Genomics Studies
Several research designs and technologies have been used over the past two decades in the effort to identify genomic regions, genes, and sequence variants associated with exercise behavior and exercise biology traits. This section reviews advances made over time and key laboratories and investigators involved in these studies. Candidate genes, genome-wide linkage scans, genome-wide association (GWA) studies, and contributions from combinations of transcriptomics and genomics are reviewed. Efforts to identify genetic differences between elite athletes and sedentary controls are also briefly summarized.
Candidate Genes
A candidate gene is a gene that has a theoretical relationship with the behavioral, physiological, or metabolic system regulating the trait of interest. Candidate genes for exercise biology traits have been defined on the basis of advances in exercise physiology studies and animal models and on purely theoretical grounds. Transgenic (overexpression of a gene or genes), knockdown (reduced expression level), and knockout (ablation of a gene or genes) mice have generated a number of candidate genes that were subsequently investigated for potential involvement in human variation. The expression level of a gene has also been used as a candidate phenotype against which DNA variants can be tested for associations. The complete transcriptome, which is the full set of ribonucleic acid (RNA) transcripts in a given cell or tissue, has been used to define new panels of candidate genes for further genomics studies (e.g., 72, 224).
Candidate genes were very prominent in the early phase of exercise genomics. Many of the earlier studies were quite simple in design and were launched primarily because of the availability of whole blood stored in freezers. They were generally case-control or cross-sectional cohort studies with unrelated subjects. In the case of continuous traits (e.g., V\od\O2max), the association with a candidate gene or marker is tested by comparing mean trait values across genotypes or between carriers and noncarriers of a specific allele. In the case-control design, testing for a relation between a trait and a candidate gene marker is based on the comparison of allele and genotype frequencies between two informative groups of subjects, one with the phenotype of interest (e.g., elite power and strength athletes - the "cases") and the other without the phenotype of interest (the "controls").
Special Case of Athletes Versus Controls
Luigi Gedda from the Gregor Mendel Institute of Medical Genetics and Twin Research at the University of Rome was perhaps the pioneer in the effort to understand the importance of genetic differences between athletes and nonathletes (67, 68) (figure 7.7). The initial efforts were based on surveys of Italian twins, specifically on 351 pairs of twins who were engaged in competitive sport. The results of the surveys indicated familial aggregation among athletes and the potential importance of zygosity in concordance of sport selection; participation in markedly different sports occurred in 85% of DZ twins in contrast to only 6% of MZ twins (68). The familial aggregation of athletes was also noted by others, such as Grebe (75, 76) of Germany, Linc and Fleischmann (117) of Czechoslovakia, and Jokl (92, 93), who at the time was in the United States. The work of Gedda and his longtime collaborator, Paolo Parisi, was the most systematic in this regard and focused on genealogical and twin methods. The research of Parisi continues today with the Italian Registry of Twin Athletes (154, 155). A higher prevalence of MZ twins among swimmers has been one of their findings (154). In a large sample of adult British female twins, athlete status (nonathlete vs. local-, county-, or national-level athlete) had a heritability of 0.66 (50).
Luigi Gedda (1902-2000), MD, Mendel Institute of Medical Genetics and Twin Research in Rome, right, and Paolo Parisi (1940 - ), PhD, professor of biology and rector, Italian University Sport and Movement, University of Rome "Foro Italico," left. Gedda was the first to confirm, through classic pedigree studies in the mid-1950s, the common knowledge that elite athletes often cluster in families and to then attempt to quantify the role of genetic factors using the twin concordance model. He studied a large sample of MZ and DZ twin pairs and found considerably higher concordance rates in the former than in the latter with respect to various aspects of sport participation, such as kind of sport practiced and performance level. In later decades, Gedda’s longtime associate Paolo Parisi conducted epidemiological and cell genetics research, also through the aid of a special registry of Italian twin athletes, in order to explore the genetic basis of athletic performance. Of particular interest to him is the relation of physical activity to health, aging, and chronic disease. One major line of research of his institute focuses on the role of exercise-induced oxidative stress in genetic regulation processes.
The first documented attempt to identify differences in genetic markers for performance-related phenotypes dates to the late 1960s. A group of geneticists, taking advantage of the 1968 Olympic Games in Mexico, investigated common blood genetic markers in an attempt to discriminate between Olympic athletes and controls (45).
Interestingly, phenylthiocarbamide nontasters were underrepresented among the athletes. The effort to document genetic differences between elite athletes and sedentary controls was continued on the occasion of the 1976 Olympic Games in Montreal (37, 41). A slightly higher frequency of the A1 allele of the ABO blood group was observed in endurance athletes participating in the 1976 Summer Olympic Games than in reference populations (37). However, the other plasma and red blood cell markers (ABO, MNSs, Rhesus, Duffy, Kell, P, LDH, MDH, phosphoglucomutase, adenosine deaminase, adenylate kinase, esterase D, haptoglobin, transferrin, hemoglobin, glucose-6-phosphate dehydrogenase, and acid phosphatase) did not differ between athletes and controls (37, 41, 45). These early attempts were all based on polymorphisms in red blood cell antigens and enzymes. Later, skeletal muscle gene product variants were screened (17, 128) and selected enzyme markers were investigated for their putative effects on a variety of skeletal muscle and cardiorespiratory endurance indicators. Genetic variants of skeletal muscle creatine kinase (CKM) and adenylate kinase 1 (AK1) were screened for using isoelectric focusing in 295 subjects (16). A variant form of both enzymes was identified with an allele frequency of 1% (CKM) and 3.5% (AK1). There was no difference in V\od\O2max between the carriers of the variant alleles and matched controls homozygous for the nonvariant allele (16). None of these studies yielded reliably significant genomic predictors of performance or fitness.
Genomic differences between athletes and untrained controls began to be investigated in a more systematic fashion with the launch of the GENATHLETE study in the laboratory of Claude Bouchard at Laval University in 1993. The main collaborators on GENATHLETE included Louis Perusse, Marcel Boulay, and the late Jean-Aime Simoneau (all from Laval University), Rainer Rauramaa (Kuopio, Finland), and particularly Bernd Wolfahrt (from Germany). The study focused on DNA sequence differences between a panel that has now attained more than 300 elite endurance athletes with a V\od\O2max no lower than 75 ml·kg-1·min-1 and more than 300 sedentary controls with a V\od\O2max no higher than 50 ml·kg-1·min-1. The participants are all white males from Canada, Germany, Finland, and the United States. Single nucleotide polymorphisms (SNPs) in several genes have been investigated to date, but none have provided strong evidence for differences in allele and genotype frequencies between athletes and controls (16, 178, 184, 185, 245-247). However, a common variant and haplotype in the hypoxia inducible factor 1, alpha subunit (HIF1A) gene were found to be more prevalent in athletes compared with controls in a 2010 report based on the GENATHLETE cohort (54), an observation that was defined in an accompanying editorial as an important milestone (194).
Another effort to delineate the differences between athletes and nonathletes is under way in Spain under the leadership of Jonatan Ruiz and Alejandro Lucia. They have used samples of Spanish Caucasian endurance athletes (about 100), power athletes (about 50), and nonathlete controls. A number of differences at candidate genes between athletes and controls, as well as between endurance and power athletes, have been reported (34, 71, 190). Taking advantage of the concept of a total genotype score developed by Williams and Folland (238), Ruiz and colleagues in 2009 attempted to predict world-class endurance athlete status based on the frequency of variant alleles at 7 candidate genes genotyped in 43 Spanish endurance athletes and 123 controls (191).
Yannis Pitsiladis from the faculty of biomedical and life sciences at the University of Glasgow has taken advantage of striking performance differences at the elite level in athletes of various ethnic backgrounds to investigate the potential role of variation in nuclear and mitochondrial DNA (figure 7.8). For instance, distance runners from Kenya and Ethiopia and sprinters from Jamaica won 25% of all track and field medals at the 2008 Olympic Games in Beijing (166). The effort to impute ethnic differences in sport performance to genetic differences is clearly challenging, particularly because only about 10% of human genetic variation can be found between the major ethnic groups. Pitsiladis’ laboratory has explored differences in sequence variants in mitochondrial DNA and in Y chromosome markers between Ethiopian athletes, Ethiopian controls, Kenyan athletes, and Kenyan controls. Mitochondrial DNA haplogroups found in Kenya were different from those found in Ethiopia (197). Differences in haplogroup distributions were observed between athletes and controls in Kenya but not in Ethiopia. The full significance of this observation needs to be established. Analyses of Y chromosome haplogroups in endurance athletes from Ethiopia revealed differences between them and samples from the more general Ethiopian population (145). Population stratification was ruled out as the explanation for such differences. One important conclusion reached by Pitsiladis and colleagues is that the clusters of Kenyan and Ethiopian elite runners are distinct from each other from a genetic point of view (166).
Yannis P. Pitsiladis (1967 - ), MMedSci, PhD, FACSM, integrative and systems biology research theme, faculty of biomedical and life sciences, University of Glasgow, Scotland. Pitsiladis is a reader in exercise physiology at the Institute of Biomedical and Life Sciences at the University of Glasgow and founding member (and previous director) of the International Centre for East African Running Science, set up to investigate the physiological, genetic, psychosocial, and economic determinants of the phenomenal success of East African distance runners in international athletics. Recent projects include the study of West African sprinters (including elite sprinters from Jamaica and the United States) and the study of world-class swimmers (e.g., why there are very few black swimmers). He is a visiting professor in medical physiology at Moi University (Eldoret, Kenya), Addis Ababa University (Addis Ababa, Ethiopia), and University of Technology (Kingston, Jamaica). Pitsiladis has a particular research interest in mitochondrial DNA and Y chromosome markers and their potential role in ethnic differences in performance.
Photo courtesy of Yannis P. Pitsiladis.
Two genes have received considerable attention in the exercise genomics literature: angiotensin I-converting enzyme (ACE) and actinin alpha 3 (ACTN3). Both have been strong favorites of exercise scientists over the past decade or so.
ACE Gene
Hugh Montgomery (1962 - ), MB, BS, BSc, FRCP, MD, FRGS, director of University College London Institute for Human Health and Performance; consultant intensivist at Whittington Hospital; and professor of intensive care medicine at University College London. Montgomery obtained his BSc degree in neuropharmacology and in circulatory and respiratory physiology in 1984 before graduating with a medical degree in 1987. He has since trained and accredited in general internal medicine, cardiology, and intensive care medicine. He is now professor of intensive care medicine at University College London, where he also directs the Institute for Human Health and Performance. In the early 1990s, he began using gene - environment interaction as a means to explore human physiology. Working with army recruits, he first identified a role for the angiotensin I-converting enzyme (ACE) I/D polymorphism in the regulation of human cardiac growth. This work was extended to broader measures of performance, including high-altitude mountaineering aptitude. Mechanistically, a role in the regulation of metabolic efficiency was suggested, leading to phase IV studies of ACE inhibition in the treatment of cancer cachexia. Studies of downstream gene variants (e.g., those of the bradykinin B2 receptor gene) have offered further mechanistic insight. Montgomery has a special interest in high-altitude performance: He was science lead for the Caudwell Xtreme Everest Expedition in 2007 and has identified genes under selection pressure in high-altitude populations.
The potential role of ACE sequence differences in human physical performance was first investigated in the 1990s by Hugh Montgomery from the University College London (figure 7.9). He and his coworkers reported in 1997 that 10 wk of physical training in British army recruits induced greater increases in left ventricular mass and septal and posterior wall thickness in the ACE D/D homozygotes than in the I allele carriers (141). A few years later, the same group of investigators confirmed the finding by reporting that the training-induced increase in left ventricular mass in another cohort of army recruits was 2.7 times greater in the D/D genotype compared with the I/I homozygotes (147). The I allele was associated with higher muscular endurance gains after 10 wk of physical training in British army recruits (142).
The story began in 1990 when it was reported that a 287 bp I/D polymorphism in intron 16 of the ACE gene was associated with plasma ACE activity (183). ACE activity was highest in D/D homozygotes and lowest in I/I homozygotes. More than 50 reports have subsequently dealt with the potential role of the ACE I/D genotype in some aspects of fitness or performance. Overall, the results are heterogeneous and often contradictory. For instance, in postmenopausal women, I/I homozygotes had a higher V\od\O2max than D/D homozygotes (79), whereas the opposite was observed in Chinese males (251). In the HERITAGE Family Study, no associations were found between the ACE I/D polymorphism and maximal and submaximal exercise V\od\O2 and power output phenotypes in healthy, sedentary blacks and whites (41).
This literature was recently reviewed by Montgomery and colleagues (206). In the aggregate, it appears that the ACE I/D polymorphism may contribute to human variation in skeletal and cardiac muscle growth and functional properties as well as in adaptation to hypoxia. However, its contribution to variation in human physical performance remains a matter of debate.
ACTN3 Gene
Kathryn North and colleagues from the Children’s Hospital at Westmead in Australia discovered a common variant in the gene for actinin-3, a structural actin binding protein found in skeletal muscle fast-twitch fibers (figure 7.10). The polymorphism (R577X) replaces an arginine residue by a premature stop codon at position 577 and results in ACTN3 deficiency. About 16% of the world population is homozygous for the XX null genotype; the frequency of the null allele is highest in Caucasians and Asians and lowest in Africans (8). The first report on ACTN3 genotype and performance was from North’s laboratory in 2003. It indicated that the frequency of the stop codon mutation was lower in sprinters than in controls and endurance athletes (248). Several studies have concluded that the stop codon variant (X577X) was lower in sprint and strength athletes than in nonathletes (55, 57, 153, 189, 192, 248), but other reports have been negative (149, 198, 249). Interestingly, a number of studies based on direct measures of muscle strength and power have also been published and the results are contradictory; that is, the X allele is negatively associated with these muscle phenotypes in some studies whereas the converse is found in others (39, 51, 66, 132, 133, 146, 150, 228). North has reviewed the biology of the ACTN3 polymorphism in recent publications (8, 120). In the aggregate, the data suggest that there is probably a role of the X allele in skeletal muscle function and muscle performance, but the data remain inconclusive at present. More studies with appropriate power and designs, as well as more extensive functional studies, are warranted.
Kathryn North (1960 - ), MD, FRACP, Director, Murdoch Childrens Research Institute, and David Danks Professor of Child Health Research at University of Melbourne, Australia. In the process of studying genes implicated in muscle disease, North and colleagues at the Children’s Hospital at Westmead discovered a common variant in the actinin alpha 3 (ACTN3) gene. ACTN3 encodes a structural protein found in fast-twitch skeletal muscle fibers. The ACTN3 gene variant results in complete deficiency of actinin-3 in almost 20% of the general population. North showed that ACTN3 deficiency is rare in sprint athletes, suggesting that this protein plays a crucial role in the function of fast-twitch muscle fibers. North and her team developed a strain of mice engineered to be completely deficient in ACTN3 and showed that the muscle of these knockout mice displays an increase in oxidative metabolism. This metabolic shift could explain why ACTN3 deficiency is detrimental to sprint activities that require fast or glycolytic metabolism. Her analysis of deoxyribonucleic acid (DNA) samples from individuals from around the world suggests that the ACTN3 deficiency provided some benefit to the ancestors of modern Europeans and Asians after their migration out of Africa, resulting in its increase in frequency due to natural selection. The benefit provided by ACTN3 deficiency may have allowed them to adapt to the more hostile environments of Eurasia. North’s group plans to use mouse and human studies to determine the effect of ACTN3 deficiency on muscle aging, response to exercise, and the progression of inherited muscle diseases such as the muscular dystrophies.
Other Candidate Genes
A series of studies has dealt with the contributions of other candidate genes to human variation in muscular strength and power. These genes were identified because of their biological relevance to muscle contractile and sarcomeric proteins, myogenesis and muscle regeneration, muscle mass atrophy and sarcopenia, hormonal regulation of muscle mass, energy transfer to sustain muscle contraction, and other properties. The evidence from candidate gene studies of muscle strength and power was recently reviewed by Thomis (220).
An example of such genes is the myostatin (MSTN) gene. Mutations in MSTN leading to inactive or defective gene product have been shown to double or even triple skeletal muscle mass in mice, cows, and sheep. Functional MSTN mutations are very rare in humans. However, a boy appeared extraordinarily muscular at birth and at 4.5 yr of age was described as muscular and very strong (196). It turned out that he was a homozygote for a G/A mutation located five nucleotides downstream of exon 1, which abolishes a normal splice donor site and activates a cryptic splice site further downstream in intron 1, resulting in a truncated, inactive myostatin peptide. A number of polymorphisms in MSTN have been investigated for their associations with muscle hypertrophy or muscle strength and power (40, 60, 107, 199, 231), but the overall findings have been inconsistent.
Another class of candidate genes of great interest for muscular strength and power is that of the insulin-like growth factors and related binding proteins. These molecules play important roles in muscle growth and repair and in muscle response to exercise. Polymorphisms in insulin-like growth factors 1 (IGF1) and 2 (IGF2) have been rather consistently associated with muscle strength and its response to training, as well as other relevant endophenotypes (52, 81, 108, 193, 195).
The laboratories of James Hagberg and Stephen Roth in the department of kinesiology at the University of Maryland in the United States have pioneered a number of candidate gene research avenues since the late 1990s (figure 7.11). Their efforts focused on sequence variants in ACE; nitric oxide synthase 3 (NOS3); angiotensinogen (AGT); ACTN3; angiotensin II receptor, type 1 (AGTR1); tumor necrosis factor (TNF); nuclear factor of kappa light polypeptide gene enhancer in B-cells 1 (NFKB1); activin A receptor, type IIB (ACVR2B); adenosine monophosphate deaminase 1 (AMPD1); vascular endothelial growth factor A (VEGFA); vitamin D receptor (VDR); and MSTN, among others (see 59, 78, 79, 81, 134, 156, 167). Hagberg has a strong interest in exercise blood pressure and other hemodynamic phenotypes and through most of his career has consistently focused on the role of exercise in health-related outcomes and therapeutic applications, with an emphasis on the importance of genetic differences at key genes. Roth, whose primary interest is in muscle size and strength, has contributed a primer on exercise genetics (187) and has a longstanding interest in the role of genetic variation in talent selection for athletic performance. Both have trained a good number of graduate students in exercise genetics and have been longtime contributors to the fitness and performance gene map series that was published in Medicine and Science in Sports and Exercise until 2009.
Stephen M. Roth (1973 - ), PhD, FACSM, associate professor, and James M. Hagberg (1950 - ), PhD, FACSM, professor, department of kinesiology, school of public health, University of Maryland. Hagberg completed his graduate studies at the University of Wisconsin. He has been on the faculty of a number of universities and at the University of Maryland for nearly 20 yr, where he is a distinguished scholar-teacher. He has been funded by the National Institute of Aging for genetics-based investigations and has published numerous papers identifying genetic markers that associate with the degree to which endurance-exercise training improves cardiovascular disease risk factors in older men and women. Stephen M. Roth completed graduate studies in exercise physiology at the University of Maryland in 2000 and postdoctoral training in human genetics at the University of Pittsburgh. In 2003 he was recruited back to the University of Maryland as director of the Functional Genomics Laboratory. He has been funded by the National Institutes of Health to perform a variety of genetics investigations. In addition to several research publications, Roth is author of a textbook titled Genetics Primer for Exercise Science and Health. The department of kinesiology at the University of Maryland was one of the first to develop a research focus in the area of exercise genomics. As a group, they have published more than 60 peer-reviewed publications in the areas of exercise genomics.
When the last version of the fitness and performance gene map was published based on the publications available by the end of 2007, there were 214 autosomal, 7 X chromosome, and 18 mitochondrial gene and other loci entries (31). Most of the entries were related to specific candidate genes and were supported by more than 350 peer-reviewed papers. With the exception of the ACE and ACTN3 common variants, few of the genes had been replicated in multiple studies. Moreover, after almost 25 yr of candidate gene studies in exercise biology, this line of research continues to be plagued by small sample sizes and lack of appropriate statistical power (31, 177). Studies with small sample sizes occasionally report positive findings, especially if they include several traits and perform multiple statistical tests for association. Fortunately, these positive findings are seldom replicated and usually fall into oblivion.
