In-vivo testing of an injectable matrix gel for the treatment of shoulder cuff muscle fatty degeneration

Abstract

Introduction:

Extracellular matrix (ECM) gels have shown efficacy for the treatment of damaged tissues, most notably cardiac muscle. We hypothesized that the ECM gel prepared from skeletal muscle could be used as a treatment strategy for fatty shoulder cuff muscle degeneration.

Methods:

We conducted experiments to 1) evaluate host biocompatibility to ECM gel injection using a rat model; and 2) examine the effect of ECM gel injection on muscle recovery following delayed repair of a released supraspinatus (SSP) tendon using a rabbit model.

Results:

The host biocompatibility to ECM gel was characterized by a transient rise (first two weeks only) in several genes associated with macrophage infiltration, matrix deposition, and inflammatory cytokine production. By eight weeks all genes had returned to baseline levels and no evidence of fibrosis or chronic inflammation was observed from histology. When gel injection was combined with SSP tendon repair, we observed a significant reduction (7%) in SSP muscle atrophy (24+3% reduction from uninjured) when compared to treatment with tendon repair only (31+7% reduction). While fatty degeneration was elevated in both treatment groups, fat content trended lower (2%) in response to combined tendon repair and intramuscular ECM injection (4.1+2.1%) when compared to tendon repair only (6.1+2.9%). Transcriptome analysis revealed adipogenesis and osteoarthritis pathway activation in the repair only group. These key pathways were abrogated in response to treatment using combined repair plus gel.

Discussion:

The findings suggest that ECM injection had a modest but positive effect on muscle mass, fatty degeneration, and key cellular signaling pathways.

In selected cases, the surgical care for damaged rotator cuffs is reattachment of the torn tendon(s) to their original boney insertion sites on the humerus ^{5, 34, 49}. Despite significant progress with arthroscopic techniques and improved fixation devices, rotator cuff surgery failure rates remain high, reported to be around 30–35% ^{23, 28, 31}. The muscles of the rotator cuff undergo progressive fatty degeneration, also termed fatty infiltration, following tendon tear ^{15, 17}. This pathological degeneration of shoulder muscle in patients with rotator cuff tears is shown to be a contributing factor to the high failure rate ¹⁹. While surgical techniques have been developed to address the torn tendinous portion of the damaged rotator cuff, there are currently no effective methods to address fatty degeneration. Fatty degeneration remains a well-recognized but currently unaddressed pathological condition that limits the success of rotator cuff surgery ^{10, 14, 41, 57}. Our goal is the development of an effective treatment for the repair of fatty degenerated rotator cuff muscle that could become an adjunct to the current tendon repair techniques.

Under the right conditions, skeletal muscle has a robust capacity for self-repair. Following mild muscle injury (for example: strains, contusions, and lacerations) cells may be injured, but the damaged underlying ECM persists at the site of injury and regeneration is robust and complete ^{24, 39}. Growth factors released from injured muscle matrix promotes satellite cells residing between myofibers to migrate to sites of injury, reenter cell cycle, proliferate, and differentiate into myoblasts, which in turn, undergo fusion to form nascent myofibers ^{2, 52}. However, when muscle tissue is lost, the structural cues produced by the ECM are missing and regeneration is instead marked by fibrosis and the formation of non-contractile scar tissue ^{3, 29, 53}. The differential response to mild and severe muscle injury suggests an important role for extracellular matrix as a trigger for muscle regeneration. To deliver extracellular matrix cues into fatty infiltrated muscle, we are proposing the utilization of an injectable, water-based, matrix gel prepared from skeletal muscle tissue. Our enthusiasm for a matrix gel approach as a muscle repair strategy was motivated not only by the recognized role of matrix during muscle healing, but also by the cardiac work of Dr. Christman’s group at the University of California at San Diego. They have developed an injectable matrix gel using cardiac tissue and demonstrated that injection of the material into infarcted cardiac muscle tissue stimulated regeneration ^49–51. A possible mechanism of repair is ECM gel promotion of vessel formation/perfusion, endothelial cell/ muscle progenitor cell infiltration as well as muscle cell proliferation ^{9, 43}.We believe the same regenerative results that have been reported in infarcted cardiac muscle tissue could be translated to fatty degenerated skeletal muscle.

We conducted this study to deeply explore the safety and efficacy of an injectable ECM gel while also deciphering whether any key wound healing mechanisms were impacted by matrix gel implantation. Specifically, we designed the study to test the hypothesis that intra-muscular delivery of matrix gel triggers the activation of pro-myogenic wound healing pathways that ameliorate fatty muscle infiltration. To test this hypothesis, we conducted experiments to 1) examine the longitudinal host response (biocompatibility) to ECM gel injection into normal muscle using a rodent model and 2) test matrix gel regenerative performance in a well-accepted ^{20, 21, 45, 46} animal model of delayed rotator cuff repair (rabbit).

2. Methods

2.1. ECM gel preparation

Human quadriceps muscle (Science Care, US) was thawed, trimmed to remove fat and connective tissue, and decellularized in 1% (wt/vol) solution of sodium dodecyl sulfate (SDS) in phosphate buffered saline (PBS) with agitation and multiple solution exchanges for up to 2 weeks ^{9, 27} (Fig. 1a and b). We then rinsed the tissue with deionized water and incubated in DNase/RNase solution overnight at 4°C with agitation. The remaining ECM was rinsed, lyophilized, flash-frozen using liquid nitrogen, and ground into fine powder. We digested the ECM powder using a pepsin solution (1 mg/ml pepsin in 0.1 M HCl) at a ratio of 10mg ECM/1ml pepsin at a pH of 2.4 (adjusted every 12h if necessary) for 48h at room temperature ^{9, 56} (Fig. 1c and d). The solubilized ECM solution was neutralized (pH=7.4) and diluted to a concentration of 10mg/ml in phosphate buffered saline (PBS). Finally, we loaded the neutralized pre-gel solution into syringes and stored it at −20°C (Fig. 1e and f).

Figure 1:

(a) Human quadriceps muscle was (b) decellularized, (c) lyophilized, (d) ground into a fine powder and (e) pepsin digested to create an injectable pre-gelled matrix solution. (f) The pre-gelled muscle matrix solution forms a gel when warmed to 37°C.

2.2. Host Response

Adult (12–14-weeks old) Sprague Dawley rats (Harlan, Indianapolis, IN, USA) (n=12), weighing approximately 325–350g, were used as the animal model to explore the host response to injected ECM. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Arkansas (#19002). We used isoflurane for anesthesia with induction dosage of 4% and maintenance dosage of 2% in oxygen. Syringes containing ECM gel were thawed and injected (200 μL; 27 gauge needle) into the TA at five sites along its distal-to-proximal length intramuscularly through the skin. We injected the contralateral TA’s with sterile PBS to serve as comparative controls. Postoperative analgesia consisted of 0.1mg/kg buprenorphine administered subcutaneously via injection twice daily for two days. A single surgeon performed all procedures. Animals also had access to anti-inflammatory medication (Rimadyl; 2mg/day) via a dietary tablet (Rodent MD’s, Bio Serv, Flemington, NJ, USA). We added tablets daily to each cage for one week following surgery in accordance with IACUC approved protocols and housed animals individually in standard-sized cages with unrestricted movement. The animals were allowed to bear weight on the operative extremity as tolerated. All animals were housed for a 3, 14, or 56 day recovery period (n=4 animals / time point).

2.3. Rotator Cuff Repair

Adult (24-weeks old) male New Zealand white rabbits (pre-surgery mass ~3.5 kg) purchased commercially (Charles River) (n=14) were used to examine the effect of ECM gel injection on recovery from torn rotator cuff repair. All animal procedures were approved by the Institutional Animal Care and Use Committee of the University of Arkansas for Medical Sciences (#3655). The rotator cuff tear and delayed tendon repair surgeries were performed following published methods ^{37, 42}. A trained orthopedic surgeon performed the surgeries. We used isoflurane for anesthesia with induction dosage of 5% and maintenance dosage of 3% in oxygen. The left shoulder was shaved and disinfected. A longitudinal incision was made over the shoulder, and dissection performed down to the deltoid. The deltoid was retracted to reveal the supraspinatus (SSP) tendon. The supraspinatus tendon was transected at its insertion onto the greater tuberosity. The cut end of the tendon was tagged with a non-resorbable No. 4–0 suture (Prolene, Ethicon, Somerville, NJ, USA) for later identification. All attachments of the tendon to the surrounding tissues, including the infraspinatus, were released, allowing the tendon to retract freely away from its insertion site. The incisions were closed with interrupted, subcutaneous No. 3–0 Vicryl sutures (Ethicon), and the skin was closed with a running 4–0 monocryl suture (Ethicon). The contralateral shoulder remained untreated to serve as a comparative control. All animals were housed for a 12-week recovery period.

Following recovery, all animals underwent SSP tendon reattachment repair surgery. The free tendon edge was identified by use of the tag suture and double stitched to the humerus using a fiber wire suture. We randomly assigned half of the animals to receive 1 mL of muscle derived ECM gel injection into the SSP muscle at five locations along its length. Following repair, the surgeon closed surgical site with a suture as previously described. The animals were maintained in the animal facility for a recovery period of 12 weeks (n= 7 / treatment group). At the end of the recovery period, a veterinarian euthanized all animals via intracardiac injection of a commercial euthanasia solution. SSP muscles were harvested, weighed, imaged, and prepared for histological sectioning and gene expression testing.

2.4. Contractile Force Measurements (Rat Only)

At each post-injection time point (3, 14 and 56 days), we measured peak tetanic TA contractile torque in-situ using published methods familiar to our group ²⁹. The ankle was flexed to 90° and the foot was secured to the lever arm of a commercial muscle physiology system (Aurora Scientific, Ontario, Canada). We measured the TA peak isometric tetanic torque by stimulating the peroneal nerve with the aid of a physiological stimulator (Grass; S88) and determined Optimal voltage (2 – 5 V) using a series of te tanic contractions (150Hz, 0.1 ms pulse width, 400 ms train). We recorded raw peak tetanic contractile force (N) from both the treated and contralateral control limb of each animal. We calculated peak tetanic force for each animal using the average of four contractions and reported the animal weight normalized force data (N/kg body weight) in the results. At the conclusion of electrophysiological testing, we harvested the TA muscles and euthanized the rats by carbon dioxide inhalation.

2.5. Histological Analysis

Muscle tissue (rat and rabbit) was flash frozen in isopentane (2-methylbutane) chilled in liquid nitrogen. Tissue was sectioned (8μm) with the aid of a cryostat (Leica BioSystems) and mounted onto microscopic slides. Slides were immuno-stained using primary antibodies (1:300, Novus Biologicals) directed against collagen III (IgG), or collagen I (IgG) with or without myosin heavy chain (MHC) co-stain, followed by incubation in the appropriate fluorescently labeled secondary antibodies (AlexaFluor, 1:300, Life Technologies). Additional tissue sections were stained using hematoxylin and eosin (H&E) or Oil red-O following manufacturer’s guidelines (Sigma-Aldrich). All sections were digitally imaged (100X, Nikon CiL).

Tissue immunoreactivity to collagen I as a percentage of total tissue area (% collagen I) was calculated with the aid of image analysis software (ImageJ) using techniques familiar to our group ²⁹. Similar image analysis methods were used to compute fiber cross-sectional area (μm²) using collagen III positive area. Oil red-O images were evaluated using an in-house developed MATLAB code to isolate oil red-O positive (fat) and negative (muscle) tissue regions within each section. The ratio of oil red-O positive tissue area to total tissue area was used to calculate the intramuscular fat area fraction (% area).

2.6. Gene Expression

RNA was extracted from tissue samples using the RNeasy Kit (Invitrogen). Commercially available TaqMan primers (Invitrogen) for Pax7, MyoD, MyoG, Col I, Col III, TGF-β1, IL-6, IL-10, CD68 and 18s ribosomal housekeeping were used to quantify the expression of desired host-response genes from rat tissue. Expression was normalized to 18S and then referenced to the contralateral saline injected limb. Commercially available SYBR Green primers (Bio-Rad) for PPAR-γ, MURF1 (Trim63), Pax7, MyoG, and RPL4 housekeeping were used to quantify the expression rabbit muscle genes. Experimental sample group (n=6–7/group) expressions were normalized to RPL4 and then referenced to the contralateral uninjured limb. Gene expression levels are reported as fold change using the 2-(^ΔΔCt) method.

The full transcriptomes of representative repair only, repair plus ECM gel, and uninjured controls rabbit SSP muscle samples (n=5/ group) were analyzed using RNA-Seq. cDNA libraries were sequenced on the NextSeq500 platform (Illumina) to a mean depth of 20 million reads per library. RNA sequencing reads were mapped to the Oryctolagus cuniculus genome (OryCun2.0) from NCBI using the 2-pass STAR protocol¹¹. Reads were quantified using FeatureCounts³⁴, followed by analysis of differential expression and normalization in edgeR⁴⁴. Differential expression was selected using a maximum false discovery rate (FDR) of 0.25 and a minimum log fold change of 1.5. Pathway level analysis was also performed using Ingenuity Pathway Analysis (IPA) (Qiagen)³².

2.7. Statistical Analysis

All data are presented as the mean and standard deviation unless otherwise noted. The effect of treatment on TA mass, TA contractile force, TA collagen I area (%), TA myofiber cross-sectional area, TA and SSP gene expression, as well as SSP mass reduction were evaluated using a two-sided Student’s t-test. The effect of treatment SSP fat area fraction, SSP collagen I area fraction, and SSP myofiber area were evaluated using ANOVA with Tukey’s post hoc test. A standard p<0.05 level of significance was used as the threshold for all statistical tests.

3. Results

3.1. Host Response

All animals tolerated the ECM gel injection treatment well. Peak tetanic contractile force values for ECM gel injected TA muscle were statistically indistinguishable from PBS controls at three days (0.97+0.41 N/kg vs. 0.72+0.18 N/kg) and at 14 days (1.72+0.87 N/kg vs.1.34+0.56 N/kg) post-injection (Fig. 2a). However, at 56 days, the study endpoint, a statistically significant (p=0.048) increase in contractile force was detected for the ECM gel injection group. Mean gel injected TA contractile force values were on average 53+37% higher than PBS injected control values (3.25+0.41 N/kg vs. 2.25+0.70 N/kg). TA mass followed a similar trend to force. At 3 days (1.93+0.23 g/kg vs. 2.14+0.16 g/kg) and 14 days (2.03 + 0.08 g/kg vs. 2.09+0.09 g/kg) post-injection TA mass did not differ between the ECM gel and PBS injected groups respectively (Fig. 2b and c). Similar to force results, at 56 days there was a statistically significant (p=0.003) increase (18+8%) in TA mass for the ECM gel injection group (2.43+0.14 g/kg) when compared to PBS controls (2.05+0.07 g/kg).

Figure 2:

(a) TA muscle peak contractile force (N/kg body weight) and (b) wet weight (g/kg body weight) data normalized to animal weight. (c) Gross morphology of TA muscles 56 days post PBS injection (left) and 3, 14, and 56 days post ECM gel injection (right). Data is presented as group means + SD; n=4/group; * denotes statistically significant (p<0.05) differences from controls.

Rat TA sections immunostained for collagen I exhibited no discernible differences between ECM gel injection and PBS controls across all time points (Fig. 3a). There were no indications of abnormal inflammatory or fibrotic responses within any of the TA tissue sections examined. Fiber cross-sectional area as indicated by collagen III-positive regions were normal in appearance with spheroid muscle fiber cross sections observed for both groups (Fig. 3b). H&E staining also appeared normal with tightly apposed bundles of myofibers and no evidence of hyperplastic regions within any of the TA sections (Fig. 3c). Collagen I-positive area fraction (Fig. 3d) and myofiber cross-sectional area (Fig. 3e) were not significantly different between either groups at any time point.

Figure 3:

TA muscle cross-sections were stained for (a) collagen I (green), b) collagen III (red), and with (c) Hematoxylin and Eosin. Representative 56 days post PBS injection (left) and 3, 14, and 56 days post ECM gel injection TA cross-sections are presented. Scale bar = 100μm unless noted. Cross-sections were quantified for (d) collagen I area fraction and (e) muscle fiber cross-sectional area. Group means + SD are presented; n=4/group.

The expression levels of genes associated with ECM and ECM regulation: Col I (0.031), Col III (p=0.017), and TGF-ß1 (p=0.012) were all significantly upregulated (~10 fold) at 3 days post-injection (Fig. 4a). By 14 days post-injection, Col III (p=0.035) and TGF-ß1 (p=0.017) remained upregulated by approximately 5-fold in the gel injected groups compared to controls until returning to basal levels of expression by 56 days. We observed a similar trend in a panel of inflammatory genes, with IL-6 (p=0.033) and IL-10 (p=0.025) being expressed approximately 15-fold higher in response to gel injection at 3 days post-injection when compared to PBS controls, while CD68 (p=0.049), a macrophage specific marker, was expressed approximately 50-fold higher in the gel group compared to PBS controls at the same time point (Fig. 4b). By 14 days all gene expression related to these targets returned to PBS control tissue levels. Myogenic gene expression (Pax7, MyoD, MyoG) is statistically indistinguishable between the two treatment groups at all post-injection time points (Fig. 4c)

Figure 4:

Comparison of relative gene expression for PBS and ECM gel injected groups determined using qRT-PCR. The expression of (a) ECM (Col I and Col III), ECM regulatory (TGFβ1), (b) inflammation (IL-6), anti-inflammation (IL-10), macrophage marker (CD68) and (c) myogenic (Pax7, MyoD, MyoG) genes were measured using muscle tissue harvested from the TA of each animal tested. Expression is presented as fold change normalized to PBS injected muscle expression. Group means + SD are presented, n=3–4/group. * denotes statistically significant (p<0.05) differences from controls.

3.2. Rotator Cuff Repair

All animals tolerated the RC surgery, bore weight normally on the treated limb by the end of the first postoperative week, and gained weight throughout the study period. Animal growth rate was examined during the twelve-weeks prior to and following SSP treatment. While growth generally slowed with increasing animal age, we did not detect any differences in growth rate between repair only and repair plus ECM gel groups during either the twelve weeks before repair or during the twelve weeks following repair. At the study endpoint (12 weeks after repair), a layer extramuscular fat remained on muscles harvested from both tendon repair only and combined tendon repair plus ECM gel treatment animals (Fig. 5a). No accumulation of extramuscular fat was observed on any of the uninjured contralateral SSP muscles. While both treatment strategies resulted in significant muscle atrophy when compared to uninjured contralateral controls, combined tendon repair with ECM gel injection treatment had a positive effect on SSP mass to decrease atrophy. Treatment with tendon repair only resulted in a 31+7% reduction in SSP mass when compared to uninjured controls. Alternatively, treatment using combined tendon repair plus ECM gel injection resulted in a 24+3% reduction in SSP mass, a decrease of 7% when compared to the repair only (Fig. 5b). The SSP mass reduction between the repair only and the combined repair plus ECM injection groups was statistically significant.

Figure 5:

(a) Gross morphology of representative SSP muscles from uninjured, repair only, and combined repair plus ECM gel groups are presented to show the appearance of surface fat accumulation on injured muscles. (b) SSP muscle wet weight data (% of uninjured) for repair only and combined repair plus ECM gel groups. Data is presented as group means + SD; n=7/group; * denotes statistically significant (p<0.05) differences between groups.

Oil red-O staining of uninjured, repair only, and repair with ECM gel injection SSP sections revealed elevated levels of intramuscular lipids within muscle tissue collected from both treatment groups (Fig. 6a). The repair only and combined repair with ECM gel sections contained a gradient of intramuscular fat with the densest regions localized nearest the tendon injury site. Qualitative examination of sections immunostained for collagen I or collagen III revealed no notable differences between either of the treatment groups or differences from uninjured controls (Fig. 6b and c). Collagen I and III immunoreactivity was uniformly distributed around muscle fibers (collagen III primarily) and multifiber bundles (collagen I) for all treated and uninjured control muscles examined. We did not detect any abnormal changes to myofiber cross-sectional morphologies between the treatment groups. H&E sections from both treatment groups consistently exhibited a lack of tight myofiber apposition when compared to sections prepared from contralateral normal tissue and further confirmed fiber atrophy due the RCT (Fig. 6d).

Figure 6:

SSP muscle cross-sections were stained for (a) Oil red-O, (b) collagen I (green) counterstained with MHC (red), (c) collagen III (green), and (d) hematoxylin and eosin. Representative 12 weeks post SSP repair images for uninjured, repair only, and combined repair plus gel groups are presented. Scale bar = 100μm unless noted. Repair only and combined repair plus ECM gel cross-sections were quantified to determine (e) area fraction collagen I, (f) area fraction oil red-O, and (g) muscle fiber cross-sectional area (μm²). Group means + SD are presented; n=6–7/group. * denotes statistically significant (p<0.05) differences between groups.

Quantitative image analysis of oil red-O stained sections revealed that both repair and repair with ECM gel treatments resulted in significant increases in intramuscular fat area fraction when compared to uninjured contralateral controls. Fat fraction of repair only group was 6.1+2.9%, while repair plus ECM gel fat fraction was 4.1+2.1%, a reduction of 2% from repair only values. The difference in fat area fraction between the repair only and combined repair plus gel treatment group did not reach statistical significance (p=0.18) (Fig. 6e). The uninjured contralateral SSP fat content was minimal and similar within the repair only (0.6 +0.2%) and the combined repair plus ECM gel (0.7+0.6%) groups. No statistical difference was detected between these groups. We did not detect any abnormal fibrosis in any of the tissue examined. Specifically, the collagen I area fraction was not effected by either repair only or combined repair plus gel treatment, when compared to normal muscle sections (Fig. 6f). We did detect a significant difference in fiber cross-section area between treatment groups. Repair only treatment resulted in a 30+9.8% decrease in muscle fiber size when compared to uninjured contralateral controls, while combined tendon repair plus ECM gel injection resulted in a more modest decrease of 10+4.3% (p= 0.005) (Fig. 6g).

The mRNA abundance (fold change) of genes associated with muscle degeneration (PPAR-γ, MURF1) were sensitive to repair only treatment (Fig. 7a). PPAR-γ was significantly upregulated (3.15+2.44 fold change) (p= 0.040) while MURF1 was strongly downregulated (0.07+0.02 fold change) (p= 0.042) when compared to contralateral uninjured tissue. These genes were not significantly different from uninjured controls when examined in the combined repair plus gel injection group. The myogenic genes examined (Pax7, MyoG), were generally downregulated for both treatment groups (Fig. 7b). Pax7 was significantly downregulated in both the repair only (0.08+0.07 fold change) and combined repair plus ECM gel group (0.05+0.04 fold change). However, although MyoD appeared downregulated when compared to uninjured controls tissue for both treatment groups, it only reached statistical significance for the repair only group (0.35+0.05 fold change).

Figure 7:

Comparison of relative gene expression for repair only and combined repair plus ECM gel injected groups using RT-PCR. The expression of (a) adipogenesis (PPAR-γ), muscle synthesis (MURF1), (b) satellite cell dynamics (Pax7) and myogenesis (MyoG) were measured using muscle tissue harvested from the SSP of each animal tested. Expression is presented as fold change normalized to contralateral uninjured muscle expression. Group means + SD are presented, n=6–7/group. * denotes statistically significant (p<0.05) differences from uninjured controls.

Analysis of RNA-Seq data in EdgeR revealed a total of 40 differentially expressed genes (all upregulated) between the repair only group relative to uninjured controls, and 1 differentially expressed (downregulated) gene (DNAse1) between the combined repair plus gel group relative to control (Supplementary Table 1). Notably, in the repair only group, IPA detected the activation of the canonical adipogenesis pathway (p<.0001) due to four differentially expressed genes (PPAR-γ, LEP, KLF5, FABP4) and the activation of the canonical osteoarthritis pathway (p<0.001) due to four differentially expressed genes (LEP, MMP-12, PPAR-γ, SSP1). IPA did not detect the activation of any pathways in muscle tissue collected from the combined repair 281 plus ECM gel injection group.

4. Discussion

Our host response findings using the rat model, which included functional, histological, and gene expression measures, suggest that intramuscular ECM gel injection was well tolerated. We detected a significant increase in the expression of inflammatory cytokines (IL-6, IL-1) and macrophage markers (CD68) at 3 days post ECM injection. This finding is consistent with normal muscle wound healing and while not a direct measure of macrophage presence, the increased expression of CD68 suggests macrophage infiltration in response to the presence of ECM fragments from the gel ^{35, 36}. At 14 days post-injection, macrophage marker and inflammatory gene expression returned to basal levels suggesting that ECM injection did not elicit a foreign body response with chronic inflammation ⁵⁴. Future studies could benefit from a more extensive analysis of macrophage dynamics at the site of ECM gel injection. We did not observe foreign body giant cells within tissue sections at any of the time points. Consistent with normal wound healing, specifically remodeling and repair, ²² we observed a characteristic upregulation of key matrix (CoL I and III) and matrix-regulatory (TGF-ß1) genes in response to intramuscular ECM matrix gel injection ⁵⁵. These genes remained upregulated into the second week but had returned to basal levels of expression at 56 days post-injection. Although matrix and matrix regulatory genes were upregulated in response to ECM injection, we did not observe any evidence of intramuscular fibrosis, suggesting normal matrix protein regulation during wound healing with balanced deposition of new matrix and degradation of damaged matrix ⁶. Interestingly, we detected a significant increase in peak muscle force and mass in response to ECM gel injection. This outcome was unexpected, especially absent a finding of increased muscle fiber size, and warrants continued investigation.

The shoulder cuff repair findings indicate that the combination of tendon repair with ECM gel injection can have a positive effect on muscle atrophy. Although significant SSP atrophy occurred within both treatment groups when compared to contralateral uninjured SSP muscle, the use of combined tendon repair with ECM injection significantly reduced the amount of atrophy when compared to the tendon repair only group. Histological evaluation revealed that the repair only group had significantly smaller myofiber cross-sectional area when compared to uninjured controls, which is consistent with the pathological effect of muscle atrophy in other chronic rotator cuff injury models ^{16, 30}. Alternatively, combining tendon repair with ECM gel injection resulted in the preservation of myofiber cross-sectional area in the SSP. Gene expression analysis using RT-qPCR provides a potential explanation for the differences in SSP atrophy and fiber size observed between the two treatments. We detected a reduction in MURF1, a well-known regulator of atrophy that acts as an important rate-limiting enzyme in protein degradation in skeletal muscle ⁴⁷, gene expression within the repair only group, while the expression was equivalent to normal uninjured muscle when examined in combined tendon repair with ECM gel injection tissue. The downregulation of MURF1 in the repair group, coupled with the absence of mass preservation of the SSP suggests that tendon repair alone may be insufficient to sustain hemostasis of muscle protein synthesis in chronic RCT injury. Examination of genes related to myogenesis revealed that Pax7, a marker of satellite cell activation, was significantly downregulated in both treatment groups. These results suggest that atrophic signaling may be inhibiting myogenic signaling in both treatment groups, which has been previously demonstrated in models of chronic RTC injury ³³.

Although it did not reach the level of statistical significance, fatty infiltration was modestly reduced (2%) in response to treatment with combined tendon repair with a single round of ECM gel injection. Cell signaling pathway analysis conducted using the transcriptome profiling results of both treated and uninjured SSP tissue detected the activation of key canonical signaling pathways in the repair only group while none were found to be activated in uninjured SSP tissue or the ECM gel injection group. This suggests that ECM gel injection is possibly having a dampening effect on muscle tissue cellular signaling. In particular, within the adipogenic signaling pathway, four distinct transcription factors (PPAR-γ, LEP, KLF5, FABP4) were found to be significantly upregulated in response to tendon repair only treatment. These results are consistent with the observed increase in adipogenic gene expression in the repair only group with PCR. This story is further corroborated by a similar finding in the upregulation of FABP4 for the repair group, a protein that is involved in fatty acid uptake, transport, and muscle metabolism ^{13, 26}. Interestingly, FABP4 has also been identified to be one of the key transcription factors affecting marbling in cattle meat, a condition with similarities to muscle fatty infiltration following rotator cuff injury ²⁵. An understanding of the transcriptional changes in FABP4 and abrogation of the adipogenesis pathway in response to ECM gel injection may explain the reduced level of fatty degeneration that we observed. In turn, it may provide a candidate target for combinatorial therapies that could be incorporated into ECM gels in order to enhance their efficacy.

Interestingly, we also detected upregulation of the canonical osteoarthritis (OA) pathway in the repair only group, but not in the ECM injection or uninjured SSP muscle groups. Specifically, matrix metalloproteinase-12 (MMP-12) was significantly upregulated in response to tendon repair only. MMPs in general, and MMP-12 specifically, are associated with a number of degenerative diseases including rheumatoid arthritis ^{7, 38, 40}. In addition to degrading a wide range of substrates including those that compose the basal lamina, MMP-12 has also been shown to activate other matrix MMPs, further promoting matrix degradation and mechanical instability of the shoulder region ^{4, 38}. This suggests that current repair techniques could potentially have the unintended consequence of creating tissue degradation and a mechanical environment that is associated with poor postoperative outcomes ⁴⁸. Abrogation of the OA pathway in the combined tendon repair with ECM injection group may warrant further investigation as it would appear to exert a protective effect on not just the injected muscle but the shoulder joint as well. We also detected significant downregulation of DNase1 within the ECM injection group. While there have been no published findings concerning the role of DNase1 in rotator cuff injuries, the finding may be significant as DNase1 delivery has been shown to reduce extracellular traps (ET) that promote inflammatory and thrombosis cascades ¹ which could play a role in shoulder health.

A limitation to this study that deserves discussion was the lack of an early post-injection time point in the rabbit model. The 12-week endpoint examined in this study, while appropriate for the examination of recovery, did not allow us to make longitudinal comparisons between the repair only and repair plus ECM gel groups. As such, we cannot conclude whether the reduction in muscle atrophy that we observed in response to combined repair plus ECM gel treatment was the result of gel induced muscle growth or simply a decrease in the rate of muscle atrophy when compared to repair only. Future studies will incorporate early post-injection time points in order to examine atrophy kinetics, as well as cellular (macrophages and other immune cells) and soluble factor (cytokines and growth factors) responses that may be responsible for the muscle atrophy differences we observed. Lastly, we did not utilize a sham injection as part of the repair only treatment group. As such we cannot separate the effect of the injection site injury from the presence of the ECM gel itself. However, the response to a sham PBS injection in the rat model was negligible, suggesting that the changes we observed in the rabbit model were predominately in response to the presence of the ECM gel.

To sum, its well accepted that chronic rotator cuff tears are associated with a variety of pathological changes to the skeletal muscle, including atrophy, fibrosis, and fatty infiltration ^{8, 18, 32}. While means to repair chronically torn shoulder cuff tendons exist, unaddressed muscle pathologies can decrease the capacity for tendon healing, prevent full recovery of shoulder function, and are generally associated with poor outcomes ⁸. We envision the use of the ECM gel material explored in this study as an adjunct to existing tendon repair strategies, in order to address both tendon and muscle injury in patients with chronically torn rotator cuffs. It is worth noting that the clinical use of ECM materials in muscle is not without precedent. Acellular ECM scaffolds have been used in to improve the strength and range of motion in patients with volumetric muscle loss injuries ¹². While we do not yet know the mechanism responsible for the improvement in muscle atrophy we observed, the ECM gel injection results when explored in a delayed SSP repair model were encouraging and warrant continued pre-clinical investigation in order to firmly establish safety and efficacy.

5. Conclusion

ECM gel injection elicited a host response that resulted in the elevation of wound repair gene expression at early time points, but did not lead to any pathological changes in muscle architecture, suggesting tolerance by the host environment. Furthermore we detected a significant increase in muscle mass and contractile force in response to ECM gel injection, which warrants continued exploration. When examined in a delayed SSP repair model we observed a modest but measurable effect when ECM gel was used in combination with tendon repair. Specifically, intramuscular ECM gel injection mitigated SSP atrophy and abrogated the activation of key adipogenic and osteoarthritic signaling pathways. The results suggest that the short-term activation of wound healing pathways observed in the rat model are complemented by a longer-term positive effect on muscle atrophy and cellular signaling.

Supplementary Material

1

Supplementary Table 1: Differentially expressed (compared to uninjured) genes detected (RNA-Seq) within the repair only group. Only one ( DNAse1) differentially expressed gene was detected in the repair plus ECM gel injection group. ( p<0.05, log ratio >1.5, and FDR>0.25)