. In asubsequent study by Thomas et al. (2009), 51geographical subpopulations were examined

for the IL28B polymorphism. The results weresimilar: the C allele frequency was highest inAsian populations, modest in European popu-lations, and the lowest in African ancestry pop-ulations. This result showed the IL28B allele fre-quency distribution in higher resolution andcorroborated the initial observations.

This correlation substantially explains thereason why different populations have signifi-cantly different treatment success rates. Up untilnow, the only gene for which there is strongevidence of an influence on HCV treatment re-sponse has been IL28B. An extensive search forother genetic factors that might contribute toHCV treatment response has been performed,but no statistically significant result for othergenes that modify the effect of IL28B has beenfound thus far.

The profile based on race to predict treat-ment success rate in the past is now proven tobe overly simplified. It is actually the IL28Bgenotype that plays a major role in determin-ing treatment response, not ethnicity, and thedifferences observed among ethnicity can beexplained merely by the allele frequency dif-ferences among geographic populations. HCVtreatment response is a great example of howallele frequency can affect treatment outcomesamong populations, and it seems highly likelythat there will be other examples like this to befound in the future.

VARIATION OF IL28B ALSO AFFECTSSPONTANEOUS CLEARANCE OF HCV

Spontaneous clearance is the clearance of virusby the immune system without the administra-

Table 1. Correlation between SVR rate and IL28B Callele frequency

SVR%

IL28B C allele

frequency

Sample

size

African-Americans 24% 0.40 191Hispanics 51% 0.58 75European-Americans 56% 0.63 871East Asians 76% 0.95 154

Linear regression r2 ¼ 0.93. Data adapted from Ge et al.

(2009).

Y.-F. Lu et al.


ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



tion of additional drugs. Based on studies ofthe natural history of HCV, 20%–30% of infect-ed patients can spontaneously clear the virus,whereas the other 70%–80% become chroni-cally infected and require drug therapy. Thespontaneous clearance rate was estimated tobe 36% in patients of non-African ancestryand 9% in patients of African ancestry (Thomaset al. 2000). Soon after the discovery of geneticassociation with treatment response for HCV,IL28B again was shown to be associated withthe spontaneous clearance of HCV. Thomasand colleagues examined the IL28B polymor-phism in six independent patient cohorts withthe diagnosis of HCV infection. Patients werecategorized as being chronically infected orhaving spontaneously cleared HCV by at leasttwo blood tests separated by an interval of atleast 6 months. Strikingly, the C allele of IL28B(rs12979860) also favors HCV clearance in thesecohorts consisting of both European- and Afri-can-Americans. Individuals with the lL28B C/Cgenotype were, once again, two to three timesmore likely to clear the virus than the non-C/Cpatients. This result was similar to what hadbeen observed in drug-induced HCV clearance.This finding suggests that IL28B has a universaleffect on HCV resolution in natural settingswithout the administration of drugs, an impor-tant biological clue.

Because there is clear evidence of IL28Bassociation with both treatment-induced andspontaneous viral clearance, it would be in-triguing to know whether IL28B is also associ-ated with the geographic distribution of HCVprevalence. However, the prevalence of HCV inmajor continents and IL28B frequency do notseem to be highly correlated. Although in mostAfrican countries, the prevalence rates are .3%of the total population (.3% is consideredhigh prevalence for HCV), many East Asiancountries (Asian populations have the highestrate of protective IL28B C allele) also comprisethe majority of HCV chronic infections world-wide. For example, there is a 3.2% seropreva-lence rate in China, which accounts for a majorglobal HCV-infected population (Shepard et al.2005). Many believe that country-specific fea-tures of the health care system itself may play

a major role in determining the likelihood ofHCV exposure. For example, the availability ofsafe injections dramatically decreases the chanceof exposure. Nevertheless, because HCV wasfirst discovered in 1989, it has been impossibleto obtain actual data of global HCV prevalencebefore industrialization.

Thanks to advances in tools for humangenetic research, GWAS methods provide uswith insight on common variants and infec-tious disease. In many carefully controlled clin-ical trials for HCV treatments, clear and consis-tent correlation between treatment success andthe presence of the IL28B polymorphism hasbeen shown. The genetic discovery shatters thelong-lasting myth that race plays a role in HCVclearance. In fact, most of the difference in SVRrate can be explained solely by the frequencydifferences of IL28B alleles among populations.HCV infection, therefore, provides a salutaryexample of how common variation affects dis-ease susceptibility and drug response.

IMPLICATIONS FOR HOW TO THINKABOUT DISCOVERY AND ITS CLINICAL USE

The relative importance of rare and commonvariants in the traits that impose the greatestpublic health burden in the developed worldremains unclear. Many believe that most casesof common disease are influenced by variantsdistributed across many genes, each with smalleffect, interacting with the environment in wayswe do not yet understand. But we also knowthat sometimes cases of relatively commonand certainly complex diseases can be causedby rare genetic changes of large effect. Amongthe best examples of the latter is neuropsychiat-ric disease, including conditions such as autism,epilepsy, and schizophrenia, in which rare, largegenetic rearrangements (so-called “copy-num-ber” variants) collectively account for a smallbut significant fraction of cases (Murdoch andState 2013). Another illustrative example of rarevariants with large effects is epilepsy. In a recentreport on two classical epileptic encephalopa-thies (infantile spasms and Lennox–Gastautsyndrome), researchers have discovered statisti-cally significant enrichment of de novo muta-



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



tions, that is, new variants that arise in the germ-line of the patient’s parents, in specific gene sets.Some of these genes have significantly more denovo mutations in the patient cohort thanwould be expected by chance. This finding dem-onstrates that de novo mutations (occurring atone of several different genes) can have a stronginfluence on the risk of epilepsy (Epi4K Con-sortium, Epilepsy Phenome/Genome Project2013).

A related question concerns the proportionof the functional genetic variation that is pre-sent in the human population as a result of someform of carrier advantage, as is clearly impor-tant in sickle cell anemia, or some sort of mu-tation-selection balance (in which the arisal ofde novo mutations is balanced by their loss be-cause they reduce the fitness of the mutationbearer) as is clearly responsible for the “copynumber” variants mentioned above. In the lat-ter, genetic rearrangements can lead to changesin human cognitive potential, but in ways thatwe cannot yet predict with a high degree of con-fidence.

Despite the clear similarity of human pop-ulations as described by Lewontin, the few ex-amples listed here show that no matter what sortof evolutionary tradeoffs existed in the past ofthe human species, the genetic bases of medi-cally relevant traits can be profoundly differentat both the individual and population levels.The nature of these differences will obviouslydepend, in large part, on what sort of geneticvariation causes most diseases. More generally,we still do not exactly know whether the major-ity of important variants is generally deleteriousand present because of mutation-selection bal-ance, or whether the important ones are morenuanced in their effects in that they are some-times helpful, sometimes harmful.

It is clear that geneticists cannot assume thataspects of human genetic disease and othermedically relevant traits can be understood bystudying only one population because humangroups are neither homogeneous nor geneticstudies fruitful unless they are population com-parative. As such, research programs that con-centrate their discovery efforts in populations ofEuropean ancestry alone, as most genomics ef-

forts have performed to date (Need and Gold-stein 2009; Bustamante et al. 2011), are ineffi-cient and incomplete. If we are to fulfill thepromise of the Human Genome Project, en-hance biological discovery, and begin to bringour knowledge of population genetics to bear onlong-standing health disparities, then we mustunderstand and appreciate the enormous rangeof variation within our species. As the poetAudra Lorde wrote, “It is not our differencesthat divide us. It is our inability to recognize,accept, and celebrate those differences.”

REFERENCES

Aidoo M, Terlouw DJ, Kolczak MS, McElroy PD, ter KuileFO, Kariuki S, Nahlen BL, Lal AA, Udhayakumar V. 2002.Protective effects of the sickle cell gene against malariamorbidity and mortality. Lancet 359: 1311–1312.

Alfonso-Sanchez MA, Perez-Miranda AM, Garcia-ObregonS, Pena JA. 2010. An evolutionary approach to the highfrequency of the DF508 CFTR mutation in Europeanpopulations. Med Hypotheses 74: 989–992.

Ashley-Koch A, Yang Q, Olney RS. 2000. Sickle hemoglobin(HbS) allele and sickle cell disease: A HuGE review. AmJ Epidemiol 151: 839–845.

Bertranpetit J, Calafell F. 1996. Genetic and geographicalvariability in cystic fibrosis: Evolutionary considerations.Ciba Found Symp 197: 97–114; discussion 114–118.

Bray SM, Mulle JG, Dodd AF, Pulver AE, Wooding S, WarrenST. 2010. Signatures of founder effects, admixture, andselection in the Ashkenazi Jewish population. Proc NatlAcad Sci 107: 16222–16227.

Bustamante CD, Burchard EG, De la Vega FM. 2011. Geno-mics for the world. Nature 475: 163–165.

Cavalli-Sforza LL, Menozzi P, Piazza A. 1994. The history andgeography of human genes. Princeton University Press,Princeton, NJ.

Cirulli ET, Goldstein DB. 2010. Uncovering the roles of rarevariants in common disease through whole-genome se-quencing. Nat Rev Genets 11: 415–425.

Cochran G, Hardy J, Harpending H. 2006. Natural history ofAshkenazi intelligence. J Biosoc Sci 38: 659–693.

Epi4K Consortium, Epilepsy Phenome/Genome Project.2013. De novo mutations in epileptic encephalopathies.Nature 501: 217–221.

Fernandes Filho JA, Shapiro BE. 2004. Tay–Sachs disease.Arch Neurol 61: 1466–1468.

Ge D, Fellay J, Thompson AJ, Simon JS, Shianna KV, UrbanTJ, Heinzen EL, Qiu P, Bertelsen AH, Muir AJ, et al. 2009.Genetic variation in IL28B predicts hepatitis C treat-ment-induced viral clearance. Nature 461: 399–401.

Ghany MG, Strader DB, Thomas DL, Seeff LB, AmericanAssociation for the Study of Liver Diseases. 2009. Diag-nosis, management, and treatment of hepatitis C: Anupdate. Hepatology 49: 1335–1374.

Y.-F. Lu et al.


ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



Hirszfeld L, Hirszfeld H. 1919. Essai d’application desmethodes au problem des races [Testing the applicationof methods to the question of race]. Anthropologie 29:505–537.

Lewontin R. 1972. The apportionment of human diversity.In Evolutionary biology (ed. Dobzhansky T, Hecht M,Steere W), pp. 381–398. Appleton Centuary Crofts, NewYork.

Liu CH, Liu CJ, Lin CL, Liang CC, Hsu SJ, Yang SS, Hsu CS,Tseng TC, Wang CC, Lai MY, et al. 2008. Pegylated inter-feron-a-2a plus ribavirin for treatment-naive Asian pa-tients with hepatitis C virus genotype 1 infection: A mul-ticenter, randomized controlled trial. Clin Infect Dis 47:1260–1269.

Lorde A. 2007. Sister outsider: Essays and speeches. Ten SpeedPress, New York.

Manolio TA, Collins FS, Cox NJ, Goldstein DB, HindorffLA, Hunter DJ, McCarthy MI, Ramos EM, Cardon LR,Chakravarti A, et al. 2009. Finding the missing heritabil-ity of complex diseases. Nature 461: 747–753.

Murdoch JD, State MW. 2013. Recent developments in thegenetics of autism spectrum disorders. Curr Opin GenetDev 23: 310–315.

Need AC, Goldstein DB. 2009. Next generation disparities inhuman genomics: Concerns and remedies. Trends Genet25: 489–494.

Nei M, Roychoudhury AK. 1972. Gene differences betweenCaucasian, Negro, and Japanese populations. Science177: 434–436.

Ostrer H, Skorecki K. 2013. The population genetics of theJewish people. Hum Genet 132: 119–127.

Poolman EM, Galvani AP. 2007. Evaluating candidate agentsof selective pressure for cystic fibrosis. J R Soc Interface 4:91–98.

Ratjen F, Doring G. 2003. Cystic fibrosis. Lancet 361: 681–689.

Shepard CW, Finelli L, Alter MJ. 2005. Global epidemiologyof hepatitis C virus infection. Lancet Infect Dis 5: 558–567.

Slatkin M. 2004. A population-genetic test of founder effectsand implications for Ashkenazi Jewish diseases. Am JHum Genet 75: 282–293.

Suppiah V, Moldovan M, Ahlenstiel G, Berg T, Weltman M,Abate ML, Bassendine M, Spengler U, Dore GJ, Powell E,et al. 2009. IL28B is associated with response to chronichepatitis C interferon-a and ribavirin therapy. Nat Genet41: 1100–1104.

Tanaka Y, Nishida N, Sugiyama M, Kurosaki M, Matsuura K,Sakamoto N, Nakagawa M, Korenaga M, Hino K, Hige S,et al. 2009. Genome-wide association of IL28B with re-sponse to pegylated interferon-a and ribavirin therapyfor chronic hepatitis C. Nat Genet 41: 1105–1109.

The 1000 Genomes Project Consortium. 2010. A map ofhuman genome variation from population-scale se-quencing. Nature 467: 1061–1073.

Thomas DL, Astemborski J, Rai RM, Anania FA, SchaefferM, Galai N, Nolt K, Nelson KE, Strathdee SA, Johnson L,et al. 2000. The natural history of hepatitis C virus infec-tion: Host, viral, and environmental factors. JAMA 284:450–456.

Thomas DL, Thio CL, Martin MP, Qi Y, Ge D, O’Huigin C,Kidd J, Kidd K, Khakoo SI, Alexander G, et al. 2009.Genetic variation in IL28B and spontaneous clearanceof hepatitis C virus. Nature 461: 798–801.

Torkamani A, Pham P, Libiger O, Bansal V, Zhang G, Scott-Van Zeeland AA, Tewhey R, Topol EJ, Schork NJ. 2012.Clinical implications of human population differences ingenome-wide rates of functional genotypes. Front Genet3: 211.

Via M, Gignoux C, Burchard EG. 2010. The 1000 GenomesProject: New opportunities for research and social chal-lenges. Genome Med 2: 3.



ww

w.p

ersp

ecti

vesi

nm

edic

ine.

org



published online July 24, 2014Cold Spring Harb Perspect Med Yi-Fan Lu, David B. Goldstein, Misha Angrist and Gianpiero Cavalleri Personalized Medicine and Human Genetic Diversity

Subject Collection Human Variation

DiversityPersonalized Medicine and Human Genetic

Yi-Fan Lu, David B. Goldstein, Misha Angrist, et al.

Race in Biological and Biomedical ResearchRichard S. Cooper

http://perspectivesinmedicine.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright © 2014 Cold Spring Harbor Laboratory Press; all rights reserved


http://perspectivesinmedicine.cshlp.org/cgi/collection/

http://perspectivesinmedicine.cshlp.org/cgi/collection/


personalized medicine and human genetic...

Documents