Annals of Biomedical Engineering, Vol. 39, No. 5, May 2011 ( 2011) pp. 15351545 DOI: 10.

1007/s10439-011-0254-0

Low Stress Tendon Fatigue is a Relatively Rapid Process in the Context of Overuse Injuries
, and EVE LANGELIER GABRIEL PARENT, NICOLAS HUPPE

de Sherbrooke, Sherbrooke, QC J1K 2R1, Canada PERSEUS, Department of Mechanical Engineering, Universite
(Received 21 October 2010; accepted 14 January 2011; published online 2 February 2011) Associate Editor Michael R. Torry oversaw the review of this article.

AbstractTo stimulate healing and prevent tendinosis through optimized physical exercise, it is important to elucidate the tendon response to repetitive mechanical loading. However, the study of this response is challenging due to complex cellmatrix interactions. In an initial approximation, the authors examined tendon mechanical response only, and did not consider cellular activity. The authors investigated the hypothesis that mechanical degradation occurs relatively rapidly (<24 h) even at very low stress levels. The authors subjected rat tail tendons to mechanical loadings oscillating between 0 and 1.5 MPa up to one of three fatigue levels: 4% strain, 8% strain, or rupture. Using non-destructive mechanical tests, changes in tendon strain and compliance over the entire fatigue testing period were evaluated. Using microscopy techniques, the structural evidence of mechanical degradation was examined. The changes in tendon strain and compliance progressed nonlinearly and accelerated before rupture which took place around the 15-h mark. Histological analyses revealed a higher degree of alteration in the collagen network at increased fatigue levels. At rupture, local zones of damage with low bril density were evident. These results imply that a repair process must act rapidly at critical sites; otherwise, enzymatic degradation could cause further damage in the manner of a vicious cycle. KeywordsMechanical degradation, Biomechanics, Strain, Compliance, Histology.

INTRODUCTION Tendinopathy is a common health problem that often results from occupational or sports overuse injury. It usually takes the form of tendinosis, an asymptomatic degeneration without inammation that exposes the tendon to chronic symptomatic

Address correspondence to Eve Langelier, PERSEUS, Depart de Sherbrooke, ment of Mechanical Engineering, Universite Sherbrooke, QC J1K 2R1, Canada. Electronic mail: eve.langelier@ usherbrooke.ca

degeneration.17,30 Tendinosis can also lead to spontaneous tendon rupture.17 To date, researchers have been unable to identify a treatment that would foster adequate healing of tendinosis.2 However, one very encouraging treatment is eccentric strengthening,2 which capitalizes on mechanobiology principles, i.e., the tissue response to mechanical loading. According to these principles, mechanical loading applied to the tissue is perceived by the cells as cellular deformation,14 nuclear deformation,4,31 or uid ow.6 These local stimuli signal the cell to remodel the tissue, i.e., to structurally reorganize it. In response, the cell synthesizes or destroys (via protease synthesis and activation) the molecules of the extracellular matrix. The new matrix will have different mechanical properties, thus modifying the local stimuli at the cell level. Finally, matrix damage can also occur in response to excessive stress. Therefore, eccentric strengthening capitalizes on the dynamics of repair and degradation in the extracellular matrix by attempting to maximize the repair process via mechanical loading. This, however, is a very ne level of optimization. To prevent tendinosis and encourage healing via optimized physical exercise, it is important to elucidate the tendons response to repetitive mechanical loading. An interesting approach is to design a mechanobiological model of tendon repairdegradation dynamics. This type of model will allow us to predict the tissues condition in relation to its loading history. However, designing this model is a complex task, due to interactions between the cells and the extracellular matrix. The initial conservative approach is to examine tendon response as a material only, i.e., without cellular activity. This approach highlights the mechanical degradation process in tendons. This process plays an important role in tendon mechanobiology because it alters signal mechanotransduction pathways, and thus cell response. For
1535
0090-6964/11/0500-1535/0

2011 Biomedical Engineering Society

1536

PARENT et al.

example, a high level of mechanical degradation could limit local stimuli at the cell level and thus induce enzymatic degradation.3,23 Knowing the rate and structural evidence of mechanical degradation would, therefore, be useful for developing more complex models. To date, many research groups have studied the mechanical degradation process with tendons under cyclic mechanical stress.11,29,33,35 Unfortunately, because our current knowledge of the effect of specimen preservation and testing conditions on mechanical degradation is incomplete, certain questions may arise regarding the experimental results: (1) the studies were conducted on frozen-thawed tendons29,33,35 which may have altered mechanical properties, such as ultimate tensile strength,8,12,26 ultimate tensile strain,12 and Youngs modulus;8,26 and (2) the studies were conducted at room temperature29,33 far below body temperature, which has been shown to inuence both creep time to rupture32 and the mechanical properties of connective tissues such as storage modulus,7 cyclic load relaxation,20,34 and hysteresis.34 Moreover, temperature has been shown to inuence viscoelastic properties of polymers.25 However, the study by Fung et al.11 showed that cyclic loading at approximately 35% of the amplitude of the ultimate tensile strength (UTS) leads to rupture in about 15 min,11 which is very rapid. In the quest to prevent and heal tendon damage, it is also important to study mechanical degradation in the lower range of normal activities. For tendons, the load range for normal activities seems to vary between tendons as reviewed by An.1 Other authors have suggested that the physiological loading range for tendons lies inside the toe region of the stressstrain curve,10 or under one-eighth of the tendon UTS.19 Therefore, in this study, stress under 5% UTS was considered to be at the lower range of normal activities. This article focuses on the mechanical degradation process of the tendons mechanobiological response to repetitive loading. The authors hypothesized that (1) mechanical degradation occurs relatively fast (time scale <24 h) even at very low stress levels; (2) compliance amplitude increases with increasing fatigue levels; (3) alterations to the collagen network increase with fatigue levels. The authors investigated these hypotheses in vitro. The authors submitted rat tail tendons to mechanical fatigue inside a bioreactor and measured the changes in tissue compliance. Using three microscopy techniques offering different magnication ratios and structural

information, the authors examined the structural demonstration of mechanical degradation. With its emphasis on low stress mechanical degradation, this in vitro study complements the study of Fung et al.11 on higher stress mechanical degradation and Arcnozky et al.3 on under stimulation of tendon cells. Together, the results of these studies will help build an initial conservative model for tissue degeneration and elucidate the complex dynamics between degradation and repair, which result either in tendon homeostasis, degeneration, or repair under repetitive mechanical loading. The results will, therefore, contribute to fundamental research in developing prevention and healing strategies for tendinopathies.

MATERIALS AND METHODS Tissue Explant Isolation and Preparation5 All animal care and handling procedures were approved by the Animal Care Committee at the Uni de Sherbrooke. Four Sprague Dawley rats versite 46 months old were sacriced using carbon dioxide asphyxiation. All manipulations presented in this section were conducted in cold D-PBS 19 (311-410-CL; Wisent Inc., St. Bruno, Canada) containing 1% antibioticantimycotic (15240-062; Invitrogen Corporation, Burlington, Canada). Seven tendons (6 for the stimulation protocol plus 1 as histological control on day 0) were isolated from each rat tail within an hour of resection (Fig. 1). The tendons located on the ventral side of the tail were isolated using the following procedure: rst, the skin was cut along the tail between the dorsal tendon groups and removed. The distal tail end was cut 23 vertebrae shorter using a scalpel. With each group exposed, all the tendons were gently removed with forceps. For each vertebra removed, six tendons were extracted, four of which were associated with the ventral side of the tail. The distal tail end was then cut one vertebra shorter to expose new tendons, and tissue removal was repeated. Following isolation, the tendon cross-sectional areas were evaluated using an optical micrometer27 along with the prole reconstruction algorithm21,24 (areas = 6.09 0.02 9 1022 mm2). They were then rinsed ve times under the biosafety cabinet. For mechanical characterization and stimulation, the tendons were transferred into custommanufactured bioreactors, essentially multi-specimen traction machines for independent characterization and long-term stimulation of six specimens in tissue culture conditions (37 C, 95% humidity). (Depending on the source, the authors found rectal temperature in rats to vary between 36.8 and 39 C.13,16) In this

Low Stress Tendon Fatigue

1537

FIGURE 1. Number and distribution of samples for each of the four rats.

system, the force was measured using a load cell, and the displacement was measured using an optic encoder. The tendon extremities were wound around cylinder-shaped anchors and allowed to dry briey on the top face of the anchors, where a small drop of ethyl cyanoacrylate (10300; Krazy Glue, Columbus, USA) was applied. The tendons were different lengths at the time of harvest due to tail anatomy. During mounting, the tendon was installed to obtain a distance of 60 2 mm between the two anchors. For mechanical characterization, elongation was normalized using this distance. Tissue Culture Conditions During stimulation, the tendons were all maintained at 37 C in D-PBS containing 1% antibioticantimycotic and protease inhibitors: 1 lg/mL Aprotinin, 1 lg/mL Leupeptin, 1 lg/mL Pepstanin A, and 1 mM PMSF (APR600, LEU001, PEP605, and PMS123; Bioshop Canada, Burlington, Canada).9 Bioreactor Design Two bioreactors were used in this study for the stimulation and characterisation of 6 tendons simultaneously. Each bioreactor is contained inside an incubator (Fig. 2a). A hermetic frame supports the testing compartments, and the instrumentation composed of load cells (Tedea Huntleig 1022 3 kg, Vishay Precision Group, Malvern, USA) and optical encoders (EM1 LIN 50-6, USDigital, Washington, USA). A hole was made in the incubators wall to allow more space for the actuators. The testing compartment maintains the tendon in culture medium (Fig. 2b). It is made of biomaterials
FIGURE 2. Bioreactor design. (a) Schematic view of the bioreactor showing the testing compartment, actuator, and instrumentation. (b) Schematic cut view of the testing compartment showing the labyrinth channel for unbiased force measurement.

well suited for autoclave sterilization. In this testing compartment, the tendon is xed at one end to a mobile anchor connected to the actuator, and at the other end, to a xed anchor connected to the load cell. Gas exchange is possible through a labyrinth channel, thus limiting the risks of contamination. This labyrinth

1538

PARENT et al.

channel also makes it possible to acquire unbiased force measurement. Movement of the mobile anchor is achieved with an electromagnetic linear actuator (P01 9 23 9 80, Linmot, Spreitenbach, Switzerland) and its controller (E1 100-RS, Linmot, Spreitenbach, Switzerland). This type of actuator has a wide range of operation and a long operating life. Load cell and optical encoder data acquisition and processing are done in real time with an FPGA (PCI-7833R, National Instruments, Austin, USA). This hardware provides rapid and robust data acquisition and analysis for up to eight axes (8 tendons). Closed-loop control of the actuator is accomplished through two interrelated control loops. In the primary loop, the actuators controller governs the motion with feedback from its own internal position sensor. In the secondary loop, the FPGA governs the instructions to the controller with feedback from the load cell and the optical encoder. In our design, the FPGA controls four axes independently. In other words, four tendons at the same time with completely dierent protocols could be stimulated. Regarding environmental control, an incubator (MIR 153, Sanyo, Japan) regulates the temperature of the bioreactor. A commercial system (ProCO2, Biospherix, Lacona, USA) was added for the control of CO2 level when needed and humidity was generated using a water tray. The medium changes were performed inside the bioreactor. In order to do so, a commercial access device designed for needle-free administration of drugs was used (Clave LifeShield CLAVE Needle-Free Connector, Hospira, Lake Forest, USA). It is simple to use, and the authors could not observe any contamination for up to 10 days of tissue culture. Finally, regarding mechanical characterization, the bioreactor has the following capabilities: 12.7-lm resolution for the optical encoders, 25-lm resolution for the vision system, and 45-lN resolution for the load cells. Mechanical Stimulation The six tendons from each tail, which were assigned to the stimulation protocol were transferred into the bioreactor. After temperature stabilization, the initial distance between the two anchors (L0) and zero strain were dened by achieving a tension stress of 0.75 MPa at equilibrium, i.e., for 5 min. The tendons were then subjected to a low stress fatigue loading protocol. Each sample was cycled between 0 and 1.5 MPa using a sine wave pattern at 1 Hz. The 1.5-MPa maximum stress lies inside the toe region of the typical stressstrain curve of healthy rat tail tendons (<4% ultimate tensile strength15,18 + data not shown). (To our knowledge,

the physiological stress range of rat tail tendons is yet to be identied.) The 1-Hz frequency is commonly used in tendon fatigue experiments because it falls within the physiological range of human activity. Similar to the study by Fung et al.,11 a continuous increase in the one loading cycles mean anchor-toanchor strain during preliminary testing was observed:  is the mean distance between  L0 =L0 where L  e L the two anchors measured using the optic encoder of the bioreactor. Therefore, to examine the progression of the mechanical degradation process, the authors tested different fatigue levels by dividing the tendons under stimulation into three groups corresponding to three limit values of  e (Fig. 1). In the rst group (identied as 4%) (N = 8), stimulation was discontinued after reaching the established limit of  e 4%. In the second group (identied as 8%) (N = 8), stimulation was discontinued at  e 8%, and in the third group (identied as R) (N = 8), stimulation was discontinued upon tendon rupture. The tendons were distributed equally in each group: two tendons per tail in each group. Mechanical Characterization By means of the data recorded during fatigue loading, the authors evaluated dynamic compliance, J, e a measure of exibility, expressed as J r where e is the strain, and r is the stress. The authors chose not to evaluate storage compliance or loss compliance since difculties were experienced when measuring phase shifts. Moreover, compliance was chosen instead of modulus simply by denition since stress was applied, and strain was measured. This method is nondestructive and thus allows subsequent characterization tests (histological analysis) to be carried out on the fatigued samples. During fatigue loading, the distance between the two anchors (L) and the force (F) were recorded in a series of 10-s time segments. An 8-order Butterworth low-pass lter was applied to the data. The cutoff frequency was set to maintain 90% of the signal energy. Strain (e) and stress (r) were calculated as et Lt L0 =L0 ; rt Ft=A0 where A0 is the initial cross-sectional tendon area. Dynamic compliance (J) was calculated using an equation adapted for experimental data: s psdfe0 g1 Hz J s psdfr0 g1 Hz where e0 detrendfes : s 10 sg and r0 detrendfrs : s 10 sg. In this equation, psd stands for the power spectral density of the signal (at 1 Hz) evaluated in each 10-s time segment (Fragmentation of

Low Stress Tendon Fatigue

1539

the signal into 10-s segments allowed us to evaluate dynamic compliance at different points in time during mechanical degradation.): and detrend is a Matlab function that removes linear trends and isolates the sinusoidal portion of the signal (Matlab 7.5, Mathworks, Natick, USA). Histology Structural characterization was performed on all control and loaded tendons. Biopsies for optical microscopy (OM), transmission electron microscopy (TEM), and scanning electron microscopy (SEM) were taken randomly from the middle portion of each sample. Biopsies destined for OM were rst xed in 10% neutral buered formalin for at least 24 h and then embedded in paran. Longitudinal sections 6 lm thick were processed in hematoxylin and eosin stain and observed under light and polarized light microscopy (TE2000-U, Nikon, Tokyo, Japan). For each biopsy (observed under light microscopy), three regions of interest (ROIs) were captured taking care not to include damages caused by the cutting blade. The ROIs were then analyzed with Vision assistant software (Version 7.1 National Instrument, Austin, TX, USA). The spaces between bers were found by contrast and separated into two categories: bril (black); and background (red). The space density (qs) was calculated as: qs % # pixels ! background 100: # pixels ! image

qf %

# pixels ! fibril 100: # pixels ! image

For SEM, the biopsies were immersed in liquid nitrogen (77 K) for freeze-fracture (protocol adapted from Provenzano et al.28) before dehydration. The biopsies were then placed on a pre-cooled microscope slide + work surface. Under an optical microscope, the tissues were carefully fractured longitudinally using a pre-cooled micro-surgical scalpel blade. After dehydration, the biopsies were critical point dried, mounted on 10-mm aluminum mounting blocks with a doublefaced carbon sticker, gold/palladium sputter-coated (100 mTorr Argon, 10 mA, 1 min), and stored in a vacuum container. The biopsies were then ready to be qualitatively observed under a scanning electron microscope (S-3000N, Hitachi Ltd., Tokyo, Japan) at 3 kV of accelerating voltage. Statistics A statistical analysis was conducted on compliance measures. For each of the four rats (N = 4), six tendons were submitted to fatigue testing up to 4% mean strain, 8% mean strain or rupture. Therefore, compliance measures at 0 and 4% mean strains were obtained for all six tendons. However, compliance measures at 8% mean strain were obtained for only four of six tendons and compliance measures at rupture were obtained for only two of six tendons (cf. Figs. 1, 3). Since compliance measures are paired (two to four measures were obtained from one tendon), and since the number of data varies from one group to the other, the t-test (assuming a Gaussian distribution) and Wilcoxon matched-pairs signed rank test (assuming a non Gaussian distribution) were chosen to detect signicant differences among the groups. To do so, mean compliances were calculated for all paired group. Six statistical tests were conducted on nine mean compliances. The threshold of signicance was set at p < 0.05. Measurements resulting from optical micrographs were also analyzed statistically. For each of the four groups and each of the four rats (N = 4), the mean density and the mean area of spaces between bers were calculated. They were then analyzed to detect signicant differences among the groups using a repeated measure ANOVA with Bonferronis multiple comparison post test. The threshold of signicance was set at p < 0.05. Finally, a statistical analysis was also conducted based on measurements resulting from transmission electron micrographs. For each of the four groups and each of the four rats (N = 4), the mean bril density

The areas of each space were also evaluated by the software and averaged. Biopsies destined for TEM and SEM were xed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate pH 7.4 buer; post-xed in a solution containing 1% osmium tetroxide buered with 0.1 M sodium cacodylate at pH 7.4; and dehydrated in graded ethanol solutions. For TEM, the biopsies were inltrated and embedded with Epon epoxy resin. Transversal sections of 7080-nm thicknesses were stained with uranyl acetate and lead citrate and mounted on copper grids for observation with a transmission electron microscope (H-7500, Hitachi Ltd., Tokyo, Japan) at 60 kV of accelerating voltage. For each biopsy, three ROIs were randomly captured and analyzed. The images were morphologically bottom-hat ltered (Matlab 7.5, Mathworks, Natick, USA) to distinguish bril pixels from background pixels. Then, pixel intensity was separated into two categories: bril (black); and background (white). The bril density (qf) was then calculated as:

1540

PARENT et al.

FIGURE 3. Number and distribution of samples for the statistical analysis of compliance measurements. The paired characteristic of samples and the varying numbers of samples between conditions motivated the use of a paired t-test (assuming a Gaussian distribution) and Wilcoxon matched-pairs signed rank test (assuming a non Gaussian distribution).

was calculated. However, since data are missing in this case (problem with tissue sample preparation), the available data were analyzed using a t-test (assuming a Gaussian distribution) instead of using a repeated measure ANOVA to detect signicant differences among the groups. A Wilcoxon matched-pairs signed rank test was also conducted (assuming a non Gaussian distribution). The threshold of signicance was set at p < 0.05.

RESULTS Mechanical Properties The low stress fatigue loading applied to the tendon samples resulted in an increased mean anchor-toanchor strain through the three dened fatigue levels:  e 4%; 8% and R (Fig. 4). The strain increase was rapid in the rst minutes of loading and was thenceforth nearly exponential until rupture. Dynamic compliance also showed nonlinear progression (Fig. 4). On this typical curve, dynamic compliance decreased during approximately the rst 4 h and increased during approximately the last 4. Between these two zones of change, dynamic compliance was nearly stable. These observations are supported by Fig. 5 which shows that differences between dynamic compliances (0% vs. 4%; 0% vs. 8%; 4% vs. 8%; and

FIGURE 4. Progression of mechanical parameters. Typical progression of strain ( e) and dynamic compliance (J) during low stress fatigue loading to point of rupture (1.5 MPa sine wave, 1 Hz, rat #4). The inset shows a magnication of the rst 2 h.

Low Stress Tendon Fatigue

1541

Microscopy Under SEM, longitudinal fractures observed at low magnication (Figs. 6a6d) revealed an increasing degree of alteration to the collagen network with increasing fatigue levels. At rupture, when alteration was greatest, local zones of damage with low bril density were evident. Under OM, longitudinal sections observed at higher magnication (Figs. 6e6h) again showed that the extent of alteration to the collagen network increased along with the fatigue level, and that it was signicant at rupture. Typically, the undulations of the collagen brils increased and the space between the brils grew (Fig. 6h). Statistically however, there was no difference between space densities and areas measured at different fatigue levels (Fig. 7). At still higher magnication, the transversal sections observed under TEM revealed variations in collagen bril density (Figs. 6i6l) and orientation (not shown). Typically, as the fatigue level increased, areas of lower bril density increased in number and size. Statistically, however, there was no difference between bril densities measured at different fatigue levels (Fig. 8). From a histological perspective, tendon degradation progressed over time in a nonlinear pattern as did mechanical properties. Degradation also progressed irregularly in space. Table 1 provides a numerical summary of the experimental results.
FIGURE 5. Dynamic compliance for the three fatigue levels. Compliance (J) follows a U curve. It decreases from 0 to 4%, and then increases at 8% and rupture. Asterisk indicates a signicant difference between two groups according to the paired t-test (*p < 0.05). No signicant differences were observed according to the Wilcoxon matched-pairs signed rank test.

DISCUSSION In this study, the mechanical degradation process of the tendons mechanobiological response to low amplitude repetitive loading was examined. It was noticed that mechanical degradation occurs relatively fast (time scale <24 h) even at very low stress levels. Moreover, our mechanical analysis allowed the authors to observe a nonlinear progression of strain and compliance as seen in higher stress fatigue loading for strain and stiness.11 Finally, using different microscopic methods, an increased disorganization of the collagen structure with increasing fatigue level was observed again, as seen in higher stress fatigue loading for strain and stiffness.11 These observations implied that in live tissues under repetitive loading, repair of the extracellular matrix by the cells ought to be active to counterbalance mechanical degradation even at very low stresses, as otherwise tendon would degenerate. These observations are especially interesting in the context of tendinosis which consists of chronic lesions without inammation that develop slowly and which,

8% vs. Ruptures) were signicantly different when analyzed with a paired t-test (but not when using a Wilcoxon matched-pairs signed rank test). It is worth noting that discontinuities visible on the straintime graph occurred at the time of load cell zeroing (Fig. 4). Since the tendon stimuli were stresscontrolled, the slowly drifting load cell had to be reset every 3 h to maintain a correct nominal stress over the total testing period. Therefore, over a 24-h period, a total of eight reset procedures were performed. This recalibration process did not affect the compliance estimates indicated in Fig. 4. The load cell drift, however, decreased the actual stress that was exerted on the tendon within the 3-h period. This of course induced a step in strain and compliance results in the low mean strain region (at time = 3 h) but had negligible effect beyond that.

1542

PARENT et al.

FIGURE 6. Tendon structure from a representative set of SEM, OM, TEM images of the three fatigue levels compared to control. SEM images (ad) show longitudinal fractures of tendons revealing the general view of bril organization. OM images (eh) show longitudinal sections of tendons revealing bril alignment. TEM images (il) show transverse sections of tendons revealing collagen bril density (bar 5 ad: 200 lm, eh: 100 lm, il: 10 lm).

in certain cases, can lead to spontaneous tendon rupture.17,30 In our study, the observed degradation could have originated through various factors other than mechanical fatigue. The rst of these other factors is the enzymatic degradation caused by cell death in a non-nutritive saline solution. To limit enzymatic action, the authors supplemented the solutions with protease inhibitors. The eciency of this method was noticed in a study conducted by the authors group on live tissues under very low amplitude mechanical stimulation.9 The second factor is contamination which was avoided by using good practices in tissue culture as well as antibiotics/antimycotics. The nal factor is time which can modify the tissue via thermodynamics. Unlike the other factors, the authors could not control for the time factor. However,

knowing that both strain and dynamic compliance progress in a similar fashion to Fungs observations11 conducted under a very short timescale (13 9 min vs. 882 204 min), it is most likely that the factor examined was stress-induced mechanical fatigue. With regard to the preceding observations, the very low stresses would not protect the rat tail tendons against mechanical degradation. The authors can therefore put forward the hypothesis that even if overall mechanical loading has very low amplitude, it can be signicant enough locally to cause local breakage, likely due to material heterogeneity. One must, however, be careful not to overgeneralize with this hypothesis, since all tendons do neither have the same properties, nor the same physiological activity range. Three microscopic techniques were used to verify if such heterogeneity manifested itself at the structural

Low Stress Tendon Fatigue

1543

TABLE 1. Numerical summary of the results at the end of each fatigue level compared to the control. Control Time (h) Final length (mm) Compliance (GPa21) Mean space density (%) Mean space area (pixel) Mean bril density (%) 4% 2.5 7 0.2 5.7 8% 9.9 65 1.6 6.4 R

0 5.0 61 3 64 1.9 0.3 1.5 3.8 2.5 6.3 26 9 62 4

3.8 14.7 3.4 5 71 17 0.2 2.2 0.2 5.6 10.0 4.8 65 24 49 15

42 9 65 7

36 6 60 10

FIGURE 7. Spaces between brils for the three fatigue levels compared to control. (a) Mean space density varied between 3.8 and 10.0%. (b) Mean space area varied between 26 and 65 pixels. No signicant differences were observed.

FIGURE 8. Collagen bril density for the three fatigue levels compared to control. Fibril density varied between 49 and 65%. No signicant differences were observed.

level of the mechanically degraded tendons under low stress. The techniques that were used enabled the authors to detect structural changes starting at 4% tendon strain (increasing undulations of the collagen brils and of the space between the brils). However, changes were clearly evident just before rupture because the structural degradation appeared to occur exponentially over time.

From a spatial viewpoint, damage accumulated heterogeneously at localized sites as had been speculated. It is worth noting that in the case of optical and scanning electron microscopy, these observations could have been amplied by our thorough preparation methods. Other microscopic techniques such as the one used by Fung et al.,11 could be used to validate our observations. In the case of transmission electron microscopy, the technique is classic, relatively standardized for our sample type and used by other important laboratories. Moreover, the observed density decrease corroborates with the increase in tissue diameter observed by Lanir et al.22 This rst investigational study examined mechanical degradation in a tendon under low stress. In future study, the authors will also include the eect of cells in the mechanobiological tissue response, thus adding the repair and enzymatic degradation processes to the mechanical degradation process, which has already been examined. To maintain tissue homeostasis, the repair process must counterbalance the mechanical and enzymatic degradation processes. In view of the temporal and spatial heterogeneities of mechanical degradation, a repair process should act rapidly at critical sites. Otherwise, at degraded sites, the mechanotransduction pathways will be modied, which could induce enzymatic degradation from the aected cells.3,23 Degradation would then increase over time and space through protease action that weakens neighboring brils and causes further damage in the manner of a vicious cycle. Prevention of overuse injuries will be facilitated by establishing points of reference for the levels of stress amplitudes allowed in our tendons and the acceptable time lengths for tasks or activities. A better understanding of the dynamics between mechanical degradation, enzymatic degradation and extracellular matrix repair is thus required. The data that were acquired, along with those from Fung et al.,11 allowed us to trace a graph of tendon fatigue life (Fig. 9), which can be used as a preliminary point of reference. On this

1544

PARENT et al. Andres, B. M., and G. A. C. Murrell. Treatment of tendinopathy: what works, what does not, and what is on the horizon. Clin. Orthop. Relat. Res. 466:15391554, 2008. 3 Arnoczky, S. P., M. Lavagnino, and M. Egerbacher. The mechanobiological aetiopathogenesis of tendinopathy: is it the over-stimulation or the under-stimulation of tendon cells? Int. J. Exp. Pathol. 88:217226, 2007. 4 Arnoczky, S. P., M. Lavagnino, J. H. Whallon, and A. Hoonjan. In situ cell nucleus deformation in tendons under tensile load: a morphological analysis using confocal laser microscopy. J. Orthop. Res. 20:2935, 2002. 5 Bruneau, A., C. Champagne, P. Cousineau-Pelletier, G. Parent, and E. Langelier. Preparation of rat tail tendons for biomechanical and mechanobiological studies. J. Vis. Exp. 41, 2010. 6 Butler, S. L., S. S. Kohles, R. J. Thielke, C. Chen, and R. Vanderby, Jr. Interstitial uid ow in tendons or ligaments: a porous medium nite element simulation. Med. Biol. Eng. Comput. 35:742746, 1997. 7 Chae, Y., G. Aguilar, E. J. Lavernia, and B. J. Wong. Characterization of temperature dependent mechanical behavior of cartilage. Lasers Surg. Med. 32:271278, 2003. 8 Clavert, P., J. F. Kempf, F. Bonnomet, P. Boutemy, L. Marcelin, and J. L. Kahn. Effects of freezing/thawing on the biomechanical properties of human tendons. Surg. Radiol. Anat. 23:259262, 2001. 9 Cousineau-Pelletier, P., and E. Langelier. Relative contribution of mechanical degradation, enzymatic degradation, and repair of the extracellular matrix on the response of tendons when subjected to under- and over-mechanical stimulations in vitro. J. Orthop. Res. 28:204210, 2010. 10 Fung, Y. C. Biomechanics: Mechanical Properties of Living Tissues (2nd ed.). New York: Springer-Verlag, 568 pp., 1993. 11 Fung, D., V. Wang, D. Laudier, J. H. Shine, J. BastaPljakic, K. J. Jepsen, M. B. Schafer, and E. L. Flatow. Subrupture tendon fatigue damage. J. Orthop. Res. 27:264 273, 2009. 12 Giannini, S., R. Buda, F. Di Caprio, P. Agati, A. Bigi, V. De Pasquale, and A. Ruggeri. Effects of freezing on the biomechanical and structural properties of human posterior tibial tendons. Int. Orthop. 32:145151, 2008. 13 Gordona, C. J., E. Puckettb, and B. Padnosa. Rat tail skin temperature monitored noninvasively by radiotelemetry: characterization by examination of vasomotor responses to thermomodulatory agents. J. Pharmacol. Toxicol. Methods 47:107114, 2002. 14 Guilak, F., A. Ratcliffe, and V. C. Mow. Chondrocyte deformation and local tissue strain in articular cartilage: a confocal microscopy study. J. Orthop. Res. 13:410421, 1995. 15 Haraldsson, B. T., H. Langberg, P. Aagaard, A. M. Zuurmond, B. van El, J. Degroot, M. Kjaer, and S. P. Magnusson. Corticosteroids reduce the tensile strength of isolated collagen fascicles. Am. J. Sports Med. 34:1992 1997, 2006. 16 Harkness, J. E., and J. E. Wagner. The Biology and Medicine of Rabbits and Rodents (3rd ed.). Philadelphia: Lea and Febiger, 241 pp., 1989. 17 zsa, L., and P. Kannus. Human Tendons: Anatomy, Jo Physiology and Pathology. Champaign: Human Kinetics, 574 pp., 1997. 18 Kato, Y. P., D. L. Christiansen, R. A. Hahn, S. J. Shieh, J. D. Goldstein, and F. H. Silver. Mechanical properties of collagen bres: a comparison of reconstituted and rat tail tendon bres. Biomaterials 10:3842, 1989.
2

FIGURE 9. Fatigue life of tendons. This graph shows an extrapolation of fatigue life between our data and those obtained by Fung et al.11 As stress amplitude increases, time to rupture decreases.The readers are, however, cautioned to be careful while interpreting this graph.

graph, the authors observed that repetitive tasks executed at 10 and 15% of the UTS could cause tendon rupture in approximately 9 and 7 h, respectively, if mechanical degradation only was considered. Of course, this graph must be interpreted carefully, because tendons have different mechanical properties and physiological activity ranges depending on the species and anatomical site. Moreover, the shape and dimensions of the curve will vary with the addition of cellular activity (the processes of repair and enzymatic degradation) which the authors will try to evaluate in future studies just as has been done with strain-controlled stimulation.9 Those studies will be more complex and will require more advanced modeling techniques, because other variables such as the incorporation of rest time will eventually need to be included.

ACKNOWLEDGMENTS This study was supported by NSERC Grant 299280 and by an IRSST scholarship awarded to Gabriel Parent. The authors also thank Charles Bertrand and Melina Narlis for the preparation of specimens for and Paule Cousineauhistology; and Nicolas Huppe Pelletier for their assistance with the bioreactor and tissue isolation, respectively.

REFERENCES
1

An, K. N. In vivo force and strain of tendon, ligament, and capsule. In: Functional Tissue Engineering, edited by F. Guilak, D. L. Butler, S. A. Goldstein, and D. Mooney. New York: Springer-Verlag, 2003, pp. 96105.

Low Stress Tendon Fatigue

19

1545

Ker, R. F., R. Mc. N. Alexander, and M. B. Bennett. Why are mammalian tendons so thick? J. Zool. 216:309324, 1988. 20 Lam, T. C., C. G. Thomas, N. G. Shrive, C. B. Frank, and C. P. Sabiston. The effects of temperature on the viscoelastic properties of the rabbit medial collateral ligament. J. Biomech. Eng. 112:147152, 1990. 21 Langelier, E., D. Dupuis, M. Guillot, F. Goulet, and D. Rancourt. Cross-sectional proles and volume reconstructions of soft tissues using laser beam measurements. J. Biomech. Eng. 126:796802, 2004. 22 Lanir, Y., E. L. Salant, and A. Foux. Physico-chemical and microstructural changes in collagen ber bundles following stretch in-vitro. Biorheology 25:591603, 1988. 23 Lavagnino, M., S. P. Arnoczky, M. Egerbacher, K. L. Gardner, and M. E. Burns. Isolated brillar damage in tendons stimulates local collagenase mRNA expression and protein synthesis. J. Biomech. 39:23552362, 2006. 24 Lee, T. Q., and S. L.-Y. Woo. A new method for determining cross-sectional shape and area of soft tissues. J. Biomech. Eng. 110:110114, 1988. 25 McCrum, N. G., C. P. Buckley, and C. B. Bucknall. Principles of Polymer Engineering (2nd ed.). Oxford: Oxford University Press, 447 pp., 1997. 26 Ng, B. H., S. M. Chou, B. H. Lim, and A. Chong. The changes in the tensile properties of tendons after freeze storage in saline solution. Proc. Inst. Mech. Eng. H 219: 387392, 2005. 27 Parent, G., and E. Langelier. Accuracy and precision of algorithms in estimating cross-sectional area of rat tail

tendons. Meas. Sci. Technol., 2011. doi:10.1088/0957-0233/ 21/12/125802. 28 Provenzano, P. P., C. Hurschler, and R. J. Vanderby. Microstructural morphology in the transition region between scar and intact residual segments of a healing rat medial collateral ligament. Connect. Tissue Res. 42:123 133, 2001. 29 Schechtman, H., and D. L. Bader. Fatigue damage of human tendons. J. Biomech. 35:347353, 2002. 30 Scott, A., K. M. Khan, J. L. Cook, et al. Human tendon overuse pathology: histopathologic and biochemical ndings. In: Tendinopathy in Athletes, edited by S. L.-Y. Woo, P. A. F. H. Renstro m, and S. P. Arnoczky. Malden: Blackwell Publishing, 2007, pp. 6984. 31 Screen, H. R., D. A. Lee, D. L. Bader, and J. C. Shelton. Development of a technique to determine strains in tendons using the cell nuclei. Biorheology 40:361368, 2003. 32 Wang, X. T., and R. F. Ker. Creep rupture of wallaby tail tendons. J. Exp. Biol. 198:831845, 1995. 33 Wang, X. T., R. F. Ker, and R. M. Alexander. Fatigue rupture of wallaby tail tendons. J. Exp. Biol. 198:847852, 1995. 34 Woo, S. L., T. Q. Lee, M. A. Gomez, S. Sato, and F. P. Field. Temperature dependent behavior of the canine medial collateral ligament. J. Biomech. Eng. 109:6871, 1987. 35 Wren, T. A. L., D. P. Lindsey, G. S. Beaupre, and D. R. Carter. Effects of creep and cyclic loading on the mechanical properties and failure of human Achilles tendons. Ann. Biomed. Eng. 31:710717, 2003.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.