Although inflammatory cells and their products are involved in various pathological processes, a possible role in tendon dysfunction has never been convincingly confirmed and extensively investigated. The goal of this study was to determine whether or not an acute inflammatory process deprived of mechanical trauma can induce nonspecific damages to intact collagen fibers. To induce leukocyte accumulation, carrageenan was injected into rat Achilles tendons. We first tested the effect of leukocyte recruitment on the concentrations or activities of matrix metalloproteinases and tissue inhibitors of matrix metalloproteinases. Second, we analyzed at the biochemical, histological, and biomechanical levels the impact of leukocyte invasion on tendons. Finally, collagen bundles isolated from rat-tail tendons were exposed in vitro to mechanical stress and/or inflammatory cells to determine if mechanical loading could protect tendons from the leukocyte proteolytic activity. Carrageenan-induced leukocyte accumulation was associated with an increased matrix metalloproteinase activity and a decreased content of tissue inhibitors of matrix metalloproteinases. However, hydroxyproline content and load to failure did not change significantly in these tendons. Interestingly, mechanical stress, when applied in vitro, protected collagen bundles from inflammatory cell-induced deterioration. Together, our results suggest that acute inflammation does not induce damages to intact and mechanically stressed collagen fibers. This protective effect would not rely on increased tissue inhibitors of matrix metalloproteinases content but would rather be conferred to the intrinsic resistance of mechanically loaded collagen fibers to proteolytic degradation.
- matrix metalloproteinases
- tissue inhibitors of metalloproteinase
dampening of inflammation has proven efficient in alleviating a number of pathological states and remains one of the most investigated therapeutic strategies (9). Specifically, the administration of nonsteroidal anti-inflammatory drugs or cryotherapy are common clinical practices to control the inflammatory process following connective tissue injury (21, 22). However, there are contradictory conclusions on the potential of anti-inflammatory strategies to either prevent nonspecific damages or accelerate healing after acute tendon or ligament trauma (1, 4, 18, 24). This controversy is clearly fed by the opposite roles of leukocytes regarding healing and by the intrinsic capacity of tendons to minimize collagen degradation after an acute trauma.
The first phase of inflammation is thought to be highly catabolic. Accumulating leukocytes perform extensive phagocytosis and release a plethora of potentially damaging enzymes. Among them, matrix metalloproteinases (MMPs) are believed to induce nonspecific damages to the extracellular matrix (ECM). For example, gelatinases MMP-2 and MMP-9 are released and activated following acute trauma such as cardiac puncture (27). However, direct contribution of those enzymes to tissue injury is still a matter of debate. Different defense mechanisms are in place to minimize nonspecific damages to the ECM following an acute trauma. Indeed, tissue inhibitors of metalloproteinases (TIMPs) can counterbalance the actions of MMPs and limit ECM damages under proinflammatory conditions (10, 30). In addition, thrombospondin can inactivate MMPs and accelerate their clearance from the extracellular space (12). The cleavage and internalization of MMPs following recognition by specific cellular receptors are other means to neutralize their activity (3, 7).
Cumulative evidences also show that, apart from releasing catabolic enzymes, inflammatory cells can ultimately favor tissue healing through different biological processes. Indeed, after the clearance of pathogens and other signals of danger, neutrophils will undergo apoptosis and be engulfed by phagocytic macrophages. Macrophages will then release growth factors that can stimulate fibroblast proliferation, collagen synthesis, and angiogenesis during the early phase of healing (5, 8, 15, 19). Because inflammatory cells can favor healing and defense mechanisms can counterbalance ECM degradation, additional support is needed to validate whether or not the inflammatory response induces nonspecific damages to tendon collagen bundles following an acute injury.
Based on this rationale, this investigation examined the contribution of leukocyte invasion to tendon damage. More specifically, we tested the hypothesis that foreign body-induced accumulation of leukocytes would be associated with biomechanical impairment of collagen bundles. Tendon inflammation was therefore achieved by injecting carrageenan (Carr), a vegetal polysaccharide devoid of endogenous proteolytic activity, in intact tendons. In this experiment, we were able to determine that massive accumulation of neutrophils and macrophages did not cause a decrease of hydroxyproline content, a marker of collagen, which translated into unaltered mechanical properties. Ex vivo experiments using tendon explants showed that mechanical loading per se can protect tendon collagen bundles from damages related to inflammation.
MATERIALS AND METHODS
Animals and tendon inflammation model.
To establish the necessary amount of Carr to induce extensive accumulation of inflammatory cells in the Achilles tendon of rats, we injected increasing doses (0–800 μg) of Carr within and around the Achilles tendon, using a previously validated procedure (17). Two hundreds micrograms of Carr in a volume of 30 μl was finally selected for the remaining of the experimental protocol because it produces an accumulation of neutrophils and macrophages that is histologically and quantitatively similar to what was observed in our model of tendon injury where collagenase was injected in the Achilles tendon (17). Briefly, female Wistar rats weighing 175–200 g were anesthetized using a cocktail containing ketamine (Pfizer Canada; Kirkland, Quebec, Canada) (87.5 mg/kg) and xylazine (Bayer; Toronto, Ontario, Canada) (12.5 mg/kg). Three groups of rats were used for these experiments; for all animals, skin was surgically prepared, and both Achilles tendons were percutaneously injected with type IV λ-carrageenan (Sigma; St. Louis, MO) (Carr group) autoclaved and dissolved in phosphate-buffered saline (PBS) at a pH of 7.4, or with an equal volume of PBS (Veh group). All experimental animals were euthanized at 1 or 3 days postinjection for histological, biochemical, or biomechanical analyses. Ambulatory control rats matched for age (Ctr group) were also used for comparison. All animal care and handling was approved by the Council of Animal Protection at Laval University.
Immunolabeling of inflammatory cells.
At death, hindlimbs were removed, positioned in a dorsiflexion position, and immersed in a zinc-based fixation solution for up to 48 h. After fixation, tendons were dissected and embedded in paraffin (18). Longitudinal sections were performed at a 5-μm thickness and overlaid on positively charged microscope slides for immunolabeling of neutrophils (anti-neutrophil from Accurate Chemical; Westbury, NY) and macrophages (anti-CD68 from Serotec; Raleigh, NC). Immunolabeling was performed as previously described (24). Briefly, sections were rehydrated, digested in 0.25% trypsin, quenched in 3% H2O2, blocked for 30 min, incubated with primary antibody for 2 h, washed with PBS, and incubated for 1 h with biotinylated anti-rabbit or anti-mouse IgG antibodies (Vector Laboratories; Burlington, Ontario, Canada). After washing with PBS, tissues were incubated with avidin-coupled horseradish peroxidase (Vector Laboratories), and labeling was revealed with AEC kit (Vector Laboratories). All steps were performed at room temperature.
Positive cells were counted under a light microscope and expressed relatively to the volume of tissue examined to obtain a cell density. We used a systematic sampling procedure where ∼15% of total surface is counted. Three distinct lanes of contiguous fields covering the whole width of the tendon were systematically selected at each third of the total length of the longitudinal section. Cell densities obtained with this sampling method provided directly proportional values and a correlation coefficient larger than 0.9 when plotted against data generated by counting 100% of the tissue section.
In-gel zymography and Western blotting.
Immediately after death, Ctr, Veh, and Carr tendons were harvested, frozen in liquid nitrogen, and stored at −80°C. On the day of the experiment, whole tendons were thawed, minced with a sharp blade, and put in 25 μl of extraction buffer per milligram of tendon. For in-gel zymography, buffer consisted of 0.5 M Tris, pH 6.8, 10% glycerol, 2% SDS. For Western blotting, lysis buffer contained 10 mM Tris, pH 7.5, 10 mM NaCl, 0.1 mM EDTA, 0.5% Triton X-100, 0.02% NaN3, 0.2 mM PMSF dissolved in DMSO, and a cocktail of protease inhibitors 1:1,000 (Sigma). Tendons were homogenized using a mortar in which small glass beads were added. After homogenization, tendon extracts were vigorously vortexed, incubated on ice for 30 min, and spun for 5 min at 16,000 g (8). Supernatant was collected, divided in aliquots, and stored at −80°C. Protein content was evaluated with the Lowry-type bicinchoninic acid protein assay (Pierce; Rockford, IL) to ensure equal protein loading.
Gelatin zymography was performed by migrating equal amount of proteins by means of SDS-PAGE. Gels contained 9% polyacrylamide and 1 mg/ml porcine gelatin (Sigma). After electrophoresis, gels were washed three times in 2.5% Triton X-100 for renaturation of proteins. Gels were equilibrated and incubated overnight at 37°C in digestion buffer containing 50 mM Tris, pH 7.5, 5 mM CaCl2, 1 μM ZnCl2. Gels were stained for at least 1 h in 0.25% Coomassie brilliant blue and destained in acetic acid-methanol solvent. Lysis bands, corresponding to MMP activity, were quantified using the image analysis software MetaMorph (Universal Imaging; Sunnyvale, CA) (33).
For Western blotting, equal amounts of proteins were separated by SDS-PAGE in gels containing 15% polyacrylamide. After electrophoretic separation, proteins were transferred onto polyvinylidene difluoride membranes and labeled overnight at 4°C with a mouse anti-human TIMP-1 (AB800 from Chemicon; Temecula, CA) or mouse anti-TIMP-2 (AB801 from Chemicon). Membranes were then incubated with horseradish peroxidase-coupled anti-rabbit IgG (GE Healthcare; Baie d'Urfé, Québec, Canada) and revealed with Western Lightning (PerkinElmer; Boston, MA). Films (GE Healthcare) were digitalized, and optical density of the bands was quantified using MetaMorph. The same rat myocardium or tendon extract was loaded on every gel for TIMP-1 or TIMP-2 immunoblotting, respectively, as internal controls.
To evaluate the effect of Carr on the biomechanical properties of rat Achilles tendons, load to failure was measured as described previously (18). In brief, Achilles tendons were clamped into metallic jaws attached to a MTS 858 Mini Bionix II device (MTS Systems; Eden Prairie, MN) consisting of a hydraulic-driven linear voltage differential transformer connected to a 0.5-kN load cell. Initial length was manually set at a force of 1–2 N. The test was conducted with model 793.00 System Software version 3.2d (MTS Systems) to obtain tension-elongation curves using a strain rate of 10% of initial length per second, until rupture. Force and elongation were monitored at a frequency of 10 Hz. Tension-elongation curves were plotted and discarded if there was an absence of midsubstance failure and of a clear rupture point. Load to failure was defined as the maximal tension a tendon could sustain before showing a rapid and drastic loss of tensional strength after the linear phase of the tension-elongation curve (23).
Determination of hydroxyproline content.
Hydroxyproline content was quantified as previously described (18). Briefly, Achilles tendons were dehydrated for 24 h at 60°C, and dry mass was noted. After dehydration, tendons were hydrolyzed in 6 N HCl at 130°C for 3 h, acidity was neutralized with NaOH, and samples were further diluted with water in a final volume of 10 ml. A fraction of this volume was used for the determination of hydroxyproline content by the method of Woessner et al. (32).
Rat-tail tendon assays.
To test the hypothesis that inflammatory cells and their by-products would preferentially damage loose collagen fibers rather than stressed ones, collagen bundles were isolated from rat-tail tendons and divided in two 3-cm portions. These two segments of the same bundle were placed in culture in Ham's F12 K medium containing a cocktail of antibiotics and antimycotics (Invitrogen Canada; Burlington, Ontario, Canada) plus 2% fetal bovine serum (Wisent; St. Bruno, Quebec, Canada) (see ⇓⇓⇓⇓⇓⇓Fig. 7A for conceptual details). Although we determined that the absolute loads to failure were not different between both segments of a same bundle (not shown), an equal number of proximal and distal segments of the collagen bundles were alternatively stressed or let loose for the 24-h incubation period. A tension sufficient to remove collagen's crimping (10 g, which corresponded to approximately 1–2% strain; deformation per unit length) was applied in the stressed group. For each individual experiment, the paired stressed and loose tendons were then placed for 24 h in a same dish containing the activated leukocytes (8 × 105/ml) or culture medium alone. Leukocytes were harvested from rat air pouches 8 h after the injection of 2 mg of Carr with 2 lavages of 10 ml culture medium; they were spun and resuspended in 5 ml Ham's F12 K.
Following the 24-h incubation period, collagen bundle extremities were then clamped in hydrostatic forceps that were attached to a dual-mode (force and displacement) lever arm (Cambridge Technology; Cambridge, MA) that allows carrying out tension elongation tests at a rate of 10% of initial length per second. Tests were driven using DMC/DMA Software (Aurora Scientific; Aurora, Ontario, Canada). All data are presented relative to a control group in which no inflammatory cells were added.
Scheffé's test was used to compare groups when significant F ratios were obtained following the one-way or two-way ANOVAs. For the in vitro experiments in which we tested whether or not inflammatory cells would preferentially damage unloaded tendons, unpaired t-test was used to compare the means of the two experimental treatments. All results are presented as means ± SE, and the level of significance was set at P < 0.05.
Injection of Carr and inflammatory cell accumulation.
The injection of increasing doses of Carr induced a significant accumulation of neutrophils and macrophages at 24 h postinjection that reached a plateau at a dose of 200 μg (Fig. 1A). Maximal density was approximately 2 × 105/mm3 and 3.5 × 105/mm3 for neutrophils and macrophages, respectively. The density of neutrophils (Fig. 1B) significantly decreased 3 days postinjection while the density of macrophages (Fig. 1C) remained constant. Leukocytes were mainly located into the paratenon and between collagen fibers and tissues appeared edematous following Carr injection (Fig. 2C).
MMP activity and TIMP content in inflamed tendon.
Zymography was performed to confirm that inflammatory cell accumulation was associated with an increase in MMP activity (Fig. 3). Lysis bands were visible at ≈92, 74, 72, 66, and 45 kDa. The 74- to 72-kDa and 66-kDa bands correspond to pro- and active-MMP-2, respectively (6). MMP-9 (92 kDa) and MMP-13 (45 kDa) activities were insignificant compared with MMP-2 at those time points (Fig. 3B). Our findings indicate that the injection of 200 μg of Carr significantly increased by fourfold the amount of active MMP-2 (66 kDa) compared with Veh-injected tendons, suggesting that leukocyte recruitment would provide an additional source of MMPs or that they stimulate tenoblasts to produce MMPs (Fig. 3). We could not detect any significant modulation of the rat collagenase MMP-13 using in-gel zymography (not shown).
Western blots were also performed on extracts from Ctr and Veh- and Carr-injected tendons to test whether TIMP content would parallel the increase in MMPs (Fig. 4, A and B). The injection of Veh had a tendency to reduce the concentration of TIMP-1 at 1 day posttrauma (P < 0.11), but it rapidly returned to Ctr level on day 3. Carr injection reduced the concentration of TIMP-1 and TIMP-2 on day 1 compared with Ctr. TIMP-1 and TIMP-2 concentrations still remained significantly reduced, by ∼45%, 3 days postinjection, compared with Veh and Ctr groups.
Hydroxyproline content and biomechanical properties following Carr injection.
Since MMP activity increased and TIMP concentrations decreased in Carr-injected tendons, we tested the possibility that accumulation of leukocytes may alter the integrity of the ECM. Hydroxyproline content of Ctr tendons was ∼8.2% of dry mass. Three days following the injection of 200 μg of Carr, the concentration of hydroxyproline was identical between Ctr, Veh-, and Carr-injected tendons, which strongly suggests that collagen was not significantly degraded in all experimental conditions (Fig. 5). Mechanical analyses were performed to further confirm that collagen content and integrity were preserved and not affected by the presence of a high density of leukocytes. Absolute load to failure of Ctr tendons was ≈75 N. The injection of 200 μg Carr did not significantly reduce the load to failure 1 and 3 days posttrauma compared with Veh-injected or Ctr tendons (Fig. 6). Experiments where rat-tail tendons were isolated and cultured for 24 h in presence or absence of inflammatory cells (Fig. 7A) showed that the load to failure of loose tendons was significantly decreased by ∼25% compared with the stressed ones (Fig. 7B).
Leukocytes, especially when chronically present, may damage normal tissue. A number of clinical conditions such as ischemia-reperfusion injury (25), genetically inherited muscle pathologies (29), cystic fibrosis (28), and emphysema (26) have been linked to leukocyte-mediated tissue damage. In these pathologies, the chronically infiltrated tissue will often end up with scar, fibrosis, and a decreased functional capacity. However, the formation of scar tissue and fibrosis is a prerequisite for tendon healing following injury and raises the question as to whether or not leukocytes are beneficial or detrimental for tendon healing. Indeed, a possible injurious role for inflammatory cells has never been convincingly confirmed in healthy and intact tendons. Our results show that accumulation of inflammatory cells is associated with an increase of MMP-2 activity, which does not lead to tendon degradation and biomechanical dysfunction, and that tensile stress can play a protective role in preventing loss in mechanical function.
Injection of Carr induces the accumulation of inflammatory cells.
Carr is a linear sulfated polysaccharide extracted from seaweed that has no proteolytic activity and is experimentally used to induce reversible inflammatory response. Injection of Carr into rat Achilles tendon was previously shown to stimulate the infiltration of inflammatory cells and to impair rat ambulation (11). This biological response was rapidly reversible, supporting the transient nature of the inflammatory process induced by Carr injection. However, the subset of leukocytes recruited and the functional impact of these cells have never been fully investigated in tendon. The present findings showed that increasing doses of Carr induce levels of leukocyte accumulation of various magnitudes. No matter the intensity of the proinflammatory stimuli, macrophages were the predominant cell type that accumulated at 24 and 72 h posttrauma. This result contrasted with the model of collagenase-induced tendon injury in which the accumulation of neutrophils was more important than macrophages at 24 h postinjury (17). This discrepancy may be due to the fact that collagenase induces the generation of neutrophil-directed chemotactic collagen fragments (31) as well as vascular disruption and bleeding, while Carr induced an inflammatory response that relies mainly on diapedesis with apparently unruptured blood vessels. Since Carr generates an inflammatory response in the absence of a primary mechanical insult, we believe that this model of tendon inflammation is well suited to dissect out the intrinsic role of inflammatory cells in inducing nonspecific damages.
Carr induced an increase in gelatinolytic activity without functional tendon deficit.
Although the implications are obvious, very little is known regarding the regulation of MMP activities in normal and pathological tendons. The present findings demonstrated that normal tendons contain basal levels of MMP-2, as well as very low amounts of MMP-9, which is consistent with data published by Koskinen et al. (10). Under various conditions, neutrophils and macrophages can release various MMPs, such as MMP-2 as well as MMP-9, and induce tissue degradation (33). Despite early neutrophil and macrophage accumulation, the injection of Carr was not associated with an increased MMP-9 activity but with the activation of MMP-2 only. This increased gelatinolytic activity neither altered collagen content nor influenced load to failure 1 and 3 days postinjection, even in the presence of a slight MMP-13 like activity. Although unexpected, these results are consistent with the observation that MMP-2 deficiency did not protect myocardium but exacerbated inflammation in a model of cardiomyopathy (20). Thus the detrimental role of MMP as a causal agent leading to nonspecific tissue damage has yet to receive strong support and remains to be proven in many animal models of injury.
Carr injection modulates TIMP-1 and TIMP-2 concentrations.
Although the most likely protective mechanism against MMP activity in vitro is the binding of TIMPs to MMPs (13), our in vivo results clearly illustrate that the amount of TIMP-1 and TIMP-2 is decreased following Carr injection. A decrease in the expression of TIMP-1, -2, -3, and -4 is frequently observed under various tendon pathological states (6, 16). For example, Fu et al. (6) showed that TIMP-1 expression was altered in tenoblasts isolated from tendons displaying features of tendinosis compared with healthy tissue, indicating a possible link with the degenerative processes in tendon pathology. On the other hand, TIMP-1 and TIMP-2 are released from tendons following application of a physiological stress, like treadmill running (10). Clearly, the mechanisms controlling TIMPs synthesis in tendons under different pathophysiological conditions are still far from being understood. Our findings suggest that downregulation of TIMP-1 and TIMP-2 is not necessarily associated with collagen degradation and that other protective mechanisms are in place to prevent tendon collagen from degradation in inflamed but uninjured tendons.
Tensile stress protects tendons from inflammatory cell-induced degradation.
A key finding of the present experiments is the observation that a constant but minimal mechanical stress on rat-tail tendon reduced significantly the damage induced by activated leukocytes in vitro. Previous results have shown in vitro that mechanical stress can modulate the MMP/TIMP balance. For instance, mechanical stress can decrease MMP-13 synthesis in rat tenocytes/tenoblasts (2), whereas unloading can increase collagenase expression (14). On the other hand, proinflammatory stimuli can cooperate synergistically with mechanical stress to increase MMP-3 activity (2). Nabeshima et al. (23) also showed that tensile stress-derived protection against bacterial collagenase would rather be conferred by the architectural organization of collagen bundles. They further speculated that the three-dimensional conformation of stressed collagen might mask the proteinase cleavage sites, which suggests that independently of any cellular activity, fibrillar collagen is protected from degradation when submitted to mechanical stress. Thus mechanical stress can obviously protect against self-proteolysis but the cellular and biological mechanisms remain elusive.
In summary, our results show that endogenous release and activation of MMPs during the inflammatory phase in vivo does not inevitably impair the biomechanical properties of tendons. Data suggest that mechanical tension, rather than massive TIMP release, would be the primary mechanism for that reduced susceptibility to enzymatic degradation. On the other hand, the possibility remains that inflammatory cells may cause damage if collagen continuity is lost and/or mechanical loading on those fibers cannot be applied. Finally, the percentages, duration, and frequency of strain that would be physiologically relevant for ruptured and nonruptured tendons during acute or chronic inflammation should certainly be better defined to accelerate tendon repair and improve the rehabilitation treatment.
This research was supported by Fonds de Recherche en Santé du Québec, Natural Sciences and Engineering Research Council of Canada, Canadian Foundation for Innovation, and Canadian Institutes of Health Research.
We thank Marie-Hélène Tremblay for expertise and technical assistance.
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.
- Copyright © 2007 the American Physiological Society