Aged Tendons Lack Adaptive Response to Acute Compressive Injury

Abstract

Rotator cuff tendinopathy has a multi-factorial etiology, with both aging and external compression found to influence disease progression. However, tendon’s response to these factors is still poorly understood and in vivo animal models make it difficult to decouple these effects. Therefore, we developed an explant culture model that allows us to directly apply compression to tendons and then observe their biological responses. Using this model, we applied a single acute compressive injury to C57BL/6J flexor digitorum longus tendon explants and observed changes in viability, metabolic activity, matrix composition, matrix biosynthesis, matrix structure, gene expression, and mechanical properties. We hypothesized that a single acute compressive load would result in an injury response in tendon and that this effect would be amplified in aged tendons. We found that young tendons had increased matrix turnover with a decrease in small leucine-rich proteoglycans, increase in compression-resistant proteoglycan aggrecan, increase in collagen synthesis, and an upregulation of collagen-degrading MMP-9. Aged tendons lacked any of these adaptive responses and instead had decreased metabolic activity and collagen synthesis. This implies that aged tendons lack the adaptation mechanisms required to return to homeostasis, and therefore are at greater risk for compression-induced injury. Overall, we present a novel compressive injury model that demonstrates lasting age-dependent changes and has the potential to examine the long-term response of tendon to a variety of compressive loading conditions.

Introduction

There are over 26 million musculoskeletal injuries annually in the United States, with a disproportionately high prevalence in the aged population.¹ Rotator cuff tendinopathy is especially common, affecting over 50% of the population above the age of 60.² Despite the commonality of rotator cuff tendinopathy, the exact etiology is multi-factorial and most of these factors are still poorly understood.^2–4 Previous studies have identified that external compression of rotator cuff tendons influences disease progression.^2,5 Tendons experience tensile forces along their longitudinal direction as a response to muscle contractions, but some tendons also experience localized compressive forces where they pass around neighboring bony structures. The supraspinatus tendon in the rotator cuff is one such tendon, as it experiences compression due to its positioning underneath the acromioclavicular (AC) joint.² In aged subjects, there are multiple degenerative changes in the AC joint, such as narrowing of the sub-acromial space, degeneration of cartilage, and osteophyte formation, that could impact efficient tendon function.^6,7

To study the degenerative changes that occur during the multi-factorial onset of tendon disease, models are necessary to isolate specific variables. Most commonly, the onset of tendinopathy is studied using in vivo animal models. These studies have used a variety of methods to induce a tendinopathic phenotype, such as overuse exercise,^5,8–12 injection of pro-inflammatory agents,^13,14 and the application of non-homeostatic tension.^15–17 These models are useful because they allow for the observation of compositional and mechanical changes in tendon without disrupting the native tendon structure or its interactions with neighboring tissue. However, in vivo tendon models are complex due to the interactions between tissues (tendon, muscle, and bone), infiltration of inflammatory cells from neighboring tissue, and the geometry. It is also difficult to control the magnitude and direction of mechanical loading, particularly for compressive and shear loading. When the effects of aging are considered as well, age-related changes such as muscle weakness, poor posture, and soft tissue tightness can further complicate these models.² Due to these limitations, some studies have looked at the tenocyte-specific responses using explant model systems. Although extensive research has looked at the effects of compression on similar tissues like cartilage and ligament,^18–24 loading studies in tendon have mainly investigated the biological response to changes in tension.^25–27

Therefore, the purpose of this study is to develop an in vitro explant culture model to explore age-related differences in the response to a single acute compressive injury. We previously developed a murine flexor tendon explant model that allows for the analysis of cell behavior without disruption to the native extracellular matrix, cell-cell, and cell-matrix connections. Building upon this, we designed and built a custom-designed biaxial loading bioreactor, allowing for the application of controlled loads in tension and compression simultaneously. Now, we can observe the response of tendon to compression and decouple age-related changes in other organ systems from local responses. We hypothesized that an acute compressive load would result in an injury response, with a loss of collagen fiber organization, transition to a more cartilage-like phenotype, and increased expression of markers associated with matrix degradation and inflammation. We also hypothesized that the aged group would show a larger injury response with larger and earlier increases in inflammation and degradation markers.

Methods

Sample Preparation

Directly following sacrifice, flexor digitorum longus (FDL) tendon explants were harvested from the hind limbs of young (4 months; n=88 animals) and aged (22–24 months; n=79 animals) C57BL/6J male mice using previously described methods²⁸ per approved animal use protocol (BU IACUC PROTO202000046).

After harvest, all explants were immediately loaded into grips with a 10-mm gauge length (starting from the base of toes) and placed into a custom-built multiaxial loading bioreactor (Figure 1A–B). The bioreactor consists of a miniature load cell (5g), linear stepper motor, and linear variable differential transformer (LVDT) for each axis to enable precision in both load- and displacement-controlled experiments. For all experiments in this study, we utilized displacement control where the resolution of our displacement system is approximately 5µm. All explants were then loaded to 3% static tensile strain at 1% strain/sec and held at this tension for the duration of the experiment. Immediately after tensioning, the compression (“C”) group was then preloaded in compression to 20g to ensure contact between compressor and tendon, compressed to 50% strain over 1 second, and then the compression was released. The indenter is flat with a radius of 8mm to ensure that the entire width and most of the length of the tendon is compressed. The non-compression (“NC”) group received only the static tensile strain. All tendons were tensioned within 1 hour of harvest. Mice were randomly allocated to C and NC groups.

Figure 1.

(A) Biaxial loading device used to maintain tension while inducing acute compressive injury at D0 with (B) a close-up of a single well showing a tendon being compressively loaded and (C) the compressive loading profile. (D) Study design outline with young and aged mice either receiving static tension and acute compression (Compression or “C”) or just static tension (Noncompression, or “NC”). Tissue from each group were collected for assays on Days 0, 1, and 7.

Explant Culture

Throughout culture, explants were kept in culture medium consisting of low glucose Dulbecco’s Modified Eagle’s Media (1 g/l; Fisher Scientific) supplemented with 10% fetal bovine serum (Cytiva, Marlborough, MA), 100 units/mL penicillin G, 100 µg/mL streptomycin (Fisher Scientific), and 0.25 µg/mL Amphotericin B (Sigma-Aldrich). Medium was changed every other day and spent culture medium was collected to measure matrix metalloproteinase (MMP) and secreted protein activity. Explants were collected at day 0 (D0), day 1 (D1), and day 7 (D7) to assess viability, metabolic activity, matrix composition, matrix biosynthesis, matrix structure, gene expression, and mechanical properties.

Viability, Metabolism, Biosynthesis, and Composition

Cell viability was assessed through live/dead staining in 1x PBS containing fluorescein diacetate (4 mg/mL; Fisher Scientific) and propidium iodide (1 mg/mL; Fisher Scientific). Tendons (n=4/group/day) were imaged using a FLUOVIEW FV3000 confocal laser scanning microscope (Olympus Scientific Solutions, Waltham, MA) at 10x magnification to obtain z-stacks for each sample. A custom MATLAB script was then used to combine the z-stacks into a maximum intensity z-projection and quantify the live and dead cells using a fluorescent intensity based automated thresholding algorithm. Explant cell metabolism was measured with a resazurin reduction assay as previously described.²⁹ Following a 3-hour incubation with a resazurin and culture medium solution, intensity of the reduced product, resorufin, was measured at 554/584 nm. Values were then normalized to those of control wells without explants, such that a value of 1 represents a well without viable tissue. Synthesis of DNA, sulfated glycosaminoglycans (sGAG), and total protein was measured by 24-hour incorporation of ³H-thymidine (1 μCi/ml), ³⁵S- sulfate (20 μCi/ml), and ³H-proline (1 μCi/ml) respectively (Perkin-Elmer, Norwalk, CT). Following culture, tendon wet weight was obtained by immersing tendons in 1x PBS for 1 minute, dabbing excess PBS, and then weighing tendon. Tendons were then lyophilized for 3 hours and tendon dry weight was obtained. All weights were obtained in triplicate and percent water content was calculated by dividing the dry weight by wet weight, subtracting the resulting value from 1, and then multiplied by 100. Tendons were then digested overnight with 5 μg/mL proteinase K (Sigma-Aldrich, St. Louis, MO). Radiolabel incorporation rate was measured using a liquid scintillation counter (Perkin-Elmer).³⁰ DNA content was measured through the PicoGreen dye-binding assay.³¹ sGAG content was measured using the dimethylmethylene blue (DMMB) assay.³² Total collagen content was measured using the hydroxyproline (OHP) assay.³³ Biosynthesis and composition data were normalized to tendon dry weight.

MMP Activity

Activity of MMPs (1,2,3,7,8,9,10,13,14) was determined via analysis of spent culture medium (n= 5/group/day) using a FRET-based generic MMP activity kit (SensoLyte 520 Generic MMP Activity Kit Fluorimetric, Anaspec, Fremont, CA).

Histology

Following culture, explants (n=4-6/group/day) were embedded in optimal cutting temperature (OCT) compound (Fisher Scientific) and 10 µm thick cross sections were collected on glass slides. Prior to staining, sections were fixed using 10% buffered formalin (Fisher Scientific). For hematoxylin and eosin (H&E) staining, slides were immersed in hematoxylin (Harris Hematoxylin; Electron Microscopy Sciences) for 4 min and eosin (Electron Microscopy Sciences) for 10 seconds. For toluidine blue staining, slides were immersed in toluidine blue (Fisher Scientific) for 1 min while being gently agitated up and down. Both H&E and toluidine blue slides were imaged at 20x with a VS120 virtual slide scanner (Olympus Scientific Solutions, Waltham, MA). H&E images were evaluated using a custom MATLAB script for cell density and cell aspect ratio. Toluidine blue images were evaluated using a custom MATLAB script that calculates the percentage of the total stained region that is filled with purple hues, which is indicative of sGAG content. Second harmonic generation (SHG) images were obtained with an Ultima 2-photon laser scanning microscopy system (Bruker Fluorescence Microscopy). Two-photon excitation was provided by a Chameleon Ultra femtosecond Ti:Sapphire laser (Coherent) tuned to 1190 nm. SHG images were evaluated for collagen straightness and alignment using the CT-FIRE software from the University of Wisconsin-Madison³⁴, which extracts individual collagen fibers from SHG images. For all histological analyses, the ends of the tendon were excluded to ensure that analyses only included the indented area of the tendon.

Gene Expression

Explants at D0 and D7 (n=5-7/group/day) were flash frozen in liquid nitrogen then placed at −80℃ until RNA extraction. Samples were placed in Trizol reagent, homogenized with a bead homogenizer (Benchmark Scientific), and then separated using phase-gel tubes (Qiagen). The supernatant was then purified according to the Zymo Quick-RNA purification kit protocol (Zymo Research). The RNA was then converted into cDNA with reverse transcription and qPCR was performed with the Applied Biosystems StepOne Plus RT-PCR (Applied Biosystems, Foster City, CA). Expression for each gene was calculated using the ΔΔCT method and was normalized to the housekeeping gene β-actin and day 0 values.

Secreted Protein Analysis

Protein levels of proinflammatory cytokines (IFN-γ, IL-10, IL-12p70, IL-1β, IL-6, KC/GRO, TNF-α) were determined via analysis of spent culture medium (n= 3-5/group/day) using a V-Plex Proinflammatory Panel kit (Meso Scale Discovery, Rockville, MD).

Mechanical Properties

After culture, tendons (n=6-8/group/day) were gripped with sandpaper at a 5-mm gauge length and loaded into a custom-built tank filled with PBS. All samples were loaded on a uniaxial testing system (Model 5944, Instron, Norwood, MA) with a protocol consisting of preload (0.02 N), preconditioning (10 cycles between 0.02 and 0.04 N), a 4% stress relaxation, and a ramp to failure at 0.1% strain per second. Load measurements were obtained with a 10 N load cell with a resolution of 0.01 N. Cross-sectional area was determined from post-preload images assuming an elliptical tendon cross-section. Maximum stress and strain, percent relaxation, stiffness, and modulus of the linear region were calculated using a custom MATLAB software.

Statistics

Values more than 2 standard deviations from the mean were considered outliers and removed. Statistical analyses were performed in GraphPad Prism 8 (GraphPad, San Diego, CA) via two-way ANOVAs with Bonferroni post-hoc corrected t-tests where significance was set at p<0.05. All data are presented as mean ± 95% confidence interval.

Results

Neither compression nor aging caused a significant difference in cell viability (Figure 2A). In both age groups, metabolic activity increased over time in culture, with a significant difference at D0 between young and aged tendons. Compressive injury also decreased metabolic activity in the aged group when comparing to NC at day 7 (Figure 2B). DNA content decreased over time in culture for both young and aged explants, with lower DNA content in aged samples (Figure 2C). There was no effect of compression on DNA content at D7. Cell proliferation increased from D1 to D7 in both groups but was stimulated by compressive injury in the young group only (Figure 2D).

Figure 2.

Measurements of (A) cell viability, (B) metabolic activity, (C) DNA content, and (D) cell proliferation in the young (blue, circles) and aged (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance between age groups is denoted with a star (*) and significance between time points is denoted with a bar (−) with p < 0.05.

sGAG content decreased over time in young explants, while remaining constant in aged tendons. Compressed young tendons had increased sGAG content compared to non-compressed tendons by day 7, with no differences observed in the aged group (Figure 3A). sGAG synthesis was increased as a function of both time and compression in the young group, but no differences were found in the aged group (Figure 3B). Collagen content displayed a similar response in both age groups over time, with an increase in content at D1 before returning to baseline levels by D7. Compression increased collagen content in the aged explants compared to NC aged tissues at D7 (Figure 3C). Collagen synthesis remained relatively constant over culture in both aged groups, with higher synthesis in young tendons at D1. Compression suppressed collagen synthesis in aged tendons (Figure 3D). MMP activity increased over culture in the compressed young group and was higher at all time points compared to NC tendons (Figure 3E). Aged tendons had no differences over time in the compressed group, but a decrease from D4 to D6 in the NC group. There was also higher MMP activity in non-compressed aged tendons at D2 only (Figure 3F).

Figure 3.

Measurements of (A) sGAG content, (B) sGAG synthesis, (C) collagen content, and (D) collagen synthesis in the young (blue, circles) and aged (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance between age groups is denoted with a star (*) and significance between time points is denoted with a bar (−) with p < 0.05.

There were no qualitative differences in H&E staining between any of the groups (Figure 4A). Quantitative analysis of the images found no difference in cell density (Figure 4C). Nuclear aspect ratio remained constant in young samples, but it was increased due to compression in aged samples (Figure 4D). A qualitative assessment of toluidine blue images finds more staining in the D7 compressed images at both age groups. We also observed much more stain concentrated in the cytoplasm instead of inside the nuclei in the young D7 compressed images compared to other groups (Figure 4B). The quantitative analysis observed similar results with higher sGAG content in the young samples in D7 compressed images compared to D0 and D7 NC images (Figure 4E). Our SHG images showed no changes in collagen straightness or collagen alignment (Supplemental Figure S1).

Figure 4.

Representative images of (A) H&E and (B) toluidine blue for each group, with quantifications of (C) cell density and (D) nuclear aspect ratio from H&E images and (E) toluidine blue score for the young (blue, circles) and aged (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance between age groups is denoted with a star (*) and significance between time points is denoted with a bar (−) with p < 0.05.

Gene expression of type 1 collagen decreased by D7 when compared to D0 control samples in both aged groups. No changes were observed because of compression (Figure 5A). Type 2 collagen was downregulated due to compression in the young group only. The young D7 compressed group and the aged D7 non-compressed group both decreased from D0 (Figure 5B). Sox9 expression decreased due to compression in the young group (Figure 5C). Among proteoglycans, aggrecan was upregulated and fibromodulin and biglycan were downregulated by compression in the young group only. All groups for decorin and fibromodulin decreased in expression from D0 to D7. Aggrecan expression decreased in the young non-compressed group and increased in the young compressed group (Figure 5D–G). Among the MMPs, compression downregulated MMP-3 and upregulated MMP-9 in young tendons. Expression of MMP-3 in both young groups, MMP-9 in young compressed and both aged groups, and MMP-13 in both young groups and aged non-compressed all increased by D7 compared to D0. Expression of MMP-3 was also downregulated in the aged non-compressed group over time in culture (Figure 5H–K).

Figure 5.

Measurement of mRNA levels of (A) collagen I, (B) collagen II, (C) SOX9, (D) decorin, (E) aggrecan, (F) fibromodulin, (G) biglycan, (H) MMP-3, (I) MMP-8, (J) MMP-9, (K) MMP-13 for the young noncompressed (dark blue, circles), young compressed (blue, squares), aged noncompressed (dark green, circles), and aged compressed (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance from Day 0 is denoted with a star (*) and significance between loading conditions is denoted with a bar (−) with p < 0.05.

We also looked at the gene and protein levels of inflammatory cytokines to investigate injury responses (Figure 6). In the young group, TNF-α expression was upregulated from day 0 in the non-compressed group and further upregulated with compression (Figure 6A). There were no significant differences in IL-6 or IL-1β. Aged tendons exhibit elevated expression of IL-1β and TNF-α in all groups over time in culture (compared to baseline day 0 values), while IL-6 was increased in the non-compressed group and decreased in the compressed group. Interestingly, aged groups showed downregulation of IL-6 with compression. For the protein cytokine levels released into the medium, Il-10, IL12p70, IFN-γ and IL-1β levels were too low to be detected. Compression increased protein levels of IL-6 and TNF-α released from young tendons. Aged compressed tendons had lower IL-6, TNF-α, and KC/GRO than young compressed tendon (Figure 6B).

Figure 6.

Measurement of (A) mRNA levels and (B) protein levels of inflammatory cytokines for the young noncompressed (dark blue, circles), young compressed (blue, squares), aged noncompressed (dark green, circles), and aged compressed (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance from Day 0 is denoted with a star (*) and significance between groups is denoted with a bar (−) with p < 0.05.

Finally, we looked at what mechanical changes occurred because of the alterations to protein content and gene expression following compressive injury. We found that young tendons had increased stress relaxation compared to aged tendons at D7 (Figure 7A). We found no differences in modulus or stiffness in any groups (Figure 7B–C). Maximum stress was higher in D7 tendons compared to D0 tendons in young tendons but no changes were found in maximum strain (Figure 7D–E)

Figure 7.

Measurements of tensile mechanical properties, (A) percent stress relaxation, (B) modulus, (C) stiffness, (D) maximum stress, and (E) maximum strain, in the young (blue, circles) and aged (green, squares) explants. Data are presented as mean ± 95% confidence interval. Significance between age groups is denoted with a star (*) and significance between time points is denoted with a bar (−) with p < 0.05.

Discussion

This is the first study to examine how aging affects the response of tendon to compressive loading in a tendon explant model. Previous studies have primarily looked at how properties, such as mechanics and collagen content, of tendons directly change due to aging,^35–38 but they have not examined how aging effects the response and adaptation of tendon to controlled compressive loads. Our study demonstrates lasting age-dependent changes in response to just a single acute compressive injury. Like previous studies, aged tendons exhibited reduced cell content and metabolic activity than young tendons at baseline.³⁰ Reduced cell function or cell number in aged tendons could limit their capacity for adaptive response compared to that of young tendons. This might explain why we see a decrease in MMP protein levels from D4 to D6 in aged tendons. Compressive injury appeared to stimulate cell proliferation without causing cell death in young tendons, perhaps suggesting cell turnover. This proliferation could also indicate activation of cells to adapt to the compressive injury as tenocytes are naturally somewhat quiescent at homeostasis. Despite affecting many other markers of cell health, we found no changes in cell viability with any of our groups. The viability at D0 is lower than expected, but pilot experiments found that this reduction in viability is due to the loading of tendons into grips. Importantly, there is no reduction in viability over time in culture for any groups. It is also important to note that our cell viability does not correlate with DNA content as our DNA assay measures all double stranded DNA in both live and dead cells. We did, however, find a decrease in DNA content from D0 to D7. Given the lack of cell death, it is unlikely this change in DNA content is due to apoptosis occurring throughout culture but rather due to the removal of cell material from necrosis during the initial harvest. Interestingly, metabolic activity also decreased due to compression, but in aged tendons only. Generally, metabolic activity is increased in healing tendon immediately following an acute injury,³⁹ therefore this decreased metabolism could have a detrimental effect on the healing response of aged tendons to compressive load. However, little is known about how metabolism regulates tenocyte function to restore homeostasis following mechanical injury.

In addition to differences in markers of cell health, we also found differences between groups in the major ECM components, collagen and proteoglycans/sGAGs. We found that compression stimulated sGAG synthesis in young tendons, while having no effect on aged tendons. While excess sGAGs are often associated with tendinopathy, sGAG accumulation is also an expected response to excessive compressive loading since regulation of fluid flow by sGAGs can help tendon better resist compressive loading.⁴⁰ This is further supported in our work by an upregulation of aggrecan, the primary proteoglycan responsible for resisting compressive loads in cartilage, in young tendons. Interestingly, young tendons also exhibited sGAG loss following compressive injury, mirroring previous studies in cartilage.⁴¹ Therefore, increased sGAG synthesis could be compensating for the injury-associated loss. We also see decreased expression of fibromodulin and biglycan, two of the more common proteoglycans in tendon, so it is likely we see injury-associated loss in these small leucine-rich proteoglycans (SLRPs) and then tenocytes adapt by increasing aggrecan synthesis. This shift in proteoglycan expression could also account for the increase in stress relaxation, a measure of viscoelastic properties. Previous studies have found that the combined knockdown of decorin and biglycan leads to an increase in stress relaxation, so we may be seeing a similar phenomenon here as a result of decreased SLRPs due to compression.⁴² Importantly, compressive injury did not alter sGAG turnover or increase aggrecan expression in aged tendons, demonstrating a lack of adaptive response that would allow tendon to properly resist repeated loading.

There was also a decrease in MMP-3 expression with compression in young tendons. This is somewhat surprising because MMP-3 is responsible for degrading proteoglycans, but we see a decrease in SLRP expression. Due to the decrease in sGAG content from D0 to D7, it is possible that the proteoglycan degradation occurs earlier following the injury, and by D7 the expression of MMP-3 has decreased, and sGAG synthesis increases to make up for the injury associated sGAG loss. Future studies should examine the gene expression changes earlier in the adaptation process to observe the immediate gene expression response following injury. Some studies have found a decrease in MMP-3 levels in torn human rotator cuff models, but it is unknown if the MMP-3 protein levels in this model decrease via the same mechanism as our model, as the MMP-3 decrease in vivo occurs over a much longer time scale and likely represents a failure to properly remodel.^43,44

Our results in collagen content and synthesis also point towards a differing adaptive response to compressive loading between young and aged tendons. Aged tendons exhibited increased collagen content in compressed tendons compared to the non-compressed group, suggesting a potential fibrotic response. Fibrosis leads to an excess of disorganized ECM that is not a strong as native tendon tissue.⁴⁵ A fibrotic response in our aged tendons implies that they do not heal as well as young tendons and this is commonly seen clinically in humans with older patients having higher rates of tendon re-tear.⁴⁶ We also see some positive adaptations in the young tendons, like higher collagen synthesis at D1 and increased MMP-9 levels, that we do not see in the aged group. MMP-9 is responsible for degradation of small collagen fibers, and together with the increased synthesis and increased total MMP protein levels, implies that young tendons have increased collagen remodeling due to compression. MMP-13, which is responsible for degrading type I collagen, however, does not change with compressive injury in our study. Tendon rupture and tear models also tend to see an upregulation of MMP-13,⁴³ so perhaps this single compressive injury is not enough stimulus to cause large-scale collagen degradation.

While we see these adaptive changes, there are some changes that are common in tendon overuse models and signify a transition to a more cartilage-like phenotype, that we do not see because of a single acute compressive injury. Overuse models generally see an increase in type II collagen, the primary collagen in tissues that mainly experience compressive loads like cartilage, and sox9, a cartilage marker.^8,9 However, surprisingly we found a decrease in both of these markers. The reason for this decrease is currently unknown, but it is likely that a single acute compressive load is not enough to induce the cartilage-like phenotype. Future studies will see if dynamic compression over an extended loading period can cause changes similar to those seen in overuse models.

Compressive injury not only causes structural changes to the ECM, but we also see changes in the morphology of tenocytes themselves, with an increase in nuclear aspect ratio in aged tendons due to compression. Since we see no corresponding change in young tenocytes, this implies that the ECM of young tendons has remodeled to resist deformation due to the same compressive load. It is possible that the initial increase in sGAG content immediately following injury and the increase in aggrecan expression by D7 in young tendons serves to mediate the strain transfer to cells and prevent deformation due to loading. These results are consistent with previous studies in rat Achilles and tail tendon that found an increase in nuclear aspect ratio with age.^47,48 However, these studies found a baseline difference in aspect ratio without loading, while we only saw a difference with compression. It is surprising though that although the ECM in young tendons appears to resist more compression than that of aged tendons, we see a larger adaptive response in young tenocytes.

We also found differences in the inflammatory profile between our groups. Surprisingly, IL-6 expression was decreased by compression in the aged group. IL-6 has been linked with collagen synthesis in the past⁴⁹, so the decreased IL-6 could be responsible for the decreased collagen synthesis we see in the aged group. We also see higher TNF-α gene expression and protein levels and higher IL-6 protein levels due to compression in the young group, which implies an increased inflammatory response. This confirms an injury response as some of the previously discussed adaptations appear to be beneficial to the tendon. This is strangely not accompanied by increased expression of IL-6 and IL-1β, as many other tendon injury models observe an increase in all 3 genes.⁵⁰ IL-6 and IL-1β are both generally recruited at this point in the healing cascade, but IL-1β is released by tenocytes, in addition to infiltrating neutrophils and macrophages, so it’s surprising to not see any change.⁵¹ Since we see increased IL-6 protein levels but no increased expression, perhaps gene expression levels were increased earlier in the experiments, but have started to return to baseline levels by D7. Future studies will look at earlier PCR timepoints to see if inflammatory cytokine expression is elevated earlier in the compressive injury response. It is also possible that this single compressive injury is not enough to stimulate inflammatory signaling to incite a large-scale healing response, but instead only stimulates signaling molecules involved in healthy ECM turnover. It is important to note that our model only considers the tendon-specific response and ignores the infiltration of inflammatory cells from other tissues that normally occurs during tendon injury, so our inflammatory changes are not entirely representative of the changes that would occur in vivo.

In addition to the lack of infiltrating inflammatory signals from neighboring cells and tissues, our model has a few other limitations. First, our sGAG and collagen radiolabeling assays only measure sulfated GAGs and total protein content respectively. Future studies aim to use more specific proteomic assays to determine the timing and profile of ECM turnover. Also, this study uses the FDL tendon due to the more suitable size and shape for bioreactor experiments. Future studies plan to look at the effects of compression on more clinically relevant tendons such as the supraspinatus or Achilles tendon to address site-specific differences. In addition, this study only looks at changes in male mice, so future studies will determine the role of sex in the response to compressive injury. This study also only looks at the effects of a single acute compressive injury at one specific strain value and does not examine the effects of different loading conditions. Therefore, future studies will examine the effects of different loading durations, frequencies, and strain levels. Finally, this study only investigates the response for 7 days, so future studies will look further out to identify if these adaptations provide long lasting alterations.

Regardless, we present a novel compressive injury model to study aberrant matrix turnover in mouse tendon explants. We demonstrate lasting changes in the response to a single acute injury that appear to be age dependent. We found that aged tendons had fewer adaptive changes to compressive injury, implying that they lack the adaptations required to properly remodel in response to repeated compressive loads and return to homeostasis. This presents one novel mechanism by which aged tendons may be more prone to accumulation of matrix damage that could lead to tendon degeneration or tendinopathy. In addition, the changes in young tendons show the ability of healthy tendon to adapt and remodel to just a single acute load and perhaps show a mechanism by which tendons can be strengthened to better bear future loads. In the future, we will capitalize on our explant system to zero in on tendon’s long-term response to a variety of compressive loading conditions without the added complexities of in vivo models, thus quantifying the initiating factors in compression-related tendon damage and enabling the development of targeted therapeutics to prevent age-related tendon degeneration.