Article Text
PDF
Abstract
Background Tendinopathy commonly occurs in tendons with large in vivo loading demands like the Achilles tendon (AT) and supraspinatus tendon (SST). In addition to differences in their local anatomic environment, these tendons are designed for different loading requirements because of the muscles to which they attach, with the AT experiencing higher loads than the SST. One possible factor in the progression of tendinopathy is the interplay between mechanical loading and the regulation of enzymes that degrade the extracellular matrix (matrix metalloproteinases (MMPs)) and their inhibitors (tissue inhibitor of metalloprotienases (TIMPs)). Thus, overuse injuries may have different biological consequences in tendons designed for different in vivo loading demands.
Aim In this study, the tendon-specific regulation of MMP-13, MMP-3 and TIMP-2 expression in rat AT and SST exposed to two different mechanical environments was investigated.
Methods Rat AT and SST were exposed to stress deprivation (ie, detached from attachments) and intermittent cyclic hydrostatic compression (with attachments intact). Levels of MMP-13, MMP-3 and TIMP-2 mRNA were evaluated in time-zero control, attached, stressdeprived and “compressed” tendons.
Results Stress deprivation led to elevated expression of MMP-13, MMP-3 and TIMP-2 in both tendons, although the magnitude of the increase was greater for the SST than the AT. Intermittent cyclic hydrostatic compression of attached tendons increased expression of MMP-13 in the SST, but not the AT.
Conclusions The results of this study suggest that stress deprivation may be one contributor to the progression of tendinopathy in AT and SST, where the tendon designed for the lower in vivo loading demand (SST) was the most affected by a change in mechanical loading. The unique upregulation of MMP-13 with hydrostatic compression supports the impingement injury theory for rotator cuff tears.
Statistics from Altmetric.com
Tendons are discrete bands of dense collagenous tissue that connect muscle to bone. Tendons transmit load created in muscle to bone and function to make joint motion possible.1 Overuse injuries comprise 30% to 50% of all sports-related injuries.2 Tendon injuries account for a large portion of overuse injuries.1 3 The common sites of tendinopathy are energy-storing tendons with large magnitude loading demands in vivo: Achilles tendon (AT) of the ankle, patellar tendon of the knee, extensor carpi radialis brevis tendon of the elbow, and rotator cuff tendons of the shoulder, including the supraspinatus tendon (SST).1 Microinjury of the tendon structure from repetitive loading may create a partially damaged tendon where the damaged portions are unloaded (stressdeprived) and the remaining intact portions are overloaded.
Different tendons are loaded to different stresses based on the muscles to which they attach.4 The human AT functions at 50% to 100% of its ultimate tensile strength in activities ranging from walking to running; whereas, the human SST functions normally within 25% to 30% of its ultimate tensile strength.5 Thus, tendons that are designed for different in vivo loading demands may have different responses to changes in loading due to injury. For instance, Huang et al6 found that the same exercise loading protocol, which caused significant injury in the rat SST,7 did not injure the rat AT.
In addition to the tensile load that a tendon is designed to transmit, some tendons are exposed to compressive load due to overlying structures in their local anatomic environment. In the supraspinatus tendon, tendinopathy may occur secondary to subacromial impingement, that is, from the repetitive compression of the supraspinatus tendon between the humeral head and the acromion and coracoacromial ligamentous complex.8 Although not proven,9 this impingement process is thought to cause inflammation, fibrosis, and ultimately rupture of the SST that leads to clinical shoulder pain, weakness, and shoulder impairment.10 Thus, imposing a chronic compressive load on a tissue designed for tensile loading may generate a biological response.
Tendon maintenance/remodelling, growth and development are highly dependent on a series of tightly regulated biological events that include matrix molecule extracellular proteolysis. Matrix metalloproteinases (MMPs) are a family of zincdependent endopeptidases which collectively degrade essentially all the components of the extracellular matrix.11,–,15 Collagenases (MMP-1, MMP-8, MMP-13 in humans) are a subset of the MMP family and are primarily responsible for the cleavage of the triple helix fibrillar collagens.16 In particular, MMP-1 and MMP-13 both cleave collagen type I, the primary collagen of tendon, with comparable efficiency.17 The stromelysins (MMP-3, MMP-10, MMP-11) are another subset of the MMP family and have a broad range of substrate specificity. Tissue inhibitor of metalloprotienases (TIMPs) are endogenous inhibitors of MMPs.15 MMPs and TIMPs have been implicated in the pathogenesis of tendinopathy.18,–,20
In ruptured SST, Lo et al19 and Riley et al18 found increased MMP-13 expression and MMP-1 activity, respectively. In ruptured AT, MMP-1321 and MMP-120 expression were upregulated. Although collagenases were increased in these ruptured tendons, MMP-3 and TIMP-2 were decreased.18,–,20 Of those markers, Jones et al20 found that only MMP-3 was decreased in degenerated AT which had yet to progress to rupture. The study of de Mos et al22 similarly documented significantly downregulated MMP-3 in degenerated AT but also a trend for upregulated MMP-13. Collectively, these results suggest a role for MMPs/TIMPs in extracellular matrix degradation and/or remodelling in tendinopathy, with the timing and magnitude of these changes relating to the progression of the degeneration.
Our purpose in this study was to investigate the response of MMPs/TIMPs to altered tendon mechanical loading, either by detaching an attached tendon (simulating the unloaded region of a partially damaged tendon) or “compressing” an attached tendon by subjecting it to intermittent cyclic hydrostatic compression (ICHC; simulating a component of the loading during impingement23). Thus, our primary hypothesis was that in vitro stress deprivation and ICHC of SST and AT would lead to an upregulation of MMP-13 and a downregulation of MMP-3 and TIMP-2. Our secondary hypothesis was that the responsiveness of the SST would be greater than that of the AT.
Methods
In this study approved by the University of Calgary Animal Care Committee, 34 AT and SST from 23 4-month-old (mass: 478 (11 g)) male Sprague–Dawley rats (Charles River, Saint- Constant, Quebec, Canada) were assigned to four groups: timezero control (n=5), attached (n=9), stress-deprived (n=10), or attached plus subjected to ICHC (n=10). For the attached tendons, shoulder and ankle joints were excised. The AT and SST were exposed while maintaining their tendon attachments. Under RNase-free conditions, joints were immersed in sterile, serum-free Dulbecco's modified Eagle medium (DMEM, GIBCO, Grand Island, New York, USA) with antibiotics (antibiotic–antimycotic, 1%, GIBCO) contained in a cylinder which was surrounded by a water bath at 39°C. After 4 h in culture, the tendons were excised, weighed, snap-frozen in liquid nitrogen, and maintained at −70°C until RNA extraction. For the stress-deprived tendons, AT and SST were detached from their attachments before culturing for 4 h as described above. For the time-zero controls, AT and SST were excised and processed immediately post mortem without exposure to culture. For the “compressed” tendons, ICHC was performed using a custom-designed Plexiglas cylinder and accompanying stainless steel piston mounted on a servohydraulic testing system (MTS Systems, Minneapolis, Minnesota, USA).24 Tendons were exposed while maintaining their tendon attachments and immersed in DMEM with antibiotics contained in the cylinder surrounded by a 39uC water bath. Hydrostatic pressure of 1 MPa was applied cyclically at 0.5 Hz for 1 min. This loading was repeated every 15 min for 4 h. After 4 h, AT and SST were processed as above for RNA extraction.
RNA extraction and semiquantitative RT-PCR
Semiquantitative RT-PCR was performed using techniques developed and optimised in our laboratory as described previously.19 25,–,28 In brief, total RNA was extracted from rat tendons using the Qiagen RNeasy Kit (Qiagen Sciences, Germantown, Maryland, USA), treated with DNase according to the manufacturer's instructions and quantified.25 27 Total RNA (1 mg) from all samples in an experiment to avoid potential variation were reverse transcribed at the same time to generate single-stranded cDNA using the Qiagen Omniscript RT Kit (Qiagen Sciences). PCR of cDNA (1.2 ml) was performed in a total volume of 50 ml containing 106PCR buffer, 10 mmol/l deoxyribonucleotide triphosphate mixture, 50 mmol/l MgCl2, 0.5 mmol/l of each primer and 1.25 U of Taq DNA polymerase. The validated primer sequences for amplified genes MMP-13,29 MMP-3,30 TIMP-231 and 18S32 are shown in table 1. Rats do not have a homologue of the human MMP-1 gene; hence, MMP-13 in rats has the role of MMP-1 and MMP-13 in humans.33 The PCR was performed and normalised to 18S levels which served as the housekeeping gene and which did not vary in any of the experimental conditions. All PCR products were analysed by electrophoresis of 20 ml of each PCR product mixture in a 2% agarose gel. After staining with ethidium bromide, a densitometric analysis (ChemiDoc XRS and Quantity 1 1-D analysis, BioRad Laboratories, Mississauga, Ontario, Canada) was used to determine mRNA levels. Conditions were in the linear range for both the PCR amplification and the image analysis detection system.19 28 Sample gene expression band densities were normalised to their corresponding 18S values before comparisons to yield a semiquantitative assessment.26 28 The no-RT controls were negative; thus, genomic DNA contamination was undetectable.28
RT-PCR primers
In situ hybridisation
The MMP-13 RNA probe was developed from the RT-PCR products achieved above and in situ hybridisation performed similar to that of Lavagnino et al.34 Briefly, PCR products were purified (Qiagen Gel Extraction Kit, Qiagen Sciences) and subcloned into a plasmid vector (PCR-Script Amp Cloning Kit, Stratagene, La Jolla, California, USA). DH5a cells were transformed and incubated overnight on L-AMP 100 plates. Colonies with inserts were selected, amplified and confirmed by sequencing to verify the MMP-13 probe specificity. Finally, plasmids were linearised to allow transcription of antisense probes for use in the in situ hybridisation studies. In vitro transcription was performed using the appropriate T7 RNA polymerase in the presence of digoxigenin (DIG)-linked UTP in the reaction mixture (DIG RNA Labeling Kit, Roche Diagnostics, Indianapolis, Indiana, USA). In situ hybridisation was performed on frozen histologic sections to determine MMP13 expression in the supraspinatus tendon. Care was taken to preserve RNA during tissue manipulation. The sample sections were digested by proteinase K (10 mg/ml at 37uC for 30 min) and hyaluronidase (10 mg/ml at 37°C for 30 min). Sections were rinsed in PBS, and acetylation was performed (acetic anhydride in 0.1 M triethanolamine). Then, sections were rinsed in 2 × saline sodium citrate (SSC; 1 × SSC with 150 mM NaCl and 15 mM sodium citrate). Antisense probes for rat MMP-13 were diluted in hybridisation solution to a concentration of 200 ng/ml, incubated for 10 min at 90uC and put on ice for 5 min. Sections were exposed to the hybridisation solution overnight in a humid chamber at 42°C. Sections were washed in 2×SSC at 40°C for 30 min followed by RNaseA (10 mg/ml) incubation to remove all unhybridised probe. Then, sections were incubated with blocking solution (Roche Diagnostics) at room temperature for 1 h. The digoxigenin-labeled hybrids were detected by antibody incubation performed according to the manufacturer's instructions. Sections were counterstained with 0.01% Saffranin O, dehydrated and mounted with EuKit (Calibrated Instruments, Hawthorne, New York, USA). Then, sections were examined and photographed (Ziess Axioskop2 plus and AxioCam, Cologne, Germany).
Statistical analysis
To evaluate the effects of stress deprivation, time-zero control, attached and stress-deprived tendons were analysed using multiple linear regression with random effects to account for mixed pairing. Next, linear contrasts were performed to determine specific p-values for the comparisons of interest, with significance set at p<0.05. To evaluate the effects of ICHC, time-zero control and “compressed” tendons were analysed using Student t-tests. In order to compare the AT and SST, the fold change from time-zero control of stressdeprived tendons or “compressed” tendons was calculated. Subsequently, AT and SST were compared using paired t-tests.
Results
Stress deprivation of AT for 4 h led to significantly elevated levels in mRNA for MMP-13, MMP-3 and TIMP-2 compared to time-zero controls (fig 1A, p<0.01). Similarly, mRNA levels for all molecules investigated were greater when comparing stressdeprived to attached AT (fig 1A, p<0.008). Time-zero control and attached AT were not statistically different from each other (fig 1A). Stress deprivation of SST for 4 h led to an apparent upregulation of mRNA expression for MMP-13, MMP-3 and TIMP-2 when compared to time-zero control and attached SST (fig 1B, p<0.002). Time-zero control and attached SST values were not statistically different from each other (fig 1B). In situ hybridisation of SST tissue demonstrated an upregulation of MMP-13 mRNA in the stress-deprived versus time-zero control and attached tendons (fig 2), consistent with the PCR results. Interestingly, some cells in the stress-deprived tendon (fig 2C) were clearly positive and others clearly negative, indicating that not all cells respond equally to stress deprivation.
MMPs/TIMP-2 mRNA expression in (A) Achilles Tendon and (B) Supraspinatus Tendon. *Stress-deprived tendon different than timezero control and attached tendons (p<0.01).
In situ hybridisation of MMP-13 in supraspinatus tendon: (A) Time-zero control; (B) Attached; (C) Stress-deprived. Scale bar is 20 mm. Black arrow points to cell exhibiting signal. White arrow points to cell exhibiting no signal.
The AT and SST were compared using the parameter, fold change in mRNA expression from time-zero control. Although stress deprivation led to increased expression in both tendons, the increases were larger in the SST (fig 3A, p<0.02). For fold in the SST which was statistically greater than the fourfold increase in AT (p = 0.02).
Comparison of Achilles and supraspinatus tendon MMPs/TIMP-2 mRNA expression when exposed to (A) Stress deprivation and (B) Intermittent cyclic hydrostatic compression (ICHC). *Stress-deprived or “compressed” tendon different than time-zero control tendon (p<0.01). **Supraspinatus tendon different than Achilles tendon (p<0.04).
Exposing the attached SST to ICHC led to significantly increased mRNA levels for MMP-13 compared to time-zero control tendon (p = 0.009), while MMP-3 and TIMP-2 mRNA levels did not change (fig 3B). Tendons exposed to ICHC were only compared to time-zero control tendons because the genes of interest were not different comparing attached tendons to time-zero control tendons (fig 1). For the AT, ICHC did not lead to alterations in mRNA levels of MMP-13, MMP-3 or TIMP-2 compared to the time-zero controls (fig 3B). This type of compressive loading resulted in a threefold upregulation of MMP-13 expression in the SST which was statistically greater than the levels in the AT which did not change (fig 3B, p = 0.04).
Discussion
In the current study, stress deprivation led to an apparent upregulation in the expression of MMP-13, MMP-3 and TIMP-2 in rat SST and AT. The magnitude of the upregulation was greater in the SST than the AT, suggesting that a tendon designed for lower in vivo loading demands was more affected by the loss of loading. Interestingly, ICHC also led to an upregulation of MMP-13 expression in the SST, but not in the AT. The other genes assessed (MMP-3 and TIMP-2) were not affected by ICHC for either tendon. This unique upregulation of MMP-13 in the SST exposed to ICHC supports a potential impingement injury theory for rotator cuff tears.
Stress deprivation led to an upregulation of MMP-13 mRNA levels in both tendons, although the magnitude was greater in the SST (17-fold) compared to the AT (fourfold). The upregulation of MMP-13 in these two tendons is consistent with findings of Arnoczky and colleagues in rat tail tendons that were stress-deprived for 24 h34 or 7 days.35 These authors reported that in rat tail tendons, where collagen fibrillar damage was caused by subrupture tensile loading, MMP-13 staining was located in the regions of damage.34 Furthermore, stress-deprived rat tail tendons had lower ultimate stress and modulus compared to normal tendons.35 If repetitive loading creates regions of damage in the supraspinatus and Achilles tendon, a stress deprivation-related increase in MMP-13 could contribute to matrix degradation and corresponding mechanical property deterioration.
Ruptured human SSTs have been demonstrated to have increased MMP-13 mRNA expression19 and MMP-1 activity.18 Because rat MMP-13 takes the role of human MMP-1 and MMP-13,33 the findings of the current study suggest that some of the biological changes in clinically torn SST may have, in part, resulted from stress deprivation. MMP-13 upregulation with tendon stress deprivation may also explain the common progression of partial-thickness tears to full-thickness tears in rotator cuff tendons. In humans, degenerated AT had a nearsignificant trend towards upregulation of MMP-1322 and ruptured AT had upregulated MMP-1321 and MMP-1.20 The current study suggests that stress deprivation may also be a factor in Achilles tendinopathy. Thus, an upregulation of MMP13 following stress deprivation may contribute to the progression of tendinopathy in the SST and AT.
Applying ICHC significantly increased MMP-13 expression in the attached SST but not the attached AT. The findings in the AT are consistent with a study of medial collateral ligament scars exposed to ICHC.24 Using a rabbit model where MMP-1 was the only collagenase investigated, ligament scars did not exhibit increased MMP-1 expression with ICHC compared to time-zero controls. Interestingly, ligament scars did have significantly increased MMP-1 expression with stress deprivation. The unique upregulation of MMP-13 mRNA levels with ICHC in SST may be related to matrix turnover and, as such, could support the impingement injury theory for rotator cuff tears.
Unlike the downregulation of MMP-3 and TIMP-2 in clinically ruptured tendons, stress deprivation, in the current study, resulted in an upregulation of MMP-3 and TIMP-2 mRNA in both the SST and AT. Archambault et al found increased expression of cartilage genes (eg, Col2a1, aggrecan, Sox9) but no significant changes in expression of MMP genes in an in vivo model of overuse injury in the rat supraspinatus tendon.36 Comparing their findings to the increased MMP expression in ruptured tendons and late-stage tendinopathy, the authors commented that their model evaluated the early phases of overuse where these enzymes might not yet be regulated.36 In the current study, the apparent discrepancy in gene regulation with stress deprivation may also be explained by timing, where 4 h of stress deprivation may not capture the later-stage changes demonstrated in clinically torn tendons. ICHC caused no change in MMP-3 and TIMP-2 expression in either tendon. In the SST, where MMP-13 mRNA levels did change, this suggests that ICHC is not a regulator of MMP-3 and TIMP-2 in the SST. In the AT, where no gene of interest changed, this suggests that the magnitude of loading might be not sufficient or that the type of loading might not be a principal injury factor for the AT.
What is already known on this topic
▶. Tendinopathy occurs in tendons with high in vivo loading demands like Achilles and supraspinatus tendons, yet each tendon is designed for different loads because they attach to different muscles.
▶. Matrix metalloproteinases may play a role in tendinopathy because their upregulation has been detected in degenerated and ruptured tendons.
What this study adds
▶. Stress deprivation caused tendon-specific upregulation of matrix metalloproteinases with increased responsiveness in supraspinatus versus Achilles tendons, suggesting that the tendon designed for the lower in vivo loading demand was more affected by the loss of mechanical loading.
▶. Intermittent cyclic hydrostatic compression uniquely upregulated matrix metalloproteinase-13 in the supraspinatus tendon.
The SST exhibited a greater magnitude of upregulation for MMP-13, MMP-3 and TIMP-2 with stress deprivation compared to the AT. Additionally, the expression of MMP-13 was uniquely upregulated in the SST with ICHC, whereas, this loading did not affect the AT. These tendon-specific responses are supported by observations from in vivo animal models of tendinopathy. When the same exercise loading protocol was used, the rat SST developed an overuse injury7 but the rat AT did not.6 Huang et al suggested that the reason for the differences between the SST and AT may be due to differences in tendon anatomy, mechanical demands, load magnitude/frequency, or injury mechanisms.6 In the current study, tendonspecific findings may also be attributed to similar factors.
In summary, these findings suggest that some changes associated with chronic tears of the SST and AT may be secondary to stress deprivation. Furthermore, stress deprivationrelated MMP changes may also explain, in part, the common progression of partial-thickness tears to full-thickness tears of the human rotator cuff. The unique upregulation of MMP-13 in response to hydrostatic compression of the SST supports that impingement contributes to rotator cuff tendinopathy. The greater upregulation of the genes of interest with stress deprivation in SST versus AT suggests that tendon-specific responses to changes in mechanical loading may be related to the in vivo loading that the tendon is designed to transmit. Importantly, these findings may also have implications for the definition of controls in future studies involving molecular analysis during in vitro tendon loading.
Acknowledgments
The authors gratefully acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC), Canadian Institutes of Health Research (CIHR), Alberta Heritage Foundation for Medical Research (AHFMR) and the McCaig Fund.
References
Footnotes
-
Competing interests None.