| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Division of Orthopaedics, Department of Neuroscience and Locomotion, Faculty of Health Sciences, Linköping University, Linköping, Sweden
Submitted 23 November 2006 ; accepted in final form 2 March 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
loading; hysteresis; unloading; creep
The overall effects of both exercise and disuse on tendons are much slower than the effects on muscle or bone, probably because of less vascularity and slower metabolism (21). Earlier studies on disuse have shown that the collagen appears disorganized after disuse and that tensile strength and stiffness are reduced (11, 26).
The tendons mainly consist of collagen fibrils and proteoglycans. The collagen fibrils are stiffer than whole tendons, and most of tendon strain and mechanical behavior therefore takes place in the proteoglycan-rich interfibrillar ground substance, which embeds the collagen fibrils (19). This view, however, appears to conflict with morphological data on fibril length (18). Glycosaminoglycan (GAG) chains interconnect the collagen fibrils at proteoglycan binding sites, and proteoglycan seems to mediate the organization of the tendon ultrastructure (15). This proteoglycan substance is likely to have damping properties and function as a shock absorber. The extracellular matrix (ECM) is under constant remodeling, in which matrix metalloproteinases (MMPs) play an important role (13). The MMPs are involved in collagen as well as proteoglycan degradation. When loading conditions are changed, the ECM is adapted to meet the demands of the new situation (20). For example, if compressive load is applied, the tendon forms fibrocartilage in direct relation to compressive strain distribution (4).
There are hardly any studies of the changes in viscoelastic properties in the tendon after disuse. However, increased stress relaxation of collagen fascicles has been shown after stress shielding in a surgical rabbit model (25). We made several attempts to demonstrate a reduction in tendon strength following unloading of the Achilles tendon in rats by botulinum toxin injections in the calf muscles. We were unable to find any reduction despite almost total muscle atrophy. However, we found that creep was reduced after disuse and designed the present study to confirm this finding. The aim of this study was therefore to evaluate how disuse affected the mechanical properties (primarily creep and hysteresis), as well as the proteoglycan content in the Achilles tendon. Presumably, disuse leads to proteoglycan and collagen degradation by MMPs because of less demand for them. The second aim was therefore to evaluate if systemic treatment with a broad-spectrum MMP inhibitor, doxycycline (5, 13), at a dose that is known to disturb tendon healing (12), could prevent changes that may arise after disuse of the tendons.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The animals were anesthetized with isoflurane gas (Forene, Abbott Scandinavia, Solna, Sweden), and the skin was shaved on the left hindleg. Botulinum toxin type A (1 U; Botox, Allergan, Irvine, CA) was injected into the midportion of each of the two gastrocnemius muscle heads and into the midportion of the soleus muscle, giving a total amount of 3 U botulinum toxin per animal. Thereafter the rats were allowed unrestricted activity. Half of the botulinum toxin-treated group was given doxycycline hyclate (Sigma, St. Louis, MO), 80 mg·kg–1·day–1, in their drinking water, starting at day 0. This gives a clinically relevant serum concentration (12). Day 0 was defined as 3 days after botulinum toxin injection, when a complete paralysis has developed. Water bottles were weighed every day to confirm that a proper amount of medication was given to the doxycycline-treated group. During the ongoing experiment, we examined the pliability of the muscle by dorsal flexion every week, including at time of death, to exclude muscle contracture, which could affect tendon biomechanics. We also observed rats weekly for signs of muscle paralysis, e.g., inability of plantar flexion, to confirm that the effect of the botulinum toxin treatment was permanent throughout the study period. Biomechanical and biochemical evaluation was performed after 1 and 6 wk, and the selection of rats for each follow-up time was random.
Mechanical testing. On the evaluation day, the animals were killed with CO2 gas. The tendons were dissected free from extraneous soft tissue and harvested together with the calcaneal bone and parts of the gastrocnemius and soleus muscle complex. Sagittal and transverse diameters were measured with a slide caliper halfway between the calcaneal insertion and the most distal muscle. The cross-sectional area was calculated assuming an elliptical geometry. During tissue preparation and mounting in the materials testing machine, the tendons were kept moist using gauze with physiological solution.
For clamping, the muscle was carefully scraped off the proximal tendon by blunt dissection to produce a fan of tendon fibers. These tendon fibers were attached between fine sandpaper and fixed in a metal clamp. In the other end, the calcaneal bone was fixed in a custom-made clamp, in 30° dorsiflexion, relative to the direction of traction. The distance between the clamp and the calcaneus was measured as an estimate of tendon length. The tendons were finally mounted vertically in a materials testing machine (100R, DDL, Eden Prairie, MN) and tested at first with a cyclic phase and afterward a monotonic phase with stretching until failure. The machine pulled at a constant speed of 0.1 mm/s. The cyclic phase consisted of 20 cycles between 1 and 20 N (Fig. 1). Data were transferred to an Excel sheet, and hysteresis was measured as the sum of the areas under all curves for increasing strain minus the areas under all curves for decreasing strain. The difference was normalized to the increasing strain curves to define hysteresis as percentage mechanical energy loss. We also measured the hysteresis of the 20th cycle, regarding the preceding cycles as a conditioning phase. Creep was defined as the increase in specimen length (in %) from the first to the last (20th) loading of 1 N. Tension to failure was done immediately after the cyclic testing. Peak force (N), energy uptake until the curve had dropped to 90% of the highest point, and stiffness (N/mm) were calculated by the software of the testing machine after the investigator had marked a linear portion of the elastic phase of the curve for stiffness calculation. As a secondary check, the investigator placed a line parallel to the curve at 20 N to get a measurement for slope at that point. This measurement yielded identical information as the other definition of stiffness and is therefore not reported.
|
GAG measurements. GAG was quantified by dimethylmethylene blue (DMMB) assay for sulfated GAG and normalized to tissue dry weight (3). The tendons were lyophilized for 24 h, weighed, and diced. Papain (500 µl; Sigma) buffer solution (125 µg/ml in 1x PBE, where PBE is phosphate-buffered saline with 5% BSA and 5 mM EDTA) was added to the tendon pieces, and 150 µl solution was added to the insertions. The samples were incubated in 60°C for 18 h, or until everything was dissolved. The samples were then carefully mixed with a vortex until there was a homogeneous solution. Forty microliters of sample (diluted or concentrate) was pipetted into triplicate wells of a 96-well plate, and 200 µl DMMB solution (SERVA Electrophoresis, Heidelberg, Germany) was added. The plate was immediately read at 525 nm and compared with a standard curve made from chondroitin sulfate C (Sigma).
Statistical analysis. Data were analyzed using two-way ANOVA with time and treatment as independent factors (Statview 5.01 for Windows). Only if there was a significant interaction between these did we compare treatment effects at the time points separately. P values refer to the ANOVA. Post hoc testing using Fisher's protected least significant difference test was done for all variables to define which groups differed (i.e., control vs. disuse, control vs. disuse plus doxycycline, or disuse vs. disuse plus doxycycline). All statements in RESULTS refer to significant effects with P < 0.05 if not otherwise stated. Results are presented as means (SD).
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
90% (unpublished observation). In the present study, the muscles appeared similarly atrophic. There were no limitations of the range of passive motion at death, and the muscle remnants appeared fully pliable. Difficulties during mechanical testing, mainly slippage or tear at the clamp, were the only reason for exclusions: three rats were excluded in the control group after 1 wk and one after 6 wk. Two rats were excluded in the doxycycline group after 1 wk and three after 6 wk. No exclusion was made in the unloaded group. Two-thirds of the tendon specimens failed at the bony insertion and most of the others at the proximal clamp. Failure mode was not influenced by treatment. The body mass of the rats had increased by 22% from 1 to 6 wk. The increase was slightly less in the two botulinum toxin groups (Table 1).
|
Disuse mainly affected viscoelasticity. Disuse led to a 19% decrease (P = 0.004) in hysteresis (all cycles). This was seen already after 1 wk and did not change further at 6 wk. The first loading cycle showed the largest viscoelasticity. Analyzing cycle 20 alone, there was still a significant reduction in hysteresis (P = 0.01). Creep was also decreased after disuse (P = 0.007). It was reduced by 12% after 1 wk, and there was also a time-dependent increase in reduction to reach a 15% decrease after 6 wk. Doxycycline had no effects on these changes (i.e., there was no difference between disuse and disuse plus doxycycline groups). Percent water content and normalized GAG content were unaffected by both time and treatment.
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Both elastic modulus and hysteresis of ankle extensor tendons appeared to be independent of animal size in a study on 18 mammal species, ranging from 0.04 to 500 kg body wt (17). The elastic modulus in our rats was only one-third of the average in that species comparison. This can have several explanations. There may be methodological errors leading to overestimation of transverse area and strain, the latter due to clamp slipping. However, the values had a small variation. It is more likely that the low elastic modulus may be due to some properties of our rats. Because they are growing, the tendons are immature, and much of the collagen is new and less cross-linked. It is well recognized that immature tissues have a lower modulus (14, 23). In a study on the rat Achilles tendon, where modulus was measured in the midsection, it was found similar to ours (6). Hysteresis in our study, on the other hand, was similar to that of the above-mentioned study on a wide range of species (17). It should be noted that, as in other studies (7), our measurement of hysteresis was done before the hysteresis loop was entirely stable (Fig. 1). In an additional experiment on nine similar rats with 80 cycles, reaching stability, no significant difference in hysteresis between cycles 20 and 80 could be detected (data not shown). From a practical standpoint, therefore, the value from cycle 20 represents a tissue property suitable for group comparisons and very closely related to what a perfect method would have yielded.
In contrast to the material properties, the actual storage of energy in the ankle extensor tendons during locomotion is dependent on species size, because in large animals these tendons experience much higher stress than in small animals (16). Therefore, energy saving in tendon during locomotion may be unimportant in the rat. Still, rat Achilles tendons show a low hysteresis. It seems relatively "easy" for the body to produce a tendonlike structure with a low hysteresis. As an example, we found a hysteresis as low as 15% in immature tendon regenerates 3 wk after Achilles transection in sheep (data not shown). Possibly, the production of a tissue with a higher energy dissipation capacity is a more complex process, which requires mechanical input.
Most animal models for evaluation of the effect of disuse on tendon properties use external fixation or surgical interventions such as denervation and disarticulation (1, 9, 11, 26, 27). These are often demanding procedures, and immobilization could have undesired side effects due to loss of joint movement. Muscle disuse also results in atrophy, which dramatically reduces the size of the muscle. Therefore, an immobilizing cast has to be changed regularly (11). Denervation of the muscle also denervates the tendon and the surrounding tissues responsible for its nutrition. This could influence tendon metabolism. We believe it is better to paralyze the gastrocnemius and soleus muscle complex with botulinum toxin to create disuse without surgery or external fixation devices. This is a less time-consuming procedure and probably results in fewer side effects on the animals. However, there is a possibility that the tendons in our rats have been strained on rare occasions. This would occur when the knee was fully extended and the ankle joint dorsiflected toward 90°. In this position, the musculotendinous complex was taut at the time of specimen harvest, as well as in normal anesthetized rats. This is an unphysiological position, that would rarely occur in real life, but it cannot be fully excluded that such events are responsible for the observed maintenance and growth of the tendon.
The botulinum toxin treatment was efficient, and after 6 wk the gastrocnemius and soleus muscle complex had decreased to a negligible size. Moreover, the muscle remnants appeared to remain fully pliable, as tested by dorsal flexion of the foot. So why did not the rat Achilles tendons deteriorate in strength and stiffness on disuse, as do rabbit or human tendons (11, 20, 26)? Moreover, we have previously found in models for tendon repair that similar botulinum toxin treatment had a drastic negative effect on tendon callus strength: 2 wk after Achilles tendon transection in rats, force at failure was reduced to <30% of transected but loaded controls (22). In the present experiment, the tendons also continued to grow. This may be the clue: The results suggest that as long as the animals grow (and rats always grow), tendon size and strength are under systemic control and require little mechanical stimulation for their maintenance. However, hysteresis and creep were negatively affected by disuse, and modulus increased, suggesting that these properties are always dependent on loading history.
Damping probably takes place in the noncollagenous material in the interfibrillar space. At least in rats, the response to unloading obviously occurs first or most in this space, reducing the damping properties, so that the more purely elastic spring properties of the undisturbed collagen fibrils become prominent. Also in rabbit patellar tendons there is evidence to suggest a stronger response to unloading in the ground substance than in the collagen (24). This loss of damping properties may lead to less efficient strain dissipation and an increased risk of localized microinjury after suddenly vigorous movements. Earlier results have also shown that it takes a long time to recover after even a short time of tendon disuse, and it appears that complete stress shielding leads to a longer time required for recovery than immobilization (27). It is possible that studies of the metabolic response to tendon disuse would miss important information if they concentrate on collagen only. However, our measurements of gross GAG content were inconclusive. Earlier studies of GAG after immobilization have shown a decrease (27). We saw no effects of doxycycline. Previously, we have been able to show a negative effect of doxycycline on rat Achilles tendon repair with similar experimental setup and dosing. In that study we also found adequate serum levels of doxycycline (12). Therefore, the results suggest that MMP activity is not crucial for the observed viscoelastic changes.
As in earlier studies, it was hard to get reliable peak force values of the tendon, since very few specimens failed at midsubstance (11, 27). This makes it difficult to evaluate whether there was a reduction in failure stress in the tendon substance after disuse. However, the clear increase in peak force from 1 to 6 wk and that they ruptured at the calcaneal bone rather than at the clamp suggest that the measurements are valid. Moreover, size, stiffness, and modulus are all unaffected by failure mode. It should be pointed out that the statistical analysis in this study is based on treatment or time effects in the whole data set and does not separately compare each group with each other one, except when there was a significant dependency between the two factors.
Our findings contrast with earlier studies in several species, including humans, showing reduced stiffness after unloading. With regard to the Achilles tendon in rats, hindlimb suspension for 3 wk led to a reduction of modulus by one-half (1). Very recently, however, Arruda et al. (2) reported that hindlimb denervation increased the modulus of rat tibialis anterior tendons, without affecting collagen content or concentration (2). This is in line with our findings. They suggest that the discrepancy between their results and those of others is due to methodological errors in previous experiments, mainly related to variations in the response to unloading along the length of the tendon. Our study mainly describes the free portion of the Achilles tendon and not the aponeurosis. The aponeurosis and the Achilles tendon may adapt differently to the loading history. Indeed, all changes induced by unloading in a recent human study occurred in the aponeurosis only (7). Two other studies attempted to show improved biomechanical properties of the rat Achilles tendon after training, and both were unable to find any effect on tendon size or modulus (6, 8). However, Legerlotz et al. (8) found increased creep with high strength training, which is an unexpected finding, in line with ours. In rabbit patellar tendons, unloading with a cerclage wire led to fast and dramatic swelling and a reduction in modulus and failure stress, although bulk properties may have been unchanged (26). However, we have been unable to find such effects in rats with similar unloading by cerclage (data not shown). Human volunteers subjected to prolonged bed rest showed 58% reduction in Achilles tendon stiffness after 90 days (11). A main difference between these studies and ours is that we used rats. It seems, however, unlikely that the mechanical control of tendon behavior should be so drastically different between mammalian species. Rather, we believe that the important difference is that rats continue to grow their whole life. Probably, hormonal or other growth signals override the effects of loading history, with exception for damping properties. This can be tested by studies on growing and adult animals of a species that does not continue to grow.
Tendinosis might be initiated by microdamage due to localized fatigue. It is therefore likely that the condition starts at sites with local stress concentration. It is possible that damping dissipates the mechanical energy of an impact load over time and also evens out the local stress between different sites within the tendon. Our results suggest that the genetic design for growing tendons provides a high energy-storage capacity but low damping ability. Under some conditions, the effect of loading or training might then rather be to provide signals that lead to increased hysteresis and better dashpot function. If untrained people's tendons are less damped and have a higher modulus than the athlete's, this might perhaps explain why sudden efforts to exercise so often lead to tendon symptoms in untrained people.
In conclusion, disuse decreased hysteresis and creep, and increased modulus in our model. The findings contrast to previous results, but if their validity can be confirmed in other models, they might open for new insights in the effects of training.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
