INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CILIARY NEUROTROPHIC FACTOR (CNTF) has trophic effects on both neuronal (31) and muscular tissues (5, 8, 18). Although CNTF synthesis is associated with peripheral nerves (27), a specific binding subunit of the CNTF receptor (CNTF receptor-) is required for CNTF activity (3). Moreover, CNTF receptor- is abundantly expressed in skeletal muscle (14). As such, recent studies have examined the roles of CNTF and CNTF genotype on neuromuscular disease and muscle function. One study demonstrated that sciatic nerve CNTF levels are associated with swimming performance and muscular strength in rats (8). In that same study, CNTF levels declined with age, and exogenous CNTF administration in older rats increased muscular strength and muscle fiber area to the level observed in younger adult animals (8). More recently, CNTF administration has been shown to prevent losses of soleus muscle mass and function after hindlimb suspension in rats (6). Takahashi et al. (29) first described a CNTF gene variant (A allele) in humans predicted to result in a nonfunctional protein, although individuals homozygous for this "null" mutation (A/A) did not have altered rates of neuromuscular disease. Muscular strength differences were not examined in that investigation.

Because of the apparent influence of CNTF on skeletal muscle function (6) and the fact that lower levels are associated with lower muscular strength in rats (8), we hypothesized that humans heterozygous or homozygous for the CNTF A allele would exhibit significantly lower muscular strength compared with individuals homozygous for the more common allele (G/G). We investigated the role of CNTF genotype on muscular strength and quality in a large sample (~500) of men and women across the entire adult age span (20-90 yr) from the Baltimore Longitudinal Study of Aging (BLSA). The known effects of age, gender, body mass, and race on muscular strength (16, 17, 20) were accounted for to specifically target the relationship of CNTF genotype to muscular strength and quality.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Volunteers (n = 494; 250 men and 244 women, age range 20-90 yr) from the BLSA (28) participated in the study. Of the total sample, 413 subjects were Caucasian, 63 were African-American, and 18 were of other races. All subjects received a complete medical history and physical examination, and subjects with clinical cardiovascular or musculoskeletal disorders known to be adversely affected by exercise testing were excluded. Detailed exclusion criteria are outlined elsewhere (16, 17). The results of a physical activity questionnaire indicated that only a small proportion of the subjects (<1%) participated in regular resistive exercise training, with no differences in participation by age or gender (16). After receiving a complete explanation of all procedures and risks associated with the study, all subjects gave their written informed consent. The experimental protocols were approved by the Institutional Review Boards for Human Subjects at Johns Hopkins Bayview Medical Center (Baltimore, MD), the University of Maryland (College Park, MD), and the University of Pittsburgh (Pittsburgh, PA).

Total body mass and fat-free mass. Body mass and height were measured for each subject to the nearest 0.1 kg and 0.5 cm, respectively, using a Detecto medical beam scale. A subset of the participating subjects was assessed for total body fat-free mass (FFM; n = 379) and dominant leg FFM (n = 321). In these subjects, a total body scan was performed using dual-energy X-ray absorptiometry (DEXA) as described previously (16, 17). Based on the total body DEXA scan, nonosseous total body FFM and dominant leg FFM (estimated muscle mass) were determined (16). FFM measured by DEXA has been shown to correlate strongly with muscle mass in humans across the age span (9, 13).

Strength and muscle quality. Peak torque (strength) was measured using the Kinetic Communicator isokinetic dynamometer (Kin-Com model 125E, Chattanooga Group, Chattanooga, TN). Concentric (Con) and eccentric (Ecc) peak torques were measured at angular velocities of 0.52 rad/s (30°/s) and 3.14 rad/s (180°/s), respectively, for the dominant knee extensors (KE) and flexors (KF). Maximal voluntary isometric torque was measured for the KE at 2.09 rad (120°). For each test, subjects performed three maximal efforts, separated by 30-s rest intervals, from which the highest value of the three trials was accepted as the peak torque. Detailed procedures regarding subject positioning and stabilization, warm-up, testing order, gravity correction, and Kin-Com calibration are described elsewhere (16, 17). Correlation coefficients for test-retest reliability ranged from 0.96 to 0.99 (16). Muscle quality was estimated for the KE by calculating the ratio of Con peak torque at both 0.52 rad/s and 3.14 rad/s to the dominant leg estimated muscle mass.

Genotype. Genomic DNA was extracted from whole blood samples (21). CNTF genotype was determined as outlined previously (29, 30). Subjects were categorized as exhibiting the G/G, G/A, or A/A genotype.

Statistics. Analysis of physical characteristics by CNTF genotype consisted of one-way ANOVA with least squares difference (LSD) post hoc, when appropriate. The various measures of peak torque were used as response variables by means of analysis of covariance (ANCOVA), with age, gender, and either total body mass, total body FFM, or dominant leg muscle mass as covariates (with LSD post hoc). Muscle quality was compared by genotype group by use of ANCOVA, with age and gender as covariates. Body mass and FFM variables were also compared by genotype group with ANCOVA, with age and gender covaried. To address the question of whether the results were affected by population stratification or subdivision, a separate analysis was performed on Caucasian subjects only (n = 413). The physical characteristics of those subjects who did not have measures of FFM (and were thus excluded from analyses in which FFM was used as a covariate) did not differ significantly from the subjects included in those analyses. Regression analysis was used to estimate the proportion of the variance in strength that was accounted for by CNTF genotype. Data are reported as least squares means ± SE. Statistical significance was accepted at P  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject characteristics are outlined in Table 1. No significant differences existed among CNTF genotype groups for any physical characteristic. Of the 494 total subjects in the study, 389 exhibited the G/G genotype (78.7%), 95 exhibited the G/A genotype (19.2%), and only 10 exhibited the rare A/A genotype (2.0%). The G allele frequency was 88.3%, and the A allele frequency was 11.6%. These values did not differ significantly from predicted Hardy-Weinberg equilibrium (² analysis, P > 0.05). We further characterized the genotype and allele frequencies based on race. Of the 413 Caucasian subjects, 316 exhibited the G/G genotype (76.5%), 88 exhibited the G/A genotype (21.3%), and 9 exhibited the rare A/A genotype (2.2%). The G allele frequency was 87.2%, and the A allele frequency was 12.8%. Of the 63 African-American subjects, 57 exhibited the G/G genotype (90.5%), 5 exhibited the G/A genotype (7.9%), and 1 exhibited the rare A/A genotype (1.6%). The G allele frequency was 94.5%, and the A allele frequency was 5.5%.

Table 2 outlines the KE Con and Ecc peak torque values for each genotype group, with age, gender, and body mass used as covariates. Con KE peak torque at 3.14 rad/s was significantly greater in the G/A compared with the G/G genotype group (P < 0.05), whereas Ecc KE peak torque at 0.52 rad/s was significantly lower in the A/A compared with both the G/G and G/A genotype groups (P < 0.05). No significant differences were observed for Con KE peak torque at 0.52 rad/s or Ecc KE peak torque at 3.14 rad/s among the genotype groups (Table 2). When total body FFM was used in place of total body mass as a covariate, KE Ecc peak torque at 0.52 rad/s was significantly lower in the A/A compared with the G/A and G/G genotype groups (P < 0.05), but no other significant differences were observed. However, when the dominant leg muscle mass was used as a covariate in place of total body FFM, Con KE peak torque at 3.14 rad/s was significantly greater in the G/A vs. the G/G genotype group (107.9 ± 2.8 vs. 101.2 ± 1.3 N · m, respectively; P < 0.05). Despite these differences, regression analysis indicated that <1% of the variance of these strength differences was explained by CNTF genotype (r = 0.04). Maximal isometric KE torque did not differ significantly among the genotype groups (Table 2). Similar results occurred in both men and women when the data were analyzed within gender.

Peak torque and genotype relationships in Caucasian subjects revealed results similar to those of the entire sample. Con peak torque at both 0.52 and 3.14 rad/s was significantly greater in the G/A genotype group compared with both the G/G and A/A groups (P  0.05) in Caucasians. Ecc peak torque at 0.52 rad/s was significantly lower in A/A genotype compared with G/G and G/A Caucasians (P < 0.05).

Significant group differences were also noted for several KF peak torque values, as outlined in Table 3, when age, gender, and total body mass were covaried. KF Con peak torque at 3.14 rad/s was significantly higher in the G/A vs. the G/G genotype group, KF Con peak torque at 0.52 rad/s was significantly greater in the G/A vs the A/A genotype group, and both G/G and G/A genotype groups exhibited significantly greater KF Ecc peak torque at 0.52 rad/s compared with the A/A genotype group (all P < 0.05).

An analysis of KE Con muscle quality revealed a significant difference between the G/G and G/A genotype groups at 3.14 rad/s when age and gender were covaried (P < 0.05; Table 2). No such difference was noted for KE Con muscle quality at 0.52 rad/s. Similar results were observed for KE Con muscle quality in Caucasians, with Con muscle quality at 3.14 rad/s showing a strong trend for lower values in the G/G vs. the G/A genotype group (12.1 ± 0.2 vs. 12.8 ± 0.3 N · m · kg leg FFM¹; P = 0.06). No such relationship was observed at 0.52 rad/s.

A subanalysis of KE peak torque and muscle quality was completed for subjects 60 yr and older (average age 72 yr), based on previous data suggesting that age-related strength declines begin to become significantly lower than for young individuals at ~60 yr (16). This analysis of older subjects revealed trends similar to those outlined above for the entire cohort (Table 4). For example, G/A individuals >60 yr demonstrated significantly greater Con peak torque at 3.14 rad/s (age, gender, and body mass covaried) and Con muscle quality at 3.14 rad/s (age and gender covaried) compared with the other genotype groups (P < 0.05; Table 4). In addition, KE Con peak torque at 3.14 rad/s was significantly greater in the G/A genotype group compared with the G/G genotype group when dominant leg muscle mass was used in place of body mass as a covariate (102.2 ± 5.2 vs. 85.2 ± 2.9 N · m, respectively; P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Because of the apparent influence of CNTF on skeletal muscle function (6) and the fact that lower levels are associated with lower muscle strength in rats (8), we hypothesized that a CNTF gene variant that is predicted to result in a nonfunctional protein (29) would be associated with significantly lower muscle strength in humans. Contrary to our hypothesis, individuals heterozygous for the CNTF null allele (G/A) exhibited significantly greater KE and KF Con peak torque and muscle quality using KE peak torque at the fast velocity (3.14 rad/s) when age, gender, and total body mass were used as covariates. Furthermore, when the dominant leg muscle mass was used in place of total body mass as a covariate, quadriceps Con peak torque at the fast velocity was also significantly greater in subjects with the G/A genotype compared with the more common G/G genotype group. However, in support of our hypothesis, subjects homozygous for the rare A allele did exhibit lower strength values generally compared with the other genotype groups; however, the sample size for this group was low (n = 10). These results were also observed when the analysis was performed with data from Caucasian subjects only.

Although the trophic effect of CNTF on motoneurons is well established (26) and CNTF is clearly associated with motoneuron survival and regeneration (31), the trophic effects of CNTF on skeletal muscle are less clear. Although some studies have indicated a trophic effect of CNTF on skeletal muscle (5, 18) and CNTF administration has been shown to partially prevent the atrophy and reduced tetanic and twitch tensions of denervated skeletal muscle (11), other research has indicated that CNTF administration results in cachexia and muscle atrophy (12, 19). With regard to muscle function, however, data are more consistent. Guillet et al. (8) reported that sciatic nerve levels of CNTF correlated closely with swimming performance in rats and that both parameters decreased similarly with increasing age. In addition, these investigators showed that CNTF administration improved twitch and tetanic force in the soleus muscle of older rats to the level of adult rats. Moreover, CNTF administration in the older animals resulted in a 17% increase in muscle fiber area, providing further evidence of a trophic effect of CNTF on skeletal muscle (8). More recently, Fraysse et al. (6) reported that exogenous CNTF prevented soleus muscle atrophy and strength loss resulting from hindlimb unweighting. Thus existing data suggest that CNTF has trophic effects on both neural and muscular tissues, as well as effects on muscle function.

The possible mechanism for our observation of higher strength and muscle quality in heterozygous (G/A) individuals compared with either A/A or G/G individuals is unclear. The observation of Takahashi et al. (29) that heterozygous individuals express both the normal and mutant allele at the RNA level, but that immunoblot analysis reveals only the normal protein, suggests that no protein product is produced by the null allele. CNTF signaling is mediated by binding to a CNTF-specific receptor lacking an intracellular domain and stimulation of intracellular signaling by interaction of this complex with either gp30 or leukemia inhibitory factor receptor-, which possess tyrosine kinase activity (22). If a truncated version of CNTF is produced by the null allele, competitive interaction with the receptor complex could lead to altered signaling. It has been proposed that the CNTF receptor is released from the cell surface by phosphatidylinositol-specific phospholipase C cleavage of the glycosylphosphatidylinositol linkage, which anchors the receptor to the cell surface and that the CNTF/soluble CNTF receptor complex mediates biological responses not provoked by CNTF alone (3). The presence of a truncated version of CNTF could alter the kinetics of CNTF interaction with its soluble receptor. The nonadditive effect of the CNTF null allele on muscle strength and quality in the present study thus justifies a reexamination of the issue of protein expression from the null allele.

The muscular strength of A/A homozygotes was lower than that of the other two genotype groups, despite similar age, body mass, and total body and leg muscle mass. These results would appear to provide support for our hypothesis regarding individuals homozygous for the CNTF null allele, although the genotype frequency in this study was low (n = 10). Therefore, there is a need for further investigation into the relationship between strength and CNTF A/A genotype with the use of larger sample sizes. The results of Takahashi et al. (29) suggested that individuals possessing the A/A genotype were not at increased risk of neurological disease, but the relationship of the null mutation to muscular strength was not examined in that investigation.

In the present sample, 2.2% of Caucasian subjects exhibited the rare A/A genotype, whereas 76.5 and 21.3% exhibited the G/G and G/A genotypes, respectively. These genotype frequencies differ from previous studies. Takahashi et al. (29) observed that 61.9% of 391 Japanese subjects exhibited the G/G genotype, whereas 35.8 and 2.3% exhibited the G/A and A/A genotypes, respectively. Münzberg et al. (23) reported that ~28.4% of 429 German subjects carried the G/A genotype, and Gelernter et al. (7) showed that 30% of 90 European-American subjects from Connecticut exhibited one mutant allele (G/A), with an allele frequency of 19%. The current results may differ from these previous reports because of differences in sample size, different ethnic/racial populations, and/or a different American subpopulation (6). The Caucasian subjects in the current investigation included North American men and women from throughout the United States but mostly from Maryland and the District of Columbia (28). The genotype and allele frequencies of the African-American subjects in the current investigation differed from those of the Caucasian subjects (A allele frequency 5.5 vs. 12.8%, respectively), results similar to those reported previously (7).

The present results may indicate an importance for CNTF genotype in the context of aging and muscle function. For example, quadriceps concentric strength at 3.14 rad/s was 5.8% higher in the G/A vs. the G/G individuals in the present study. Furthermore, quadriceps concentric strength differences between these genotype groups were more profound in individuals over 60 yr, with G/A individuals exhibiting an 11% greater peak torque than G/G individuals. Rothstein et al. (25) have reported a strong correlation between KE Con peak torque at 2.10 rad/s (120°/s) and KE power (r = 0.95 to 0.97). Kannus and Järvinen (15) reported a significant correlation between KE Con peak torque at 3.14 rad/s and KE power; however, the measure of power in that study is suspect, because leg speed appears to have been held constant. Nevertheless, the present data support the hypothesis that CNTF genotype may be associated with muscular power. This seems consistent with the role of CNTF in neurological function, because muscle power is associated with both muscular and neurological factors (10). Because the loss of muscle power with age (1, 20) has important functional implications for older individuals (2, 24), the present results further indicate the importance of studying CNTF genotype and muscle power in older individuals. Muscle quality is thought to be an important factor for functional ability in the elderly (4, 16, 17), and the present results suggest an association of CNTF genotype and muscle quality. Thus the findings of this study strongly suggest the need for a longitudinal investigation into the effects of CNTF genotype on the loss of muscle power and muscle quality with age.

In summary, the present results indicate the potential importance of CNTF genotype on muscle strength, power, and quality and are the first data to suggest this relationship in healthy adult men and women across the entire adult age span. Given the apparent relationship between peak torque at relatively fast contraction speeds and muscular power and between muscular power and functional abilities, the present data indicate the importance of further studying the role of CNTF genotype on muscle function in older individuals.