226 S.M. Roth Table 1 Lowest and highest concentric knee extension strength in each decade of the adult life span in 1,283 men and women from the Baltimore Longitudinal Study of Aging (Shock et al. 1984; Lindle et al. 1997; Lynch et al. 1999; Ferrucci 2008) Age Range (year) 20–29 30–39 40–49 50–59 60–69 70–79 80–96 Men (N = 661) 101–248 (N = 21) 57–317 (N = 60) 37–411 (N = 102) 55–205 (N = 156) 38–330 (N = 114) 19–178 (N = 117) 16–239 (N = 90) Women (N = 622) 28–126 (N = 22) 29–151 (N = 73) 27–134 (N = 102) 20–240 (N = 168) 11–136 (N = 125) 17–140 (N = 83) 12–117 (N = 49) Data are isokinetic peak torque values (Nm) at 180 deg/s. 227Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia twins on the other hand similarly share the intrauterine and external environment through young adulthood, but share only approximately 50% of their genetic variation. Thus, correlations performed between monozygotic and dizygotic twin pairs can be compared and estimates of genetic contribution, termed heritability, can be determined (Bouchard et al. 1985, 1997; Roth 2007). When traits exhibit closer correlation in monozygotic compared to dizygotic twins, the assumption is that genetic factors are contributing to the closer correlation in monozygotic twins and heritability can be calculated from the extent of difference observed in the cor- relation values. Clark reported one of the first heritability studies with relevance to skeletal muscle in 1956 (Clark 1956). In that report, a series of anthropometric traits were compared in monozygotic and dizygotic twins, including measures of arm and calf circumference both of which were greater than 60% heritable. Later studies pro- vided more direct measures of skeletal muscle traits. For example, the heritability of grip strength was estimated between 30% and 50% in several early studies (Montoye et al. 1975; Venerando and Milani-Comparetti 1970; Kovar 1975). In a study of older twins, genetic factors accounted for 65% of the variance in grip strength in 260 mono- and dizygotic twins (59–69 year), even after adjusting for body weight, height and age (Reed et al. 1991). More recently, twin studies have revealed heritability values for muscle strength phenotypes ranging from 30% to 85% depending on the conditions of the strength measure (e.g., limb, contraction angle, velocity, and type) (Thomis et al. 1998a, 2004; Perusse et al. 1987a, b; Huygens et al. 2004a; Karlsson et al. 1979; Reed et al. 1991; Thomis et al. 1998a; Arden and Spector 1997; Zhai et al. 2004; Ropponen et al. 2004). Skeletal muscle fiber type composition has also been shown to be a heritable trait (Komi et al. 1977; Simoneau and Bouchard 1995), though variability in the biopsy technique and heterogeneity of fiber type distribution within skeletal muscle make these estimates remarkably challenging. The hypothesis that genetic factors may influence muscu- lar strength is also supported by data from rats in which a 1.5- to 5.2-fold diver- gence between the muscular strength of 11 different strains with the lowest and highest strength levels has been reported (Biesiadecki et al. 1998). With regard to skeletal muscle mass, evidence for significant heritability has been identified across a number of traits, with the first studies reporting heritabil- ity of limb circumferences (Clark 1956, Huygens et al. 2004b; Loos et al. 1997; Susanne 1977; Thomis et al. 1997). The first direct study of lean body mass (LBM) was performed by Bouchard et al. (1985) who reported 80% heritability of LBM by hydrodensitometry in twin pairs. Later Forbes et al. (1995) reported 70% heritability of LBM by the potassium 40 counting method, and Seeman et al. (1996) and Arden et al. (1997) provided the first estimates (50–80%) using dual energy x-ray absorptiometry (DXA). Other studies have reported similar findings (Nguyen et al. 1998; Loos et al. 1997; Thomis et al. 1998b; Livshits et al. 2007; Karasik et al. 2009) and recently Prior and colleagues (2007) reported significant heritability of lean mass and calf cross-sectional area (CSA) in families of African-descent, providing the first evidence of heritability values in this race group, which is known to higher muscle mass and strength traits compared to 228 S.M. Roth subjects of European descent (Aloia et al. 2000; Gallagher et al. 1997; Jones et al. 2002; Visser et al. 2000a; Newman et al. 2006). Across these studies, heritability estimates greater than 50% are not uncommon for muscle mass measurements. Perhaps most relevant for this discussion are the various studies that have examined heritability within older subjects. In addition to the study of grip strength by Reed and colleagues (1991) discussed above, several other reports have demon- strated significant heritability values for muscle strength in older individuals (Frederiksen et al. 2002, 2003; Tiainen et al. 2004, 2005, 2009; Zhai et al. 2004, 2005). For example, Frederiksen and colleagues (2002) showed heritability of grip strength at 50% across several age groups from 46 to 96 year. The change in muscle strength with advancing age has also been found to be heritable (Carmelli et al. 2000; Zhai et al. 2004), though some studies indicate that the contribution of envi- ronmental factors appears to increase at older ages (Carmelli and Reed 2000; Tiainen et al. 2004). With regard to the more general trait of functional perfor- mance, the results are more mixed with moderate heritability for lower-extremity function in older male twins (Carmelli et al. 2000), low heritability reported for age-related functional impairment in male twins (Gurland et al. 2004), and low but significant heritability for older female twins in the rate of change of physical function with age, with a non-significant genetic component in older male twins (Christensen et al. 2002, 2003). These findings are consistent with the idea that more general, multi-component traits are likely to be influenced by a wider range of environmental factors, especially in older individuals (Tiainen et al. 2005; Harris et al. 1992). Overall, genetic variation explains a significant fraction of the inter-individual variability in skeletal muscle phenotypes, including muscle traits in older individuals. While there is strong evidence for a heritable component to muscle phenotypes, the genetic analysis of muscle architecture is in its infancy. 3 Linkage Analysis and Skeletal Muscle Traits After the familial aggregation and heritability of a trait is firmly established, until recently the next step in genetic analysis was to perform linkage analysis studies in families. The goal of linkage analysis was to rely on the shared genetic variation with families to identify chromosome locations that harbor genes and gene variants that contribute to trait variation. By determining several hundred genotypes spread across the genome in each of the individuals of several families, linkage analysis would identify those regions most closely correlated with the trait of interest. Significantly correlated regions are assumed to harbor genetic variation relevant to the trait of interest, though these identified regions are often quite extensive, with many potential genes. Thus, linkage analysis is useful for narrowing the potential list of candidate genes from many thousand to several hundred, but considerable work remains even after a linkage study to confidently determine the specific contributing genes. 229Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia In the first genome-wide linkage analysis for genes related to muscle mass, Chagnon et al. (2000) examined microsatellite markers in the Quebec Family Study, which consisted of 748 subjects from 194 families. Fat-free mass (FFM) was calculated from percent body fat determined by hydrostatic weighing. Significant linkages were observed for a CA-repeat within the insulin-like growth factor 1 receptor (IGF1R) on 15q25-26, and at two markers at 18q12; moderate linkage was noted on 7p15.3, with the authors noting possible candidate genes of neuropeptide Y (NPY) and growth hormone-releasing hormone (GHRH) receptor in that location. A second study by Chagnon et al. (2001) examined body composition in 364 sib- ling pairs from 99 families from the HERITAGE Family Study before and after 20 weeks of aerobic exercise training. In that analysis, no significant loci were identi- fied for baseline FFM, though change in FFM in response to aerobic exercise train- ing was linked to loci at IGF1, 1q22, and 8q24.12. Livshits and colleagues (2007) reported significant linkage with LBM in 3180 female twin pairs at chromosomes 12q24.3 and 14q22.3. Most recently, Karasik et al. (2009) reported significant link- age in 1346 adults from 327 families from the Framingham study for leg lean mass measured by DXA. Two loci (12p12.3-12p13.2 and 14q21.3-22.1) were identified as having bivariate linkage with both leg lean mass and bone phenotypes. Two studies have examined strength-related phenotypes in family-based linkage analysis. De Mars and colleagues (2008a) reported significant linkage signals for torque-velocity ratios of the knee flexors and extensors (strongest signal at 15q23), as well as for the torque-velocity slope of the knee extensors. The same group reported significant linkage for the torque-length relationship of the knee flexors (strongest signal at 14q24.3) and isometric knee torque in 283 male siblings from 105 families (De Mars et al. 2008b). A few linkage studies have been performed in a more focused manner, isolating a small number of regions in order to better identify potential candidate genes. In the HERITAGE Family Study, Sun et al. (1999) performed a focused linkage analy- sis around a microsatellite marker in the IGF1 locus. In 502 individuals from 99 families, the IGF1 locus was not significantly linked with baseline FFM, though was significantly associated with the change in FFM after aerobic exercise training, consistent with the genome-wide linkage results of Chagnon and colleagues (2000) described above. Huygens et al. (2004c) performed a gene-specific linkage analysis for the RB1 locus in 329 young Caucasian male siblings from 146 families for trunk strength and identified multiple linkage peaks for trunk flexion measures with no evidence of linkage for trunk extension measures. In a second study, Huygens and colleagues (2004c) performed a gene-targeted single-point (one marker per gene) linkage analysis in the myostatin pathway (across 10 genes) in the same young male cohort for various measures of muscle mass and strength. Significant linkage was reported with markers D2S118, D6S1051, and D11S4138 for knee extension and flexion peak torque measures. These markers are in the MSTN (myostatin, formerly GDF8), CDKN1A, and MYOD1 genes, respectively. Huygens et al. (2005) then performed an expanded multi-point (multiple markers per gene) linkage analy- sis in 367 young Caucasian male siblings from 145 families with nine genes involved in the myostatin signaling pathway and various measures of muscle 230 S.M. Roth strength. Significant linkages were reported on four chromosomal regions with knee muscle strength measures: chromosome 13q21 (D13S1303), chromosome 12p12-p11 (D12S1042), chromosome 12q12-q13.1 (D12S85), and chromosome 12q23.3-q24.1 (D12S78). Only one linkage study has targeted older individuals in particular. In 2008, Tiainen et al. (2009) examined 397 microsatellite markers in 217 female twin pairs aged 66 to 75 years from the Finnish Twin Study on Aging. Significant linkages were reported for knee extensor isometric strength on chromosome 15q14, for leg extensor power on chromosome 8q24.23, and for calf muscle CSA on chromosomes 20q13.31 and 9q34.3. Importantly, the linkage noted at 9q34 was similarly observed by Chagnon and colleagues (2001) for change in FFM in response to exercise training, providing some of the first evidence of replication of a locus related to skeletal muscle mass across different linkage studies. Recently, linkage analysis studies have given way to genome-wide association studies that can be used to identify specific gene regions in unrelated individuals by use of high-density single nucleotide polymorphism microarrays, which allow as many as 1 million genotypes to be determined and used in association analyses. These studies have been successful at identifying a clinically relevant candidate gene for age-related macular degeneration (Klein et al. 2005), and have provided important novel targets for other health-related traits (Lindgren et al. 2009; Graham et al. 2009). Only one such study has been performed for skeletal muscle traits to date. In 2009, Liu and colleagues examined 379,319 polymorphisms across the genome in nearly 1,000 unrelated U.S. whites for association with LBM measured by DXA. In the initial genome-wide analysis, two polymorphisms were identified as statistically significant (with Bonferroni corrected p values at 7 × 10 −8 ) and another 146 polymorphisms approaching statistical significance. The two significant poly- morphisms are both located in the TRHR gene, which encodes the thyrotropin- releasing hormone receptor. These two polymorphisms were then genotyped in three replication cohorts consisting of over 6000 total white and Chinese subjects and consistent significant associations were observed in those analyses. Because of the importance of thyroid hormone in skeletal muscle development (Larsson et al. 1994; Norenberg et al. 1996; Soukup and Jirmanova 2000), the TRHR gene is thus recog- nized as an important candidate gene for future investigation. Though currently unpublished, other research groups have genome-wide association data available and additional findings are expected before the end of 2010. 4 Genetic Variation and Skeletal Muscle Traits The ultimate goal of linkage and genome-wide association studies is the identification of specific genes and gene variants with clinically relevant influences on skeletal muscle traits important to physical function. The advent of genome-wide association studies provides an important technical improvement in the ability to identify specific 231Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia loci for in-depth investigation, though as mentioned above, only one has been published to date for skeletal muscle traits. As loci are replicated across studies, specific gene variants will be identified and their clinical relevance determined. In the next sections, specific genes and gene sequence variants that have been associated with skeletal muscle phenotypes will be discussed, including those associated with muscle strength, muscle mass, and sarcopenia in particular. Genes related to skeletal muscle adaptation will only be discussed briefly as this is not a focus of this chapter. While the reference lists for these sections will be comprehensive, only those genes examined in multiple investigations or otherwise shown to be functional in some way will be discussed in detail. Replication of genetic associations, especially those of generally weak genetic influence, is generally considered the gold standard for considering a gene important to a trait, though other approaches exist (Khoury et al. 2005, 2007). 4.1 Genetic Variation and Skeletal Muscle Strength The identification of genetic factors important to skeletal muscle strength is remarkably difficult owing to the fact that multiple strength variables are com- monly measured in different studies, including different muscle groups (forearm, knee extensor, leg), contraction types (isometric, isotonic, isokinetic), and measure- ment instruments. Moreover, different genes are likely to contribute to different aspects of strength that may not be reflected across the different measurement types. Additionally, some studies have included measurements of muscle quality or muscle power given their importance to physical function, especially for the elderly (Dutta et al. 1997; Bassey et al. 1992). All this means that for a particular gene or genotype of interest, the chances of finding replication across multiple studies for the same trait are small. This has both positive and negative implications: though few studies demonstrate replication and thus few studies have found evidence of the importance of any one gene, when genes are found to be important across multiple, different strength measurements the likelihood the gene is truly important to muscle strength improves. Table 2 summarizes the genes that have been studied in relation to skeletal muscle strength measurements, focusing on genes associated with baseline strength values; genes related to muscle strength adaptation to exercise training are dis- cussed in a later section. Genes that have been studied in only one paper or that have not been replicated in some way and are not discussed here in detail include: COL1A1 (Van Pottelbergh et al. 2001, 2002); BDKRB2 (Hopkinson et al. 2006); DIO1 (Peeters et al. 2005); MYLK (Clarkson et al. 2005b); IL6 (Walston et al. 2005); TNF (Liu et al. 2008a); NR3C1 (van Rossum et al. 2004; Peeters et al. 2008); AR (Walsh et al. 2005); and IL15 and IL15RA (Pistilli et al. 2008). Angiotensin Converting Enzyme (ACE) ACE and its insertion/deletion (I/D) polymorphism is arguably the most studied of genes with regard to exercise 232 S.M. Roth Table 2 Genes and gene sequence variants associated with skeletal muscle strength phenotypes in multiple studies Gene References Variants Examined Subjects Skeletal Muscle Strength Measurements ACE Woods et al. (2001) I/D polymorphism 83 postmenopausal women Change in isometric strength of adductor pollicis in response to HRT Hopkinson et al. (2004) I/D polymorphism 103 COPD patients Quadriceps isometric strength Williams et al. (2005) I/D polymorphism 81 young men Quadriceps isometric strength Moran et al. (2006) ACE I/D and haplotype 1,027 adolescents Handgrip strength and vertical jump in females Wagner et al. (2006) I/D polymorphism 62 young men and women Contraction velocity and isometric force in multiple muscles Yoshihara et al. (2009) I/D polymorphism 431 older Japanese men and women Hand grip strength and walking speed ACTN3 Clarkson et al. (2005) R577X 602 young men and women Biceps isometric strength in females Delmonico et al. (2007) R577X 157 older men and women Knee extensor peak power in women Vincent et al. (2007) R577X 90 young men Isokinetic knee extensor strength Delmonico et al. (2008) R577X 1,367 older men and women Physical limitation and walk performance Walsh et al. (2008) R577X 848 men and women Isokinetic knee extensor strength in women CNTF Roth et al. (2001) rs1800169 494 men and women Isokinetic knee extensor and flexor strength and muscle quality Arking et al. (2006) rs1800169 and CNTF haplotype 363 older women Hand grip strength 233Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia CNTFR Roth et al. (2003) C-1703T, T1069A, C174T 465 men and women Isokinetic knee extensor and flexor strength De Mars et al. (2007) C-1703T, T1069A, C174T, and others 493 older men and women Knee flexor and extensor strength measurements IGF2 Sayer et al. (2002) ApaI 693 older men and women Hand grip strength in men Schrager et al. (2004) ApaI 596 men and women Isokinetic strength of multiple muscle groups MSTN Seibert et al. (2001) K153R 286 older women Composite isometric strength score Corsi et al. (2002) K153R 450 older men and women Composite isometric strength score Kostek et al. (2009) K153R, A55T 23 young African Americans Isometric biceps strength VDR Geusens et al. (1997) BsmI 501 older women Isometric quadriceps and handgrip strength Grundberg et al. (2004) BsmI, poly A repeat 175 young women Isokinetic knee flexor strength Roth et al. (2004) FokI 302 older men Isometric knee extensor strength Wang et al. (2006) ApaI, BsmI, TaqI 109 young Chinese women Multiple knee and elbow strength measures Windelinckx et al. (2007) BsmI, TaqI, and FokI 493 older men and women Multiple quadriceps strength measures Hopkinson et al. (2008) FokI and BsmI 107 COPD patients; 104 control men and women Isometric quadriceps strength 234 S.M. Roth performance phenotypes (Jones et al. 2004) and several investigations have targeted skeletal muscle traits in particular. Folland and coworkers (2000) first reported no significant association between ACE genotype and quadriceps iso- metric strength in 33 young males, though differences in muscle strength response to strength training were observed. Woods et al. (2001) found that the rate of change in muscle force in response to hormone replacement therapy (HRT) was stronger in I/I compared to D/D genotype carriers in a study of 83 older post- menopausal women. Thomis and colleagues (1998b) found that the ACE I/D polymorphism was not significantly associated with elbow flexor strength in a study of 57 young male twins. Hopkinson et al. (2006) reported significantly higher knee extensor maximal strength in chronic obstructive pulmonary disease (COPD) patients carrying the D-allele compared to I/I patients, though the asso- ciation was not observed in 101 age-matched healthy controls. Williams et al. (2005) examined quadriceps muscle strength in 81 young Caucasian men and reported that baseline isometric strength was significantly associated with ACE genotype, with I-allele homozygotes showing the lowest strength values. Moran and colleagues (2006) examined handgrip strength and vertical jump in 1,027 Greek adolescents and reported higher handgrip strength and vertical jump scores in females carrying the I/I genotype. No significant associations were observed in males. The authors performed haplotype analysis of the ACE gene region using three polymorphisms and determined that the I/D polymorphism explained the bulk of the explained genetic variance. Pescatello and co-workers (2006) studied the I/D genotype in relation to elbow flexor strength in 631 young men and women and reported no association with muscle strength in either arm. Wagner et al. (2006) examined leg press strength variables in 62 young men and women. They showed that no single muscle phenotype was consistently associated with ACE I/D genotype, but that combinations of traits including contraction velocity, isometric force, and optimum contraction velocity differed among the three geno- type groups in both men and women with I/I genotype carriers exhibiting lower maximum and optimum contraction velocity compared to I/D and D/D carriers. McCauley and colleagues (2009) did not observe any associations between ACE I/D genotype and knee extensor isometric or isokinetic torques in 79 young males, though serum ACE activity was associated with ACE genotype as expected. Charbonneau et al. (2008) reported higher quadriceps muscle volume in D/D carriers in a study of 225 older men and women, but no genotype differ- ences were observed for muscle strength (1RM). Finally, Yoshihara et al. (2009) recently reported that the I/D polymorphism was associated with physical func- tion in 431 elderly Japanese subjects, with higher hand grip and 10 m maximum walking speed in D/D carriers. In summary, ACE genotype has been associated with muscle strength variables in a number of studies, but those associations are balanced by several studies showing no association or inconsistencies among find- ings. There is little evidence to suggest that ACE genotype is a strong contributor to inter-individual variation in skeletal muscle strength. Alpha Actinin 3 (ACTN3) The ACTN3 gene and its nonsense R577X polymor- phism has generated considerable attention following a number of cross-sectional 235Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia investigations in elite athletes that pointed to a considerable disadvantage for X/X carriers in sprint and power related activities (Yang et al. 2003; Niemi and Majamaa 2005; Roth et al. 2008). Several groups then moved to examine quantitative traits to determine the underlying phenotype impacted by the alpha actinin 3 protein deficiency resulting from the X/X genotype. Clarkson and colleagues (2005a) reported that X/X women had lower baseline isometric strength than the R/R women in a study of 602 young men and women. No association was observed in men. Delmonico and coworkers (2007) examined knee extensor concentric peak power in 157 older men and women. Contrary to expectation, women X/X carriers exhibited greater relative peak power than both R/X and R/R genotypes. In men, no genotype differences were observed. Both men and women participated in a strength training program that indicated a stronger adaptation for R/R carriers com- pared to X/X carriers. Vincent and colleagues (2007) studied the R577X polymor- phism in relation to isometric and isokinetic knee extensor strength in 90 young men and reported lower concentric peak torque at 300 deg/s in X/X compared to R/R homozygotes. The authors also reported a lower proportion of type IIx muscle fibers in X/X vs R/R homozygotes. In a study of 1,367 older adults (70–79 year), Delmonico et al. (2008) reported greater losses of 400 m walk time performance over 5 years in male X/X vs R-allele carriers, while X/X women had a 35% greater risk of lower extremity physical limitation compared to R/R women. Walsh et al. (2008) examined knee extensor shortening and lengthening peak torque values in 848 adults (22–90 year) and reported that X/X women displayed lower knee exten- sor strength values compared with R/X + R/R women. No genotype-related differ- ences were observed in men. Women X/X homozygotes also displayed lower levels of FFM, as described in the next section. Some studies have not been able to con- firm these genotype differences. For example, Norman and colleagues (2009) reported no significant associations with muscle power or torque-velocity relation- ships among ACTN3 genotypes in a study of 120 moderately to well-trained men and women. They were also unable to confirm the difference in fiber type propor- tion reported by Vincent and colleagues (2007). Similarly, McCauley and col- leagues (2009) did not observe any associations between ACTN3 genotype and knee extensor isometric or isokinetic torques in 79 young males. The general con- sensus among these studies is that ACTN3 X/X carriers may have modestly lower skeletal muscle strength and power in comparison to R-allele carriers, with the work of Delmonico and colleagues (2008) indicating potential clinical importance for the X/X genotype in older men and women. Ciliary Neurotrophic Factor (CNTF) Three studies have examined genetic variation in the CNTF gene and/or its receptor, CNTFR. Roth and colleagues (2001) first reported that a null mutation (rs1800169; A/G: A = null allele) in the CNTF gene was associated with muscle strength and muscle quality in 494 men and women across the adult age span. Homozygotes of the rare null allele (A/A) had lower strength while heterozygotes had higher strength than G/G carriers across multiple muscle strength and muscle quality measurements. Arking et al. (2006) examined eight polymorphisms surrounding the CNTF locus, including the rare rs1800169 nonsense polymorphism in 363 older Caucasian women. Haplotype . strength and muscle quality Arking et al. (2006) rs1800169 and CNTF haplotype 363 older women Hand grip strength 233Genetic Variation and Skeletal Muscle Traits: Implications for Sarcopenia CNTFR. lowest strength values. Moran and colleagues (2006) examined handgrip strength and vertical jump in 1,027 Greek adolescents and reported higher handgrip strength and vertical jump scores in. (2006) ACE I/D and haplotype 1,027 adolescents Handgrip strength and vertical jump in females Wagner et al. (2006) I/D polymorphism 62 young men and women Contraction velocity and isometric force