Tendon stem cells were isolated from Sprague-Dawley rats and cultured according to procedures described in our previous study (Zhang et al., 2020). Achilles tendons that were free of peritendinous tissues were harvested from euthanized rats. To reduce blood contamination, the samples were thoroughly washed with sterile phosphate-buffered saline (PBS) (Solarbio, Beijing, China). After digestion with trypsin (Beyotime, Shanghai, China), the specimens were cut into 1-mm³ blocks and were immersed in low-glucose Dulbecco’s Modified Eagle’s Medium (LG-DMEM) (Gibco, Grand Island, NY, United States) containing 3% w/v collagenase type I (Invitrogen, Carlsbad, CA, United States) and 4% w/v neutral proteinase (PeproTech, Rocky Hill, NJ, United States). The samples were incubated with gentle horizontal shaking for 1.5 h at 37°C, then passed through a 70-μm filter to obtain a single-cell suspension. After centrifugation, the cell pellet was resuspended in complete medium composed of LG-DMEM, fetal bovine serum (Biological Industries, Beit HaEmek, Israel) and penicillin plus streptomycin (Beyotime). The primary cells were cultured in T25 flasks in a humid environment with 5% CO₂ at 37°C. The medium was changed for the first time on day 5 and replaced every 2 or 3 days thereafter. When they reached 80% confluence, the attached cells were released by trypsinization, identified and expanded for use in experiments.

The following six TSC CM were prepared: CM1, LG-DMEM only; CM2, LG-DMEM with TSCs; CM3, LG-DMEM with 10 ng/ml HGF (PeproTech) + TSCs; CM4, LG-DMEM with 20 ng/ml HGF + TSCs; CM5, LG-DMEM with 40 ng/ml HGF + TSCs; and CM6, LG-DMEM with 80 ng/ml HGF + TSCs. The same number of TSCs were used in each preparation. After 24 h of culture, the supernatant from CM2–6 was collected and centrifuged, and the new supernatant was filtered to remove cell debris. The filtrate was divided into 2 equal parts. One portion was used in in vitro experiments and the other was concentrated using 3-kDa ultrafiltration centrifuge tubes (Millipore, Billerica, MA, United States) at 6000 × g for 1 h at 4°C for in vivo studies.

Based on cytokine antibody chip technology, the proteomic analyses of the soluble components of CM2–6 were conducted. All experimental procedures referred to the instruction of a Rat XL Cytokine Array Kit (R&D Systems, Minneapolis, MN, United States). A 1.5 ml solution containing 1.0 ml CM (i.e., the preparations described in section “Preparation of CM”) was dropped onto the surface of a blocked array chip. After overnight incubation at 4°C, the array chip was rinsed with washing buffer and then incubated with 1.5 ml diluted Detection Antibody Cocktail for 1 h at room temperature with gentle shaking. After washing, the array was incubated with horseradish peroxidase (HRP)–streptavidin for 30 min. Immunoreactive spots were detected with the Chemi Reagent Mix and the soluble secretory products of TSC were identified. By comparing the gray values of the corresponding spots on different array chips, the regulatory effect of HGF on the paracrine of TSC was verified. The experimental results were analyzed by cluster analysis and a protein–protein interaction (PPI) network was constructed using STRING v11.0¹. The protein pairs with parameters of interaction relationships greater than 0.7 were listed.

Male Sprague-Dawley rats weighing 160–180 g were purchased from the Laboratory Animal Center of Harbin Medical University. Animal experiments were carried out in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals, and were approved by the Ethics Committee of Harbin Medical University. Proper biosafety measurements were also adopted.

To establish the complete Achilles tendon rupture model, amputation was performed at the midpoint of the calcaneal insertion and musculotendinous junction with a sharp surgical knife. The injured tendon was reconstructed by double-cross suturing with a 4-0 Vicryl absorbable suture (Johnson & Johnson, New Brunswick, NJ, United States) under a microscope. The skin incision was closed and reinforced with biological tissue glue (B. Braun, Melsungen, Germany). The rats were freely fed with food and water after they recovered from anesthesia. After 2 weeks, the naturally healed tendons of six rats were harvested for the isolation of tendon fibroblasts and all the other 63 rats used for in vivo experiments were sacrificed for sample harvest. Fifty one of the 63 rats were treated with CM2-6 (Group A–E). The condensed CM (0.1 ml) was postoperatively administered by local injection three times per week for 14 days. The other 12 rats (Group F) were left untreated as a control. The specific animal amount for each experiment was preciously introduced below.

The harvested naturally healed tendons were rinsed with cold sterile PBS. After mincing the tissue into 1- to 2-mm³ pieces, an explant culture was established to obtain primary cells. The composition of the culture medium and schedule for medium changes were the same as described for TSC culture. Third-generation long fusiform cells were used for in vitro experiments.

Tendon fibroblasts (1 × 10⁴) were seeded in a 96-well plate. After overnight incubation, the culture medium was replaced with 100 μl of CM1–6. The effect of CM on cell viability was evaluated with the CCK-8 assay (Dojindo, Kumamoto, Japan) according to the manufacturer’s instructions. The absorbance at 450 nm was measured at five time points (2, 4, 8, 12, and 24 h). The experiment was performed in triplicate and the data were averaged. Results for groups treated with CM2–6 are presented relative to the values for the CM1-treated group.

As in our previous study (Han et al., 2019), tendon fibroblasts were cultured in 6-well plates. When the cells reached 100% confluence, a sterile 200-μl pipette tip and ruler were used to draw a smooth scratch mark across the bottom of each well. After washing off the scraped cells with sterile PBS, 2 ml CM was added to each well. Photos were obtained with an optical microscope (Olympus, Tokyo, Japan) at five time points (0, 12, 24, 48, and 60 h). The cell migration rate at each time point was calculated based on the corresponding acellular zone at 0 h.

Tendon fibroblasts were cultured with CM for 24 h, then lysed in precooled radioimmunoprecipitation assay buffer (Solarbio) supplemented with phenylmethylsulfonyl fluoride (Beyotime) for 30 min. Cell debris was cleared by centrifugation at 12,000 × g for 20 min at 4°C. The total protein concentration was determined with the BCA Protein Assay Reagent kit (Beyotime). The protein samples were diluted with 5 × sodium dodecyl sulfate loading buffer (Beyotime) and 20 μg protein per lane was electrophoretically separated on a 10% polyacrylamide gel (Beyotime). The proteins were transferred to polyvinylidene difluoride membranes and blocked with 5% non-fat dry milk (BD Biosciences, Franklin Lakes, NJ, United States) for 2 h at room temperature. Then, the membranes were separately probed overnight at 4°C with primary antibodies against collagen (COL)III (1:500), fibronectin (1:2000), matrix metalloproteinase (MMP)-1 (1:2000), MMP-9 (1:500) (all from Abcam, Cambridge, United Kingdom) and β-actin (1:1000) (WanleiBio, Shenyang, China). The membranes were rinsed with Tris-buffered saline (Solarbio) with 0.1% Tween-20 (Solarbio) and then incubated with the appropriate HRP-conjugated secondary antibodies (Boster Bio, Wuhan, China) for 1.5 h at room temperature. Protein bands were visualized using enhanced chemiluminescence substrate reagent (HaiGene, Harbin, China). The gray value was quantified with ImageJ software (NIH, Bethesda, MD, United States).

Extracellular matrix (ECM) organization at the tendon lesion site following CM treatment was evaluated by immunohistochemistry. Tendon specimens (1 cm long) harvested from 15 CM treated rats (n = 3) and 3 untreated ones were immediately immersed in 4% paraformaldehyde solution (BioSharp, Hefei, China). After dehydration through a graded alcohol series, samples were embedded in paraffin and serial sections of the tissue blocks were cut at a thickness of 5 μm parallel to the long axis of the tendons on a microtome (Leica, Wetzlar, Germany). After antigen retrieval, tissue sections mounted on glass slides were treated with 3% hydrogen peroxide to quench endogenous peroxidase activity, blocked with 10% goat serum (Solarbio) for 15 min at room temperature and incubated overnight at 4°C with primary antibodies against COLIII (1:100), fibronectin (1:100), α-smooth muscle actin (SMA) (1:100), MMP-2 (1:100), MMP-9 (1:100), tissue inhibitor of metalloproteinase (TIMP)-1 (1:100), and vascular endothelial growth factor (VEGF) (1:100) (all from Abcam). The slides were then incubated in the dark with species-specific secondary antibodies. Diaminobenzidine (Solarbio) was used to visualize the bound antibody complex. After dehydrating the sections and covering the slides with glass coverslips, the sections were examined and photographed under a microscope (Olympus) mounted with a camera. Positive signals were quantified with ImageJ software.

Tendon samples were harvested from 15 CM treated rats (n = 3). Normal tendons were also collected. After dehydration and embedding in paraffin, the samples were also cut into 5-mm thickness serial sections. Then, the slices were stained with hematoxylin and eosin (H&E) and Masson’s trichrome and observed under a light microscope. The organization of tendon fibers was evaluated according to a previously described parallel fiber alignment scoring method (Jiang et al., 2016) by investigators who were blinded to group assignment. The scoring scale was as follows: 0, 0–25% parallel fiber alignment; 1, 25–50%; 2, 50–75%; and 3, 75%–100%.

The expression of the pro-inflammatory cytokine interleukin (IL)-6 and anti-inflammatory cytokine IL-10 was detected by immunofluorescence analysis. The sample collection was same as the description in section “Immunohistochemistry analysis”. Tendon tissue samples were embedded in optimal cutting temperature compound and frozen serial sections were cut at a thickness of 10 μm. After antigen retrieval, the sections were blocked with 10% goat serum for 15 min at room temperature, then incubated overnight at 4°C with primary antibodies against IL-6 (1:100) and IL-10 (1:100) (both from Abcam), followed by Cy3-conjugated goat anti-rabbit IgG (Beyotime) for 60 min at room temperature. Nuclei were stained with DAPI (Beyotime). Images were acquired under a fluorescence microscope (Olympus), and the area of positive signal was determined using ImageJ software.

We evaluated the effectiveness of the CM with the greatest therapeutic effect in all the above experiments [sections “Cell Counting Kit (CCK)-8,” “Wound healing assay,” “Western blotting,” “Immunohistochemistry analysis,” “Histological analysis,” and “Immunofluorescence analysis”] in promoting functional recovery in the rat model of Achilles tendon rupture. Eighteen tendon-calcaneal complexes were carefully dissected from six CM treated rats, six naturally healed rats, and six intact animals (i.e., without injury). They were used for biomechanical testing as previously described (Chamberlain et al., 2019). Briefly, the transverse diameter (TD) and anteroposterior diameter (AD) of the healed sections were measured with electronic vernier calipers. The tissue sample’s cross-sectional area (S) was approximated as an oval, and was calculated with the equation S = πTD × AD/4. A universal tensile testing machine (Hengruijin, Jinan, China) was used for testing. The proximal muscle wrapped in a piece of 100 × sandpaper and the distal naked paw were affixed to the two testing clamps. An axial 0.1 N preload was applied along the long axis of the tendon, with elongation at a speed of 5 mm/min. The maximum tensile load was recorded. Stiffness and Young’s modulus were calculated based on the linear slope of the strain–stress curve.

Data are presented as mean ± SD. Differences between groups were evaluated by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test if necessary or by two-way ANOVA using Prism v8.3.0 software (GraphPad, La Jolla, CA, United States). P < 0.05 was considered statistically significant.

Primary TSCs were spindle-shaped, reaching 40–50% confluence on Day 10 after seeding (Figure 1A). The parts stained by crystal violet were the cell colonies (Figure 1B). Flow cytometry analysis revealed that 99.7% of cells expressed the cell surface marker cluster of differentiation (CD) 90 and 99.1% expressed CD44. Meanwhile, only 3.9% and 7.4% expressed CD11 and CD106, respectively (Figure 1C). TSCs had multi-differentiation potential, as determined by the detection of lipid droplets (Figure 1D), mineralized nodules (Figure 1E), and acid mucopolysaccharide (Figure 1F).

A total of 12 molecules were detected in the CM2 with the proteome profiler antibody array. HGF contained in the CM3–6 were depleted from the culture medium after 24 h of culture. As a result, it caused variations in the concentrations of CM components but no generation of new products (Figure 2A). Almost all the secretion of these bioactive molecules can be inhibited by a high concentration of HGF (i.e., CM6 with 80 ng/ml HGF). Compared with CST3, CCL2, Serpin E1, VEGFa, MMP-2, and WISP-1, the concentration variations of LGALS3, NOV, and TNFRSF11b were sensitive to the dose of HGF. While, only when the concentration of HGF was 40 mg/ml, the contents of IGFRP-6, IGFBP-3, and SPP1 changed obviously (Figure 2B). PPIs of the 12 molecules are shown in Figure 2C. HGF interacted with 6 proteins. SPP1, CCL2, and MMP-2 interacted with eight proteins. In maximum, there were 10 PPIs for VEGFa. WISP-1 only interacted with MMP-2 but not with other proteins; the same was true for IGFBP-6 and IGFBP-3, NOV and VEGFa. The strength of the interactions are shown in Figure 2D. LGALS3, IGFBP-3, IGFBP-6, NOV, and WISP-1 had a low degree of interaction (<0.7), whereas HGF and VEGFa, VEGFa and Serpin E1, HGF and Serpin E1, MMP-2 and VEGFa, CST3 and SPP1 showed strong interactions (>0.9).

Conditioned medium treatment had no effect on the viability of tendon fibroblasts at 2, 4, 8, and 12 h, but the viability was doubled at 24 h. No statistical significance was observed among CM2-6 (Figure 3A). Likewise, CM accelerated the migration of tendon fibroblasts (Figures 3B,C). The effect showed a dose-dependent relationship with the concentration of HGF. Obviously, the wound closure of the fibroblasts with CM1-5 increased gradually. The maximum rate of migration was observed with CM5 (64.48% ± 0.80% wound closure after 60 h). When the concentration of HGF reached at 80 ng/ml (i.e., CM6), no further increase of wound closure was observed. In fact, it yielded a result similar to that obtained with CM4.

In vitro experiment, MMP-2 and MMP-9 were upregulated following the application of CM (Figures 4A,D,E). The expression of MMP-2 was highest in tendon fibroblasts treated with CM5 and CM6 compared to the other CM formulations, with no difference between the two groups. Similar results were obtained for MMP-9, which showed the highest expression in the CM5-treated group. With the increase of HGF concentration, their expressions of COLIII and fibronectin decreased (Figures 4A–C). All the above results, upregulated MMPs and downregulated ECM components, were verified in vivo experiment by immunohistochemistry (Figures 5A,B,D,E). Although theirs levels were lower than that of Group F, the expression of α-SMA was increased among experimental groups treated with CM2-6 (Figure 5C). Meanwhile, the expression of TIMP-1 first decreased and then increased with increasing concentrations of HGF (Figure 5F). The turning point came at Group D. VEGF, which promotes angiogenesis, was upregulated in the presence of CM, with the highest level detected in tissues treated with CM6 (i.e., Group E) (Figure 5G).

The effect of CM on tendon fiber alignment was evaluated by histological analysis. Achilles tendon tissue sections treated with CM5 and CM6 (i.e., Group D and Group E) showed a much more regular and compact alignment of collagen fibers compared to those treated with CM2 and CM3 (i.e., Group A and Group B), which had disorganized fibers (Figure 6A). Samples treated with CM4 (i.e., Group C) had an intermediate phenotype. These results were also confirmed by the fiber alignment score, which was higher for samples treated with CM5 and CM6 than for the other three groups (Figure 6B).

IL-6 level was highest in rats with Achilles tendon injury that were left untreated (Group F) and decreased with CM treatment (Figure 7A). The capacities of inhibiting pro-inflammatory cytokine expression in Group A and Group B (i.e., treatments with CM2 and CM3) were weaker than those in Group C, Group D, and Group E. However, the difference between Group C and Group D was not statistically significant (Figure 7B). The opposite trend was observed for IL-10, which showed the lowest expression in Group F (Figure 7C). With the treatment of CM, anti-inflammatory cytokine expressions increased. The effect was greater at higher concentration of HGF. It maximized in Group D. When it came to Group E, no further increase of anti-inflammatory cytokine expression was observed. In contrast, it dropped to the same value as that in Group B (Figure 7D).

Based on the above results, tendons from Group D (i.e., treatment with CM5) were used for biomechanical testing. Healed tendon from Group D and F had a cross-sectional area nearly three times that of uninjured tissue (Figures 8A,B). Although there was functional improvement after CM treatment, the maximum tensile load was still lower than normal tendon (Figure 8C), corresponding to 60% recovery. What was worse, naturally healed tendons in Group F only recovered 30% of the normal maximum tensile load. The stiffness and Young’s modulus of samples in Group D and Group F were still far less than that of normal tissues. There were also no differences between CM treated tendons and naturally healed ones (Figures 8D,E).