Temporal changes in the muscle extracellular matrix due to volumetric muscle loss injury

Abstract

Purpose/Aim:

Volumetric muscle loss (VML) is a devastating orthopaedic injury resulting in chronic persistent functional deficits, loss of joint range of motion, pathologic fibrotic deposition and lifelong disability. However, there is only limited mechanistic understanding of VML-induced fibrosis. Herein we examined the temporal changes in the fibrotic deposition at 3, 7, 14, 28, and 48 days post-VML injury.

Materials and Methods:

Adult male Lewis rats (n=39) underwent a full thickness ~20% (~85mg) VML injury to the tibialis anterior (TA) muscle unilaterally, the contralateral TA muscle served as the control group. All TA muscles were harvested for biochemical and histologic evaluation.

Results:

The ratio of collagen I/III was decreased at 3, 7, and 14 days post-VML, but significantly increased at 48 days. Decorin content followed an opposite trend, significantly increasing by day 3 before dropping to below control levels by 48 days. Histological evaluation of the defect area indicates a shift from loosely packed collagen at early time points post-VML, to a densely packed fibrotic scar by 48 days.

Conclusions:

The shift from early wound healing efforts to a fibrotic scar with densely packed collagen within the skeletal muscle occurs around 21 days after VML injury through dogmatic synchronous reduction of collagen III and increase in collagen I. Thus, there appears to be an early window for therapeutic intervention to prevent pathologic fibrous tissue formation, potentially by targeting CCN2/CTGF or using decorin as a therapeutic.

INTRODUCTION

Pathologic fibrosis is a common sequela of severe injuries and diseases affecting various tissues, including skeletal muscle. It is characterized by an excessive accumulation of extracellular matrix (ECM) components, namely collagen, and in skeletal muscle is generally seen in conjunction with atrophy.¹ The combination of fibrosis and atrophy limits muscle function by reducing range of motion and maximal force production, effectively reducing the overall quality of life for those affected. In minor acute injuries to skeletal muscle, such as strains and contusions, there is a transient period of inflammation and ECM remodeling, which is ultimately resolved and replaced by repaired and/or regenerated skeletal muscle fibers.² However, in more severe pathologies there appears to be a tipping point where muscle regeneration is repressed, and ECM deposition progresses uninhibited.³

A critical component of the fibrotic repair process is the timely activation and balance of pro-inflammatory M1 and anti-inflammatory M2 macrophages. M1 macrophages are proposed to be responsible for clearing necrotic debris and activating satellite cells, while M2 macrophages control matrix remodeling and fusion of satellite cells.² Notably, M1 and M2 macrophages exist at the ends of a continuum sequence of possible states without clear boundaries, and typically transition in the direction from M1 to M2.⁴ The imbalance or prolonged activation of either end of the continuum will likely result in detrimental effects to the tissue. For example, the initial inflammatory response is crucial for tumor necrosis factor (TNF) mediated apoptosis of fibro/adipogenic progenitors (FAPs), a major source of ECM producing fibroblasts.⁵ However, too early of a transition to an anti-inflammatory state increases transforming growth factor-β1 (TGF-β1) expression, resulting in decreased TNF-induced FAP apoptosis and thus fibroblast accumulation.⁵ Overexpression of TGF-β1 further contributes to fibrotic development by stimulating fibroblasts and regenerating myofibers to produce collagen.⁶ As the area of ECM increases, so does the distance from muscle fibers to capillaries, resulting in a hypoxic local environment. In fact, hypoxia coupled with TGF-β1 expression has recently been found to stimulate myofiber secretion of connective tissue growth factor (CCN2/CTGF), further increasing fibrosis and creating a cycle of progressive fibrotic development.⁷

While muscle pathology research has primarily emphasized the promotion of muscle fiber growth and regeneration without regard to fibrosis, it is becoming more recognized that the microenvironment in which muscle fibers reside is crucial to successful regeneration. The composition of ECM layers in healthy skeletal muscle is fairly well understood,⁸ yet how the ECM becomes altered from various fibrotic pathologies is not as clear. Therefore, a better understanding of the skeletal muscle microenvironment in specific pathologic conditions is necessary to increase the probability of improving regeneration and functional outcomes. Collagen types I and III are the most abundant fibrillar isoforms in skeletal muscle ECM, with the perimysium and epimysium layers predominately composed of type I.⁹ Type III collagen appears to be more abundant in the endomysium layer, yet still a component of all layers.⁹ Type IV collagen is the most abundant non-fibrillar isoform, primarily making up the basement membrane of muscle fibers, while type VI collagen provides a link from the basement membrane to the ECM.¹⁰ Recent work has indicated that type VI collagen is vital for maintaining satellite cell viability in skeletal muscle.¹¹ Moreover, proteoglycans residing within the ECM, namely decorin and biglycan, play an important role in regulating collagen fibril formation and stabilization of the neuromuscular junction. Genetic deficiencies in either decorin or biglycan have been shown to cause structural abnormalities in collagen fibers.^12,13

Conditions presenting with known pathologic fibrosis of skeletal muscle include Duchenne muscular dystrophy (DMD), cerebral palsy (CP), and polymyositis (PM). Interestingly, though they differ in etiology (genetic, congenital, and autoimmune, respectively), all three conditions display fairly similar fibrotic characteristics. Biopsies from DMD, CP, and PM patients are observed to have an increase in the proportion of collagen III in skeletal muscle.^14,15,16 Recently, Smith et al. revealed an increase in decorin content but a decrease in biglycan content in patients with CP,¹⁵ supporting a plausible role in abnormal neuromuscular junctions found in CP,¹⁷ as biglycan plays a role in maintaining synapse stability.¹⁸ Yet, findings in DMD patients have not been as clear, with some reports indicating decreases in decorin and biglycan mRNA,^19,20 although most have found significant increases in mRNA and protein content of both.^21,22,23

Volumetric muscle loss (VML), defined as the “traumatic or surgical loss of skeletal muscle with resultant functional impairment”,²⁴ similarly generates a robust and lasting fibrotic response but is less understood. As VML is the result of an acute traumatic injury rather than a chronic underlying condition, the fibrosis observed after injury may display temporal differences compared to the aforementioned conditions. Two major tissue level pathologies occur with VML injuries. First, there is a loss of muscle fibers, and specifically a loss of contractile tissue. Second, there is an accumulation of pathologic fibrotic tissue, which impedes muscle healing and regeneration, causing destruction of muscle architecture and altering the microenvironment of the muscle. Both are key contributors to strength loss for which there is no current clinically therapy. At the tissue level, it is clear the lost skeletal muscle is incapable of endogenously regenerating and the defect area is replaced by an aberrant overproduction of ECM accumulation.^25,26 Yet at the molecular level, there is limited understanding of the makeup of the resulting fibrotic mass in terms of collagen proportions and organization, which could be transformative to develop and evaluate anti-fibrotic treatments for skeletal muscle. Previous evaluation of VML-injured muscle has indicated an overall increase in collagen content and upregulation of fibrotic genes such as TGF-β1, CCN2/CTGF, and various collagen isoforms (e.g., types I and III) within the defect area and immediately adjacent remaining muscle;^27,28,29 however, understanding of content and ratio of these proteins over time is lacking. Moreover, the extension of fibrosis into more distal regions of the remaining muscle has not been thoroughly observed. Herein the temporal fibrotic response after VML injury was evaluated in a rat model of VML injury to test the hypothesis that ECM components become proportionally altered after injury compared to healthy muscle.

MATERIALS AND METHODS

Ethics approval

All protocols were approved by the Institutional Animal Care and Use Committee at the University of Minnesota; in compliance with the Animal Welfare Act, the Implementing Animal Welfare Regulations and in accordance with the principles of the Guide for the Care and Use of Laboratory Animals.

Experimental design and animal model

Male Lewis rats (Charles River, n=39) underwent unilateral VML injury to the left tibialis anterior (TA) muscle and were randomized to terminal groups of 3, 7, 14, 21, or 48 days post-VML. TA muscles were harvested from injured (left) and contralateral (right) hindlimbs at the given terminal time points for histological and biochemical analysis. Muscles from the contralateral limbs were used as controls for all outcome measures and collapsed into one group. Rats were given at least one week to acclimate to the facility prior to study. All rats were given food and water ad-libitum throughout the duration of the study and maintained on a 14:10 hour dark:light cycle. At the time of surgery and randomization, rats were on average at 3.7 ± 0.2 months of age and 379 ± 21g of weight. At the terminal time point, rats were euthanized by overdose of sodium pentobarbital (>100mg/kg i.p.) while under isoflurane anesthesia and skeletal muscle was harvested and stored for further analysis.

Surgical creation of volumetric muscle loss injury

All rats in the study underwent surgery to create a full thickness VML injury to the left TA muscle, as previously described.^30,31 Rats received analgesic injection (SR buprenorphine, 1.2mg/kg s.q.) roughly 2 hours prior to the start of surgery. Rats were then anesthetized with isoflurane (1.5–3.0%) throughout the duration of the surgery. Briefly, an anterior incision was made through the skin, and the fascia was cut to expose the TA muscle. A metal plate was inserted behind the TA, and a 6mm punch biopsy was taken to remove ~20% (85 ± 13mg) of the TA muscle. Proline sutures were used as a reference to identify the defect area at the time of harvest. Following VML injury, the fascia was sutured and the skin was stapled. Rats were monitored twice daily for three days post-surgery, and once a week thereafter for signs of pain and complications. There were no incidence of adverse events in the study and rats healed and recovered ambulatory ability promptly after surgery.

Histological and image analyses

For each terminal time point, the whole TA muscle from both left and right hindlimbs was harvested, blotted dry and subsequently weighed. The distal half of the VML defect area and approximate area on the contralateral limb was mounted on cork using tragacanth gum for histological evaluation. The tissue was placed in 2-methylbutane cooled by liquid nitrogen to freeze, and subsequently stored at −80°C until further analysis. The TA muscle was sectioned at 10μm using a Leica cryostat and microtome and placed on glass slides.

Broad qualitative analysis of muscle tissue was obtained using Masson’s trichome stain (Thermo Scientific Richard-Allan Scientific Chromaview Kit; 87019) following manufacturer instructions. Once stained, images were obtained using a 20X objective (0.75 NA, 0.5 μm/pixel resolution) on the brightfield Huron TissueScope LE slide scanner (Huron Digital Pathology, St. Jacobs, ON, Canada).

Visualization and quantification of collagen orientation was obtained using a picrosirius red stain under polarized light. Briefly, frozen sections were fixed in 4% PFA for 10 minutes, rinsed in deionized water and subsequently placed in picrosirius red solution (0.1% Direct Red 80, 1.2% picric acid; Abcam, ab246832) for 60 minutes. Sections were then quickly rinsed in 3 changes of 0.5% acetic acid (~10 dips each), dehydrated in 2 changes each of 95% and 100% alcohol (1 minute each) and cleared in xylenes before being immediately mounted with DPX mounting medium (Electron Microscopy Sciences). The slides were then viewed and imaged using a Nikon Eclipse 200 light microscope with diascopic cross-polarized filter at 40x magnification (E Plan 40x/0.65 OFN20) using digital camera interfaced to Nikon Elements D software. Regions of interest (ROI) for analysis included the VML defect area, the remaining uninjured muscle, and two images along the border region of the defect and remaining muscle. For the contralateral control limbs, analogous regions were imaged, however a single control value is presented for comparison. During all imaging gain, offset, gamma, and exposure were constant. The area of red and green staining within ROI boxes were obtained using FIJI.³² Specifically, two thresholds were applied using the ‘a* channel’ in the CIELAB color space. A green threshold was made by setting the upper threshold to 128 and lower threshold to 0, while adjusting the lightness value to ~30. The upper threshold in the ‘b* channel’ was set to 220 in order to remove yellow pixels. An orange/red threshold was then made by setting the ‘a* channel’ lower threshold to 128 and upper threshold to 255, while adjusting the lightness value to ~45. A yellow threshold was added to the calculations of orange/red by setting the upper threshold of the ‘a* channel’ to 128, and lower threshold of the ‘b* channel’ to 220. The images were then binarized in their respective thresholds and measured for area. Data is presented in two ways, first, as the orange/red/yellow and green area independently and second, as a ratio of the orange/red/yellow to green area. A lower ratio indicates collagen that is aligning in parallel to muscle fibers and greater ratios indicate disorganized collagen alignment to the longitudinal axis of the muscle fibers as previously described.³³

The TA was also stained with anti-CCN2/CTGF. Briefly, sections were fixed, washed and blocked prior to probing with anti-CCN2/CTGF (Abcam: ab5097, 5 μg/ml) primary antibody. When blocked by the immunising peptide a single band of 38kD molecular weight on Western blot is identified with this antibody (manufacturer’s technical information). Sections were stained with an Alexa-Fluor 488 conjugated secondary (Invitrogen: A-11034, 1:200) and DAPI (1 μg/ml) was added to stain for nuclei. Imaging and analysis of muscle ROI’s followed the same strategy as described for the picrosirius red. Images were taken using an automated Nikon C2 upright confocal microscope (Nikon Instruments Inc., Melville, NY) at a magnification of 20X (PlanApo 0.75 NA, 0.31 μm/pixel resolution). For semi-quantitative analysis, images were binarized and the image intensity threshold was determined using the Otsu algorithm (FIJI) to measure the CCN2/CTGF area per ROI. Data are presented as a percentage of the total image area.

In all cases, the expected staining patterns in normal skeletal muscle were observed and the specificity of anti-labeling (CCN2/CTGF) was confirmed by the absence of staining outside expected structures and was consistent with manufacturer’s technical information. The CCN2/CTGF antibody was further validated by blocking with the immunising peptide (Abcam: ab13741) before staining, with negligbile positive CCN2/CTGF observed when imaged with identical parameters (gain, offset, exposure, pinhole size, etc.). Broadly across the literature the staining patterns of CCN2/CTGF in skeletal muscle and there is both intra- and extra-cellular staining present and expected due to the being an excreted protein from fibroblasts. For display purposes only, images were down-converted, without introducing any changes in brightness or contrast and produced in Adobe Photoshop (Adobe Systems Inc.). During all fluorescent imaging, the laser intensity and gain were kept consistent across samples. Investigators were blinded during all imaging and post-imaging analyses.

Biochemical analyses

Following harvest, the uninjured portion of TA muscle tissue distal to the VML defect area, and approximate area on the contralateral limb, was cut vertically into two pieces (medial/lateral), snap-frozen using liquid nitrogen and stored at −80°C until analysis. The lateral piece of each muscle was used to measure specific collagen isoforms, in addition to the protein content of decorin. Quantification of collagen I, III, IV, and VI (MyBioSource, COL1α2-MBS2023826, COL3α1-MBS2023591, COL4α1-MBS7226846, COL6α1-MBS102349) was conducted by ELISA. Briefly, tissues were weighed and homogenized in a 10-mM phosphate buffer at a ratio of 1:100 (mg/μl), using a glass pestle tissue grinder over ice. The resulting homogenate was centrifuged for 5 minutes at 10,000 g, and the supernatant was aliquoted for immediate storage at −80°C. Total protein content was analyzed using the Protein A280 application (1 Abs = 1 mg/ml) on a NanoDrop One spectrophotometer (Thermo Scientific) after one freeze/thaw cycle. Samples were analyzed for specific collagen isoforms using the respective ELISA kits as indicated by the manufacturer and normalized to protein content. This methodology is expected to quantify soluble proteins, no attempt was made to specifically extract proteins for analysis of insoluble components of the ECM The distal-medial piece of each TA muscle was then used to measure total collagen content. This was determined by quantifying the content of hydroxyproline in the muscle sample, as previously described.^34,35

Immunoblot analyses

Fifteen μg of total protein were separated by 4–20% SDS-PAGE, transferred onto a PVDF membrane, and immunoblotted. Anti-decorin (Proteintech: 14667-1-AP; 1:500) primary antibody was detected using a corresponding horseradish peroxidase-conjugated secondary antibody (Cell Signaling: #7074, 1:1000). Protein detection was obtained using Clarity Max ECL Western Blotting Substrate (Bio-Rad). Immunoblots were visualized with stain-free and chemiluminescent imaging using a ChemiDoc System (Bio-Rad Laboratories, Hercules, CA). First, using the image acquired from the stain-free blot the total protein in each lane was quantified. Second, following chemiluminescent imaging the single decorin was identified at 70 kDa, as expected and consistent with manufacturer’s technical information. The intensity of the decorin band was then normalized to total protein in each respective lane using Bio-Rad Laboratories Image Lab software (Hercules, CA).

Statistical analysis

Analysis was performed with JMP Pro statistical software (version 14.2 SAS Institute, Cary, NC, USA). Dependent variables for controls, day 3, 7, 14, 21, and 48 were analyzed by one-way ANOVA across time points. When significance was found, Tukey’s HSD post-hoc analysis was used to ascertain specific time point significance. Data are presented as mean±standard deviation (SD). Statistical significance level was set at p≤0.05. During all evaluation the research team was blinded to the experimental groups.

RESULTS

Animals

Following VML surgery, all rats recovered promptly and without complications. Overall body weights steadily increased with time after injury, as expected, with rats in the 48 day group 15% heavier at the terminal time point than at surgery. In parallel, the weight of the control (right, uninjured) TA muscle increased over time post-injury, as expected. Conversely, the weight of the VML-injured TA muscle steadily decreased at all time points after day 3. Specifically, at 48 days post-VML the injured TA muscle was ~18% smaller than the control TA muscle (Table 1).

Table 1.

Whole body and skeletal muscle characteristics

		Days post-VML
	Contralateral control (n=9)	3 (n=8)	7 (n=7)	14 (n=8)	21 (n=8)	48 (n=8)	p-value
Surgical body weight (g)	- -	383.4 ± 26.1	377.0 ± 21.5	374.4 ± 12.3	394.8 ± 22.5	385.6 ± 12.1	0.291
Terminal body weight (g)	- -	368.8 ± 22.2 §‡	368.7 ± 22.5 §‡	376.3 ± 19.0 §‡	406.9 ± 26.4 §	444.6 ± 12.5	<0.0001
Injured TA weight (mg)	- -	689.1 ± 83.5 †	616.9 ± 85.0	582.5 ± 37.4	601.0 ± 65.9	599.1 ± 80.4	0.049
Contralateral TA weight (mg)	690.9 ± 67.6	639.0 ± 33.1 §‡	645.2 ± 75.5 §	704.9 ± 32.4	719.6 ± 27.1	730.5 ± 23.4	0.001
TA/body weight (mg/g)	1.73 ± 0.11 §	1.88 ± 0.31 §‡†	1.67 ± 0.18 §	1.55 ± 0.03	1.49 ± 0.22	1.35 ± 0.17	<0.0001
TA protein content (μg/mg)	93.44 ± 15.56	80.74 ± 8.18 §	68.50 ± 13.93 *	73.12 ± 11.90 *§	72.89 ± 5.60 *§	54.26 ± 9.99 *	<0.0001

Contralateral controls from a sub-set of animals are collapsed over time points for data from TA muscle.

Data analyzed by one-way ANOVA; differences are indicated as

^*different than control;

^†different than day 14;

^‡different than day 21;

^§different than day 48.

Histologic Presentation of Fibrosis following VML Injury

Gross histologic presentation of the TA muscle following VML injury was qualitatively evaluated using Masson’s Trichrome staining (Figure 1). Qualitatively there is significant fibrotic deposition that begins to develop as early as 7 days post-VML, as well as substantial cellular infiltration into the VML defect area (see also Figure 3). By 48 days post-VML the fibrosis appears to spread more prominently into the remaining muscle.

Figure 1.

Masson’s trichrome staining was conducted across time points to evaluate the progressive fibrotic deposition following VML injury. Representative images from each time point are displayed, connective tissue is blue and muscle fibers are red. Scale 1mm.

Figure 3.

A) Representative whole TA muscle stained for CCN2/CTGF (green) and DAPI (blue); example is from 21 days post-VML. Scale 1mm. B) Representative images taken from the border region of defect area across time points and analogous control region. Scale 50μm. Percent area positive for CCN2/CTGF across post-VML time points are plotted from C) remaining muscle (p<0.0001), D) border region (p<0.0001), and E) defect area (p<0.0001). Dashed line indicates averaged control area. Images were randomly selected from respective regions for analysis. One image was analyzed from remaining muscle and defect area, while two images were analyzed from each border region and plotted as separate points. Data analyzed by one-way ANOVA with Tukey’s HSD post-hoc; *different than control; ¥ different than day 3.

Staining for picrosirius red was used to visualize and evaluate collagen packing (Figure 2A–B). Using cross-polarized filters, staining exploits the birefringent properties of collagen, displaying a spectrum of red to green hues and eliminating non-collagenous structures. Historically, hue differences were thought to distinguish collagen I from collagen III based on thickness, with collagen I (thicker) staining red/orange/yellow and collagen III (thinner) staining green.³⁶ However, it is unlikely hue differences are specific to collagen types, given that immature collagen I may be relatively thin. Rather, it is likely indicative of collagen packing, whereby loosely packed collagen appears green and densely packed collagen appears red/orange/yellow. Therefore, the area fractions of collagen stained red/orange/yellow and the area stained green were quantified in the TA muscles. Muscle sections were evaluated at three systematically selected ROIs: the remaining muscle, the defect area, and the border between the defect and remaining muscle. The total fraction of the polarized area increased over time in the defect and border regions post-VML. Specifically, the green area fraction increased significantly at 7, 14, and 21 days in both regions before declining at 48 days (p≤0.045), while the red/orange/yellow fraction continued to significantly increase through 48 days (p≤0.0001; Figure 2C–E). Thus, the ratio of red/orange/yellow to green in defect and border regions was decreased at early time points post-VML before significantly increasing from 21 to 48 days (p≤0.007; Figure 2G–H). Remarkably, the total fraction area in the remaining muscle was also significantly increased after 7 days post-VML. While the red/orange/yellow area was significantly increased at all time points after 7 days, the green area was significantly increased at 21 days post-VML. The ratio of red/orange/yellow to green in the remaining muscle was slightly lower than controls following VML, and significantly so at 3 and 21 days post-VML (p≤0.018; Figure 2F), indicating an increased relative fraction of green, or loosely packed collagen.

Figure 2.

A) Representative whole TA muscle stained with picrosirius red; example is from 21 days post-VML. Areas staining green represent loosely packed collagen, while areas staining red/orange/yellow represent more densely packed collagen. Scale 1mm. B) Representative images taken from the border region of defect area across time points and analogous control region. Scale 50μm. Total area fraction of collagen separated into red/orange/yellow and green fractions across post-VML time points are shown for the C) remaining muscle (p<0.0001), D) border region (p<0.0001), and E) defect area (p<0.0001). Ratio of orange/red/yellow fraction to green fraction are plotted from F) remaining muscle (p<0.009), G) border region (p<0.0001), and H) defect area (p<0.0001). Dashed line indicates averaged control ratio. Images were randomly selected from respective regions for analysis. One image was analyzed from remaining muscle and defect area, while two images were analyzed from each border region and plotted as separate points. Data analyzed by one-way ANOVA with Tukey’s HSD post-hoc; *different than control; ¥ different than day 3; †different than day 14; ‡different than day 21; §different than day 48.

Staining for CCN2/CTGF was used to evaluate the fibrotic deposition in the muscle in the same ROIs (Figure 3A–B). The percentage of area positive for CCN2/CTGF was significantly increased (p≤0.004) as early as 3 days post-VML throughout the muscle and defect, and further stabilized or increased through 48 days post-VML. In the defect region, the CCN2/CTGF positive area was 1.4-fold greater from 3 to 7 days post-VML. Similar increases were seen in the border region. Notably, throughout the remaining muscle CCN2/CTGF positive area did not increase further at any time point after 3 days post-VML but remained significantly increased.

Variations in Collagen, Isoforms, and Signaling following VML Injury

Total collagen, as determined by hydroxyproline content, was not significantly increased at any time post-VML in the remaining muscle (p=0.092; Figure 4A), likely due to greater than expected sample variability. A statistically non-significant ~56%1-fold increase in collagen content in VML compared to control was observed at 48 compared to 3 days post-injury.

Figure 4.

A) Total collagen content from injured TA normalized to muscle weight (p=0.092). Protein content of B) collagen I (p<0.0001), C) collagen III (p<0.0001), D) collagen IV (p<0.0001), and E) collagen VI (p=0.003), normalized to total protein content of the sample are plotted across time points post-VML. Data analyzed by one-way ANOVA with Tukey’s HSD post-hoc; *different than control; ¥ different than day 3; †different than day 14; ‡different than day 21; §different than day 48.

Protein levels of collagen III (normalized to protein content) were significantly increased at all time points following VML injury compared to control (p=0.026; Figure 4C). At 3 and 14 days post-VML, collagen III was increased by 2- to 3-fold, respectively, before a gradual decrease in protein levels by day 48. Conversely, collagen I was not significantly increased until 48 days post-VML, where it rose ~2.5-fold compared to control (p<0.0001; Figure 4B). Accordingly, the collagen I:III ratio was reduced at 3, 7, 14, and 21 days post-VML, yet increased at 48 days. Protein levels of collagen IV and VI were both significantly increased at 7 and 48 days post-VML (p≤0.031; Figure 4D–E).

Content of decorin within the remaining muscle was significantly increased (p=0.005) at 3 days post-VML compared to control (Figure 5). However, by 48 days post-VML decorin content was 32.7% lower than control and significantly reduced compared to day 3.

Figure 5.

A) Relative expression of decorin across time points post-VML (p=0.005). Dashed line indicates control value of 1. B) Representative immunoblot images for decorin (70 kDa) at each time point and corresponding stain-free blot used to quantify the total lane protein, in which each decorin band is normalized to the protein quantified per lane. Data analyzed by one-way ANOVA with Tukey’s HSD post-hoc; *different than control; ¥ different than day 3.

DISCUSSION

Pathologic fibrosis is detrimental to skeletal muscle tissue by increasing stiffness and reducing maximal force production. Nevertheless, the implications of fibrotic progression after VML injury on the overall goal of restoring muscle function are often overlooked. Herein the early wound healing phase after VML injury was highlighted by increased proportions of collagen III in the remaining muscle and loosely organized collagen in the VML defect area, presumably following a normal muscle repair process of transient ECM remodeling. However, these observations were abruptly transformed by 48 days post-VML into a muscle environment dominated by collagen I and a heavily aligned and densely packed collagenous defect area. Similarly, decorin, a known inhibitor of fibrosis, was initially increased at 3 days before decreasing to levels slightly lower than control muscle by 48 days post-VML, possibly allowing the overproduction of collagen I observed at that time point. The ability to prevent or mitigate fibrosis through pharmacologic or regenerative medicine approaches may benefit through this understanding of the variations in the ECM and muscle microenvironment following VML injury.

Picrosirius red staining combined with polarized light microscopy has been widely used for collagen evaluation previously, because the birefringent properties of collagen become enhanced and can provide more information than brightfield images. However, a review of the literature reveals multiple variations in the interpretation of hue differences seen using this method. Many have reported hue as corresponding to collagen thickness or “packing” in a given area,^37,38 with green hues indicating thinner or more loosely packed collagen and orange hues indicating thicker or densely packed collagen. Yet, others report intensity of staining indicates collagen thickness, while hue indicates direction of collagen fiber alignment relative to the incident light.^33,39 In this case, increasing intensity represents more densely packed fibers, while green hues indicate collagen aligning parallel with the path of light (and muscle fiber axis) and orange hues indicate collagen aligning at a larger angle to the muscle fiber axis. Given that areas staining green were generally of lower intensity when observed qualitatively, it may be possible both interpretations are valid; hue corresponds to both collagen packing and alignment. Thus, cross-sections of the mid-belly of the TA muscle, directly at the mid-point of the VML injury, were histologically stained to evaluate collagen organization and alignment. Changes in hue were quantified within the VML defect, in the remaining muscle, and in the border region between defect and remaining muscle. Findings herein reveal a heterogenous mixture of collagen packing and alignment through 21 days post-injury in all three regions evaluated, suggesting a more disorganized ECM during the remodeling phase after injury. When compared to healthy muscle, there appears to be an increased proportion of collagen fibers aligning in parallel with the muscle fiber axis. By aligning in this fashion, collagen would bear more of the stress applied to the muscle and take on a greater functional role in force transmission, in addition to protecting muscle fibers from stretch induced damage. However, clear regional differences were observed after 48 days post-VML injury. The defect and border regions shifted to a more uniform alignment of densely packed collagen at a larger angle relative to the muscle fiber axis, likely changing tensile strength of the muscle. Moreover, densely packed collagen possibly influences overall stiffness of the muscle tissue,³⁸ which has been shown to increase over time following VML.⁴⁰ Curiously, that trend was not observed in the remaining muscle, which maintained a heterogeneous collagen organization through 48 days post-VML.

All biochemical measurements of protein contents were taken from the remaining muscle (i.e., not initially injured by the VML) directly distal to the defect area from 3 to 48 days following VML injury. Therefore, data herein indicate significant fibrotic changes are not confined to the site of injury and immediate surrounding area, but which are pervasive throughout the muscle. Recent work by Southern et al. further supports this notion as substantial disruptions in the mitochondrial network were observed well into the uninjured muscle.⁴¹ Moreover, as expected the proportion of collagen III in the remaining muscle increased during early time points post-VML. However, rather unexpected was the seemingly synchronized increase in collagen I and reduction in collagen III by 48 days post-VML. The switch in collagen proportions could highlight an important phase in wound healing for intervention. The late increase in collagen I likely indicates a more heavily cross-linked ECM in the remaining muscle, and taken together with histologic findings, a more mature densely packed fibrotic scar tissue in the defect area. Addressing pathologic fibrosis through pharmacologic administration and/or physical rehabilitation would preferably occur before fibrosis becomes denser, as heavily packed and cross-linked collagen is harder to degrade.⁴²

Histologically CCN2/CTGF was quantified to assess its role in fibrotic signaling after VML injury. Importantly, CCN2/CTGF is thought to act synergistically, yet downstream of TGF-β1 in the development of fibrosis. TGF-β1 is a popular anti-fibrotic target due to its well-known role as a pro-fibrotic factor; however, it also plays an important role in the myogenic response and thus preserving TGF-β1 signaling may prove beneficial.^43,5 To that extent, this study found robust increases in CCN2/CTGF histologically at all time points through 48 days post-VML, suggesting it could be an ideal target for inhibition of fibrosis independent of TGF-β1. Indeed, a monoclonal antibody that inhibits CCN2/CTGF, known as Pamrevlumab, was evaluated in mdx mice as an anti-fibrotic treatment.⁴⁴ It was observed to reduce fibrosis, increase muscle strength, and improve satellite cell engraftment. Anti-fibrotic drugs used in VML animal models have similarly been successful in the reduction of fibrosis, but unexpectedly led to a reduction of force output.^27,40 The loss in force from fibrotic reduction underlines the role of fibrosis in force transmission, or functional fibrosis as previously described.⁴⁵ Therefore, treatments successful in reducing pathologic fibrosis would likely need to be coupled with a subsequent myogenic or passive range of motion³⁵ treatments to maintain and eventually increase force output and muscle function.

Recently, the ability of autologous minced muscle grafts to aid in muscle regeneration post-VML injury has been broadly demonstrated, yet muscle strength deficits were not able to fully recover.^46,47,48,49 Transcriptome profiling revealed similar overexpression of inflammatory and fibrotic genes in muscle graft treated and non-treated VML injured groups, suggesting myogenic treatments alone do not mitigate these processes but remain hindered by them. Furthermore, Boldrin et al. demonstrated that satellite cells derived from mdx mice retain their regenerative capabilities when placed in a permissive microenvironment.⁵⁰ This would suggest that, at least in part, the accumulation of ECM components in fibrotic muscle inhibits regenerative capabilities of endogenous satellite cells as well as injected stem cells or muscle grafts. Therefore, restoring skeletal muscle the ECM to a less fibrotic state would allow greater likelihood of successful satellite cell engraftment, in addition to improved endogenous regeneration. The observation of a significant reduction in decorin content from 3 to 48 days post-VML reveals a potential treatment avenue in that regard. Decorin is known to regulate collagen fibrillogenesis by binding to collagen I,¹³ and also has binding sites for CCN2/CTGF, TGF-β1 and myostatin, effectively inhibiting their signaling potential.⁵¹ Thus, decorin has the ability to simultaneously inhibit fibrotic signaling while stimulating myogenic signaling. However, degradation of decorin can result in the release of its bound pro-fibrotic factors and further increase the pathologic signaling.⁵² Various animal models using decorin as a therapeutic have shown promising results, which range from improved muscle healing in an injury model to tumor suppression in cancer models.^53,54,55

Disability following VML injury is associated with pathologic fibrosis that impairs function directly through contracture and diminished ranged of motion, as well as indirectly by inhibiting muscle regeneration and restricting axon growth.^{46,40,56,31,26} A canonical fibrotic response has been identified across various etiologies in which synchronous replacement of loosely packed collagen III fibers are replaced with densely packed collagen I fibers forming fibrous scar tissue.^57,58 The current study identifies that in a rodent model of VML injury, this conversion occurs approximately three weeks after injury. Thus, interventions to temporize this response to allow improved regeneration and ECM organization may be most needed in the acute to subacute period post-injury, when other likely comorbidities are of greater focus, such as fracture fixation and infection control.⁵⁹ Specifically this work supports further exploration of CCN2/CTGF inhibitors or upregulation of decorin to mitigate pathologic fibrous tissue formation after VML injury.