A Genetic Predisposition Score Associates with Reduced Aerobic Capacity in Response to Acute Normobaric Hypoxia in Lowlanders

Abstract

Masschelein, Evi, Joke Puype, Siacia Broos, Ruud Van Thienen, Louise Deldicque, Diether Lambrechts, Peter Hespel, and Martine Thomis. A genetic predisposition score associates with reduced aerobic capacity in response to acute normobaric hypoxia in lowlanders. High Alt Med Biol 16:34–42, 2015—Given the high inter-individual variability in the sensitivity to high altitude, we hypothesize the presence of underlying genetic factors. The aim of this study was to construct a genetic predisposition score based on previously identified high-altitude gene variants to explain the inter-individual variation in the reduced maximal O₂ uptake (ΔVo_2max) in response to acute hypoxia. Ninety-six healthy young male Belgian lowlanders were included. In both normobaric normoxia (Fio₂=20.9%) and acute normobaric hypoxia (Fio₂=10.7%–12.5%) Vo_2max was measured. Forty-one SNPs in 21 genes were genotyped. A stepwise regression analysis was applied to detect a subset of SNPs to be associated with ΔVo_2max. This subset of SNPs was included in the genetic predisposition score. A general linear model and regression analysis with age, weight, height, hypoxic protocol group, and Vo_2max in normoxia as covariates were used to test the explained variance of the genetic predisposition score. A ROC analysis was performed to discriminate between the low- and high ΔVo_2max subgroups. A stepwise regression analysis revealed a subset of SNPs [rs833070 (VEGFA), rs4253778 (PPARA), rs6735530 (EPAS1), rs4341 (ACE), rs1042713 (ADRB2), and rs1042714 (ADRB2)] to be associated with ΔVo_2max. The genetic predisposition score was found to be an independent predictive variable with a partial explained variance of 23% (p<0.0001). A ROC analysis showed significant discriminating accuracy (AUC=0.78, 95% confidence interval=0.64–0.91) between the low- and high ΔVo_2max subgroups. This six-SNP based genetic predisposition score showed a significantly predictive value for ΔVo_2max.

Introduction

Ascent to high altitude is associated with a decline in the partial pressure of oxygen due to drop of barometric pressure. As a result, aerobic working capacity in lowlanders is impaired at high altitude. Interestingly, lowlanders exhibit a striking variation in their adaptive physiological processes, which in turn contributes to a differential ability to tolerate high altitude (Martin et al., 2010). Despite many investigations, assessment of high-altitude tolerance in healthy individuals is difficult and research fails to predict high-altitude performance (Bartsch et al., 2001). Research has been done on elite high-altitude climbers, but profiling of their physiological adaptations failed to identify a signature that could justify their extent of performance impairment (Oelz et al., 1986). Until now, no single test has the potential to differentiate between ‘well’ and ‘poor’ responders to hypoxia. The high degree of differential adaptation can at least partly be explained by differences in the degree of prior acclimatization (Lyons et al., 1995; Schneider et al., 2002), rate of ascent (Schneider et al., 2002), age (Hackett et al., 1976; Roach et al., 1995), exertion (Roach et al., 2000), and body composition (Hirata et al., 1989).

The high inter-individual variability in the sensitivity to high altitude drives research to search for genetic factors underlying these differences. Twin studies estimated the presence of genotype-environment (hypoxia) interaction or the heritability of hypoxia-induced responses in lowlanders [e.g., sensitivity to high-altitude illness or hypoxic ventilatory response (MacLeod et al., 2013)]. Recently, our group found the hypoxia-induced change in maximal oxygen uptake (Vo_2max) to be highly similar in monozygotic twins (ICC=0.80, P<0.05) (Masschelein et al., 2014). Although shared environmental factors cannot be ruled out to explain the high similarity, this high intra-pair correlation is at least suggestive for a genetic component to explain part of the observed individual differences. A study in lowlanders demonstrates an association between genetic markers used to discriminate ancestry and Vo_2max decrement in response to hypoxia (Brutsaert et al., 2003). Also a frequently studied variant that confers exercise performance benefits at high altitude in lowlanders is the insertion allele of the angiotensin-converting enzyme (ACE) I/D polymorphism (Tsianos et al., 2005).

To date, research has primarily focused on identifying genetic variants underlying differential susceptibility to high-altitude illness. Variants in 17 genes, such as nitric oxide 3 (NOS3), b2-adrenergic receptor (ADRB2), angiotensinogen (AGT), aldosterone synthase (CYP11B2), and heat-shock protein genes (HSP), have shown some association with high-altitude illness (MacInnis et al., 2010; Martin et al., 2010; Stobdan et al., 2008). In contrast to studies on the adaptations of lowlanders to high altitude, studies on highlanders have revealed many new insights. Epidemiological studies made in Tibetan, Ethiopian, and Andean natives report major population-specific differences in adaptive mechanisms to protect against hypoxia in their highland habitat (Beall, 2006; Beall, 2007). Genome-wide association studies (GWAS) of Tibetans revealed several variants in genes, including Endothelial PAS domain-containing protein 1 (EPAS1), Egl nine homolog 1 (EGLN1), and peroxisome proliferator-activated receptor alpha (PPARA) (Beall et al., 2010; Hanaoka et al., 2012; Jeong et al., 2014; Peng et al., 2011; Petousi and Robbins, 2014; Simonson et al., 2010; Yi et al., 2010).

Phenotypes related to high-altitude adaptation are multifactorial in nature and many genetic variants with most probably small effects contribute to the overall phenotype. Moreover, the common weakness of high-altitude studies is that they are based on small sample sizes, which are insufficient to detect genotype-phenotype associations with small effect sizes (Bouchard, 2011). Therefore, the development of a genetic predisposition score (GPS) offers one approach to study the effect of a set of single nucleotide polymorphisms (SNPs) simultaneously, for which a larger overall effect size may be expected.

To the best of our knowledge, the GPS approach has not been used to explain the inter-individual variability in the physiological responses to hypoxia. Therefore, the aim of this study was to define a GPS based on previously identified gene variants for different high-altitude phenotypes (high-altitude illness, gene variants in high-altitude populations, genes within hypoxia-signaling pathways) to explain the inter-individual variation in two phenotypes: 1) aerobic capacity (Vo_2max) in normoxia, and 2) the change (Δ) in Vo_2max in response to acute hypoxia (hypoxia-induced ΔVo_2max). The significance of a GPS for 41 literature-based SNPs in 21 genes was tested in predicting Vo_2max in normoxia and hypoxia-induced ΔVo_2max (8 genes within “oxygen sensing”, 10 genes within “vascular homeostasis” and 3 genes within “oxidative stress” pathways). In addition, the power of the GPS to discriminate the subgroup of subjects with the smallest and largest ΔVo_2max was evaluated by receiver-operating-characteristic (ROC) curve and calculated areas under curves (AUC) analysis. We hypothesized a significant contribution for the GPS to explain, at least in part, the variability in Vo_2max decrement in response to hypoxia.

Materials and Methods

Study population

Ninety-six healthy young male Belgian lowlanders (<300 m altitude) were included in this study. Subjects included in this study were selected from different earlier studies in our laboratories (Masschelein et al., 2012; 2014; and unpublished data), but genotypic data presented here are original. All aforementioned studies (Masschelein et al., 2012; 2014; and unpublished data) were approved by the local ethics committee and were in accordance with the Declaration of Helsinki. In total, the study cohort consisted of 96 healthy male subjects. Inclusion criteria were (1) non-smoking, (2) no history of cardiovascular or respiratory disease, and (3) no residence at>1.500 m during 6 months prior to the study. All subjects gave written consent following a medical examination and after being informed of all experimental procedures.

Exercise testing

All subjects performed a maximal incremental exercise test to volitional exhaustion on a cycle ergometer (Avantronic Cyclus II, Leipzig, Germany) in both normoxia and acute normobaric hypoxia. As indicated above, subjects were selected from three different studies in which slightly different exercise protocols and hypoxic conditions were used. In study 1 (S1, unpublished data) subjects (n=61) initial workload was set at 60 W and was increased until exhaustion by 35 W every 3 min. In study 2 (n=20, S2, Masschelein et al., 2014) and study 3 (n=15, S3, Masschelein et al., 2012), initial workload was set at 50 W and was increased by 20 W per min. These two protocols are expected to give a similar maximal oxygen uptake (Vo_2max) results (Astorino et al., 2000). Furthermore, subjects were exposed to either 12.5 % Fio₂ (S1, ∼4.000 m simulated altitude), 10.7 % Fio₂ (S2, ∼5.300 m simulated altitude), or 11 % Fio₂ (S3, ∼5.000 m simulated altitude). Although all exposures could be defined as acute, timing of the maximal incremental exercise test at altitude was after 15 min rest in the hypoxic chamber (S1), after 1 h of rest at altitude (S3), or after a 5-h incremental ascent to 5300 m and 1 h rest at this altitude (S2). During all exercise tests, gas-exchange parameters were measured using an open circuit ergospirometry device (Cortex Metalyzer II, Leipzig, Germany) to determine oxygen uptake (Vo₂). A gas calibration of the O₂ and CO₂ analyzers, volume calibration of the volume transducer, and calibration of the pressure analyzer were performed before all tests according to the manufacturer. Expired gases were collected via a face mask. Vo_2max was defined as the highest average of Vo₂ measured over 30 sec during the final stage of the test.

Genotyping

Anonymously coded blood samples were drawn from each subject. We genotyped 40 SNPs using a matrix-assisted laser desorption ionization time-of-flight (MALI-TOF) mass spectrometry iPLEX Gold platform (Sequenom Inc., San Diego, CA, USA). We assessed the angiotensin-converting enzyme (ACE) rs4341 SNP, using TaqMan SNP Genotyping Assay (#4351379, Life Technologies, Gent, Belgium), as a marker for ACE Insertion(I)/Deletion(D) polymorphism because it is previously found that these two polymorphisms are in linkage disequilibrium (Glenn et al., 2009; Tanaka et al., 2003). We assessed 41 SNPs in 21 genes potentially relevant for high-altitude tolerance or aerobic capacity. We chose these polymorphisms on the basis of evidence from previously reported associations and with reported minor allele frequencies>5% (Supplementary Table S1; supplementary material is available online at www.liebertpub.com/ham).

Statistical analysis

Data were analyzed using SAS Enterprise Guide 4.3 (SAS Institute Inc, Cary, NC). To test whether the observed genotype frequencies were in the Hardy-Weinberg equilibrium a χ² test with one degree of freedom was used. To model for the three different hypoxic protocol groups, we introduced two dummy variables. Comparisons between the different study samples for age, height, weight, Vo_2max in normoxia and hypoxia-induced ΔVo_2max (hypoxia – normoxia) were made by ANOVA with Tukey's HSD post-hoc comparisons if appropriate. The association of each individual SNP was tested for Vo_2max in normoxia adjusting for age, weight and height and hypoxia-induced ΔVo_2max adjusting for age, weight, height, Vo_2max in normoxia and hypoxic protocol group (two dummy variables) (Tukey's HSD post hoc test). Gene polymorphisms with a minor allele frequency of less than 2% were analysed as carriers vs. noncarriers of the minor allele.

Genetic predisposition score

A genetic predisposition score (GPS) was calculated to study multiple contributing gene variants simultaneously. This approach reflects the polygenic nature of the complex hypoxia-response phenotype under study and lowers the chance of false negative findings. Individual SNPs were recoded as 0, 1, and 2 according to the number of protective-hypoxic-adaptor alleles for that particular SNP. The definition of a protective allele was based on the original association findings as the allele protecting against high-altitude sickness, or the allele most frequently present in high-altitude populations, etc. (for original references see Supplementary Table S1).

A stepwise regression analysis was first applied to detect a subset of SNPs to be associated with the two phenotypes of interest (Vo_2max in normoxia and hypoxia-induced ΔVo_2max) (Bouchard et al., 2011; Buxens et al., 2011; Thomaes et al., 2013). These two subsets of gene variants were included in the GPS for Vo_2max in normoxia and hypoxia-induced ΔVo_2max, by adding the 0's, 1's, or 2's for the contributing SNPs to the respective GPS scores. If a GPS is based on six gene variants, subjects can therefore vary from 0—no protective allele at all—to a maximum GPS score of 12, carrying all protective alleles. Equal weights were given for each increasing SNP because no well-defined effect sizes were known for the different SNPs and weighing of increasing alleles might only have limited effects. Differences in Vo_2max in normoxia with age, weight, and height as covariates and hypoxia-induced ΔVo_2max with age, weight, height, Vo_2max in normoxia, and hypoxic protocol group (two dummy variables) as covariates according to the number of protective alleles in the GPS score were tested in a general linear model approach (ANCOVA). In a general linear regression model the predictive value of its GPS for Vo_2max in normoxia with adjustment for age, weight, and height and hypoxia-induced ΔVo_2max with adjustment for age, weigh, height, Vo_2max in normoxia, and hypoxic protocol group (two dummy variables) were tested.

Receiver operating characteristic curve

Two subgroups were made according to the hypoxia-induced ΔVo_2max: a subgroup (25% of the group) with the smallest hypoxia-induced ΔVo_2max and a subgroup (25% of the group) with the largest hypoxia-induced ΔVo_2max. A logistic regression model was used to test the power of the GPS to discriminate between the low- versus high- ΔVo_2max subgroups and to produce the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC). Adjusted odds ratios (ODs) were calculated to estimate the effect of the GPS as the odds per increasing allele to belong to the low ΔVo_2max subgroup. All statistical tests were performed two-sided at a significance level of 5%. Data were reported as means±SE.

Results

Descriptive statistics

Descriptive statistics of the three different study cohorts are presented in Table 1. No significant differences in the three study groups for weight, height, BMI, and Vo_2max in normoxia were found. However, a small age difference was found between the studies (p<0.05). As expected, Vo_2max in hypoxia for S1 was significantly higher compared with S2 and S3 due to the higher Fio₂ used (p<0.05).

Table 1.

Subjects Characteristics of the Entire Study Cohort and the Three Different Study Groups

Variable	Total N=96	Study 1 N=61	Study 2 N=20	Study 3 N=15
Age (years)	22.5±0.4	21.0±0.3	25.4±1.0^*	24.1±1.0^*
Weight (kg)	73.6±0.9	73.0±1.2	76.0±1.8	72.3±2.5
Height (cm)	179.3±0.7	179.1±0.9	181.4±2.0	178.8±1.4
BMI (kg/m²)	22.8±0.3	22.6±0.3	23.7±0.5	22.2±0.6
Vo_2max normoxia (l/min)	4.26±0.07	4.27±0.09	4.08±0.14	4.44±0.15
Vo_2max hypoxia (l/min)	2.94±0.05	3.11±0.05	2.46±0.08^*	2.88±0.14^*

Data are presented as means±SE. ^*p<0.05, different from study group N=61

Genotyping

Supplementary Table S1 shows the 41 SNPs in 21 genes selected for genotyping. Two SNPs with a genotype efficiency <95% were excluded from the analysis. Four SNPs were flagged as being out of Hardy-Weinberg equilibrium (p<0.01) and were removed from further analyses. Supplementary Table S2 shows the single-SNP association analyses. Significant associations for Vo_2max in normoxia could be observed for 6 out of 35 investigated polymorphisms. For hypoxia-induced ΔVo_2max, significant associations were found for 7 out of 35 polymorphisms.

Genetic predisposition score

Stepwise regression analysis revealed subsets of SNPs yielding significant association with Vo_2max in normoxia and hypoxia-induced ΔVo_2max. In particular, we identified two SNPs for Vo_2max in normoxia [rs1678607 (VHL) and rs1799998 (CYP11B2)] and six SNPs for hypoxia-induced ΔVo_2max [rs833070 (VEGFA), rs4253778 (PPARA), rs6735530 (EPAS1), rs4341 (ACE), rs1042713 (ADRB2), and rs1042714 (ADRB2)] (Supplementary Table S3). For each phenotype, the significant SNPs were included in the respective GPS. The number of subjects in the lowest and highest GPS group was very small for hypoxia-induced ΔVo_2max. We therefore combined them in a≤5 protective allele group and a≥9 protective alleles group.

Figure 1 and 2 show, respectively, the ANCOVA analyses of the GPS influence on Vo_2max in normoxia with age, weight, and height as covariates and hypoxia-induced ΔVo_2max with age, weight, height, Vo_2max in normoxia, and hypoxic protocol group as covariates. ANCOVA analysis showed significant differences between GPS and Vo_2max in normoxia (p<0.0001) and hypoxia-induced ΔVo_2max (p<0.0001). For Vo_2max in normoxia, when GPS was entered in a linear regression model with all covariates, the overall model was significant explaining a total variance of 27% (p<0.0001). However, GPS alone had no significant contribution (partial correlation (r)=0.03, p=0.75, Table 2). The main determinant for Vo_2max in normoxia was weight (r=0.34, p=0.0009, Table 2). The linear regression model for hypoxia-induced ΔVo_2max showed a total explained variance of 77% (p<0.0001) with GPS as a significant independent determinant (r=0.48, p<0.0001). Also age (r=0.44, p<0.0001), Vo_2max in normoxia (r=0.81, p<0.0001) and hypoxic protocol group (r=0.42, p<0.0001) has a significant contribution to hypoxia-induced ΔVo_2max. For each additional ‘protective’ allele, the drop in hypoxia-induced ΔVo_2max was less severe, with a protective effect of 0.11 l O₂/min per additional allele.

FIG. 1.

Genetic predisposition score for Vo_2max in normoxia. Left Y-axis: number of subjects in each increasing alleles group (bar graph). Right Y-axis: Baseline Vo_2max for each increasing alleles group (square dots±SE) corrected for age, weight and height. X-axis: GPS, Number of protective alleles. Regression line for Vo_2max in normoxia.

FIG. 2.

Genetic predisposition score for the drop in Vo_2max in response to hypoxia. Left Y-axis: number of subjects in each increasing alleles group (bar graph). Right Y-axis: Change in Vo_2max in response to hypoxia (ΔVo_2max) for each increasing alleles group (square dots±SE) corrected for age, weight, height, Vo_2max in normoxia and hypoxic protocol group. X-axis: GPS, Number of protective alleles. Regression line for hypoxia-induced ΔVo_2max.

Table 2.

Linear Regression Model of VO₂max in Normoxia and ΔVO₂max in Response to Hypoxia

	r	b coefficient	P value
V o _2max normoxia
Intercept		−0.98	0.57
GPS	0.03	0.02	0.74
Age	0.11	−0.02	0.30
Weight	0.34	0.03	0.0009
Height	0.19	0.02	0.06
R²	0.28		<0.0001
Hypoxia-induced ΔVo_2max
Intercept		−0.07	0.92
GPS	0.48	0.11	<0.0001
Age	0.44	−0.04	<0.0001
Weight	0.05	−0.002	0.64
Height	0.19	0.009	0.08
Vo_2max normoxia	0.81	−0.59	<0.0001
d1	0.42	−0.32	<0.0001
d2	0.16	−0.12	0.13
R²	0.77		<0.0001

d, dummy variable; GPS, genetic predisposition score; r, partial correlation.

Receiver operation characteristic analysis

Next, we evaluated the ability of the GPS to discriminate between the low- versus high- ΔVo_2max subgroups by calculating the AUC value of the ROC curve. GPS contributed significantly to the distinction between both groups (AUC=0.78, 95% confidence interval [CI]=0.64–0.91; OR=2.55, CI=1.39–4.69; Fig. 3). We also tested whether Vo_2max in normoxia would have an additive effect on the discriminating power of GPS. When Vo_2max in normoxia was added to the model together with GPS, the predictive value increased (AUC=0.98, CI=95–1.00; OR=4.01, CI=1.25–12.92; Fig. 3). The observed threshold value with highest sensitivity and specificity was found for GPS-score 8 (sensitivity=0.88, specificity=0.58). Logistic regression showed that subjects with 8 or more protective alleles had a 9.8 times increased chance of belonging to the low ΔVo_2max subgroup (OR=9.8, 95% CI=2.28–42.06, p<0.0001) when compared to those subjects with a GPS of 7 or less.

FIG. 3.

Two different models to predict low- versus high- ΔVo_2max in response to hypoxia. Model GPS (AUC: 0.78; 95% CI: 0.64–0.91). Model GPS and Vo_2max in normoxia (AUC: 0.98; 95% CI: 0.95–1.00).

Discussion

Aerobic working capacity in lowlanders is impaired at high altitude. However, there is a considerable individual variation found in high-altitude sensitivity. This study aimed to contribute to the understanding of the genetic etiology of high-altitude sensitivity. The main goal of the study was to quantify the degree of explained variance of a genetic predisposition score for the decrease in maximal aerobic capacity in response to hypoxia (hypoxia-induced ΔVo_2max). Furthermore, this was the first attempt to identify a genotype predictive model that might allow to discriminating the subgroups with the lowest versus highest ΔVo_2max. We found a combination of six SNPs to explain 23% of the variability in hypoxia-induced ΔVo_2max and to discriminate significantly between the low- versus high- ΔVo_2max subgroups.

Most studies report data on just one polymorphism and explained variance is relatively small. In this regard, the combined influence of several gene variants is likely to explain a larger amount of individual variability in the phenotype of interest. Recent research showed the possibility of polygenic profiles for athletic performance, as proposed by studies of Ruiz et al. (2009), Santiago et al. (2010), and Williams and Folland (2008), or for responses to endurance training (Bouchard et al., 2011) or cardiorespiratory rehabilitation (Thomaes et al., 2013). To the best of our knowledge, apart from some genetic–epidemiological observations regarding altitude illness and genetic selection in high-altitude populations, data from well-controlled hypoxia intervention studies to prove genotype-dependence of the diminished aerobic capacity in response to hypoxia are lacking. However, there is one study in lowlanders demonstrating an association between genetic markers used to discriminate ancestry and Vo_2max decrement in response to hypoxia (Brutsaert et al., 2003). Our study is the first study to search for combinations of gene variants, based on previously identified gene variants for different high-altitude phenotypes, to explain variability in the decrease in aerobic capacity in response to hypoxia. We based the set of candidate-gene variants on earlier identified gene variants for high-altitude illness, GWAS-based identifications of gene selection in high-altitude populations, known-oxygen sensing mechanisms, and endurance exercise performance. A subset of SNPs was identified by stepwise regression analysis, a procedure previously been applied by other groups (Bouchard et al., 2011; Buxens et al., 2011; Thomaes et al., 2013).

In calculating a GPS based on this subset of SNPs, every polymorphism was considered to have an equal influence on the phenotype, which is a simplified approach compared to an effect-size weighting approach (Belsky et al., 2013; Janssens et al., 2007). Our main finding was that within the selected set of 35 gene variants, based on a stepwise multiple regression analysis, we were able to identify six polymorphisms [rs6735530 (EPAS1), rs4253778 (PPARA), rs833070 (VEGFA), rs4341 (ACE), rs1042713 (ADRB2), and rs1042714 (ADRB2)], to be associated with the decrease in maximal aerobic capacity in response to hypoxia.

Polymorphisms in the EPAS1 and PPARA gene have recently been selected as candidate genes in genome-wide association studies involved in the evolutionary adaptation in the high-altitude Tibetan population (Beall et al., 2010; Hanaoka et al., 2012; Jeong et al., 2014; Peng et al., 2011; Petousi and Robbins, 2014; Simonson et al., 2010; Yi et al., 2010). Both genes are implicated in the hypoxia-inducible factor (HIF) oxygen signaling pathway. EPAS1, which encodes HIF-2α, is an element of the transcription factor HIF2 that seems to be the primary mediator of erythropoietin expression in kidneys in response to hypoxia (Kapitsinou et al., 2010). A mutation in this gene was previously associated with polycytosis (Percy et al., 2008) and, interestingly, several gene variants in this gene were found to be associated with lower hemoglobin levels amongst Tibetans adapted to living at high altitude (Beall et al., 2010). This suggests that Tibetans have evolved a blunted erythropoietic response to hypoxia. Simultaneously, variants of PPARA were also associated with lower hemoglobin levels in Tibetans (Simonson et al., 2010). PPARA encodes for a major transcriptional regulator of lipid metabolism, yet recently has been considered to be one of the candidate genes for high-altitude adaptation in Tibetans (Simonson et al., 2010). PPARA has been shown to interact with the HIF-pathway [e.g., a HIF-dependent downregulation of PPAR expression was found during hypoxia in mice (Narravula and Colgan, 2001)]. In addition, the antidiabetic agent tesaglitazar, a PPARA agonist, decreased hemoglobin levels in patients with type 2 diabetes (Wilding et al., 2007). Besides the involvement of these genes in regulating erythropoiesis, they might also play a role in regulating energy metabolism (Ge et al., 2012).

The EPAS1 haplotype that was previously associated with decreased hemoglobin concentration, was also highly associated with increased lactate concentration and the putatively advantageous PPARA haplotype was correlated with decreased activity of fatty acid oxidation (Ge et al., 2012). The current study for the first time demonstrates the importance of EPAS1 and PPARA for hypoxia-induced drop of maximal aerobic capacity in a lowland population during acute exposure to hypoxia.

VEGFA is a key factor involved in angiogenesis and vascular permeability, and is upregulated in hypoxia (D'Hulst et al., 2013; Levy et al., 1995). Based on the known function of VEGFA in hypoxic conditions, several studies have assessed the function of VEGFA in the pathogenesis of high-altitude illness (Buroker et al., 2012; Ding et al., 2011; 2012; Hanaoka et al., 2009). The haplotype (rs1413711, rs833070, and rs3025000) in the VEGFA gene has previously been shown to be associated with acute mountain sickness (AMS) (Ding et al., 2012). Here we report for the first time a significant contribution of a VEGFA polymorphism (rs833070) to the inter-individual variation in diminished maximal aerobic capacity in response to hypoxia.

The β2-adrenergic receptor (ADRB2) is a G-protein coupled cell surface epinephrine receptor that acts as the primary epinephrine receptor in the lungs. At high altitude, the epinephrine response is mainly mediated through the activation of β2-adrenergic receptors facilitating oxygen delivery (Mazzeo and Reeves, 2003). A recent study has indicated an association between polymorphisms in the ADRB2 gene (rs1042713 and rs1042714) and high altitude pulmonary edema in Indian highland population (Stobdan et al., 2010). Furthermore, there is suggestive evidence that the ADRB2 gene may be associated with endurance performance (Wolfarth et al., 2007; Sarpeshkar and Bentley, 2010). We demonstrated a link between the polymorphisms rs1042713 and rs104271 and the decreased aerobic capacity in response to acute hypoxia in lowlanders. Many studies have investigated the insertion (I)/deletion (D) polymorphism of the ACE gene (ACE I/D polymorphism) in relation to lowlanders' adaptation to high altitude. Studies showed the I allele to be associated with aspects of endurance performance, with higher observed frequencies in elite distance runners (Meyerson et al., 1999) and mountaineers (Montgomery et al. 1998; Tsianos et al., 2005). The I allele has been reported to be advantageous in maintaining higher arterial oxygen (Woods et al., 2002). We demonstrated an association between the ACE I/D polymorphism and hypoxia-induced ΔVo_2max.

Based on the aforementioned six polymorphisms, a GPS model was created. The GPS model (Fig. 2) was significantly associated and a higher number of ‘increasing/protective alleles’ resulted in a better hypoxic tolerance (i.e., smaller decrease in maximal aerobic capacity in response to hypoxia). Furthermore, GPS was found to be an independent predictive variable with a partial explained variance (R²) of 23% (p<0.0001, Table 2). For each additional ‘protective’ allele, the drop in hypoxia-induced ΔVo_2max was less severe, with a protective effect of 0.11 l O₂/min per additional allele. Moreover, the GPS discriminated between low- versus high- ΔVO₂max subgroups (AUC=0.78, CI=0.64–0.91 and OR=2.55, CI=1.39–4.69, Fig. 3).

Additionally to the primary purpose of this study, we also examined the influence of our set of 35 genetic variants on Vo_2max in normoxia. A number of studies have reported heritabilities ranging from 0.40 to 0.70 (Bouchard et al., 1998; Fagard et al., 1991; Sundet et al., 1994), suggesting genetic influence. In the present study, selection of SNPs was largely based on the high-altitude phenotype. Therefore, we expected low genetic influence on Vo_2max in normoxia. Indeed, stepwise regression analysis identified two polymorphisms [rs1678607 (VHL) and rs1799998 (CYP11B2)]. However, GPS did not significantly contribute to Vo_2max in normoxia (R²=0.001, p=0.74). The variability in Vo_2max was partially explained by the covariate weight (R²=0.11, p=0.0009).

Our findings are exploratory in nature, given the small available sample and the use of three slightly different hypoxia exposure protocols. Also the pre-set selection or ‘snapshot’ of selected gene variants among the various polymorphisms can potentially affect altitude sensitivity. We are aware that the selection of gene variants may have left out other SNPs that might be associated with aerobic capacity in hypoxia. Furthermore, we assumed allelic effects to be co-dominant, as well as each SNP to have an equal additional effect. Our model does not take into account potential complex interactions between gene variants. The significance of a GPS based on previously identified gene variants for different high-altitude phenotypes (e.g. high-altitude illness) was tested in predicting the diminished maximal aerobic capacity in response to hypoxia. Although there is no formal evidence of pleiotropic gene actions related to high-altitude illness and reduced exercise capacity at altitude, common hypoxia-induced physiological pathways might be active in both phenotypes.

We observed a 11% larger drop in ΔVo_2max in subjects affected by AMS versus AMS-resistant subjects during acute exposure to hypoxia (Masschelein et al., 2014), suggesting similar physiological pathways between maximal aerobic capacity and altitude illness may be involved. Although this six-SNP based genetic predisposition score showed significant exploratory/predictive power, additional gene variants (and their interaction) as well as environmental factors conceivably may contribute to explaining inter-individual variability in exercise intolerance at high altitude. An individual application of this ‘screening set’ of SNPs is certainly not yet in place.

Conclusion

In conclusion, we found that a genetic predisposition score, composed of six polymorphisms, was significantly associated with a diminished maximal aerobic capacity in response to acute hypoxia. The GPS explained 23% of the overall variance in hypoxia-induced ΔVo_2max and also discriminated the low- versus high- ΔVo_2max subgroups. These findings improve our understanding of the genetic predisposition to high altitude and provide additional evidence for genetic factors to be implicated in high-altitude adaptation. However, the genetic information currently available remains insufficient to explain the overall variance observed. Results must be evaluated in a prospective study to see the predictive utility of the GPS score and in cohorts with larger sample sizes.

Footnotes

Acknowledgments

The authors thank Monique Ramaekers for skillful technical assistance during the experiments. The participation of the subjects is also gratefully acknowledged.

Author Disclosure Statement

The authors do not have any financial conflicts of interest.

References

Aggarwal

, Negi

, Jha

, Singh

, Stobdan

, Pasha

, Ghosh

, Agrawal

, Prasher

, and Mukerji

. (2010). EGLN1 involvement in high-altitude adaptation revealed through genetic analysis of extreme constitution types defined in Ayurveda. Proc Natl Acad Sci USA, 107:18961–18966.

Astorino

, Robergs

, Ghiasvand

, Marks

, and Burns

. (2000). Incidence of the oxygen plateau at VO2max during exercise testing to volitional fatigue. JEPonline, 3:1–12.

Bartsch

, Grunig

, Hohenhaus

, and Dehnert

. (2001). Assessment of high altitude tolerance in healthy individuals. High Alt Med Biol, 2:287–296.

Beall

. (2006). Andean, Tibetan, and Ethiopian patterns of adaptation to high-altitude hypoxia. Integr Comp Biol, 46:18–24.

Beall

. (2007). Two routes to functional adaptation: Tibetan and Andean high-altitude natives. Proc Natl Acad Sci USA, 104:8655–8660.

Beall

, Cavalleri

, Deng

, Elston

, Gao

, Knight

, Li

, Liang

, McCormack

, Montgomery

, Pan

, Robbins

, Shianna

, Tam

, Tsering

, Veeramah

, Wang

, Wangdui

, Weale

, Xu

, Yang

, Zaman

., Zeng

, Zhang

, Zhaxi

, and Zheng

. (2010). Natural selection on EPAS1 (HIF2alpha) associated with low hemoglobin concentration in Tibetan highlanders. Proc Natl Acad Sci USA, 107:11459–11464.

Belsky

, Moffitt

, Sugden

, Williams

, Houts

, McCarthy

, and Caspi

. (2013). Development and evaluation of a genetic risk score for obesity. Biodemography Soc Biol, 59:85–100.

Bouchard

. (2011). Overcoming barriers to progress in exercise genomics. Exerc Sport Sci Rev, 39:212–217.

Bouchard

, Daw

, Rice

, Perusse

, Gagnon

, Province

, Leon

, Rao

, Skinner

, and Wilmore

. (1998). Familial resemblance for VO2max in the sedentary state: The HERITAGE family study. Med Sci Sports Exerc, 30:252–258.

10.

Bouchard

, Sarzynski

, Rice

, Kraus

, Church

, Sung

, Rao

, and Rankinen

. (2011). Genomic predictors of the maximal O(2) uptake response to standardized exercise training programs. J Appl Physiol, 110:1160–1170.

11.

Buroker

, Ning

, Zhou

, Li

, Cen

, Wu

, Zhu

, Scott

, and Chen

. (2012). AKT3, ANGPTL4, eNOS3, and VEGFA associations with high altitude sickness in Han and Tibetan Chinese at the Qinghai-Tibetan Plateau. Int J Hematol, 96:200–213.

12.

Buxens

, Ruiz

, Arteta

, Artieda

, Santiago

, Gonzalez-Freire

, Martinez

, Tejedor

, Lao

, Gomez-Gallego

, and Lucia

. (2011). Can we predict top-level sports performance in power vs endurance events? A genetic approach. Scand J Med Sci Sports, 21:570–579.

13.

Brutsaert

, Parra

, Shriver

, Gamboa

, Palacios

, Rivera

, Rodriguez

, and León-Velarde

. (2003). Spanish genetic admixture is associated with larger V(O2) max decrement from sea level to 4338 m in Peruvian Quechua. J Appl Physiol, 95:519–528.

14.

D'Hulst

, Jamart

, Van Thienen

, Hespel

, Francaux

, and Deldicque

. (2013). Effect of acute environmental hypoxia on protein metabolism in human skeletal muscle. Acta Physiol (Oxf), 208:251–264.

15.

Ding

, Liu

, Hua

, Ding

, Du

, Zhang

, Li

, and Zhang

. (2011). Polymorphisms of hypoxia-related genes in subjects susceptible to acute mountain sickness. Respiration, 81:236–241.

16.

Ding

, Liu

, Hua

, Ding

, Du

, Zhang

, Li

, and Zhang

. (2012). Associations between vascular endothelial growth factor gene polymorphisms and susceptibility to acute mountain sickness. J Int Med Res, 40:2135–2144.

17.

Fagard

, Bielen

, and Amery

. (1991). Heritability of aerobic power and anaerobic energy generation during exercise. J Appl Physiol, 70:357–362.

18.

, Simonson

, Cooksey

, Tanna

, Qin

, Huff

, Witherspoon

, Xing

, Zhengzhong

, Prchal

, Jorde

, and McClain

. (2012). Metabolic insight into mechanisms of high-altitude adaptation in Tibetans. Mol Genet Metab, 106:244–247.

19.

Glenn

, Du

, Eisenmann

, and Rothschild

. (2009). An alternative method for genotyping of the ACE I/D polymorphism. Mol Biol Rep, 36:1305–1310.

20.

Hackett

, Rennie

, and Levine

. (1976). The incidence, importance, and prophylaxis of acute mountain sickness. Lancet, 2:1149–1155.

21.

Hanaoka

, Droma

, Basnyat

, Ito

, Kobayashi

, Katsuyama

, Kubo

, and Ota

. (2012). Genetic variants in EPAS1 contribute to adaptation to high-altitude hypoxia in Sherpas. PLoS One, 7:e50566.

22.

Hanaoka

, Droma

, Ota

, Ito

, Katsuyama

, and Kubo

. (2009). Polymorphisms of human vascular endothelial growth factor gene in high-altitude pulmonary oedema susceptible subjects. Respirology, 14:46–52.

23.

Hirata

, Masuyama

, and Saito

. (1989). Obesity as risk factor for acute mountain sickness. Lancet, 2:1040–1041.

24.

Janssens

, Moonesinghe

, Yang

, Steyerberg

, van Duijn

, and Khoury

. (2007). The impact of genotype frequencies on the clinical validity of genomic profiling for predicting common chronic diseases. Genet Med, 9:528–535.

25.

Jeong

, Alkorta-Aranburu

, Basnyat

, Neupane

, Witonsky

, Pritchard

, Beall

, and Di Rienzo

. (2014). Admixture facilitates genetic adaptations to high altitude in Tibet. Nat Commun, 5:3281.

26.

Kapitsinou

, Liu

, Unger

, Rha

, Davidoff

, Keith

, Epstein

, Moores

, Erickson-Miller

, and Haase

. (2010). Hepatic HIF-2 regulates erythropoietic responses to hypoxia in renal anemia. Blood, 116:3039–3048.

27.

Levy

, Levy

, Wegner

, and Goldberg

. (1995). Transcriptional regulation of the rat vascular endothelial growth factor gene by hypoxia. J Biol Chem, 270:13333–13340.

28.

Lyons

, Muza

, Rock

, and Cymerman

. (1995). The effect of altitude pre-acclimatization on acute mountain sickness during reexposure. Aviat Space Environ Med, 66:957–962.

29.

MacInnis

, Koehle

, and Rupert

. (2010). Evidence for a genetic basis for altitude illness: 2010 update. High Alt Med Biol, 11:349–368.

30.

MacLeod

, MacInnis

, Manokhina

, and Rupert

. (2013). Twin studies in altitude and hypoxia research. Aviat Space Environ Med, 84:613–619.

31.

Martin

, Levett

, Grocott

, and Montgomery

. (2010). Variation in human performance in the hypoxic mountain environment. Exp Physiol, 95:463–470.

32.

Masschelein

, Van Thienen

, Thomis

, and Hespel

. (2015). High twin resemblance for sensitivity to hypoxia. Med Sci Sports Exerc, 47:74–81.

33.

Masschelein

, Van Thienen

, Wang

, Van Schepdael

, Thomis

, and Hespel

. (2012). Dietary nitrate improves muscle but not cerebral oxygenation status during exercise in hypoxia. J Appl Physiol, 113:736–745.

34.

Mazzeo

, and Reeves

. (2003). Adrenergic contribution during acclimatization to high altitude: Perspectives from Pikes Peak. Exerc Sport Sci Rev, 31:13–18.

35.

Myerson

, Hemingway

, Budget

, Martin

, Humphries

, and Montgomery

. (1999). Human angiotensin I-converting enzyme gene and endurance performance. J Appl Physiol, 87:1313–1316.

36.

Montgomery

, Marshall

, Hemingway

, Myerson

, Clarkson

, Dollery

, Hayward

, Holliman

, Jubb

, World

, Thomas

, Brynes

, Saeed

, Barnard

, Bell

, Prasad

, Rayson

, Talmud

, and Humphries

. (1998). Human gene for physical performance. Nature, 393:221–222.

37.

Narravula

, and Colgan

. (2001). Hypoxia-inducible factor 1-mediated inhibition of peroxisome proliferator-activated receptor alpha expression during hypoxia. J Immunol, 166:7543–7548.

38.

Oelz

, Howald

, Di Prampero

, Hoppeler

, Claassen

, Jenni

, Buhlmann

, Ferretti

, Bruckner

, and Veicsteinas

. (1986). Physiological profile of world-class high-altitude climbers. J Appl Physiol, 60:1734–1742.

39.

Peng

, Yang

, Zhang

, Cui

, Qi

, Luo

, Tao

, Wu

, Ouzhuluobu, Basang, Ciwangsangbu, Danzengduojie, Chen

, Shi

, and Su

. (2011). Genetic variations in Tibetan populations and high-altitude adaptation at the Himalayas. Mol Biol Evol, 28:1075–1081.

40.

Percy

, Furlow

, Lucas

, Li

, Lappin

, McMullin

, and Lee

. (2008). A gain-of-function mutation in the HIF2A gene in familial erythrocytosis. N Engl J Med, 358:162–168.

41.

Petousi

, and Robbins

. (2014). Human adaptation to the hypoxia of high altitude: The Tibetan paradigm from the pregenomic to the postgenomic era. J Appl Physiol, 116:875–884.

42.

Roach

, Houston

, Honigman

, Nicholas

, Yaron

, Grissom

, Alexander

, and Hultgren

. (1995). How well do older persons tolerate moderate altitude?. West J Med, 162:32–36.

43.

Roach

, Maes

, Sandoval

, Robergs

, Icenogle

, Hinghofer-Szalkay

, Lium

, and Loeppky

. (2000). Exercise exacerbates acute mountain sickness at simulated high altitude. J Appl Physiol, 88:581–585.

44.

Ruiz

, Gomez-Gallego

, Santiago

, Gonzalez-Freire

, Verde

, Foster

, and Lucia

. (2009). Is there an optimum endurance polygenic profile?. J Physiol, 587:1527–1534.

45.

Santiago

, Ruiz

, Muniesa

, Gonzalez-Freire

, Gomez-Gallego

, and Lucia

. (2010). Does the polygenic profile determine the potential for becoming a world-class athlete? Insights from the sport of rowing. Scand J Med Sci Sports, 20:e188–e194.

46.

Sarpeshkar

, and Bentley

. (2010). Adrenergic-beta(2) receptor polymorphism and athletic performance. J Hum Genet, 55:479–485.

47.

Schneider

, Bernasch

, Weymann

, Holle

, and Bartsch

. (2002). Acute mountain sickness: Influence of susceptibility, preexposure, and ascent rate. Med Sci Sports Exerc, 34:1886–1891.

48.

Semenza

. (2009). Regulation of oxygen homeostasis by hypoxia-inducible factor 1. Physiology (Bethesda), 24:97–106.

49.

Simonson

, Yang

, Huff

, Yun

, Qin

, Witherspoon

, Bai

, Lorenzo

, Xing

, Jorde

, Prchal

, and Ge

. (2010). Genetic evidence for high-altitude adaptation in Tibet. Science, 329:72–75.

50.

Stobdan

, Karar

, and Pasha

. (2008). High altitude adaptation: Genetic perspectives. High Alt Med Biol, 9:140–147.

51.

Stobdan

, Kumar

, Mohammad

, Thinlas

, Norboo

, Iqbal

, and Pasha

. (2010). Probable role of beta2-adrenergic receptor gene haplotype in high-altitude pulmonary oedema. Respirology, 15:651–658.

52.

Sundet

, Magnus

, and Tambs

. (1994). The heritability of maximal aerobic power: A study of Norwegian twins. Scand J Med Sci Sports, 4:181–185.

53.

Tanaka

, Kamide

, Takiuchi

, Miwa

, Yoshii

, Kawano

, and Miyata

. (2003). An alternative fast and convenient genotyping method for the screening of angiotensin converting enzyme gene polymorphisms. Hypertens Res, 26:301–306.

54.

Thomaes

, Thomis

, Onkelinx

, Goetschalckx

, Fagard

, Lambrechts

, and Vanhees

. (2013). Genetic predisposition scores associate with muscular strength, size, and trainability. Med Sci Sports Exerc, 45:1451–1459.

55.

Tsianos

, Eleftheriou

, Hawe

, Woolrich

, Watt

, Peacock

, Montgomery

, and Grant

. (2005). Performance at altitude and angiotensin I-converting enzyme genotype. Eur J Appl Physiol, 93:630–633.

56.

Wilding

, Gause-Nilsson

, and Persson

. (2007). Tesaglitazar, as add-on therapy to sulphonylurea, dose-dependently improves glucose and lipid abnormalities in patients with type 2 diabetes. Diab Vasc Dis Res, 4:194–203.

57.

Williams

, and Folland

. (2008). Similarity of polygenic profiles limits the potential for elite human physical performance. J Physiol, 586:113–121.

58.

Woods

, Pollard

, Collier

, Jamshidi

, Vassiliou

, Hawe

, Humphries

, and Montgomery

. (2002). Insertion/deletion polymorphism of the angiotensin I-converting enzyme gene and arterial oxygen saturation at high altitude. Am J Respir Crit Care Med, 166:362–366.

59.

, Liang

, Huerta-Sanchez

, Jin

, Cuo

, Pool

, Xu

, Jiang

, Vinckenbosch

, Korneliussen

, Zheng

, Liu

, He

, Li

, Luo

, Nie

, Wu

, Zhao

, Cao

, Zou

, Shan

, Li

, Yang

, Asan, Ni

, Tian

, Xu

, Liu

, Jiang

, Wu

, Zhou

, Tang

, Qin

, Wang

, Feng

, Li

, Huasang, Luosang

, Wang

, Chen

, Wang

, Zheng

, Li

, Bianba

, Yang

, Wang

, Tang

, Gao

, Chen

, Luo

, Gusang

, Cao

, Zhang

, Ouyang

, Ren

, Liang

, Zheng

, Huang

, Li

, Bolund

, Kristiansen

, Li

, Zhang

, Li

, Yang

, Nielsen

, Wang

, and Wang

. (2010). Sequencing of 50 human exomes reveals adaptation to high altitude. Science, 329:75–78.

60.

Wolfarth

, Rankinen

, Mühlbauer

, Scherr

, Boulay

, Pérusse

, Rauramaa

, and Bouchard

. (2007). Association between a beta2-adrenergic receptor polymorphism and elite endurance performance. Metabolism, 56:1649–1651.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB

0.03 MB

0.02 MB