research article genetic adaptation of the hypoxia-inducible factor

Research

articleGenetic Adaptation of the Hypoxia-Inducible Factor Pathway toOxygen Pressure among Eurasian Human PopulationsLin-dan Ji,1,2,3 Yu-qing Qiu,2,4 Jin Xu,1,2,5 David M. Irwin,1,6,7 Siu-Cheung Tam,8 Nelson L.S. Tang,*,2,4,9,10

and Ya-ping Zhang*,1,2,11

1State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences,

Kunming, China2KIZ/CUHK Joint Laboratory of Bioresources and Molecular Research in Common Diseases, Kunming, China3Institute of Biochemistry, School of Medicine, Ningbo University, Ningbo, China4Department of Chemical Pathology, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong, China5Institute of Public Health, School of Medicine, Ningbo University, Ningbo, China6Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, ON, Canada7Banting and Best Diabetes Centre, University of Toronto, Toronto, ON, Canada8School of Biomedical Sciences, The Chinese University of Hong Kong, Hong Kong, China9Laboratory for Genetics of Disease Susceptibility, Li Ka Shing Institute of Health Sciences, The Chinese University of Hong Kong,

Hong Kong, China10ShenZhen Research Institute, The Chinese University of Hong Kong, Hong Kong, China11Laboratory for Conservation and Utilization of Bio-resource, Yunnan University, Kunming, China

*Corresponding author: E-mail: [email protected]; [email protected].

Associate Editor: Noah Rosenberg

Abstract

Research into the mechanisms of human adaptation to the hypoxic environment of high altitude is of great interest to the fieldsof human physiology and clinical medicine. Recently, the gene EGLN1, from the hypoxia-inducible factor (HIF) pathway, wasidentified as being involved in the hypoxic adaptation of highland Andeans and Tibetans. Both highland Andeans and Tibetanshave adapted to an extremely hypoxic habitat and less attention has been paid to populations living in normoxic conditions atsea level and mild-hypoxic environments of moderate altitude, thus, whether a common adaptive mechanism exists in responseto quantitative variations of environmental oxygen pressure over a wide range of residing altitudes is unknown. Here, we firstperformed a genome-wide association study of 35 populations from the Human Genome Diversity-CEPH Panel who dwell at sealevel to moderate altitude in Eurasia (N = 691; 0–2,500 m) to identify the genetic adaptation profile of normoxic and mild-hypoxicinhabitants. In addition, we systematically compared the results from the present study to six previously published genome-widescans of highland Andeans and Tibetans to identify shared adaptive signals in response to quantitative variations of oxygenpressure. For normoxic and mild-hypoxic populations, the strongest adaptive signal came from the mu opioid receptor-encodinggene (OPRM1, 2.54� 10�9), which has been implicated in the stimulation of respiration, while in the systematic survey theEGLN1-DISC1 locus was identified in all studies. A replication study performed with highland Tibetans (N = 733) and sea level HanChinese (N = 748) confirmed the association between altitude and SNP allele frequencies in OPRM1 (in Tibetans only, P< 0.01)and in EGLN1-DISC1 (in Tibetans and Han Chinese, P< 0.01). Taken together, identification of the OPRM1 gene suggests thatcardiopulmonary adaptation mechanisms are important and should be a focus in future studies of hypoxia adaptation.Furthermore, the identification of the EGLN1 gene from the HIF pathway suggests a common adaptive mechanism forEurasian human populations residing at different altitudes with different oxygen pressures.

Key words: adaptive evolution, human genetic variation, oxygen pressure, hypoxia, hypoxia-inducible factor pathway, Eurasianpopulations.

IntroductionPhysiological adaptation to hypobaric hypoxia is an impor-tant area of human research and does not only reflect thephysiology of high-altitude adaptation, but also may yieldbenefit for both sports medicine and the treatment of pa-tients with chronic hypoxia resulting from various forms ofclinical illness (Martin and Windsor 2008). Physiological re-sponses in altitude acclimatization have already been

investigated in highland newcomers (Houston and Riley1947; Vogel and Harris 1967; Reynafarje and Hurtado 1971;Hannon and Sudman 1973). The successful climb of MountEverest by Messner and Habeler without the use of supple-mental oxygen led to research into the physiological adapta-tion mechanisms to high-altitude hypoxia (Oelz et al. 1986).Various acclimatization responses were identified including:1) cardiovascular responses which increase cardiac output,

� The Author 2012. Published by Oxford University Press on behalf of the Society for Molecular Biology and Evolution. All rights reserved. For permissions, pleasee-mail: [email protected]

Mol. Biol. Evol. 29(11):3359–3370 doi:10.1093/molbev/mss144 Advance Access publication May 23, 2012 3359

Dow

nloaded from https://academ

ic.oup.com/m

be/article-abstract/29/11/3359/1146034 by guest on 05 April 2019

2) alveolar gases exchange, kidney, and CNS responses tochange the oxygen and carbon dioxide partial pressureswhich are then followed by 3) increased hemoglobin concen-tration (a result of both passive hemoconcentration and,later, active erythropoiesis) (Martin and Windsor 2008). Allthese responses are capable of restoring, or even surpassing,the arterial oxygen content and oxygen flux found at sea level,therefore, they are believed to explain the acclimatization ofboth acutely and chronically exposed newcomers (West 1998;Hornbein and Schoene 2001). However, more recent obser-vations of compromised exercise performance of lowlandersafter prolonged acclimatization and of the harmful long-termexposure to high oxygen therapy in critically ill patients haverevealed that these classical mechanisms are insufficient toexplain the acclimatization of highland newcomers (Daviset al. 1983; Hayes et al. 1994; Calbet et al. 2003; Grocottet al. 2007).

In contrast to studies of how lowlanders acclimatize tohigh altitude, studies of the mechanisms that lead to success-ful settlement of highlanders have revealed many new andunexpected findings (Beall et al. 2010; Yi et al. 2010). To date,two populations, Andeans and Tibetans, who have function-ally successfully adapted to high-altitude hypoxia, have beenphysiologically well characterized (Beall 2007). Due to im-proved maternal vascular adaptation, they both give birthto heavier babies than acutely exposed lowlanders, and thisphenomenon yields better fetal outcomes (Moore et al. 1986,2001, 2004). Despite this similarity, Andean and Tibetan pop-ulations, however, exhibit different adaptation profiles for nu-merous physiological traits, including circulation, respiration,and hematopoiesis (Beall 2007). Highland Tibetans havehigher resting ventilation (VE) and hypoxic ventilatory re-sponse (HVR) than their Andean counterparts (Beall et al.1997), but with lower oxygen saturation and hemoglobinconcentrations (Beall and Goldstein 1987; Beall et al. 1990,1998). Interestingly, highland Tibetans do not adapt byhypoxia-induced erythropoiesis leading to polycythemia asobserved in highland Andeans and newcomers; instead,their hemoglobin levels are comparatively lower than thoseof their lowland Han Chinese counterparts (Adams andStrang 1975; Beall and Reichsman 1984; Beall et al. 2010; Yiet al. 2010). These findings contradict the long-held beliefthat the restoration of arterial oxygen content and flux bypolycythemia is the best adaptation option for humans underhypobaric hypoxia (West 1998; Hornbein and Schoene 2001).These results emphasize that additional research is needed toidentify and distinguish common as well as population-specific physiological mechanisms for hypoxia adaptation.

Although the physiological mechanisms of adaptation tohigh altitude have been extensively characterized, the under-lying genetic basis has not been fully understood. Previousstudies found that genetic factors accounted for a high pro-portion of the phenotypic variance in hemoglobin concen-tration among highland Tibetan and Andean populations(Beall et al. 1998). Quantitative genetic analyses of the familialpatterning of different adaptive traits between the two pop-ulations also provided evidence of population-specific adap-tation mechanisms (Beall 2000). Recently, EGLN1 and EPAS1,

two genes coding for molecular components of thehypoxia-inducible factor (HIF) pathway, were found to havecontributed to high-altitude adaptation in Tibetans (Beallet al. 2010; Bigham et al. 2010; Simonson et al. 2010; Yiet al. 2010; Peng et al. 2011; Xu et al. 2011); while EGLN1,PRKAA1, and NOS2A, of the same pathway, were suggestedto underlie the high-altitude adaptation in Andeans (Bighamet al. 2010). The sharing association with the EGLN1 gene byboth populations suggests that a common adaptive mecha-nism in the HIF pathway occurred in both highland Andeansand Tibetans. On the other hand, genetic polymorphisms ofthe EPAS1 gene were significantly associated with hemoglobinconcentrations in highland Tibetans (Beall et al. 2010; Yi et al.2010), which partially explains the different adaptation mech-anisms of the two highland populations. Notwithstanding theabove information, for these two highland populations,whether other common and populations-specific adaptationmechanisms exist is still uncertain and their underlying ge-netic basis has not yet been fully characterized.

Multiple environmental factors vary with an increase inaltitude, including ambient temperature, circadian rhythm(Virues-Ortega et al. 2004), oxygen pressure (Beall 2007), ul-traviolet radiation (Boucher 2010), and rainfall (Brunsdonet al. 2001). Among these, since the effect of other environ-mental factors can be mitigated or prevented, the loss ofoxygen pressure due to the exponentially decreasing baro-metric pressure with altitude is deemed as a determiningfactor in high-altitude adaptation (Beall 2007). Quantitativevariations of oxygen pressure, thus, form a continuous gradi-ent of altitude-specific habitats, where human physiologicalresponses also demonstrate a clinal pattern. In normoxia atsea level, the human physiological variability is within normallimits and is considered the reference condition. When ex-posed to mild hypoxia of moderate altitude, peripheral che-moreceptors are stimulated and activity in the carotid sinus isincreased, although no obvious increase in ventilation can beobserved (Mason 2000). In addition, a well-characterizedimpairment of cognitive function due to exposure to hypo-baric hypoxia can first be detected at a moderate altitude(2,440 m) (Farmer et al. 1992). Furthermore, a significantdecline in birth weight could be observed at the range of1,500–1,999 m, suggesting that fetal growth is sensitive touterine O2 delivery even at moderate altitude (Yip 1987;Zamudio et al. 1995; Mortola et al. 2000). Work in animalmodels also found clinal patterns in both phenotypic andgenetic variations associated with the range of moderate al-titudes (0–2,500 m) (Bitner-Mathe et al. 1995; Sorensen et al.2005; Collinge et al. 2006; Soto et al. 2010), providing supportfor observations in humans. As the ascent in altitude con-tinues, the human body gradually show significant physiolog-ical changes of the respiratory, circulatory, and hematopoieticsystems (West 1998; Hornbein and Schoene 2001). From theabove information, therefore, it is reasonable to hypothesizethat human populations residing over a wide range of alti-tudes, not just those that live at high altitude, have undergonenatural selection due to environmental oxygen pressure.

Most studies on the adaptive response to reduced oxygenpressure, however, have been conducted in Andeans and

3360

Ji et al. . doi:10.1093/molbev/mss144 MBED

ownloaded from

https://academic.oup.com

/mbe/article-abstract/29/11/3359/1146034 by guest on 05 April 2019

Tibetans who inhabit altitudes above 3,500 m, a level wherethe environment is extremely hypoxic and their results arenot conductive for normal human physiological function. Lessattention has been paid to adaptation to oxygen pressure inpopulations that live in normoxia at sea level or mild-hypoxicenvironments of moderate altitude. In particular, whether acommon adaptive mechanism that responds to quantitativevariations of oxygen pressure over a wide range of residentaltitudes is unknown. Since�95% of human populations liveat low and moderate-altitude habitats (0–2,500 m) (Penalozaand Arias-Stella 2007), it is important to understand oxygenpressure adaptation in normoxic and mild-hypoxic popula-tions as this will help illuminate the underlying adaptivemechanisms used by highland inhabitants, thus, help inter-pret the overall oxygen pressure adaptation mechanisms usedby human beings.

To examine the genetic adaptation profile of normoxicand mild-hypoxic inhabitants, we conducted a genome-wideassociation study for 35 Eurasian populations in the HumanGenome Diversity-CEPH Panel (HGDP-CEPH, N = 691, dwell-ing from sea level to 2,500 m), which have extensive SNPgenotyping data (Li et al. 2008). We then compared the re-sults from our study to several recently publishedgenome-wide scans of highland Andeans and Tibetans tosystematically identify adaptive mechanisms in response toquantitative variations of oxygen pressure. Genome-wide sig-nificant signals found in normoxic and mild-hypoxic popula-tions and shared signals over a wide range of altitudes werefurther investigated in a replication study with highlandTibetans (N = 733) as well as lowland and moderate-altitudeHan Chinese (N = 748). Our study, thus, for the first timereports the genetic profile of normoxic and mild-hypoxicadaptation in populations residing at low and moderate alti-tudes, and demonstrates that the HIF pathway operates inoxygen pressure adaptation over a wider range of altitudesthan previously recognized.

Materials and Methods

Studied Populations and Genotype Data

Population genotype data was retrieved from theHGDP-CEPH data set that included 952 unrelated individualsfrom 53 populations genotyped by the Illumina 650Y plat-form (Li et al. 2008), which is freely available at ftp://ftp.cephb.fr/hgdp_supp1. In all previous studies of Tibetan adaptationto high altitude, lowland Han Chinese subjects were usuallyused as controls (Beall et al. 2010; Yi et al. 2010). Similarly,lowland Europeans were usually used as controls in studies ofAndean adaptation to high altitude (Moore et al. 2004).Using genetically closely related highland Tibetans and low-land Han Chinese (Zhao et al. 2009) as well as the highlandAndeans and lowland European residents (Hunley and Healy2011) avoids possible confounding effects due to potentialpopulation structure and thus should effectively detect ge-netic polymorphisms under altitude selection. Accordingly, all35 Eurasian populations (Burusho and Cambodian popula-tions were excluded due to lack of specific altitude data forthese populations) were included. Genotype data from these

35 Eurasian populations that inhabit an extensive altitudinalrange of Europe and Asia (N = 691, supplementary fig. S1,Supplementary Material online) were used for this study.

Geographic Information

Altitude data was used as a surrogate for ambient oxygenlevels and were obtained from the World MeteorologicalOrganization publication “1961-1990 Global ClimateNormals,” which was compiled by National Climatic DataCenter of the United States and was also freely availablefrom Hong Kong Observatory website (http://gb.weather.gov.hk) (Ji et al. 2010). Geographic coordinates of the popu-lations were from the HGDP-CEPH data set. Altitude data forthe Chinese Han populations were extracted from “TheAnnual Surface Climate Normals of InternationalExchanging Stations of China (1971-2000)” through theChina Meterological Data Sharing Service System (http://cdc.cma.gov.cn). Altitude data were transformed to naturallogarithms, which follow a normal distribution.

Subjects and Samples for Replication Study

To replicate the results from the lowland and moderate-altitude Eurasian populations, samples were collected fromhighland Tibetan as well as lowland and moderate-altitudeHan Chinese populations, both of which are genetically clo-sely related (Zhao et al. 2009). A total of 733 native Tibetanindividuals were sampled from two regions of Tibet (Naquand Rikaze, inhabiting at >3,500 m), and 748 Han Chineseindividuals were collected from Liaoning, Shandong, Sichuan,Yunnan, and Guangdong Provinces (0–2,000 m), representingnorthern, central, and southern Han Chinese populations.Blood samples were obtained with informed consent. Theprotocol for this study was reviewed and approved by theethics committee of the Kunming Institute of Zoology,Chinese Academy of Sciences.

Genotyping for Replication Study

Five SNPs (rs17084501 of OPRM1, rs6605044 of EGLN1-DISC1,rs2495718 of HIF1AN, rs11881124 of EGLN2, and rs1809461 ofEGLN3) identified in the Eurasian HGDP-CEPH populationswere genotyped in the Tibetan and Han Chinese replicationsamples. Genomic DNA was extracted from whole blood bythe standard phenol/chloroform method. SNP rs11881124 ofthe EGLN2 gene was genotyped by PCR–RFLP using mis-matched primers (forward primer 50-CTCCCAGCTTCCCTCCGGAAT-30, reverse primer 50-GGAGAGTCAGGGTGGGGG-30) with the products digested with the restriction enzymeSspI (New England Biolabs). The other four SNPs were geno-typed by the GenomeLab SNPstream Genotyping System(Beckman Coulter Inc.) according to the protocol providedby the manufacturer. Among them, rs6605044 of theEGLN1-DISC1 locus was replaced by a nearby SNP(rs10797564, D’ = 1.000) due to lack of available appropriateprimers (www.autoprimer.com, Beckman Coulter Inc.).

3361

Human Adaptation to Oxygen Pressure . doi:10.1093/molbev/mss144 MBED

ownloaded from



Data Analyses

Among the 644,258 genotyped autosomal SNPs, 503,698 weredefined as common polymorphisms and were used for sub-sequent analysis (MAFffi 0.1). Nei’s measurement of the av-erage gene diversity per locus HS was calculated for each ofthe 35 populations (Nei 1973). A genome-wide associationstudy was performed by using additive effect with a domi-nance deviation model implemented in PLINK (Purcell et al.2007). To adjust for inflation due to population stratification,principal components (PCs) were used as covariates. PCs,which explain the intrinsic population structures, were calcu-lated by principal components analysis (PCA) of all the auto-somal genotype data (Price et al. 2006). WGAViewer softwarewas used to evaluate the genomic inflation factor and to drawManhattan plots to show the locations of the significant sig-nals (Ge et al. 2008).

SNPs with significant association were mapped to specificgenes through the dbSNP dataset (Sherry et al. 2001). ForSNPs within intergenic regions, the most adjacent gene wasassigned. Genes were then mapped to their Gene Ontologyannotation terms (Ashburner et al. 2000) and KEGG path-ways (Kanehisa et al. 2006) via Entrez Gene, using the publiclyavailable program Database for Annotation, Visualization,and Integrated Discovery (DAVID) (Huang da et al.2009a,b). Like many other similar publicly available tools,DAVID adopts a common core strategy to systematicallymap a list of genes to the associated biological annotation,and then statistically highlight the most enriched biologicalannotation out of thousands of linked terms and contents.Therefore, it can help identify biological processes most per-tinent to the biological phenomena under study (Huang daet al. 2009b).

A one-way ANOVA was used in the comparison of habitataltitude with different genotypes for the OPRM1 SNP in theEurasian HGDP-CEPH populations. Statistical analyses of theallele frequencies, sex and age differences were performed by�2-test. The associations of the allele frequencies and theresiding altitudes of the five confirmed signals among theHan Chinese populations were tested through Pearson cor-relation analysis, and among Tibetan populations were testedby Spearman rank correlation analysis (due to the lack ofspecific dwelling altitude data). All of these procedures wereperformed using SPSS for Windows, 13.0. Hardy–Weinbergequilibrium (HWE) was tested with the software PEDSTATSV0.6.8 (http://www.sph.umich.edu/csg/abecasis/).

Results

Candidate Genes for Normoxic and Mild-hypoxicAdaptation in Eurasian Populations from Genome-wide Association Study

To examine the genetic adaptation profile of normoxic andmild-hypoxic inhabitants, we conducted a genome-wide as-sociation study with the extensive SNP genotyping data from35 Eurasian HGDP-CEPH populations (N = 691, living at sealevel to 2,500 m). No significant association of average genediversity with residing altitude was observed (PearsonP> 0.05, supplementary table S1, Supplementary Material

online) indicating that the effect of potential founder effectsor population bottlenecks on the following association studyis low. The 35 studied populations were comparatively morehomogeneous than a worldwide sample from the full HGDPas shown for their genotypes using the first two PCs (fig. 1),similar to what has been previously described (Reich et al.2008). Among the first 68 significant PCs (P< 0.001), accord-ing to the TW statistic (Patterson et al. 2006), the amount ofvariation explained reaches a leveling-off by PC20, therefore,the inflation factor (�GC) due to population structure wasnormalized by adjustment through the top 20 PCs(�GC = 1.0759, fig. 2). The list of candidate SNPs was alsomore stable when approximately 20 PCs, rather than onlyfew PCs (e.g., 3 PCs), were used (supplementary tablesS2–S3, Supplementary Material online).

At the genome-wide scale, the strongest signal thatachieved significance was found near the OPRM1 gene,which encodes the mu opioid receptor (rs17084501,P = 2.54� 10�9, fig. 3), with a P-value that reached the com-monly accepted genome-wide significant level of< 1� 10�7.In general, the ancestral T allele is significantly enriched withan increase in altitude (ANOVA P< 0.001, fig. 4). To furthersurvey for any other significant associations using a pathwayapproach, the top 500 SNPs with the highest genome-wideassociation scores were examined as they likely includeadditional selection signals (supplementary table S4,Supplementary Material online). These 500 SNPs map to337 unique genes. Applying the GO annotation and KEGGpathway analysis through DAVID, we failed to identify anysignificant gene categories (Bonferroni P> 0.05).

Evidence for Shared Adaptation Mechanism Over aWide Range of Altitudes

Similar to the approach used to reveal shared adaptationmechanism at extremely high altitude by using a systematicsurvey for common genes (Simonson et al. 2010), we com-bined results from six recent genome-wide scans in Andeansand Tibetans with our present study to identify shared genesshowing adaptation signals over a wide range of altitudes.Since the recently published genome-wide scans in Tibetansand Andeans usually proposed only a list of candidate genes,instead of specific SNPs, therefore, the adaptation signals werecompared through a gene-based analysis though some de-tailed information would be lost due to this procedure.

We first reappraised six recent genome-wide scans to iden-tify any shared high-altitude adaptation signals in highlandTibetan and Andean populations (Beall et al. 2010; Bighamet al. 2010; Simonson et al. 2010; Yi et al. 2010; Peng et al.2011; Xu et al. 2011), resulting in the identification of eightcandidate gene lists from highland Tibetans and one fromhighland Andeans (supplementary table S5, SupplementaryMaterial online). The gene EGLN1 was most reproducible, as itwas present in all of the Tibetan and Andean candidate genelists while EPAS1 was highly reproducible in the Tibetans,appearing in seven of the eight candidate gene lists. The ex-clusivity of the EPAS1 gene to the highland Tibetans togetherwith the finding that its genetic polymorphisms are

3362


ownloaded from



FIG. 2. QQ plot of the genome-wide association study for normoxic and mild-hypoxic adaptation. Among the first 68 significant PCs (P< 0.001), theamount of variation explained levels off by PC20. After adjustment for the top 20 PCs, the genome inflation is almost normalized (�GC = 1.0759).

FIG. 1. Principal component analysis of the 35 Eurasian HGDP-CEPH populations (N = 691) by the first two principal components. Among the first 68significant principal components (PCs, P< 0.001), the amount of variation explained levels off by PC20 (�GC = 1.0759). The 35 studied populations werecomparatively more homogeneous than a worldwide sample from the full HGDP as shown by their genotypes using the first two PCs.

3363


ownloaded from



significantly associated with hemoglobin concentrations sug-gested that it could underlie the genetic basis for differenthypoxia adaptation mechanisms between the twopopulations.

Next, the top 500 SNPs from the present study were at-tributed to 337 unique genes. These 337 genes for normoxicand mild-hypoxic adaptation were then mapped to the can-didate gene lists for highland adaptation in Tibetans andAndeans. The DISC1 gene was found to be most reproduciblewhich appeared in the present study and four of the eightTibetan gene lists, followed by HIF1AN which appeared in thepresent study and three of the eight Tibetan candidate genelists (fig. 5). The high reproducibility of the DISC1 locus ispotentially due to its close genetic linkage with the nearbyEGLN1 gene (Yi et al. 2010), which is only about 201 kb up-stream. To test this hypothesis, we examined another genethat is closely linked to EGLN1, C1orf124 (Yi et al. 2010), whichis about 11 kb upstream of EGLN1. SNP rs2437150, which is

located within C1orf124, was found to have a weak associationwith altitude (P = 0.0018, ranked 1,635th of the 503,698 SNPs,table 1), therefore it is likely that the signal at theEGLN1-DISC1 locus exists in normoxic and mild-hypoxic in-habitants and is probably due to the EGLN1 gene.

EGLN1 is a member of a family of prolyl hydroxylase encod-ing genes, whose products degrade HIF under normoxic con-ditions (Lendahl et al. 2009). The two other members of thisgene family, EGLN2 and EGLN3, also appeared as top candi-date signals (rs11881124 of the EGLN2 gene ranked 514th,while rs1809461 of EGLN3 ranked 1,947th of the 503,698SNPs, table 1), though neither was statistically significant atthe GWAS level. These results suggest that a genuine enrich-ment of prolyl hydroxylase encoding genes in normoxic andmild-hypoxic adaptation. On the other hand, the enzymeasparaginyl hydroxylase encoded by the highly reproducibleHIF1AN is another class of enzymes that can degrade HIFunder normoxia (Lando et al. 2002). Together, the genes

FIG. 4. Boxplot of the rs17084501 genotype distribution at different altitude levels among the 35 Eurasian HGDP-CEPH populations (N = 691).The ancestral T allele of rs17084501 is significantly enriched with an increase in altitude (ANOVA P< 0.001).

FIG. 3. Manhattan plot of the genome-wide association study (GWAS) for normoxic and mild-hypoxic adaptation. 503,698 genotyped autosomalcommon SNPs (MAFffi 0.1) were analyzed in GWAS study. The single genome-wide signal rs17084501 is located near the OPRM1 gene on chromo-some 6 (P = 2.54� 10�9).

3364


ownloaded from



that encode these two classes of enzymes which degrade HIFunder normoxia (Lendahl et al. 2009), EGLN1-DISC1/ EGLN2/EGLN3 and HIF1AN, appear to be candidates for normoxiaand mild-hypoxia adaptation.

Replication Study with Highland Tibetans as well asLowland and Moderate-Altitude Han Chinese (TotalN = 1,481)

The distributions of the most significant hit OPRM1, highlyreproducible genes (EGLN1-DISC1 and HIF1AN) and relatedgenes in the EGLN1 gene family (EGLN2 and EGLN3) wereexamined in a replication sample of highland Tibetans(N = 733) as well as lowland and moderate-altitude HanChinese (N = 748) who were genotyped for their SNPs usingthe GenomeLab SNPstream Genotyping System (BeckmanCoulter Inc.) and PCR–RFLP methods. The Genetic PowerCalculator (available online at http://pngu.mgh.harvard.edu/�purcell/gpc/) indicated that our sample size was largeenough to do a case–control analysis with 80% power forall five SNPs (Purcell et al. 2003). No significant deviationfrom HWE was observed for these SNPs (P> 0.001), norwere any significant sex or age differences for these popula-tions found (P> 0.05).

For the OPRM1 SNP, the frequencies of the rs17084501Tallele were nearly fixed in both populations, and no significantdifferential signal was detected in the Han and Tibetan pop-ulations (table 3). However, if the two populations were di-vided into different altitude range subgroups then the OPRM1SNP showed a significant T allele frequency increase with anincrease in altitude in the highland Tibetan populations(N = 3, Spearman rank correlation P< 0.01, fig. 6 and supple-mentary table S6, Supplementary Material online). This trendis consistent with the pattern derived from the lowland andmoderate-altitude Eurasian HGDP-CEPH populations inwhich T allele was associated with higher altitude.

The highly reproducible EGLN1-DISC1 and HIF1AN genesboth showed a significant allelic differentiation betweenTibetan and Han Chinese populations (P< 0.05, table 3).When the subgroups were divided by altitude, theEGLN1-DISC1 locus showed a significant increase in the fre-quency of the rs10797564A allele with an increase in altitudein both Tibetan and Han Chinese subpopulations (N = 3,Spearman rank correlation P< 0.01 in Tibetan populations;N = 5, Pearson correlation P = 0.0012 in Han Chinese popula-tions; fig. 7 and supplementary tables S7–S8, SupplementaryMaterial online). On the other hand, the EGLN2 SNPrs11881124 was monomorphic (A allele only) in both the

FIG. 5. Intersection results of the top 337 genes for normoxic/mild-hypoxic adaptation and the eight gene lists for extremely-hypoxic adaptation fromsix previously published studies. Results from recent genome-wide scans in Andean and Tibetan populations are compared to those of the presentstudy to identify shared adaptive signals. The most reproducible genes are DISC1 and HIF1AN. Different colors represent candidate lists as follows: darkblue for iHS candidates (PMID 20466884, 426 genes); green for XP-EHH candidates (PMID 20466884, 433 genes); red for candidates from the presentstudy (337 genes); yellow for candidates from PMID 20961960 (29 genes); light blue for candidates from PMID 20595611 (35 genes); and orange forcandidates from PMID 2053544 (105 genes).

Table 1. Enrichment of Prolyl Hydroxylase Encoding Genes in the Present Study.

SNP Gene Description Nearby candidate P-value

rs6605044 DISC1 Disrupted in schizophrenia 1 EGLN1 0.0002

rs2437150 C1orf124 Hypothetical protein LOC83932 EGLN1 0.0018

rs11881124 EGLN2 HIF proly hydroxylase 1, HIF-PH1 (Self) 0.0005

rs1809461 EGLN3 HIF proly hydroxylase 3, HIF-PH3 (Self) 0.0023

3365


ownloaded from



Tibetan and Han Chinese populations, and no significantdifference in the distributions of the EGLN3 alleles was seen(P> 0.05).

DiscussionPrevious studies have extensively examined the mechanismsof adaptation to hypoxia in highland human populations,especially those of Tibetans (Beall et al. 2010; Bigham et al.2010; Simonson et al. 2010; Yi et al. 2010; Peng et al. 2011; Xuet al. 2011) and Andeans (Bigham et al. 2009, 2010).Reviewing these genome-wide scans indicates that conver-gent evolution may have occurred in highland Tibetan andAndean populations as suggested by the shared associationwith the EGLN1 gene, which encodes a component of the HIFpathway (Bigham et al. 2010). Different physiological adapta-tion traits and the exclusively adaptive signals between thesetwo populations also suggested population-specific adapta-tion mechanisms (Beall 2007). While a considerable numberof studies have examined the adaptation of Andeans andTibetans to an extremely hypoxic habitat, less attention hasbeen paid to populations living in normoxic and mild-hypoxic

environments, especially in mild hypoxia of moderate altitudewhere many physiological responses to reduced oxygen pres-sure can be first observed (Farmer et al. 1992; Mason 2000;Mortola et al. 2000). Thus, whether a general adaptive mech-anism exists in response to quantitative variations of oxygenpressure over a wide range of residing altitudes is unknown.A SNP near the OPRM1 gene was identified as the onlygenome-wide significant signal for normoxic and mild-hypoxic adaptation in this study. More importantly, thehighly reproducible EGLN1-DISC1 locus was identifiedamong the highland Tibetans and Andeans populations (in-habiting> 2,500 m) as well as the lowland andmoderate-altitude Eurasian human populations (inhabit-ing< 2,500 m), providing evidence for a common adaptivemechanism for human populations over a wide range ofaltitudes.

Through the use of PCs as covariants in our genome-wideassociation study we were able to identify a signal near theOPRM1 gene, a mu opioid receptor gene, thus nominating itas a candidate for normoxia and mild-hypoxia adaptation.Activation of the mu opioid receptor could cause respiratory

FIG. 6. rs17084501T allele distribution in Tibetans residing at different altitudes (N = 730). The Tibetan populations are classified into three groupsaccording to their residing altitudes (3,000–4,000 m, 4,000–5,000 m, and 5,000–6,000 m). The rs17084501 shows a significant increase in the frequency ofthe rs17084501T allele with increase in altitude in Tibetan populations (N = 3, Spearman rank correlation P< 0.01).

FIG. 7. Allele distribution of the EGLN1-DISC1 signal in Tibetans (N = 729) and Han Chinese populations (N = 735) residing at different altitudes. (A) TheTibetan populations are classified into three groups according to their residing altitudes (3,000–4,000 m, 4,000–5,000 m, and 5,000–6,000 m). TheEGLN1-DISC1 locus shows a significant increase in the frequency of the rs10797564A allele with an increase in altitude in Tibetan subpopulations (N = 3,Spearman rank correlation P< 0.01). (B) The residing altitude data of five Han Chinese populations (0–2,000 m) are transformed to natural logarithms.The EGLN1-DISC1 locus shows a significant increase in the frequency of the rs10797564A allele with an increase in altitude in these five Hansubpopulations (N = 5, Pearson correlation P = 0.0012).

3366


ownloaded from



stimulation/depression and thus may contribute to differ-ences in breathing patterns, including frequency and ampli-tude, of high and low altitude populations (Shook et al. 1990;Greer et al. 1995; Johnson et al. 2008). Administration of�-opioids can cause abnormal ventilatory responses to hy-percapnia and hypoxia (Wang and Teichtahl 2007), a clinicalphenomena that is common when lowlanders travel to ahighland environment. Recent studies in animals (Mosset al. 2006) and children with obstructive sleep apnea(Brown et al. 2004) suggest that hypoxia can enhance theanalgesic effects of exogenous opioids, and perioperativeopioid requirements are significantly decreased in hypoxicchildren living at a high altitude (Rabbitts et al. 2010). Amechanism involving respiration is thus suggested for thisgene. OPRM1 acts pleiotropically and other responses mayalso explain its contribution to adaptation to hypoxia. Amongits pleiotropic effects, paraventricular nucleus (PVN) ex-pressed mu-receptors regulate the heart rate during stress,and endogenous opioids released during stress may modulateadrenomedullary responses (Kiritsy-Roy et al. 1986). Theseobservations suggest that OPRM1 acts through both a respi-ratory mechanism as well as a cardiovascular response tounderlie mild-hypoxia adaptation at moderate altitude.

In 35 lowland and moderate-altitude Eurasian populations,the ancestral T allele of the OPRM1 rs17084501 locus is sig-nificantly enriched with an increase in altitude. In the repli-cation study, the rs17084501T allele frequency also increasessignificantly with altitude ascent in highland Tibetan subpop-ulations (>3,500 m, P< 0.01). Thus, the ORPM1 signal de-rived from lowland and moderate-altitude Eurasianpopulations was confirmed in the highland Tibetan popula-tions, indicating that OPRM1 might function over a widerange of altitudes. We failed, however, to observe a significantallelic differentiation between the Tibetan and Han Chinesepopulations or find a significant association of T allele fre-quency with altitude in Han Chinese subgroups living within amoderate range of altitudes. The lack of significant popula-tion differentiation or negative association of rs17084501Tallele with altitude in Han Chinese populations may be ex-plained by the close genetic relationship between the popu-lations and the differences in the selective pressures acting onthese two populations. The C allele of rs17084501 is rare inboth populations; therefore it is even more difficult to detecta differential signal between these two closely related popu-lations than in populations where this SNP is more prevalent.This putative hypothesis may be supported as HighlandTibetans have lived in the Qinghai-Tibet Plateau for thou-sands of years (Zhao et al. 2009), and considering their ex-tremely hypoxic environment and resultant strong selectionpressure, this time period may have been long enough forselection to significantly increase the frequency of thers17084501T allele in different subpopulations. In contrast,as a rare SNP it is hard to detect allelic differences amongHan Chinese subpopulations when they reside within a mod-erate altitude range where the selective advantage of thisallele is relatively very small. However, we do not have datato document such differential selection occurred, further ex-amination of this SNP in other populations such as highland

Andean as well as lowland and moderate-altitude Europeanpopulations, where it is more prevalent, and thus will havemore statistical power to reveal difference in frequencies,should enhance our understanding of the function of thisgene in hypoxia.

The identification of the highly reproducible EGLN1-DISC1locus and HIF1AN gene in adaptation to moderate altitudeprovides additional information on adaptation mechanismsfor normoxic and mild-hypoxic human populations. Duringascent to �2,500 m, oxygen pressure decreases to �70% ofthat of sea level (Beall 2007). At this moderate altitude, pe-ripheral chemoreceptors are stimulated and activity in thecarotid sinus increases although no obvious increase in ven-tilation is observed (Mason 2000). In addition, awell-characterized impairment of cognitive function due toexposure to hypobaric hypoxia can be first detected at thisaltitude (Farmer et al. 1992). Generally, however, at this alti-tude the decrease in oxygen pressure should not generate toomuch difficulty in acclimatization for daily activities. Giventhe pivotal role of HIF in hypoxia-induced pulmonary hyper-tension (HPH) (Chen et al. 2006; Fu et al. 2008), an enhance-ment of the transcriptional activities of HIF should not benecessary; instead, enhancement may compromise fitnessunder normoxic and mild hypoxic conditions. Consistentwith this hypothesis, we did not observe any genes that en-hance the transcriptional activity of HIF in our study of pop-ulations at moderate altitudes. In contrast, HIF1AN andEGLN1/EGLN2/EGLN3, representing two groups of enzymesthat degrade HIF under normoxia (asparaginyl hydroxylaseand prolyl hydroxylase) (Lendahl et al. 2009), were foundamong the candidate gene lists generated from sea-leveland moderate-altitude populations. The association ofEGLN1-DISC1 and HIF1AN were confirmed in our replicationstudy that used highland Tibetan and moderate-altitude HanChinese populations and these results further emphasize thecontributions of these genes to normoxia and mild-hypoxiaadaptation.

With the identification of EGLN1-DISC1 and HIF1AN fromthe HIF pathway having a role in normoxia and mild-hypoxiaadaptation from the present work, we continued to system-atically compare genome-wide adaptation signals, especiallythose for the HIF pathway genes, in response to quantitativevariations of oxygen pressure over a wide range of altitudes. Acareful reappraisal of six recent genome-wide scans in high-land Tibetans and Andeans (Beall et al. 2010; Bigham et al.2010; Simonson et al. 2010; Yi et al. 2010; Peng et al. 2011; Xuet al. 2011) together with the current adaptation study forlowland and moderate-altitude Eurasian populations, weidentified EGLN1 as a gene that existed on the candidategene lists for all of these populations (table 2). In additionto EGLN1, each population also had unique genes from theHIF pathway that were involved in altitude adaptation(table 2). Examples include, EPAS1 for highland Tibetans(Beall et al. 2010; Bigham et al. 2010; Simonson et al. 2010;Yi et al. 2010; Peng et al. 2011; Xu et al. 2011), NOS2A andPRKAA1 for highland Andeans (Bigham et al. 2010), andHIF1AN for lowland and moderate-altitude Eurasian humanpopulations. These results suggest that the HIF pathway tends

3367


ownloaded from



to be inhibited under normoxia and mild-hypoxia, while ac-tivation of the HIF pathway would be preferable under ex-tremely hypobaric hypoxia. Thus, in addition to the sharedadaptation signal of EGLN1, it is interesting to note that dif-ferent mechanisms from the same HIF pathway are probablyinvolved in adaptation to different altitudes with differentoxygen pressures.

In summary, this study has shown that both theoxygen-sensing pathway and cardiopulmonary function un-derpin adaptation in Eurasian lowland and moderate-altitudehuman populations. Furthermore, the shared genetic adap-tation signal of EGLN1 emphasizes the role of the HIF pathwayin hypoxia adaptation to the continuous gradient of altitude-specific habitats, while other HIF pathway genes, which areunique to each population, suggest different mechanisms ofaltering the HIF pathway were used. The present results pro-vide new insights for future high altitude adaptation studiesas well as sports medicine and the clinical treatment of pa-tients with chronic hypoxia that resulted from various formsof clinical illness.

Links to Data Sets Used in This StudyHGDP-CEPH data set: ftp://ftp.cephb.fr/hgdp_supp1. Alsogeographic coordinates of the populations were from theHGDP-CEPH data set.

Altitude data were obtained from the WorldMeteorological Organization publication “1961–1990Global Climate Normals,” which is also available from HongKong Observatory website (http://gb.weather.gov.hk).

Altitude data for the Chinese Han populations were ex-tracted from “The Annual Surface Climate Normals ofInternational Exchanging Stations of China (1971–2000)”through the China Meteorological Data Sharing ServiceSystem (http://cdc.cma.gov.cn).

Supplementary MaterialSupplementary figure S1 and tables S1–S8 are available atMolecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).

Acknowledgments

We are grateful to the anonymous reviewers and AssociateEditor for their useful criticism and suggestions for improve-ment. This work was supported by grants from the NationalBasic Research Program of China (973 Program,2007CB411600), the National Natural Science Foundation ofChina, and Bureau of Science and Technology of YunnanProvince. This research was done in State Key Laboratoryof Genetic Resources and Evolution, Kunming Institute ofZoology, Chinese Academy of Sciences and Department ofChemical Pathology, The Chinese University of Hong Kong.

ReferencesAdams WH, Strang LJ. 1975. Hemoglobin levels in persons of Tibetan

ancestry living at high altitude. Proc Soc Exp Biol Med. 149:1036–1039.

Ashburner M, Ball CA, Blake JA, et al. (20 co-authors). 2000. Gene on-tology: tool for the unification of biology. The Gene OntologyConsortium. Nat Genet. 25:25–29.

Table 3. Genotype and Allele Frequency Distributions of the Confirmed SNPs in Tibetans and Han Chinese.

Gene SNP Genotype, no. (%) Allele, no. (%) Allelic �2 P

OPRM1 rs17084501* CC TC TT C TTibetan 2 (0.27) 7 (0.96) 723 (98.77) 11 (0.75) 1453 (99.25) P> 0.05Han Chinese 1 (0.14) 13 (1.78) 715 (98.08) 15 (1.03) 1443 (98.97)

EGLN1-DISC1 rs10797564y AA AG GG A GTibetan 86 (11.76) 326 (44.60) 319 (43.64) 498 (34.06) 964 (65.94) 2.71� 10�16

Han Chinese 31 (4.14) 247 (33.02) 470 (62.83) 309 (20.66) 1187 (79.34)

HIF1AN rs2495718 AA AG GG A GTibetan 64 (8.74) 318 (43.44) 350 (47.81) 446 (30.46) 1018 (69.54) 0.0494Han Chinese 54 (7.79) 268 (38.67) 371 (53.54) 376 (27.13) 1010 (72.87)

EGLN3 rs1809461 AA AG GG A GTibetan 107 (14.68) 376 (51.58) 246 (33.74) 590 (40.47) 868 (59.53) P> 0.05Han Chinese 104 (14.23) 379 (51.85) 248 (33.93) 587 (40.15) 875 (59.85)

*The HWE P-value for rs17084501 in Tibetans is 0.008 (two-tailed, simulation of 1000). The HWE P-value for other SNPs is> 0.05.y rs10797564 represents rs6605044 (D’ = 1.000).z rs11881124 of the EGLN2 gene is monomorphic in the two populations.

Table 2. Candidate Genes of the HIF Pathway and the Original Data Sets Used in the Systematic Survey from Genetic Adaptation Studies inHuman Populations at Different Altitudes.

Populations Genes References

Highland Tibetans EGLN1, EPAS1 Beall et al. 2010; Bigham et al. 2010; Simonson et al. 2010; Yi et al.2010; Peng et al. 2011; Xu et al. 2011

Highland Andeans EGLN1, NOS2A, PRKAA1 Bigham et al. 2010

Low/Moderate Eurasians EGLN1, HIF1AN Present study

3368


ownloaded from



Beall CM. 2000. Tibetan and Andean contrasts in adaptation to

high-altitude hypoxia. Adv Exp Med Biol. 475:63–74.

Beall CM. 2007. Two routes to functional adaptation: Tibetan and

Andean high-altitude natives. Proc Natl Acad Sci U S A.

104(Suppl. 1), 8655–8660.

Beall CM, Brittenham GM, Macuaga F, Barragan M. 1990. Variation in

hemoglobin concentration among samples of high altitude natives

in the Andes and the Himalayas. Am J Hum Biol. 2:639–651.

Beall CM, Brittenham GM, Strohl KP, et al. (12 co-authors). 1998.

Hemoglobin concentration of high-altitude Tibetans and Bolivian

Aymara. Am J Phys Anthropol. 106:385–400.

Beall CM, Cavalleri GL, Deng L, et al. (29 co-authors). 2010. Natural

selection on EPAS1 (HIF2alpha) associated with low hemoglobin

concentration in Tibetan highlanders. Proc Natl Acad Sci U S A.

107:11459–11464.

Beall CM, Goldstein MC. 1987. Hemoglobin concentration of pastoral

nomads permanently resident at 4,850-5,450 meters in Tibet. Am J

Phys Anthropol. 73:433–438.

Beall CM, Reichsman AB. 1984. Hemoglobin levels in a Himalayan high

altitude population. Am J Phys Anthropol. 63:301–306.

Beall CM, Strohl KP, Blangero J, et al. (13 co-authors). 1997. Ventilation

and hypoxic ventilatory response of Tibetan and Aymara high alti-

tude natives. Am J Phys Anthropol. 104:427–447.

Bigham A, Bauchet M, Pinto D, et al. (14 co-authors). 2010. Identifying

signatures of natural selection in Tibetan and Andean populations

using dense genome scan data. PLoS Genet. 6:e1001116.

Bigham AW, Mao X, Mei R, Brutsaert T, Wilson MJ, Julian CG, Parra EJ,

Akey JM, Moore LG, Shriver MD. 2009. Identifying positive selection

candidate loci for high-altitude adaptation in Andean populations.

Hum Genomics. 4:79–90.

Bitner-Mathe BC, Peixoto AA, Klaczko LB. 1995. Morphological variation

in a natural population of Drosophila mediopunctata: altitudinal

cline, temporal changes and influence of chromosome inversions.

Heredity 75:54–61.

Boucher O. 2010. Stratospheric ozone, ultraviolet radiation and climate

change. Weather 65:105–110.

Brown KA, Laferriere A, Moss IR. 2004. Recurrent hypoxemia in young

children with obstructive sleep apnea is associated with reduced

opioid requirement for analgesia. Anesthesiology 100:806–810.

Brunsdon C, McClatchey J, Unwin DJ. 2001. Spatial variations in the

average rainfall-altitude relationship in Great Britain: an approach

using geographically weighted regression. Int J Climatol. 21:455–466.

Calbet JA, Boushel R, Radegran G, Sondergaard H, Wagner PD, Saltin B.

2003. Why is VO2 max after altitude acclimatization still reduced

despite normalization of arterial O2 content? Am J Physiol Regul

Integr Comp Physiol. 284:R304–R316.

Chen YR, Dai AG, Hu RC, Jiang YL. 2006. Differential and reciprocal

regulation between hypoxia-inducible factor-alpha subunits and

their prolyl hydroxylases in pulmonary arteries of rat with

hypoxia-induced hypertension. Acta Biochim Biophys Sin. 38:

423–434.

Collinge JE, Hoffmann AA, McKechnie SW. 2006. Altitudinal patterns for

latitudinally varying traits and polymorphic markers in Drosophila

melanogaster from eastern Australia. J Evol Biol. 19:473–482.

Davis WB, Rennard SI, Bitterman PB, Crystal RG. 1983. Pulmonary

oxygen toxicity. Early reversible changes in human alveolar struc-

tures induced by hyperoxia. N Engl J Med. 309:878–883.

Farmer EW, Lupa HT, Dunlop F, McGowan FF. 1992. Task learning under

mild hypoxia. In: Sutton JR, Coates G, Houston CS, editors. Hypoxia

and mountain medicine. Burlington (VT): Queen City Printers.

p. 1–8.

Fu D, Dai A, Hu R, Chen Y, Zhu L. 2008. Expression and role of factor

inhibiting hypoxia-inducible factor-1 in pulmonary arteries of rat

with hypoxia-induced hypertension. Acta Biochim Biophys Sin. 40:

883–892.

Ge D, Zhang K, Need AC, Martin O, Fellay J, Urban TJ, Telenti A,

Goldstein DB. 2008. WGAViewer: software for genomic annotation

of whole genome association studies. Genome Res. 18:640–643.

Greer JJ, Carter JE, al-Zubaidy Z. 1995. Opioid depression of respiration in

neonatal rats. J Physiol. 485(Pt 3):845–855.

Grocott M, Richardson A, Montgomery H, Mythen M. 2007. Caudwell

Xtreme Everest: a field study of human adaptation to hypoxia. Crit

Care 11:151.

Hannon JP, Sudman DM. 1973. Basal metabolic and cardiovascular func-

tion of women during altitude acclimatization. J Appl Physiol. 34:

471–477.

Hayes MA, Timmins AC, Yau EH, Palazzo M, Hinds CJ, Watson D. 1994.

Elevation of systemic oxygen delivery in the treatment of critically ill

patients. N Engl J Med. 330:1717–1722.

Hornbein TF, Schoene RB. 2001. High altitude: an exploration of human

adaptation. New York: Marcel Dekker.

Houston CS, Riley RL. 1947. Respiratory and circulatory changes during

acclimatization to high altitude. Am J Physiol. 149:565–588.

Huang da W, Sherman BT, Lempicki RA. 2009a. Bioinformatics enrich-

ment tools: paths toward the comprehensive functional analysis of

large gene lists. Nucleic Acids Res. 37:1–13.

Huang da W, Sherman BT, Lempicki RA. 2009b. Systematic and inte-

grative analysis of large gene lists using DAVID bioinformatics re-

sources. Nat Protocols. 4:44–57.

Hunley K, Healy M. 2011. The impact of founder effects, gene flow, and

European admixture on native American genetic diversity. Am J

Phys Anthropol. 146:530–538.

Ji L, Xu J, Wu D, Xie S, Tang NLS, Zhang Y. 2010. Association of disease-

predisposition polymorphisms of the melatonin receptors and sun-

shine duration in the global human populations. J Pineal Res. 48:

133–141.

Johnson SM, Kinney ME, Wiegel LM. 2008. Inhibitory and excitatory

effects of micro-, delta-, and kappa-opioid receptor activation on

breathing in awake turtles, Trachemys scripta. Am J Physiol Regul

Integr Comp Physiol. 295:R1599–R1612.

Kanehisa M, Goto S, Hattori M, Aoki-Kinoshita KF, Itoh M, Kawashima

S, Katayama T, Araki M, Hirakawa M. 2006. From genomics to

chemical genomics: new developments in KEGG. Nucleic Acids

Res. 34:D354–D357.

Kiritsy-Roy JA, Appel NM, Bobbitt FG, Van Loon GR. 1986. Effects of

mu-opioid receptor stimulation in the hypothalamic paraventricular

nucleus on basal and stress-induced catecholamine secretion and

cardiovascular responses. J Pharmacol Exp Ther. 239:814–822.

Lando D, Peet DJ, Gorman JJ, Whelan DA, Whitelaw ML, Bruick RK. 2002.

FIH-1 is an asparaginyl hydroxylase enzyme that regulates the tran-

scriptional activity of hypoxia-inducible factor. Genes Dev. 16:

1466–1471.

Lendahl U, Lee KL, Yang H, Poellinger L. 2009. Generating specificity and

diversity in the transcriptional response to hypoxia. Nat Rev Genet.

10:821–832.

Li JZ, Absher DM, Tang H, et al. (11 co-authors). 2008. Worldwide

human relationships inferred from genome-wide patterns of varia-

tion. Science 319:1100–1104.

3369


ownloaded from



Martin D, Windsor J. 2008. From mountain to bedside: understandingthe clinical relevance of human acclimatisation to high-altitude hyp-oxia. Postgrad Med J. 84:622–627.

Mason NP. 2000. The physiology of high altitude: an introduction to thecardio-respiratory changes occurring on ascent to altitude. CurrAnaesthesia Crit Care. 11:34–41.

Moore LG, Brodeur P, Chumbe O, D’Brot J, Hofmeister S, Monge C. 1986.Maternal hypoxic ventilatory response, ventilation, and infant birthweight at 4,300 m. J Appl Physiol. 60:1401–1406.

Moore LG, Shriver M, Bemis L, Hickler B, Wilson M, Brutsaert T, Parra E,Vargas E. 2004. Maternal adaptation to high-altitude pregnancy: anexperiment of nature–a review. Placenta 25(Suppl. A), S60–S71.

Moore LG, Zamudio S, Zhuang J, Sun S, Droma T. 2001. Oxygen trans-port in tibetan women during pregnancy at 3,658 m. Am J PhysAnthropol. 114:42–53.

Mortola JP, Frappell PB, Aguero L, Armstrong K. 2000. Birth weight andaltitude: a study in Peruvian communities. J Pediatr. 136:324–329.

Moss IR, Brown KA, Laferriere A. 2006. Recurrent hypoxia in rats duringdevelopment increases subsequent respiratory sensitivity to fenta-nyl. Anesthesiology 105:715–718.

Nei M. 1973. Analysis of gene diversity in subdivided populations. ProcNatl Acad Sci U S A. 70:3321–3323.

Oelz O, Howald H, Di Prampero PE, et al. (12 co-authors). 1986.Physiological profile of world-class high-altitude climbers. J ApplPhysiol. 60:1734–1742.

Patterson N, Price AL, Reich D. 2006. Population structure and eigen-analysis. PLoS Genet. 2:e190.

Penaloza D, Arias-Stella J. 2007. The heart and pulmonary circulation athigh altitudes: healthy highlanders and chronic mountain sickness.Circulation 115:1132–1146.

Peng Y, Yang Z, Zhang H, et al. (15 co-authors). 2011. Genetic variationsin Tibetan populations and high-altitude adaptation at theHimalayas. Mol Biol Evol. 28:1075–1081.

Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D.2006. Principal components analysis corrects for stratification ingenome-wide association studies. Nat Genet. 38:904–909.

Purcell S, Cherny SS, Sham PC. 2003. Genetic Power Calculator: design oflinkage and association genetic mapping studies of complex traits.Bioinformatics 19:149–150.

Purcell S, Neale B, Todd-Brown K, et al. (11 co-authors). 2007. PLINK: atool set for whole-genome association and population-based linkageanalyses. Am J Hum Genet. 81:559–575.

Rabbitts JA, Groenewald CB, Dietz NM, Morales C, Rasanen J. 2010.Perioperative opioid requirements are decreased in hypoxic childrenliving at altitude. Paediatr Anaesth. 20:1078–1083.

Reich D, Price AL, Patterson N. 2008. Principal component analysis ofgenetic data. Nat Genet. 40:491–492.

Reynafarje C, Hurtado A. 1971. Adaptation to high altitude. N Engl J Med.285:862.

Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM,Sirotkin K. 2001. dbSNP: the NCBI database of genetic variation.Nucleic Acids Res. 29:308–311.

Shook JE, Watkins WD, Camporesi EM. 1990. Differential roles of opioidreceptors in respiration, respiratory disease, and opiate-induced re-spiratory depression. Am Rev Respir Dis. 142:895–909.

Simonson TS, Yang Y, Huff CD, et al. (12 co-authors). 2010.Genetic evidence for high-altitude adaptation in Tibet. Science329:72–75.

Sorensen JG, Norry FM, Scannapieco AC, Loeschcke V. 2005.Altitudinal variation for stress resistance traits and thermal adapta-tion in adult Drosophila buzzatii from the New World. J Evol Biol. 18:829–837.

Soto IM, Soto EM, Carreira VP, Hurtado J, Fanara JJ, Hasson E. 2010.Geographic patterns of inversion polymorphism in the second chro-mosome of the cactophilic Drosophila buzzatii from northeasternArgentina. J Insect Sci. 10:181.

Virues-Ortega J, Buela-Casal G, Garrido E, Alcazar B. 2004.Neuropsychological functioning associated with high-altitude expo-sure. Neuropsychol Rev. 14:197–224.

Vogel JA, Harris CW. 1967. Cardiopulmonary responses of restingman during early exposure to high altitude. J Appl Physiol. 22:1124–1128.

Wang D, Teichtahl H. 2007. Opioids, sleep architecture andsleep-disordered breathing. Sleep Med Rev. 11:35–46.

West JB. 1998. High life: a history of high-altitude physiology and med-icine. Oxford: Oxford University Press.

Xu S, Li S, Yang Y, et al. (13 co-authors). 2011. A genome-wide search forsignals of high-altitude adaptation in Tibetans. Mol Biol Evol. 28:1003–1011.

Yi X, Liang Y, Huerta-Sanchez E, et al. (70 co-authors). 2010. Sequencingof 50 human exomes reveals adaptation to high altitude. Science329:75–78.

Yip R. 1987. Altitude and birth weight. J Pediatr. 111:869–876.

Zamudio S, Palmer SK, Droma T, Stamm E, Coffin C, Moore LG. 1995.Effect of altitude on uterine artery blood flow during normal preg-nancy. J Appl Physiol. 79:7–14.

Zhao M, Kong QP, Wang HW, et al. (16 co-authors). 2009.Mitochondrial genome evidence reveals successful Late Paleolithicsettlement on the Tibetan Plateau. Proc Natl Acad Sci U S A. 106:21230–21235.

3370


ownloaded from



research article genetic adaptation of the hypoxia-inducible factor

Documents