INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

COLLAGEN, the major structural constituent of the interstitial space (14), provides a scaffold for maintaining muscle-tendon integrity and is involved in the transmission of muscular forces (3). Collagen is architecturally assembled such that it affords very little resting force in muscle at or below optimal length (L_o). However, as the muscle is stretched beyond L_o, the elastic properties of collagen are responsible in part for the passive force developed (3). Similar to other tissues, the load-deformation curve of skeletal muscle is approximately exponential (9). Several factors related to collagen may contribute to the passive viscoelastic properties of skeletal muscle, including the amount of collagen, its phenotypic distribution (2, 16), the extent of collagen cross-linking, and the architectural organization of the collagen fibrils (13, 16).

The maturation of collagen, an essential posttranslational process involving the formation of nonreducible cross-links, has been described, but factors regulating the process are poorly understood (6, 22). The rate-limiting step involves the oxidation of lysine and hydroxylysine residues by the enzyme lysyl oxidase, and the major end result in skeletal muscle is the formation of the acid-stable collagen cross-link HP (6, 22). The maturation of collagen alters the mechanical and chemical properties of the protein, leading to increased stability and tensile strength (6, 29), decreased solubility (23), and enhanced resistance to some proteases (4).

Aging is associated with significant changes in the connective tissue compartment of skeletal muscle. An increase in both concentration of collagen and extent of nonreducible cross-linking occurs with aging in both skeletal muscle (8, 33) and heart (28). The age-related changes are thought to be the result of decreased collagen turnover rates. The biochemical changes of collagen have been correlated with changes in muscle stiffness, i.e., stress/strain, such that stiffness increases in muscles of old animals (8, 10).

Exercise training is associated with increased prolyl-4-hydroxylase (PH) activity in skeletal muscle and heart without an increase in collagen concentration (11, 27), suggesting a higher collagen turnover rate. As a result of the increased rate of both collagen deposition and degradation, collagen cross-linking decreases in both skeletal muscle and heart from old trained animals (28, 33). It has been hypothesized that a reduction in collagen cross-linking in skeletal and cardiac muscle might contribute to decreased passive stiffness (28, 33); however, this hypothesis has not been tested. The purpose of this study was to determine the effects of endurance exercise training on locomotor muscle passive stiffness in both young and old rats. Our hypothesis was that exercise training of muscle in old animals, with its resulting decrease in collagen cross-linking, would similarly attenuate the age-associated increase in passive muscle stiffness.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals. Young (age 3 mo) and old (age 23 mo) male Fischer 344 rats were obtained from the National Institute on Aging (Bethesda, MD) and housed according to State University of New York at Buffalo Institutional Animal Care and Use Committee guidelines. Rats were maintained on a 12:12-h light-dark cycle and were provided with water and commercial rat chow ad libitum.

Training protocol. The young and old rats were randomly assigned to either a sedentary control [young control (YC), n = 9; old control (OC), n = 8] or exercise training group [young trained (YT) n = 9; old trained (OT) n = 8]. The training protocol was similar to that previously described (7). Each animal was habituated to treadmill walking (8-10 m/min) over a 1-wk period. Exercise training consisted of progressive treadmill incline (15%) running such that, by the end of 10 wk of training, the animals were running at ~70% of maximal oxygen consumption for 45 min/day, 5 days/wk (18).

Measurement of active muscle mechanical properties. A dual-channel length-force servo-control system (model 305B, Cambridge Technologies) was used to measure both the isometric contractile properties and the passive force-length relationship of the Sol muscle. The Cambridge system has a 500-g force transducer with a resolution of 100 mg. The servo arm has a movement excursion of ±8 mm with a resolution of 1 µm. A custom-made computer program (LabVIEW, National Instruments) controlled muscle length during measurement of passive stiffness.

Measurement of passive muscle mechanical properties. The muscles were sinusoidally oscillated between 85 and 115% of estimated optimal fiber length (L_f) (i.e., ±15% L_f) for a period of 80 s at a frequency of 1 cycle/s. L_f was calculated as two-thirds of muscle L_o, based on results from previous studies (1, 24). Small-amplitude oscillations (±0.5% L_f) at a frequency of 75 Hz were superimposed on the larger oscillations, allowing for the determination of muscle stiffness at different muscle lengths. The output signals of length (Fig. 1A) and force (Fig. 1B) were acquired at a sampling frequency of 600 Hz by a custom-made computer data-acquisition program (LabVIEW, National Instruments). The passive force developed during mechanical cycling was normalized for muscle cross-sectional area (CSA) at the given fiber length. Muscle stiffness was calculated at 111 and 115% L_f from the small-amplitude oscillations. All measurements were made on the final (80th) cycle. Stiffness (stress/strain) was calculated with the equation, F/(L/L), where F is the change in force (N/cm²), L is the change in length (cm) induced by the small oscillations, and L is the muscle fiber length (cm). The changes in length and force were determined by filtering out the slower, larger oscillations (detrending) and then measuring the amplitude of the smaller oscillations. This procedure was validated by cycling a rubber band under similar conditions and simultaneously measuring force with the Cambridge servomotor and a stationary force transducer (Grass FT03C). CSA of the muscle segment was estimated based on the calculation, CSA = MW/(L_f × 1.056), where MW is the weight of the muscle segment and 1.056 is the density (g/cm³) of muscle (21). After completion of the in vitro studies, the Sol muscle was then frozen in liquid nitrogen and stored at 70°C for analysis of collagen concentration and cross-linking.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A: length changes of soleus muscle lengthened and shortened ±15% of optimal fiber length (Lf) at 1 Hz. A smaller (±0.5% Lf) oscillation at 75 Hz is superimposed on slower, larger oscillation. B: force changes of soleus muscle during change in muscle length.

Collagen hydroxyproline (HYP) assays. HYP assays were carried out on the Sol muscle (including tendons) from all four groups of rats. The HYP concentration was measured colorimetrically by using the method of Woessner (31). Collagen concentration of each tissue sample was calculated with the assumption that collagen weighs 7.25× the measured weight of HYP (31).

HP cross-link assays. A portion of the prepared filtered acid hydrolyzate was used for HP cross-link analysis. The hydrolyzate was prepared for HP cross-link analysis by elution from a CF-1 cellulose (Sigma Chemical) column by using the procedure described by Skinner (25). The remaining hydrolyzate was concentrated, suspended in a slurry of CF-1 cellulose and n-butanol-acetic acid-water (4:1:1) mobile phase, and applied to the CF-1 cellulose column. HP, which absorbs onto organic acid-alcohol mixtures, was separated from the other amino acids and impurities eluting from the column with several column volumes of organic acid-alcohol mobile phase. HP was then eluted with water, and the fraction was dried in a Savant Speedvac for HP cross-link analysis. An aliquot of the dried hydrolysate from the CF-1 cellulose column was resuspended in 1% n-heptafluorobutyric acid (HFBA) containing pyridoxamine dihydrochloride (Sigma Chemical) as an internal standard. Pyridoxamine was used as an internal standard because of its structural and fluorescence similarities to HP. HP and pyridoxamine were detected by fluorescence spectrophotometry by using excitation at 295-nm wavelength and emission at 395-nm wavelength. The isolation and quantification of the HP cross-link and pyridoxamine were accomplished by reverse-phase, high-performance liquid chromatography. Resolution of HP and pyridoxamine was obtained by using a binary mobile phase of 15% acetonitrile in water containing 0.01 M HFBA and 85% acetonitrile with 0.01 M HFBA. Identification of the HP peak was confirmed by comparison with a pure HP standard, which we routinely prepare from bovine articular cartilage using the procedures described by Eyre et al. (5). Collagen HP cross-links were expressed as moles of HP per mole of collagen, assuming that pyridoxamine has a molar fluorescence yield 3.1× that of HP and that the molecular weight of collagen is 300,000 (31).

Statistical analyses. Group data were analyzed by using a two-way analysis of variance with post hoc (Bonferroni) analysis. P < 0.05 was considered statistically significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Before training, same-aged groups were matched for body weight to evaluate the effects of the exercise program on whole body (BW), left ventricular (LV), and Sol muscle weights. A training effect was documented in both age groups of trained rats (YT, OT) compared with their sedentary counterparts (YC, OC) by attenuated weight gain and a relative hypertrophy of both LV and Sol muscles (Table 1). Maximal isometric-specific force was significantly lower (P < 0.05) in the Sol muscle of old compared with young rats and was unaffected by training (Table 1).

Stiffness of the Sol muscle for each group is illustrated in Fig. 2. At both 111 and 115% L_o, passive stiffness was significantly higher in the OC group compared with the YC and OT groups (Fig. 2, A and B; P < 0.05) but not with the YT group. A similar trend was observed at 107% L_o. Exercise training resulted in a significant decrease in the passive stiffness of the Sol muscle from old but not young adult animals.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Passive stiffness (N/cm2) of soleus muscle at 111% (A) and 115% (B) of optimal length from young control (YC), young trained (YT), old control (OC), and old trained (OT) groups. Bars are means ± SE. * OC significantly different from YC and OT, P < 0.05.

Measurements of collagen concentration and cross-linking in the Sol muscle are shown in Figs. 3 and 4, respectively. Collagen concentration significantly increased in Sol muscle as a consequence of age. Exercise training had no effect on collagen concentration in the Sol muscle in either age group. There was a significant age-by-training interaction with respect to Sol muscle HP content. Intergroup comparisons revealed an age-associated increase in HP in OC compared with YC Sol muscle (0.77 ± 0.03 vs. 0.30 ± 0.03 mol HP/mol collagen; Fig 4). In the young rats, there was no training-associated change in HP content. In contrast, the OT Sol muscle HP content was significantly lower than that observed in the OC muscle (P < 0.0001).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Collagen concentration (means ± SE) of soleus muscle from YC, YT, OC, and OT groups. * Old significantly different from young, P < 0.05.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Collagen hydroxylysylpyridinoline (HP) cross-link concentration (means ± SE) of soleus muscle from YC, YT, OC, and OT groups. * OC significantly different from YC, YT, and OT; P < 0.05.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The collagen data from the present study are consistent with previous reports. In both heart and skeletal muscle, aging results in an increase in collagen concentration and cross-linking (8, 28, 33). For example, when compared with young adult rats, collagen cross-linking is ~357, 27, and 380% higher in Sol muscle (33), diaphragm muscle (8), and LV free wall (28), respectively, of old rats. The increase in skeletal muscle collagen concentration seen with aging is not accompanied by a change in the activities of PH or galactosylhydroxylysine glucosyltransferase (11), two enzymes involved in posttranslational modification of collagen. Moreover, the fractional synthesis rate of collagen in rat skeletal muscle decreases ~10-fold, from 1 to 24 mo of age (17). These results suggest that the increase in collagen concentration observed in skeletal muscle from old animals is a result of a decreased rate of degradation out of proportion to the reduced biosynthetic activity.

The training protocol used in this study is similar to protocols used in other studies (7, 33) that resulted in significant increases in LV mass, maximal oxygen consumption, and locomotor muscle oxidative enzyme potential. In the present study, we observed a significant increase in the LV/BW and Sol/BW ratios in both YT and OT groups, indicative of a training effect. Ten weeks of endurance treadmill running resulted in a significant decrease in collagen cross-linking in both senescent skeletal (33) and heart muscle (28). Because exercise training is associated with increased activity of PH and galactosylhydroxylysine glucosyltransferase in skeletal muscle (11) and heart (27), the reduction in collagen cross-linking in muscles of OT rats is presumably because of increased turnover rates of fibrillar collagen as a consequence of increased muscle recruitment.

We hypothesized that the reduction of collagen cross-links in trained muscle of old rats would significantly attenuate the increase in muscle stiffness. In the present study, muscle stiffness was significantly decreased in the OT compared with the OC group but was unchanged in the YT compared with the YC group. The reduced muscle stiffness in the OT group is consistent with the diminished content of HP cross-links seen in muscle from this same group. Our results differ from those reported by Kovanen and Suominen (10), who noted an increase in the passive stiffness of Sol muscles of OT rats. It is unclear why our results differ, but it may be due to a number of reasons. One possibility is that different training regimens were used (10 wk in our study vs. up to 2 yr in the study by Kovanen and Suominen). Because exercise intensity was not reported in the study by Kovanen and Suominen, it is not possible to comment on this aspect. We saw a significant reduction in the level of HP cross-links, whereas Kovanen and Suominen did not directly measure HP cross-links, but rather measured the amount of salt-soluble collagen. With the assumption that the level of HP cross-links corresponds to the amount of salt-soluble collagen, the training study by Kovanen and Suominen failed to result in a significant change in HP cross-links. Finally, all our mechanical measurements were initiated from L_o. Because Kovanen and Suominen initiated all measurements from an initial resting tension of 0.01 N rather than from L_o, it is unclear whether all muscles were at the same sarcomere length. A slight change in resting tension (or muscle length) can significantly alter the passive length-force curve.

The passive mechanical properties of skeletal muscle are similar to other soft tissues such that the passive load deformation curve is approximately exponential (9, 26). A number of variables may contribute to passive tension both at L_o and at lengths above L_o. Studies employing isolated single fibers from frogs suggest that most of the resting tension is borne within the myofibril (15) and that the protein titin is involved in this phenomenon (30). However, in another study in which a muscle was passively stretched, the force-length curve was correlated with the amount of intramuscular connective tissue (9). In addition, differences in stress/strain properties have been noted between muscles with differing functional and collagenous properties (13). For example, slow-twitch muscle (Sol) has a greater collagen concentration and more extensive cross-linking of collagen than does a fast-twitch muscle (rectus femoris) (12). These biochemical properties of slow-twitch muscle correlate with a higher ultimate tensile strength (i.e., the point at which a muscle tears when stretched) than those of fast-twitch muscle. Additionally, fast-twitch muscle is capable of greater strain (i.e., change in length before mechanical failure) than is slow-twitch muscle (13). An even stronger argument for the importance of collagen cross-linking to tissue viscoelastic properties comes from studies in which the rate-limiting enzyme lysyl oxidase is inhibited by the presence of -aminopropionitrile, thereby inducing lathyrism. Kovanen et al. (13) reported that both muscle tensile strength and stiffness significantly decrease under these conditions. Thus, although certain intracellular proteins no doubt contribute to passive viscoelastic properties of skeletal muscle, the extracellular matrix, and in particular collagen cross-linking, also has an important role in determining these parameters.

Despite the fact that OT animals were exercising at a much lower absolute workload than their younger counterparts, the training effect on Sol muscle HP cross-links was seen only in the older rats and not in the younger group, a finding observed previously (28, 33). These findings suggest that there is an optimal level of HP cross-link required for normal function of the muscle, and the level is maintained in skeletal muscle of young rats undergoing daily endurance exercise. Little is known, however, about the mechanisms that regulate collagen cross-linking in skeletal muscle. In cases of extreme copper deprivation or lathyrism, lysyl oxidase enzyme activity may be limiting. In muscle in which remodeling occurs, we have hypothesized that altered collagen fibril architecture, mediated by a small proteoglycan like decorin, may be a key collagen-binding regulator of cross-linking (20). It is also possible that mechanical load plays a role, because muscles with different functional properties have different levels of the HP cross-link (33).

It is presently unclear how much exercise (in terms of duration, frequency, and intensity) is required to achieve a reduction in muscle stiffness and HP cross-links in locomotor muscle from aged rats. The functional significance of reducing muscle stiffness in aging muscle by exercise training also has yet to be clarified. Because some studies have reported an increased susceptibility to exercise-induced injury in aging muscle (1, 32), one possible consequence of decreased stiffness may be reduced susceptibility to exercise-induced injury, particularly as a consequence of lengthening contractions. However, additional studies are required to test this hypothesis.

In conclusion, both muscle stiffness and the concentration of the mature collagen cross-link HP increases in Sol muscle of aging rats. However, regular exercise training can alter the passive mechanical properties of senescent Sol muscle, and this change is associated with concomitant changes in collagen cross-linking. These findings highlight the potential role of collagen in influencing the passive viscoelastic properties of skeletal muscle.