The Secretome From Mechanically Loaded Myoblasts Enhances Tenocyte‐Mediated Wound Healing in a 3D In Vitro Tendon Model

ABSTRACT

Exercise is well known to promote tendon healing, an effect traditionally attributed to mechanical loading‐induced responses within the tendon itself. However, skeletal muscle also functions as a secretory organ, releasing bioactive factors (secretome) during exercise that influence various tissues. We hypothesized that muscle‐derived secretome released during exercise may also contribute to tendon healing. To test this, we applied mechanical loading to cultured muscle cells (myoblasts) using the FlexCell tension system to simulate exercise in vitro. Our previous studies, using 2D‐cultured tendon cells (tenocytes), have demonstrated that secretome from statically loaded myoblasts, particularly under 2% loading, enhanced tendon healing‐related responses. Building upon these findings, we employed a 3D tendon construct model to more closely mimic in vivo healing conditions. We found that secretome derived from statically loaded myoblasts, especially at 2% loading, promoted tendon healing‐related processes as compared with the control group, which received no secretome treatment (no conditioned media). These included increased cell‐covered area, expression of the tenocyte marker scleraxis (SCX), and elevated production of Type I and III collagens at an early stage (Day 7). Additionally, a reduction in type III collagen production was found at a later stage (Day 14), suggesting a potentially accelerated healing process. These findings highlight the therapeutic potential of the muscle‐derived secretome in promoting tendon healing and may inform future strategies for rehabilitation and regenerative medicine.

Human primary myoblasts were subjected to 2% or 10% static mechanical loading for 24 h, with unloaded cells as controls. Conditioned media were collected and applied, together with normal medium, to a 3D tendon fibrin gel model. Secretome from loaded myoblasts, particularly 2% loading, enhanced tenocyte proliferation, scleraxis (SCX) expression, and both type I and III collagen production at Day 7, while reducing type III collagen expression at Day 14, indicating improved tendon healing potential.

1. Introduction

Tendons consist primarily of a dense extracellular matrix (ECM), rich in collagen, which provides tensile strength and enables body movement. Tendon ruptures are complex, multifactorial injuries that commonly occur in individuals exposed to high‐force exertion, repetitive strain, or intense athletic activity [1, 2, 3]. Tendon ruptures heal slowly and often with suboptimal outcomes, including a high risk of re‐tear. In certain patient populations, recurrence rates reach up to 90% [4, 5, 6].

Rehabilitation is widely used during the healing process. Previous studies have shown that applying low amounts of mechanical tension to an injured tendon a few weeks after injury can promote healing [7, 8, 9]. Traditionally, the prevailing view has been that tension directly deforms the ECM of the tendon cells (tenocytes), and this deformation alters gene expression through mechanotransduction pathways [10, 11]. However, emerging evidence suggests that tendon tension may not be the only mechanical factor influencing tendon functions. Skeletal muscle may also play an important regulatory role. Recent works suggest that tendon healing is shaped by a dynamic interplay between muscle and tendon tissues. For example, in rats, injury to the gastrocnemius muscle alters ECM remodeling and affects the functional properties of the Achilles tendon [12]. Likewise, in zebrafish, mechanotransduction originating from muscle contraction modulates tenocyte differentiation via transforming growth factor‐β (TGF‐β) signaling [11]. In addition, a study from our group showed that tendons in denervated muscles that are only passively exposed to tension exhibit degenerative changes [13].

Together, these findings suggest that it is not only the structural presence of muscle that influences tendon function, but also the active tension generated within muscle and its associated signaling. We therefore hypothesize that secretomes derived from mechanically loaded skeletal muscle contribute to tendon function and healing via muscle‐derived factors that enter the circulation and ultimately reach the tendon. Our previous work supports this concept: secretome from muscle cells (myoblasts) exposed to low levels of static loading enhanced tenocyte proliferation through insulin‐like growth factor‐1 (IGF‐1) signaling [14] and promoted tenocyte migration via Neuroblastoma suppressor of tumorigenicity 1 (NBL1) [15]. Cell migration and proliferation are fundamental processes in tendon healing. Migration enables the recruitment of repair cells to the injury site, while proliferation ensures an adequate supply of these cells for tissue repair. In the present study, we extend these findings by examining additional tendon‐healing–related parameters that may be influenced by secretome from mechanically loaded myoblasts.

Tenocytes are responsible for ECM synthesis and turnover and play a central role in maintaining tendon homeostasis and tenogenesis [16]. These cells express scleraxis (Scx), a key transcription factor essential for tendon development and maintenance [17, 18]. Scx directly regulates several genes involved in the early phase of healing, including those controlling tenocyte differentiation as well as ECM production and organization [19]. Proper regulation of SCX‐mediated pathways supports a tenogenic phenotype and balanced matrix remodeling; however, dysregulation of this process can shift cells toward a pro‐fibrotic lineage, characterized by myofibroblast differentiation and excessive ECM deposition [20].

This pathological shift is reflected by increased expression of actin alpha 2, smooth muscle (gene: ACTA2/protein: α‐SMA) and S100 calcium‐binding protein A4 (S100A4), which are closely associated with fibrotic healing. α‐SMA is a hallmark of myofibroblasts and drives wound contraction and ECM remodeling, but its sustained expression contributes to pathological fibrosis [21, 22, 23]. S100A4, a fibroblast‐enriched protein, is similarly upregulated in fibrotic conditions, and its inhibition has been shown to reduce scar formation and limit fibrosis‐promoting myofibroblasts [24, 25].

The composition and organization of the tendon ECM, which are determined by the phenotype and activity of resident cells, critically influence healing outcomes [26]. Type I collagen is the main structural component of healthy tendons, while type III collagen provides an early provisional scaffold during repair. Both increase during the proliferative phase, but type I collagen gradually replaces type III collagen as remodeling restores normal tendon architecture [27, 28]. Fibronectin, a glycoprotein involved in adhesion, migration, and matrix organization, is also upregulated after injury and supports fibroblast attachment and ECM assembly, contributing to the structural and mechanical restoration of the healing tendon [29, 30].

While our previous studies used 2D co‐cultures of rat myoblasts and tenocytes and proteomic profiling of the secretome, it remained unclear how mechanical loading intensity influences the pro‐healing potential of the secretome in a more physiologically relevant system. To address this, we employed a 3D fibrin‐based tendon model combining primary human myoblasts and tenocytes. In this system, tenocytes were embedded within a 3D fibrin gel. This model mimics key features of in vivo tendon healing, as fibrin is the primary provisional ECM component produced during the initial phase of healing. Fibrin rapidly covers the injury site before being replaced by newly synthesized collagen. In addition, the model generates tension by anchoring the two ends of the gel to the culture wells, allowing evaluation under mechanical conditions that more closely resemble native tissue.

We hypothesized that myoblast‐derived secretome, generated under an optimal static loading, would maximize tendon‐healing potential in our 3D model. To test this, we evaluated the effects of secretomes derived from mechanically loaded human myoblasts on tendon‐healing–related outcomes over a 14‐day culture period. We assessed metabolic activity, total protein concentration, cell coverage area, and morphological changes as indicators of cell proliferation, along with expression of SCX, S100A4, and ACTA2, and regulation of key ECM components, including type I and III collagens and fibronectin. These endpoints were measured at Days 7 and 14, with Day 7 reflecting early cellular responses and Day 14 capturing the differences in the newly formed tendon‐like constructs.

2. Materials and Methods

2.1. Human Muscle Cell Isolation

Human muscle biopsies were obtained from human residual semitendinosus muscle tissue collected during Anterior Cruciate Ligament (ACL) reconstruction surgery. The Swedish Ethical Review Authority granted ethical approval (2021‐05627‐01). All tissue donors signed a written consent form. Muscle samples were washed with HBSS (Gibco, 14170112, UK), and fibrotic and adipose tissues were carefully removed. The muscle tissue was cut into small pieces and enzymatically digested for 1 h at 37°C using Dulbecco's Modified Eagle Medium (DMEM) + GlutaMAX (Gibco, 61965‐026, USA) supplemented with 2 mg/mL collagenase (Sigma‐Aldrich, C0130, USA), and 10% fetal bovine serum (FBS; Sigma‐Aldrich, F9665, USA). After digestion, the suspension was centrifuged at 500 g for 5 min to remove collagenase, resuspended in culture medium, and incubated for 4 days. Following incubation, both the muscle debris in the culture medium and trypsin‐detached cells were collected and filtered through a 40 μm nylon filter (Corning, USA, 431750). The resulting cell suspension was centrifuged at 500 g for 5 min, resuspended in culture medium, and plated onto a collagen‐coated plate for muscle cells purification, with shaking every 5 min for 1 h. The remaining unattached cells, identified as muscle cells, were transferred to a flask and expanded until passage 3 before being cryopreserved at −80°C. Cells between passages four and six were used for experiments, with their phenotype markers measured before use.

2.2. Mechanical Loading of Muscle Cells and Collection of the Conditioned Medium

Primary myoblast from four donors (two female and two male donors, ages 19–29) was seeded onto separated FlexCell plates at a density of 1.5 × 10⁵ cells per well. After 24 h, the culture medium was changed into DMEM/F‐12 (Gibco, 11039‐021) supplemented with 1% FBS (Sigma‐Aldrich, F9665, USA). Cells were then loaded with 2% or 10% static loading for 24 h using the FlexCell Tension System. Unloaded cells served as a control.

After the loading period, conditioned medium was collected. To obtain sufficient volumes for subsequent experiments, conditioned medium from two female and two male donors exposed to the same loading intensity was pooled, centrifuged at 1000× g for 5 min at 4°C, and stored at −80°C until further use.

2.3. Human Tendon Cell Isolation Protocol

Human tenocytes isolated from semitendinosus tendons obtained from six patients undergoing ACL reconstructive surgery at Linköping University Hospital (Linköping, Sweden) were used for a 3D tendon model. Following informed consent and approval by the Regional Ethical Review Board in Linköping (2015/408‐31) and an amendment from the Ethical Review Authority in Sweden (2024‐04117‐02). Tendon tissue (2–3 mm) was enzymatically digested overnight at 37°C in 5% CO² using DMEM/F12 (Thermo Fisher Scientific, Hvidovre, Denmark, 21041025) supplemented with 1 mg/mL type II collagenase (Worthington Biochemical, NJ, USA, LS004177), 20% FBS (Thermo Fisher Scientific, A5256801), and 1% penicillin–streptomycin (Pen Strep; Thermo Fisher Scientific, 15140‐122). Following digestion for 18–24 h, cells were centrifuged, washed with cell culture medium, and cultured in flasks containing cell culture medium (DMEM/F12 medium supplemented with 10% FBS and 1% Pen Strep). For sub‐culturing, cells were detached using Trypsin/EDTA (Thermo Fisher Scientific, 15 400–054).

2.4. 3D Fibrin Gels With Tenocytes

Twenty‐four‐well cell culture plates were first coated with silicone on the bottom (Dow Chemicals, DOW SYLGARD 184, Michigan, USA), followed by sterilization through immersion in 75% ethanol for 45 min, air‐drying in a culture hood, and washed twice with sterile PBS. Primary human tenocytes from six donors were suspended at a density of 3.33 × 10⁵ cells/mL in tendon construct medium (DMEM/F‐12 supplemented with 5% FBS, 200 μM ascorbic acid) (Sigma‐Aldrich, A8960, Missouri, USA), 50 μM L‐proline (Sigma‐Aldrich, P0380), and 1% Pen Strep (Thermo Fisher Scientific, 15140‐122). A volume of 120 μL of the cell suspension was seeded per well. To promote fibrin gel formation, the cell suspension was mixed with 125 μg/mL aprotinin (Roche Diagnostics, 10236624001, Mannheim, Germany), 5 mg/mL fibrinogen (Sigma‐Aldrich, F3879), and 1 U thrombin (Sigma‐Aldrich, T6884). After approximately 1 h of incubation, allowing for fibrin polymerization, previously collected myoblast conditioned medium supplemented with 5% FBS, ascorbic acid, L‐proline, and Pen Strep was added on top of the gels. For each donor, one well was allocated per treatment group. Controls were treated with normal tendon construct medium. The medium was replaced on Days 2 and 5. To enable cell‐driven gel contraction, the fibrin gel was detached from the sides of the well on Day 2 and the samples were collected for analysis at Day 7 (Table 1).

TABLE 1.

Layout for the different treatment groups, plate type and analysis for each time‐point.

		Control	Unloaded	2% static loading	10% static loading
Day 7	24 well plate	RNA/CCK‐8	RNA/CCK‐8	RNA/CCK‐8	RNA/CCK‐8
	6 well plate	Histology/Protein	Histology/Protein	Histology/Protein	Histology/Protein
Day 14	6 well plate	RNA/CCK‐8	RNA/CCK‐8	RNA/CCK‐8	RNA/CCK‐8
	6 well plate	Histology/Protein/Imaging a	Histology/Protein/Imaging a	Histology/Protein/Imaging a	Histology/Protein/Imaging a

^a Imaging done Days 7 and 14.

2.5. Tendon Construct Formation

Human tendon constructs were assembled from primary tenocytes as previously described [31]. Six‐well cell culture plates were coated with silicone as described above. Two surgical silk sutures (Ethicon, 684G, New Jersey, USA), each approximately 3–4 mm long, were pinned 1.5 cm apart on the plates using stainless steel insect pins (0.1 mm diameter; Fine Science Tools, 26 002‐10, California, USA). The plates were then sterilized in 75% ethanol, air‐dried, and washed twice with PBS. Primary human tenocytes from 6 donors were suspended at a density of 3.33 × 10⁵ cells/mL in the tendon construct medium described above. A total volume of 600 μL of the tendon construct medium was seeded per well. The constructs were incubated for approximately 1 h to allow complete fibrin polymerization. Previously collected myoblast conditioned medium, supplemented with 5% FBS, ascorbic acid, L‐proline and Pen strep was added on top of the gel. For each donor, one well was allocated per treatment group. Controls were treated with tendon construct medium containing 5% FBS. The medium was replaced on Day 2, 5, 7, 9, and 12. To enable cell‐driven gel contraction, the fibrin gel was detached from the sides of the well on Day 2 and the samples were collected for analysis at both Days 7 and 14 (Table 1).

2.6. Gel Contraction Measurement

On Days 7 and 14, each fibrin gel and construct were photographed using a Dino‐lite microscope camera (20×, Dino‐Lite Europe, Netherlands). The gel width was measured with Fiji [32] from a representative image per treatment group for each donor. The contraction rate was calculated by subtracting the gel width from the well diameter and then dividing it by the well diameter.

2.7. Cell Morphology and Metabolism Measurement

Brightfield images of fibrin gels with tenocytes were taken at 100× magnification on Day 7. One image per treatment group for each donor was captured and used to quantify the cell‐covered area using Fiji. Within each image, 15 cells with clearly defined edges were randomly selected for measurement of cell length and width using Fiji. The mean cell length‐to‐width ratio and mean cell length normalized to the control were calculated to assess cell morphology.

Cell metabolism was measured using a CCK‐8 kit (Dojindo Laboratories, Tabaru, Japan, CK04) following the manufacturer's instructions. Briefly, 3D fibrin gels or tendon constructs were incubated with CCK‐8 reagent at 37°C with 5% CO² for 2 h. One well per treatment group was used for each donor. Following incubation, the optical density (OD) value was measured at 450 nm using a microplate reader. The OD values from six donors were then normalized to their control, and metabolic activity was expressed as a percentage relative to the control.

2.8. RNA Extraction and qPCR

Fibrin gels and tendon constructs were washed twice with PBS and then homogenized in 1 mL TRIzol (Thermo Fisher Scientific, 15596026) and stored at −80°C until use. After thawing at room temperature for 10–15 min, 300 μL of chloroform was added to the homogenate and shaken vigorously for 15 s. The tubes were then kept at room temperature for 3–5 min before centrifuging at 14 000 g for 15 min at 4°C. The upper aqueous phase was transferred to a new tube, and an equal volume of 70% ethanol was added and mixed thoroughly. RNA was then extracted using the Qiagen RNeasy Mini Kit (Qiagen, 74140, Germany). Briefly, the sample was loaded onto a RNeasy Mini spin column in a collection tube and centrifuged at 14 000 g for 25 s. The RNA was washed once with RW1 and twice with RPE buffers by centrifugation at 8000 g for 25 s. Finally, 40 μL of RNase‐free water was added to the column, and RNA was eluted into a new Eppendorf tube by centrifuging at 13 400 g for 1 min.

Subsequently, a high‐capacity cDNA reverse transcription kit (ThermoFisher, 4368813, Lithuania) was used to reverse transcribe mRNA into cDNA. TaqMan Fast Advanced Master Mix (ThermoFisher, 4444963) and cDNA were used with different probes (Table 2) to amplify the target genes. Quantitative real‐time PCR (QuantReady K9600, Life Real), with three technical replicates per cDNA sample. The results were analysed using the 2^−∆∆Ct method. Target genes and corresponding probe information used for qPCR are listed in Table 2.

TABLE 2.

List of genes used for qPCR.

Gene symbol	Gene names	Reference number	Manufactory
GAPDH	Glyceraldehyde‐3‐phosphate dehydrogenase	Hs02786624_g1	ThermoFisher
SCX	Scleraxis BHLH transcription factor	Hs03054634_g1	ThermoFisher
S100A4	S100 calcium binding protein A4	Hs00243202_m1	ThermoFisher
ACTA2	Actin alpha 2, smooth muscle	Hs00426835_g1	ThermoFisher
Col1A1	Collagen type I alpha 1 chain	Hs00164004_m1	ThermoFisher
Col3A1	Collagen type III alpha 1 chain	Hs00943809_m1	ThermoFisher

2.9. Protein Extraction, Protein Concentration Measurement, and Western Blot

Fibrin gels and tendon constructs were washed twice with PBS, collected in tubes, quickly frozen with liquid nitrogen, and stored at −80°C until further use. Upon thawing, 200 μL RIPA buffer (Thermo Scientific, 89901, USA) supplemented with a protease and phosphatase inhibitor cocktail (1861282, Thermo Scientific) was added, and the constructs were incubated on ice for 15 min. The constructs were then homogenized using a tissue disruptor with disposable probes. The homogenized solution was further incubated on ice for 30 min. Subsequently, the samples were centrifuged at 14 000 g for 15 min at 4°C. The supernatants of the samples were collected and stored at −80°C for further use.

Protein concentration was measured using the Pierce BCA Protein Assay Kit (Thermo Scientific, 23225, USA). Two technical replicates per treatment group for each donor were analyzed to assess cell density within the gels and constructs and to prepare samples for Western blot (WB) analysis. For WB, 10–20 μg of proteins were separated on SDS‐polyacrylamide gels (Bio‐Rad, 4561044, USA) and transferred to a PVDF membrane (Cytiva, 10600023, Germany). Membranes were blocked with 5% BSA for 1 h at room temperature before and incubated overnight at 4°C with primary antibodies (Table 3). After primary antibodies incubation, the membranes were washed and incubated with HRP‐conjugated anti‐rabbit IgG secondary antibodies (Table 3). Following secondary antibody staining, the membranes were washed again and treated with an enhanced chemiluminescent (ECL) HRP substrate (Thermo Scientific, 34577, USA). Signal detection and analysis were performed using the Odyssey Fc Dual‐Mode Imaging System (LI‐COR Biotechnology, Nebraska, USA).

TABLE 3.

List of antibodies used for Western blot and immunofluorescence staining.

Antibodies	Reference number	Manufactory	Concen‐tration	Species	Application
Anti‐collagen I	Ab260043	Abcam	1:1000	Rabbit	WB
Anti‐collagen III	Ab184993	Abcam	1:1000	Rabbit	WB
Anti‐fibronectin	Ab2413	Abcam	1:200	Rabbit	WB
Anti‐collagen I	C2456	Sigma	1:200	Mouse	IF
Anti‐collagen III	Ab184993	Abcam	1:200	Rabbit	IF
Anti‐a‐SMA	Ab32575	Abcam	1:200	Rabbit	IF
Donkey anti‐rabbit IgG	A31574	Invitrogen	1:400	Rabbit	IF
Goat anti‐mouse IgG	A11004	Invitrogen	1:400	Mouse	IF
Anti‐rabbit IgG	7074	Cell signal	1:2000	Rabbit	WB

2.10. Embedding and Sectioning for Histology

Gels and constructs were washed twice with PBS and subsequently embedded in OCT compound within plastic molds. The molds were placed in a container filled with 2‐propanol, immersed in liquid nitrogen to rapidly freeze the OCT‐embedded construct for approximately 5 min. The frozen constructs were stored at −80°C until use. For sectioning, the constructs were cut into 8 μm thick sections and stored at −20°C until further use.

2.11. Immunofluorescence Staining

Frozen sections were fixed by air drying for 10–15 min at room temperature. Sections were then fixed either in methanol at −20°C or 4% paraformaldehyde (PFA) at room temperature for 10 min. For intracellular proteins staining, the constructs were permeabilized with 0.3% Triton for 5 min at room temperature; this step was omitted for extracellular protein staining. The constructs were then blocked with 5% BSA for 1 h at room temperature before overnight incubation at 4°C with the primary antibodies (Table 3) diluted in 5% BSA. After four washes with PBS, the sections were incubated with secondary antibodies (Table 3) for 1 h at room temperature, followed by another four PBS washes. The sections were incubated in phalloidin for 30 min when required. After an additional four PBS washes, DAPI staining was performed for 5 min at room temperature, and finally, the sections were mounted with the mounting medium. Images were acquired using a Leica SP8 Falcon confocal microscope with a 63× objective. All images within each experiment were acquired and analyzed by a single operator who was blinded to the experimental group identities.

2.12. Statistic

Statistical analyses were performed using GraphPad Prism 10.4.2 (GraphPad Software, San Diego, CA, USA). For datasets that followed a normal distribution and showed no outliers (assessed using the Shapiro–Wilk test and ROUT method with Q = 1%), repeated measures one‐way ANOVA was applied for paired/matched data. Post hoc comparisons were performed using paired t‐tests with Holm–Šidák correction for multiple testing, and results were reported as mean ± standard deviation (SD). For datasets that did not meet the assumptions of normality or contained outliers, the Friedman test (for paired/matched data) was used. When the data were normalized to the control group, the Kruskal–Wallis test was applied instead. For nonparametric analyses, Dunn's multiple comparisons test was used for post hoc analysis, and results were presented as median with interquartile range (IQR).

The sample size of six was selected to balance biological variability with experimental feasibility. Human primary cell–based studies are inherently limited by sample availability, cell yield, and variability between donors. Using six independent biological samples provided sufficient reproducibility and statistical reliability for mechanistic investigations while remaining practical within the constraints of primary cell culture work. This number of biological samples is also consistent with previous primary‐cell studies [33, 34].

3. Results

3.1. Effects of Myoblast‐Derived Secretome on Tenocyte Growth and Metabolic Activity in 3D Fibrin Gels

Figure 1 presents a schematic overview of the experimental workflow. We first evaluated whether the myoblast‐derived secretomes influenced tenocyte proliferation by measuring cell‐covered area, cell morphology, metabolic activity, and total protein concentration, parameters that collectively indicate proliferative behavior. At Day 7, the cell‐covered area significantly increased in groups treated with secretome from the unloaded and 2% loading conditions, with median values of 32.68% (IQR: 28.21–37.82) and 32.57% (IQR: 30.05–35.82), respectively, compared with 22.32% (IQR: 17.27–26.86) in the control group (p = 0.044 and 0.022). The 10% loading group showed a median of 34.11% (IQR: 29.66–38.19), while this did not reach statistical significance (p = 0.083) compared with the control (Figure 2A). To determine whether the increased cell‐covered area reflects increased cell number or changes in cell size or shape, we examined cell shape parameters. The 2% loading group exhibited a median cell length‐to‐width ratio of 18.39 (IQR: 13.37–18.61), while the control group showed a median of 7.66 (IQR: 6.97–11.02); however, the difference was not statistically significant (p = 0.171) (Figure 2B). Cell width was unchanged (data not shown). Cell length in the 2% and 10% loading groups was 1.33‐fold (IQR: 1.23–1.94) and 1.52‐fold (IQR: 1.11–1.99) to the control, respectively, but these differences were also nonsignificant (p = 0.125 and 0.075) (Figure 2C). Figure 2D shows representative images of tenocyte density in the 3D tendon gels at Day 7. The myoblast secretome groups appeared to exhibit a higher cell density visually. However, since these observations were not quantitatively measured, they should be considered purely qualitative. Overall, these results show that conditioned media from myoblasts modestly increased cell‐covered area at Day 7, while morphology‐related parameters did not differ significantly between groups.

FIGURE 1.

Experiment workflow. (A) Human myoblasts were cultured and subjected to loading for 24 h, after which the condition medium (secretome) was collected. (B) The tenocytes were seeded in the 3D fibrin gel and treated with myoblasts‐derived secretome. (C) Expression and structure of ECM proteins, indicators of proliferation, tenocyte phenotype and morphology, as well as 3D tendon gel contraction, were analyzed at both Days 7 and 14.

FIGURE 2.

Effects of myoblast‐derived secretome on tenocyte growth and metabolic activity in 3D fibrin gels. (A) Cell‐covered area. (B) Cell length‐to‐width ratio. (C) Cell length normalized to control cells. (D) Representative images of tenocytes in the 3D tendon gels used for measured cell‐covered area. (E, G) Cell metabolism measured by CCK‐8 assay. (F, H) Total protein concentration in 3D tendon gels. Panels (A–F) represent data from Day 7, Panels G and H represent data from Day 14. (B, C) N = 4 biological replicates (independent donors), as cell edges were unclear in samples from two donors and could not be measured. (A, D, E, G) N = 6 biological replicates (independent donors); for each donor, one technical replicate was included per treatment group. (F, H) N = 6 biological replicates (independent donors); for each donor, two technical replicates were included per treatment group. *p < 0.05; **p < 0.01.

Using the CCK‐8 assay, we observed that cell metabolic activity was significantly increased in both the unloaded group and the 10% loading group, with median values 1.70‐fold (IQR: 1.33–2.58) and 1.89‐fold (IQR: 1.46–16.09) above the control, respectively (p = 0.027 and 0.014). The 2% loading group showed a 1.40‐fold change (IQR: 1.16–14.49) relative to control; however, this difference was not statistically significant (p = 0.263) (Figure 2E). Additionally, total protein concentration also increased in all three conditioned‐medium groups at Day 7, with mean values of 491.0 ± 46.86, 616.6 ± 60.86, and 617.0 ± 133.3 μg/mL for the unloaded, 2%, and 10% loading groups, respectively, compared with 388.0 ± 88.4 μg/mL in the control (p = 0.001, 0.002, and 0.005) (Figure 2F). Metabolic activity showed modest changes in the unloaded, 2%, and 10% loading groups—1.46‐fold (IQR: 0.98–2.13), 1.57‐fold (IQR: 1.20–2.12), and 1.70‐fold (IQR: 1.32–2.02) relative to control, respectively—but none of these differences reached statistical significance (p = 0.385, 0.319, and 0.195) (Figure 2G). Total protein concentrations were also not significantly different among groups at Day 14 (Figure 2H).

3.2. Sectretome Derived From Myoblasts Subjected to 2% Static Loading Enhanced Tenocyte Marker Expression in the 3D Fibrin Gels

We further investigated whether the secretome from loaded myoblasts influences tenocyte markers associated with healing. At Day 7, SCX gene expression demonstrated median fold‐change of 1.78 (IQR: 1.60–2.41), 2.54 (IQR: 1.74–2.96), and 1.64 (IQR: 1.17–2.72) in the unloaded, 2% and 10% loading groups, respectively, compared with the control. A statistically significant increase was observed only in the 2% loading group (2% p = 0.039; unloaded p = 0.205, 10% loading p = 0.443) (Figure 3A). Our results showed that the gene expression of both ACTA2 and S100A4 remained unchanged across all groups at both Days 7 and 14, indicating that myoblast‐derived secretome did not significantly alter these fibrosis‐associated markers under the conditions tested.

FIGURE 3.

Secretome derived from loaded myoblasts influenced tenocyte ECM component expression in the 3D tendon model. (A–C) Tenocyte marker expression analyzed by qPCR on both Days 7 and 14. (D) Confocal microscopy showing stress fiber (α‐SMA) formation in tenocytes within the fibrin gels and tendon constructs. (A–C) N = 6 biological replicates (independent donors); for each donor, three technical replicates were included per treatment group *p < 0.05.

Confocal microscopy of representative images suggested differences in α‐SMA and F‐actin staining among the treatment groups. At Day 7, both F‐actin and α‐SMA appeared more prominent in tenocytes treated with conditioned medium from unloaded myoblasts (Figure 3D). The 10% loading group showed relatively strong F‐actin staining, whereas the 2% loading group appeared to have weaker F‐actin staining. By Day 14, within the contracted 3D tendon constructs, F‐actin staining appeared low in the 2% loading group, while α‐SMA staining appeared broadly similar across the control, unloaded, and 2% loading groups. The representative images suggest possible variations in cytoskeletal organization and α‐SMA/F‐actin distribution, but these observations were not quantified and should be interpreted qualitatively.

3.3. Secretome Derived From Loaded Myoblasts Influenced Tenocyte ECM Component Expression in the 3D Tendon Model

We next examined the expression of key ECM proteins in tendons, focusing on type I collagens and fibronectin. qPCR analysis showed that type I collagen (COL1A1) expression at Day 7 was significantly higher in the 2% and 10% loading groups, with median fold‐changes of 2.62 (IQR: 1.74–5.00) and 2.56 (IQR: 1.54–4.68) compared to the control (p = 0.009 and 0.039, respectively). No significant differences were observed among groups at Day 14 (p > 0.999) (Figure 4A).

FIGURE 4.

Secretome derived from loaded myoblasts influenced tenocyte ECM component expression in the 3D tendon model. (A–C) ECM protein expression in tenocytes analyzed by qPCR at Days 7 and 14. (D) Representative samples from one patient demonstrate the expression of ECM proteins assessed by Western blot (WB). (E–I) Quantification of ECM protein expression from WB analysis. n = 6 biological replicates (independent donors); for each donor, one technical replicate was included per treatment group. *p < 0.05; **p < 0.01.

For type III collagen (COL3A1), median expression at Day 7 was 1.70‐fold (IQR: 1.20–1.89) in the unloaded group (p = 0.503), 2.12‐fold (IQR: 1.23–2.44) in the 2% loading group (p = 0.115), and 2.23‐fold (IQR: 1.37–2.47) in the 10% loading group (p = 0.072) relative to control. Although post hoc comparisons were not significant, the overall Kruskal–Wallis test indicated a group effect at Day 7 (p = 0.048). By Day 14, COL3A1 expression decreased to 0.56‐fold (IQR: 0.46–0.92) in the 2% loading group compared to control (p = 0.048) (Figure 4B). No differences were observed between the groups for fibronectin (FN1) at Day 7 or 14 (Figure 4C).

Figure 4D shows the representative blots for Western blot analysis. The type I collagen antibody recognized the C‐terminal peptide of pro‐type I collagen (pro‐collagen I), resulting in three bands: the upper band representing full‐length pro‐collagen I, a middle band corresponding to Cp‐collagen I after N‐terminal peptide removal, and the lowest band representing fully processed mature type I collagen (Collagen I). Quantification of Western blot data showed that at Day 7, pro‐collagen I, Cp‐collagen I, and mature collagen I levels were higher in all three myoblast secretome treatment groups compared with the control. In the unloaded group, median fold‐changes were 6.38 (IQR: 2.76–44.25), 4976 (IQR: 60.50–18 294), and 3.06 (IQR: 1.97–9.37), respectively (p = 0.031, 0.045, and 0.021). In the 2% loading group, median levels were 10.85 (IQR: 2.24–34.50), 12 690 (IQR: 48.02–28 950), and 5.08 (IQR: 2.27–8.80) (p = 0.016, 0.008, and 0.009), and in the 10% loading group, 8.88 (IQR: 2.52–77.39), 12 872 (IQR: 63.02–19 283), and 2.66 (IQR: 2.10–12.46) (p = 0.012, 0.016, and 0.031) (Figure 4E–G, upper panels). By Day 14, these differences were no longer significant (Figure 4E–G, lower panels).

For type III collagen (collagen III), median expression at Day 7 was 2.61‐fold (IQR: 2.06–4.16) in the 2% loading group and 2.77‐fold (IQR: 1.65–4.24) in the 10% loading group compared with control (p = 0.005 and 0.04), consistent with qPCR results. By Day 14, type III collagen expression tended to decrease in some samples within the 2% and 10% loading groups, but no significant differences were observed across the full dataset (Figure 4H).

For fibronectin, only the 10% loading group showed a statistically significant increase at Day 7, with a median expression 5.69‐fold higher than the control (IQR: 1.62–8.46; p = 0.022). The unloading and 2% loading groups showed a median of 4.55‐fold (IQR: 1.04–7.80) and 4.47‐fold (IQR: 0.71–6.99), respectively, but these changes were not statistically significant. By Day 14, fibronectin expression showed a decreasing trend in both the 2% and 10% loading groups, though without significant group differences (p = 0.442 and 0.705; Figure 4I).

3.3.1. Secretome Derived From Loaded Myoblasts Qualitatively Regulates ECM Structure of Tendon Constructs

Next, we examined how the myoblast secretome influenced the structural appearance of the tendon constructs. At Day 7, the gels had only begun to contract, and tenocytes were distributed throughout the fibrin matrix in all groups. At this early stage, the overall structural organization appeared similar across conditions. Type I collagen staining was visible in all groups treated with myoblast secretome. It was type III collagen staining, with no obvious visual differences between the different groups (Figure 5, upper panel).

FIGURE 5.

Secretome derived from loaded myoblasts qualitatively regulated ECM structure of tendon constructs. Representative images for type I and III collagens structure in the 3D tendon gels visualized using confocal microscopy.

By Day 14, the tendon constructs showed more defined tendon‐like structures across all conditions. In the control and unloading groups, tenocytes appeared aligned, with type I and III collagens displaying fibrous patterns. In the 2% and 10% loading groups, the constructs maintained a broader gel‐like appearance, and type I and III collagen staining appeared less compact upon visual inspection. As no quantitative analysis was performed, these findings are purely descriptive (Figure 5, lower panel).

3.4. Secretome From Myoblasts Did Not Influence the Contraction of 3D Fibrin Gels

Finally, we investigated how the secretome from myoblasts affects the contraction of 3D fibrin gels. At Day 7, the gels had only begun to contract, and no significant differences were detected among the four groups in either the contraction rate (calculated as the well width minus the gel width, divided by the well width) or in the gel width normalized to the control group (Figure 6A,B). By Day 14, the contraction rate remained comparable across all groups, with no statistically significant differences (Figure 6C,D).

FIGURE 6.

Secretome from myoblasts did not influence the contraction of 3D fibrin gels. (A, C) Gel contraction rates at Days 7 and 14. (B, D) Gel width normalized to the control gel. (E) Representative images showing 3D tendon gel contraction. n = 6 biological replicates (independent donors); for each donor, one technical replicate was included per treatment group.

4. Discussion

Our findings indicate that secretome derived from mechanically loaded myoblasts may influence early tenocyte responses in the 3D tendon model, as reflected by increased metabolic activity and elevated expression of several ECM‐related genes and proteins at Day 7. Although these changes were most apparent when compared with the control group, statistical differences between the loaded and unloaded secretome were not consistently observed. The secretome from 2% loaded myoblasts showed an increase of SCX expression at Day 7 and a decrease of type III collagen expression at Day 14 compared to control, which may indicate a shift toward accelerated healing and remodeling. Overall, the data provide preliminary evidence that mechanical loading can alter the myoblast secretome in ways that may modulate tenocyte behavior, but additional studies with larger sample sizes and direct comparisons between loading conditions are needed to confirm these patterns.

Tenocyte proliferation plays a crucial role in tendon healing. However, directly measuring cell proliferation with methods like EdU is not feasible in the 3D culture setting. The 3D tendon culture requires a 14‐day incubation period; EdU incorporation during this time could interfere with tenocyte activity and viability. For this reason, total metabolic activity assessed by the CCK‐8 assay served as the primary measure of cell growth in the 3D model. Additionally, total protein concentration, cell coverage area, and morphological changes were evaluated to further characterize proliferative behavior. Our results showed that the secretome from myoblasts subjected to 2% static loading increased the tenocyte‐covered area and total protein concentration in the fibrin gel at Day 7, while it did not induce significant morphological changes. These findings suggest that the increase in the tenocyte‐covered area was likely driven by cell proliferation. In the 10% loading group, we observed enhanced cellular metabolism and total protein concentration, which may also reflect an increase in tenocyte proliferation. These findings are in part consistent with our previous study, which demonstrated that 2% static loading of myoblasts enhanced the proliferation of co‐cultured tenocytes, while 10% loading did not have this effect [14]. The previous study was conducted using a 2D myoblast‐tenocyte co‐culture system, where myoblasts were statically loaded for 24 h. In contrast, the current study used conditioned medium collected from myoblasts subjected to static loading for 24 h. This conditioned medium was then applied to tenocytes embedded in fibrin gel for 7 days, with medium changes every 2 days. The speculated increase in tenocyte proliferation observed in the 10% loading group in the current study may be attributed to the extended treatment duration, from 1 day in the previous co‐culture study to 7 days in the current experiment. Additionally, the difference between co‐culture and conditioned medium treatment could also contribute to the observed effects. In the co‐culture system, myoblasts continuously secrete factors that are received by the tenocytes, and vice versa, allowing for dynamic cellular interactions. In contrast, in the conditioned medium treatment, only myoblast‐derived secretory factors influenced the tenocytes, without reciprocal signaling. The use of a 3D model also allows for a different cell‐ECM dynamic, which is not seen when the cells are in 2D models. Furthermore, some short‐lived secretory factors that could impact tenocytes in co‐culture may not have been effective in the conditioned medium treatment. Nonetheless, under both 2D and 3D culture conditions, the secretome from myoblasts subjected to 2% loading demonstrated a pronounced effect on tenocyte proliferation and proliferation‐related indicators.

SCX expression serves as a highly specific marker for tenogenic precursor cells and is also present in mature differentiated tenocytes. Mice lacking Scx exhibit a severe and nearly complete loss of tendons during development [35]. In rodents, tendon injury triggers a continuous migration of progenitor cells from the paratenon to the injury site, where they begin expressing Scx at 7 days post‐injury [17, 36]. This is aligned with our findings, showing that the secretome from myoblasts subjected to 2% static loading increased the expression of SCX at Day 7. In contrast, this effect was less pronounced in the unloaded and 10% loading groups. Based on these findings, we speculate that the secretome from myoblasts subjected to 2% loading contains specific molecules that create an environment that induces tenocyte phenotype change, particularly facilitating the transition of tenogenic precursor cells within the tendon cell population into healing tenocytes. Both ACTA2 and S100A4 are linked to fibrosis in tendon healing when their expression is persistently elevated [23, 37, 38]. ACTA2 enhances the contractility of tendon cells and contributes to ECM remodeling, processes that are crucial in the early healing phase, while its persistent expression leads to excessive ECM deposition and fibrosis. Similarly, knockdown of S100A4 or inhibition of S100A4 signaling reduces scar formation [23]. In our study, we did not observe changes in ACTA2 or S100A4 expression across any groups or time points. Taken together, these findings suggest that secretome from myoblasts subjected to 2% static loading may promote the formation of SCX‐positive healing tenocytes without causing fibrosis.

Tendons are characterized by a highly organized, anisotropic ECM composed predominantly of type I collagen, with smaller amounts of type III collagen [5]. During the proliferative phase of tendon healing, the expressions of both type I and III collagens increase at the wound site to support tissue repair [36, 39]. Type III Collagen initially forms a collagenous bridge over the defect space within 2 weeks post‐injury, and this is primarily synthesized by SCX‐expressing cells [35]. Over time, the deposition of type III collagen fibrils is gradually replaced by newly synthesized type I collagen fibrils during the remodeling phase (months) [17]. Our results showed that the secretome from myoblasts subjected to 2% static loading increased both type I and III collagens production at Day 7, suggesting that this treatment could potentially accelerate tendon healing during the proliferative phase. Notably, our Western blot analysis primarily detects newly synthesized type I collagen, which may explain why we did not observe increased mature type I collagen levels at Day 14. It is possible that the newly formed type I collagen requires a longer time to fully align into organized tendon fibrils [40].

Our results on ECM are consistent with previous research demonstrating that the myoblast‐derived secretome enhances tendon cell activity. Specifically, a study using both 2D and 3D cell culture models showed that conditioned media from myoblasts increased tendon cell type I collagen expression and migration while also promoting structural growth in 3D‐engineered tendon constructs [41]. However, despite these structural benefits, the study noted that collagen density and mechanical properties were compromised, suggesting an imbalance in ECM remodeling. Our results build upon this by showing that myoblast‐secreted factors, particularly under 2% static loading, not only enhance type I and III collagens production but may also accelerate the transition from the proliferative to the remodeling phase of tendon healing. The observed decrease in type III collagen mRNA levels by Day 14, without a concurrent reduction in protein levels, suggests an earlier initiation of remodeling, potentially leading to a more efficient ECM maturation process. In mice, during natural tendon healing, type III collagen typically remains elevated until 3 weeks post‐injury [36]. Type III collagen can be gradually replaced by type I collagen by 4 weeks [17, 41]. In contrast, secretome from the 10% loading group did not induce the same dynamic changes in collagen composition, indicating that while myoblast‐derived secretome can influence tendon healing, the magnitude and nature of this effect depend on the mechanical loading conditions applied to the myoblasts. These findings reinforce the concept that muscle‐derived secretome plays a crucial role in tendon regeneration and highlight the potential of mechanically tuned myoblast secretome in optimizing tendon repair.

Studies have shown that fibronectin expression is closely associated with the migration of progenitor cells toward the injury site [42]. Fibronectin is expressed in the wound area within the first week post‐injury and remains detectable in the healing tendon for up to 1 month, but it gradually disappears after an additional 2 months [43]. Our result showed that the secretome from myoblasts subjected to 10% loading increased tenocyte fibronectin protein production at Day 7. While similar trends were observed in the secretome from unloaded myoblasts and those subjected to 2% loading, these differences were not significant. This suggests that myoblast secretome, particularly from 10% loading, may influence tendon healing by modulating fibronectin production, potentially enhancing cell migration during the early healing phase. Previously, we demonstrated that the secretome from 2% statically loaded myoblasts promotes migration via NBL1, an effect not observed in the 10% loading group. Given these findings, it would be interesting to further investigate the potential role of fibronectin in migration using a 3D model, to better understand the effects of secretome from different loading conditions.

Our results showed that by Day 14, although no significant differences were observed between groups, the fibrin gels treated with 2% and 10% myoblast secretome appeared slightly wider. This observation raises the possibility that these treatments may negatively impact the remodeling of the tendon constructs. However, this widening appearance may instead be due to the increased type I and III collagens deposition at earlier time points, with type III collagen, in particular, forming a more random and loosely organized structure, requiring more space. Given that tendon remodeling occurs over months to years in vivo, the increased gel width in these groups may reflect ongoing remodeling and could potentially indicate a stronger tendon, with the increased ECM. These findings are consistent with previous research showing that myoblast secretome enhanced collagen expression and increased the cross‐sectional area of 3D‐engineered tendon constructs. While that study reported a reduction in mechanical properties due to lower collagen density, the increased construct size suggests an expansion of ECM components and possibly an increase in water content, which may parallel our observations of wider fibrin gels [41]. With longer culture periods, we may observe further structural changes that provide deeper insights into tendon healing dynamics, including whether the initially expanded ECM undergoes maturation and compaction, leading to improved mechanical properties over time.

Studies have shown that complete removal of mechanical load disrupts the organization and structure of the healing tendon [44]. Therefore, tendon rehabilitation strategies emphasize the importance of mechanical tension in the healing process. Notably, treadmill training has been found to strongly induce Scx expression in fibroblasts, which appear to migrate from the epitenon into the superficial regions of tendon fascicles [45]. Additionally, TNMD, a gene highly expressed in human tendons, is rapidly downregulated when tendons are cultured under in vitro unloading conditions [46]. Beyond generating mechanical tension, muscle functions as a secretory organ, releasing bioactive molecules that exert autocrine, paracrine, and endocrine effects on various organs [47]. Our previous studies demonstrated that secretome from mechanically loaded myoblasts influences tenocyte behavior [14, 48], suggesting that tension alone is not the only factor influencing tendon function. In line with this, our current study further confirmed that secretome from 2% statically loaded myoblasts enhances SCX expression in tenocytes. This highlights a potential muscle‐tendon crosstalk mechanism, which could be further explored for tendon regeneration strategies.

While this study provides valuable insights into the effects of secretome derived from mechanically loaded myoblasts on tenocyte behavior, it is important to clarify the rationale and scope of the 3D fibrin tendon model used. This model was specifically selected to enable controlled investigation of mechanobiological and paracrine effects under well‐defined conditions, allowing investigation of matrix composition and tenocyte behavior. Such control enhances mechanistic tractability and facilitates the analysis of direct secretome‐mediated effects on tenocyte function, which would be difficult to isolate in more complex in vivo systems. Nevertheless, this model has limitations. First, our fibrin tendon gel model does not fully replicate the complexity of an in vivo tendon wound healing environment. Unlike natural tendon healing, which involves a dynamic interplay of inflammatory cells, vascularization, and mechanical loading within a physiological context, our in vitro model primarily focuses on tenocyte proliferation and early matrix remodeling. Consequently, the findings may not fully capture the later stages of tendon repair, such as long‐term matrix maturation and tissue integration. Additionally, our study was conducted over a two‐week culture period, which is sufficient to observe early cellular responses, including tenocyte proliferation and ECM production, but does not encompass the entire proliferation and remodeling phases of tendon healing, which extend over several weeks to months in vivo. To gain a more comprehensive understanding of how the secretome influences tendon healing, longer time points should be investigated to assess how the treatment impacts later stages of ECM organization, collagen cross‐linking, and overall tendon maturation.

Additionally, our experimental design has certain limitations. We pooled secretomes derived from four donors for treatment of the 3D tendon gels, which may obscure donor‐specific variability and potentially underestimate biological heterogeneity. However, pooling was necessary because the 3D tendon model requires up to 14 days of culture with medium replacement every other day, which demands substantial volumes of culture medium. Given the limited expansion capacity, passage restrictions, and donor availability of primary human myoblasts, generating donor‐specific conditioned medium for all experimental conditions and time‐points was not feasible without compromising experimental consistency.

Our previous 2D studies using rat‐derived cells demonstrated that static loading of myoblast at 2% intensity enhanced the secretion of IGF‐1 and NBL1, promoting tenocyte proliferation and migration. Building on these findings, the present study employed a 3D model using human primary cells and focused on parameters relevant to tendon healing, including cell phenotype, collagen production and markers indicative of proliferation. Under these more physiologically relevant conditions, we observed the most prominent effects with secretomes derived from myoblasts subjected to 2% mechanical loading. Although the current study was not designed to elucidate the underlying molecular mechanisms, this represents an important next step, and mechanistic analyses will be the focus of future work.

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: Junhong Li, Pernilla Eliasson and Ludvig J. Backman; data collection: Junhong Li, Christer Thomsen; analysis and interpretation of results: Junhong Li, Christer Thomsen, Paula Milena Giraldo‐Osorno, Pernilla Eliasson and Ludvig J. Backman; draft manuscript preparation: Junhong Li, Paula Milena Giraldo‐Osorno. All authors reviewed the results and approved the final version of the manuscript.

Funding

This work was financially supported by Swedish Research Council for Sport Science (P2024‐0001, P2025‐0011), Kempe foundation (JCSMK24‐00017), Swedish Research Council (2025‐02528) and Magnus Bergvall foundation (2024‐937).

Conflicts of Interest

The authors declare no conflicts of interest.

Li J., Giraldo‐Osorno P. M., Thomsen C., Eliasson P., and Backman L. J., “The Secretome From Mechanically Loaded Myoblasts Enhances Tenocyte‐Mediated Wound Healing in a 3D In Vitro Tendon Model,” The FASEB Journal 40, no. 6 (2026): e71658, 10.1096/fj.202503624RR.

Data Availability Statement

The data that support the findings of this study are available in the materials and methods and results of this article.