Adipose stem cell-derived exosomes exert a synergic anti-inammatory effect with glucocorticoids and override their detrimental effects on rotator cuff tendons

Background: Glucocorticoids (GCs) injections are commonly used to relieve pain and improve function in patients with multiple shoulder disability, they cause detrimental effects on the rotator cuff tendons. Adipose stem cell-derived exosomes (ASC-Exos) reportedly recover impaired tendon matrix metabolism by maintaining tissue homeostasis. It is unclear whether additional ASC-Exos treatment overrides the detrimental effects of GCs without interfering with their anti-inammatory effects. Methods: The in vitro studies included inammatory analysis and cytoprotective analysis. In the inammatory analysis, rat raw cells were treated with saline, GCs, or GCs + ASC-Exos and evaluated regarding cellular proliferation, migration, and secretion of inammatory-related cytokines. In the cytoprotective analysis, rat tenocytes were treated with saline, GCs, or GCs + ASC-Exos and evaluated regarding cellular proliferation, migration, senescence, apoptosis, and transcription of tenocytic genes. In the in vivo studies, a subacromial injection of saline, GCs, or GCs + ASC-Exos was performed in a chronic injured-intact rotator cuff rat model. Histological and biomechanical analysis were performed 1 week to evaluate the protective effect of ASC-Exos against GCs-induced detriments on injured-intact in rotator cuffs. Results: In the in vitro inammatory analysis, GCs treatment signicantly decreased the proliferation, migration, and secretion of pro-inammatory cytokines in raw cells, and increased the secretion of anti-inammatory cytokines; additional ASC-Exos treatment further signicantly decreased the secretion of pro-inammatory cytokines and increased the secretion of anti-inammatory cytokines, while restoring GCs-suppressed cellular proliferation and migration. In the in vitro cytoprotective analysis, GCs treatment signicantly decreased the proliferation, migration, and transcription of tenocytic matrix molecules of tenocytes, and signicantly increased their senescence, apoptosis, and transcription of ROS and tenocytic degradative enzymes; additional ASC-Exos treatment signicantly improved the GCs-suppressed cellular proliferation, migration, and transcription of tenocytic matrix molecules, transcription of tenocytic degradative enzyme inhibitors, and signicantly decreased the GCs-induced cell senescence, apoptosis, and transcription of ROS and tenocytic degradative enzymes. In the in vivo studies, an additional ASC-Exos injection restored the impaired histological and biomechanical properties owing to GCs

Single or repeated GCs injections reportedly signi cantly reduce shoulder pain and improve shoulder range of motion in athletes with chronic pain [8]. The underlying mechanism is due to the antiin ammatory effect of GCs [9,10]. Shoulder injury is followed by a massive in ltration of in ammatory cells and the secretion of several upstream in ammatory cytokines such as TNF-α and IL-1, which triggers a pro-in ammatory reaction and ultimately induces the formation of pain-related neuropeptides [11][12][13][14]. GCs achieve their therapeutic effect by inhibiting the in ltration of in ammatory cells and the secretion of upstream pro-in ammatory factors [9,10,14].
Although GCs have potent anti-in ammatory and analgesic effects, they also have important adverse effects. There is increasing concern that GCs injections may accelerate tendon degeneration [2,6,15].
Tendon rupture after GCs injections in both the upper and lower extremities is common in clinical practice [16,17]. Although it is unclear to what extent such ruptures are attributable to GCs injections or the progression of tendinopathy itself, several studies have con rmed that GC injections are correlated with tendon detriments [2,6,15]. Haraldsson [2,15]. The adverse effects of GCs injections on the tendon are believed to be due to the negative effect of GCs on tendon metabolism homeostasis in combination with decreased cell proliferation and increased cell senescence, apoptosis, and destruction of normal collagen organization [2,6,7,15,18]. Therefore, concomitant treatment with a bioactive substance that overrides the detrimental effects of GCs on cellular metabolism homeostasis may be a practical strategy to minimize the adverse effects of GCs. However, despite considerable research, there is still no e cacious means with which to eliminate the undesirable effects of GCs while retaining their antiin ammatory effects.
Exosomes are nano-sized membrane vesicles that contain proteins, lipids, and nucleic acid and play an important role in the regulation of cellular physiological activities [19][20][21][22][23][24]. They have attracted increasing interest and have been widely studied in tissue-engineering elds in recent years [19][20][21][22][23][24]. In our previous studies, we have found that adipose stem cell-derived exosomes (ASC-Exos) were able to recover the impaired matrix metabolism of torn rotator cuff tendons by maintaining tissue homeostasis. Drawing on this, we infer additional treatment with ASC-Exos may override the detrimental effects of GCs on tenocytes without interfering with their anti-in ammatory effects.
The present study aimed to: 1) investigate the anti-in ammatory and cytoprotective effects of ASC-Exos on raw cells and tenocytes, respectively, with the concomitant treatment of GCs in vitro, and 2) evaluate the effect of ASC-Exos on injured-intact tendons with concomitant GCs injection in vivo using a rat model of chronic rotator cuff tear. Our hypothesis was that the additional ASC-Exos treatment would exert a synergic anti-in ammatory effect with the GCs and a cytoprotective effect against their adverse effects and would improve both the histological and biomechanical properties of injured-intact rotator cuffs simultaneously injected with GCs in vivo.

Isolation and characterization of human ASC-Exos
ASCs isolation and culture, and ASC-Exo isolation were performed as described in previous studies [25,26]. The morphology of ASC-Exos was observed using transmission electron microscopy (TEM; H-7650; Hitachi High-Technologies Corporation). The particle size distribution was evaluated using nanoparticle tracking analysis (NTA; ZetaView PMX 120, Particle Metrix, Meerbusch, Germany). The expression of exosomal surface markers, including CD9, CD63, and TSG-101, was detected using western-blotting.

In vitro studies
In ammation analysis Interventions and study groups Raw cells from one of the rat macrophage cell lines were used for in ammation analysis in the present study. After incubation in complete Dulbecco's Modi ed Eagle's Medium supplemented with 10% fetal bovine serum (FBS) and antibiotic solution for 24 hours to allow attachment, cells were treated with either saline (control group), 1 µM of dexamethasone (GCs group), or 1 µM of dexamethasone + 10 11 pellets/ml of ASC-Exos (GCs + ASC-Exos group), and further evaluated.

Cell proliferation
After being seeded in 96-well plates at a density of 1 × 10 3 cells/well, cells were incubated for 24 hours to allow attachment, and treated with either saline, dexamethasone, or dexamethasone + ASC-Exos. Cell proliferation was evaluated at 12, 24, and 48 hours using a CCK8 assay kit in accordance with the manufacturer's instructions (Dojindo Molecular Technologies, Inc., Kumamoto, Japan). Brie y, at 12, 24, and 48 hours, CCK8 solution was added to each well and incubated for a further 1 hour at 37°C. The absorbance of wells was measured at 450 nm using a spectrophotometric microplate reader (Bio-Rad 680, Bio-Rad, Hercules, CA, USA). The proliferation intensity of the cells was calculated as the optical density of the tested well minus the absorbance of the blank well.
After being seeded in the upper chamber of 96-well transwell plates at a density of 1 × 10 6 cells/well, cells were incubated for 24 hours to allow attachment. After this, saline, dexamethasone, or dexamethasone + ASC-Exos was added to the lower chamber. After incubation for another 24 hours, cell migration was evaluated in accordance with the manufacturer's instructions (Corning, Corning, NY, USA). Brie y, cells from the upper surface of the lter membranes were removed with a cotton swab. Cells that migrated to the lower surface of the lter membrane were xed in 4% formaldehyde for 5 minutes, and then stained with 0.5% crystal violet for 5 minutes. Migration activity was evaluated by observing the stained cells under an optical microscope.

Enzyme linked immunosorbent assay (ELISA)
After being seeded in 96-well plates at a density of 1 × 10 5 cells/well, cells were incubated for 24 hours to allow attachment and treated with saline, dexamethasone, and dexamethasone + ASC-Exos, respectively. The levels of secreted in ammation-related cytokines, including TNF-α, IL-1α, IL-1β, IL-4, and IL-10, were evaluated using an enzyme-linked immunosorbent assay kit in accordance with the manufacturer's instructions (R&D Systems). Brie y, after incubation for another 96 hours, the culture medium was centrifugated to obtain the supernatant. The supernatant was then incubated with speci c antibodies for 2 hours at 37°C in a 96-well plate. After washing, each well was incubated with the corresponding conjugate for 1 hour, and substrate solution was added to develop the color in the dark. The absorbance at 450 nm was determined using a microplate reader (Multiskan MK3; ThermoFisher Scienti c, USA).

Cytoprotective analysis
Interventions and study groups Rat tenocytes isolated from adult rats were used in the present study (Appendix). After incubation in complete Dulbecco's Modi ed Eagle's Medium supplemented with 10% fetal bovine serum (FBS) and antibiotic solution for 24 hours to allow attachment, cells were treated with either saline (control group), 1 µM of dexamethasone (GCs group), or 1 µM of dexamethasone + 10 11 pellets/ml of ASC-Exos (GCs + ASC-Exos group), and further evaluated.

Cell proliferation
After being seeded in 96-well plates at a density of 1 × 10 3 cells/well, cells were incubated for 24 hours to allow attachment and then treated with either saline, dexamethasone, or dexamethasone + ASC-Exos. Cell proliferation was evaluated at 12, 24, and 72 hours using a CCK8 assay kit as described above.
After being seeded in the upper chamber of 96-well transwell plates at a density of 1 × 10 4 cells/well, cells were incubated for 24 hours to allow attachment. Saline, dexamethasone, or dexamethasone + ASC-Exos was then added to the lower chamber. Cell migration was evaluated as described above.

Cell senescence
After being seeded in 6-well plates at a density of 1 × 10 4 cells/well, cells were incubated for 24 hours to allow attachment and then treated with saline, dexamethasone, or dexamethasone + ASC-Exos. The senescence assay was performed using a senescence β-galactosidase staining kit in accordance with the manufacturer's instructions (Solarbio Science and Technology, Beijing, China). Brie y, after incubation for another 24 hours, 1 ml/well of senescence β-galactosidase staining solution was added to each well and incubated for 15 minutes at 37°C. Senescent cells were stained with prepared dye and xed at 37°C overnight. Cell senescence was evaluated by calculating the intensity of the positively stained cells under an optical microscope using ImageJ software (National Institutes of Health).

Cell apoptosis
After being seeded in 96-well plates at a density of 1 × 10 4 cells/well, cells were incubated for 24 hours to allow attachment before being treated with saline, dexamethasone, or dexamethasone + ASC-Exos. The apoptotic assay was performed using a TUNEL apoptosis assay kit in accordance with the manufacturer's instructions (Solarbio Science and Technology, Beijing, China). Brie y, after incubation for another 24 hours, 50 µL/well of TUNEL solution was added to each well and incubated for a further 1 hour at 37°C. Apoptotic cells were stained with reaction buffer, and the uorescence intensity was evaluated using a ow cytometer with a FL3 channel.

Real-time polymerase chain reaction (Rt-PCR)
After being seeded in 6-well plates at a density of 1 × 10 6 cells/well, cells were incubated for 24 hours to allow attachment. Cells were then treated with saline, dexamethasone, or dexamethasone + ASC-Exos and incubated for a further 72 hours. The relative mRNA transcription levels of ROS, MMP-2, MMP-9, MMP-13, TIMP-1, TIMP-3, decorin, biglycan, and type I and type III collagen were evaluated as previously described [26]. Brie y, TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract the total RNA from cells. Reverse transcription was performed using script reverse transcription supermix (Bio-Rad Laboratories, Hercules, CA, USA) to reverse-transcribe the RNA into cDNA. Real-time polymerase chain reaction was conducted using the CFX Connect Real-Time PCR system (Bio-Rad) with Universal SYBR Green Supermix (Bio-Rad). The relative mRNA transcription levels of ROS, MMP-2, MMP-9, MMP-13, TIMP-1, TIMP-3, decorin, biglycan, and type I and type III collagen were calculated using the 2-ΔΔCt method. GAPDH was used as a housekeeping gene to normalize the expression of genes of interest in the present study.

In vivo studies
Power analysis and study groups The animal experimental protocol was approved by the Animal Ethical and Welfare Department of Shanghai Jiaotong University.
A power analysis was performed with α = 0.05, 1 -β = 0.8, and an assumed dropout rate of 25% to calculate the sample size required for the animal experiments. A sample size of eight shoulders was required to detect a signi cant difference in the biomechanical analysis of the ultimate load to failure. In addition, we allocated four shoulders for histological analysis in each group. Thus, a total of 36 shoulders (18 male Sprague Dawley rats aged 12 weeks) were included in the present study and divided into the following three groups: saline injection group (control group), dexamethasone injection group (GCs group), and dexamethasone + ASC-Exos injection group (GCs + ASC-Exos group). Each shoulder was randomly chosen to receive an injection of either saline, dexamethasone, or dexamethasone + ASC-Exos.

Establishment of a chronic rotator cuff injured-intact model and interventions
After the induction of general anesthesia, a 1-cm longitudinal skin incision was made over the lateral shoulder joint under sterile conditions. The supra-infraspinatus tendon was exposed by splitting the deltoid muscle. The rotator cuff injured-intact model was established by completely cutting the supraspinatus tendon from the greater tubercle to create a full-thickness injury while leaving the infraspinatus tendon intact; this was left for 4 weeks to establish a chronic injured-intact model that simulated the clinical setting. The muscle, subcutaneous tissue, and skin were closed as separate layers.
After 6 weeks, a single dose of saline, dexamethasone (0.1 mg/kg), or dexamethasone (0.1 mg/kg) + ASC-Exos (10 11 pellets/ml) of the same volume was injected into the subacromial space in each shoulder. At 1 week after injection, all rats were euthanized by CO 2 inhalation. The supra-infraspinatushumerus complex was carefully harvested. Histological and biomechanical analysis were performed after macroscopic evaluation of the harvested sample.

Histological analysis
Samples were xed in 4% buffered paraformaldehyde for 24 hours and embedded in para n. Consecutive 3-μm-thick sections were cut parallel to the long axis of the supra-infraspinatus tendon in the coronal plane.
To evaluate the general appearance, sections were stained with hematoxylin and eosin (H&E). To evaluate collagen degenerations, sections were stained with type I and III collagen antibodies. To evaluate fatty in ltration, sections were stained with oil red O.
Images of the entire section of the supra-infraspinatus were acquired by digital slide scanning (Pannoramic MIDI; 3DHISTECH Ltd.). The intensities of type I and type III collagen and fatty in ltration were quantitatively measured using ImageJ software (National Institutes of Health) as previously described [25].
Semi-quantitative analysis was also performed as previously described to comprehensively evaluate the histological properties of the tendon ( To x the isolated supraspinatus, the muscular end was gripped in a serrated clamp to prevent slippage, and the tendinous end was cross-sutured using No. 2 Orthocord suture (DePuy Mitek) and then tied to the sensor of the device. To x the infraspinatus-humerus complex, the humerus was mounted in a cylindrical holder using six interference screws, and the muscular end was gripped in a serrated clamp connected to the sensor system. The musculotendinous part of the infraspinatus was placed at an angle of approximately 90° to the humerus to mimic the anatomic direction of tensile force.
After preloading for 5 minutes using a 0.5 N tensile force, the sample received 10 cycles of loading ranging from 0.5 N to 2 N to minimize the viscoelastic effects. Immediately after preconditioning, the load-to-failure test started with uniaxial tension at 1 mm/minute. The ultimate load to failure was de ned as the rst signi cant decrease in the load-displacement curve, and the stress was calculated by dividing the ultimate load to failure by the initial cross-sectional area. The mode of failure was also recorded.

Statistical analysis
All data are expressed as mean ± SD. Statistical analysis was performed using SPSS software (version 15, SPSS). One-way analysis of variance with post hoc testing using the Bonferroni method was carried out. Signi cance was set at P < 0.05.

Characterization of ASC-Exos
The TEM and NTA revealed spherical vesicles with a particle size distribution of 50 to 150 nm. Westernblotting analysis showed that these vesicles had strong surface expression of exosomal markers, including CD9, CD63, and TSG-101 ( Figure 1).

In ammation analysis
Cell proliferation Dexamethasone treatment signi cantly decreased the proliferation intensity of rat raw cells to 0.82-, 0.89-, and 0.88-fold at 12, 24, and 48 hours, respectively, compared with the control group.
To our surprise, additional treatment with ASC-Exos did not further decrease the proliferation intensity, but overrode the detrimental effect of dexamethasone on the proliferation of raw cells, with the GCs + ASC-Exos group showing no signi cant differences in proliferation intensity compared with the control group ( Figure 2A).
Dexamethasone treatment signi cantly decreased the migration of rat raw cells to 0.69-fold at 24 hours, compared with the control group.
Consistent with the cell proliferation results, additional treatment with ASC-Exos did not further decrease cell migration, but overrode the detrimental effect of dexamethasone on the migration of raw cells, with the GCs + ASC-Exos group showing no signi cant differences in migration compared with the control group ( Figure 2B, 2C).
In contrast to the cell proliferation and migration results, additional treatment with ASC-Exos further signi cantly downregulated the secretion of TNF-α, IL-1α, and IL-1β to 0.54-, 0.76-, and 0.33-fold, respectively, compared with the control group.
Signi cant differences were also detected between the GCs group and the GCs + ASC-Exos group, indicating that additional treatment with ASC-Exos might exert a stronger anti-in ammatory effect than dexamethasone alone ( Figure 3A, 3B, 3C).

Anti-in ammatory cytokines
Dexamethasone treatment signi cantly upregulated the secretion of IL-4 to 1.39-fold, and signi cantly downregulated the secretion of IL-10 to 0.81-fold, respectively, compared with the control group.
Additional treatment with ASC-Exos further signi cantly upregulated the secretion of IL-4 and IL-10 to 3.68-and 2.31-fold, respectively, compared with the control group.
Consistent with the pro-in ammatory cytokine results, signi cant differences were also detected between the GCs group and the GCs + ASC-Exos group, indicating that additional treatment with ASC-Exos might exert a stronger anti-in ammatory effect than dexamethasone alone ( Figure 3D, 3E).

Cell proliferation
Dexamethasone treatment signi cantly decreased the proliferation intensity of rat tenocytes to 0.74-, 0.91-, and 0.70-fold at 12, 24, and 48 hours, respectively, compared with the control group.
Additional treatment with ASC-Exos not only overrode the detrimental effect of dexamethasone on the proliferation of rat tenocytes, but further signi cantly promoted their proliferation intensity to 1.17-, 1.22-, and 1.13-fold at 12, 24, and 48 hours, respectively, compared with the control group ( Figure 4A).

Cell migration
Dexamethasone treatment signi cantly decreased the migration of rat tenocytes to 0.55-fold at 24 hours, compared with the control group.
Additional treatment with ASC-Exos not only overrode the detrimental effect of dexamethasone on the migration of rat tenocytes, but further signi cantly increased their migration to 1.13-fold at 24 hours, compared with the control group ( Figure 4B, 4C).

Cell senescence
Dexamethasone treatment signi cantly increased the senescence intensity of rat tenocytes to 1.92-fold at 24 hours, compared with the control group.
Additional treatment with ASC-Exos overrode the detrimental effect of dexamethasone on the senescence intensity of rat tenocytes, as shown by the lack of a signi cant difference in the senescence intensity between the GCs + ASC-Exos group and the control group ( Figure 5A, 5B).

Cell apoptosis
Dexamethasone treatment signi cantly increased the apoptosis percentage of rat tenocytes to 1.34-fold at 24 hours, compared with the control group.
Additional treatment with ASC-Exos overrode the detrimental effect of dexamethasone on the apoptosis percentage of rat tenocytes, as shown by the lack of a signi cant difference in the apoptosis percentage between the GCs + ASC-Exos group and the control group ( Figure 5C, 5D).

Transcription of ROS
Dexamethasone treatment signi cantly upregulated the transcription of ROS of rat tenocytes to 2.40-fold of rat tenocytes at 72 hours, compared with the control group.
Additional treatment with ASC-Exos not only overrode the effect of dexamethasone on the ROS transcription by rat tenocytes, but further signi cantly decreased their transcription to 0.41-fold at 72 hours, compared with the control group ( Figure 6A).

Transcription of degradative enzymes and their inhibitors
Dexamethasone treatment signi cantly upregulated the transcription of MMP-2, MMP-9, and MMP-13 by rat tenocytes to 1.79-, 2.33-, and 3.62-fold, respectively, compared with the control group. Additional treatment with ASC-Exos overrode the effect of dexamethasone on the transcription of degradative enzymes of rat tenocytes, as shown by the lack of signi cant difference between the GCs + ASC-Exos group and the control group. Dexamethasone treatment signi cantly downregulated the transcription of type I collagen by rat tenocytes to 0.27-fold at 72 hours compared with the control group. Transcription of type III collagen was also downregulated in the GCs group to 0.78-fold compared with the control group, but this difference did not reach statistical signi cance. Additional treatment with ASC-Exos not only overrode the effect of dexamethasone on the transcription of type I and type III collagen by rat tenocytes, but further signi cantly increased their transcription to 4.90-and 3.23-fold at 72 hours, respectively, compared with the control group ( Figure 7C, 7D).
Dexamethasone treatment also signi cantly downregulated the type I/III transcription ratio to 0.34-fold compared with the control group. Additional treatment with ASC-Exos not only overrode the effect of dexamethasone on the type I/III transcription ratio of rat tenocytes, but further signi cantly increased their transcription ratio to 1.52-fold at 72 hours compared with the control group ( Figure 7E).

Macroscopic evaluation of the supra-infraspinatus-humerus complex
The torn supraspinatus tendon was connected to the humerus through brous tissue, and the connection between the infraspinatus tendon and the humerus remained complete. There was no evidence of infection.

Histological analysis
Injured supraspinatus tendon Generally, increased cellularity, vascularity, thinning, separation, and disorganization of collagen bers, and fatty in ltration were observed in all three groups. These signs were most signi cant in the GCs group ( Figure 8).
Quantitatively, the GCs group exhibited signi cantly higher fatty in ltration, and severe collagen degeneration (signi cantly lower type I collagen and higher type III collagen density) than both the control and GCs + ASC-Exos group. The immunohistochemical variables did not signi cantly differ between the control and GCs + ASC-Exos groups ( Figure 9A, 9B, 9C).
Semi-quantitative analysis results showed that the histological properties in the GCs group were signi cantly worse than those in the control and GCs + ASC-Exos groups. There were no signi cant differences between the control and GCs + ASC-Exos groups ( Figure 9D).

Intact infraspinatus tendon
Generally, slightly increased cellularity, vascularity, thinning, separation, and disorganization of collagen bers were observed in all three groups. These signs were relatively most signi cant in the GCs group.
There was no evidence of fatty in ltration in any of the three groups ( Figure 10).
Quantitatively, the GCs group exhibited severe collagen degeneration (signi cantly lower type I collagen and higher type III collagen density) than both the control and GCs + ASC-Exos groups. There were no signi cant differences between the control and GCs + ASC-Exos groups with regard to collagen degeneration ( Figure 11A, 11B).
Semi-quantitative analysis results showed that the histological properties in the GCs group were signi cantly worse than those in the control and GCs + ASC-Exos groups. There were no signi cant differences between the control and GCs+ ASC-Exos groups ( Figure 11C).

Injured supraspinatus tendon
The cross-sectional area of the supraspinatus tendon did not signi cantly differ between the three groups ( Figure 12A).
The GCs injection signi cantly decreased the ultimate load to failure and the ultimate stress to failure to 0.66-and 0.71-fold, respectively, compared with the control group ( Figure 12B, 12C). Concomitant injection of ASC-Exos and GCs overrode the GCs-induced detrimental effects on the biomechanical properties of the injured supraspinatus tendon, as shown by the lack of signi cant differences between the GCs + ASC-Exos group and the control group in the ultimate load to failure and the ultimate stress to failure ( Figure 12B, 12C).
All specimens failed at the middle of the tendon.

Intact infraspinatus tendon
The cross-sectional area of the intact infraspinatus tendon did not signi cantly differ between the three groups ( Figure 13A).
The GCs injection signi cantly decreased the ultimate load to failure and the ultimate stress to failure to 0.64-and 0.74-fold, respectively, compared with the control group ( Figure 13B, 13C). Concomitant injection of ASC-Exos and GCs overrode the GCs-induced detrimental effects on the biomechanical properties of the intact infraspinatus tendon, as shown by the lack of signi cant differences between the GCs + ASC-Exos group and the control group in the ultimate load to failure and the ultimate stress to failure ( Figure 13B, 13C).
All specimens failed at the tendon-to-bone interface.

Discussion
The most important nding of the present study was that additional treatment with ASC-Exos not only facilitated the anti-in ammatory effect of GCs, but overrode the negative effect of GCs on tenocytes,

alleviating the detrimental effect of a GCs injection on an injured-intact rotator cuff. The in vitro results
showed that additional treatment with ASC-Exos and GCs exerted a synergic anti-in ammatory effect mainly by suppressing the secretion of pro-in ammatory cytokines and increasing the secretion of antiin ammatory cytokines from raw cells; the cytoprotective effect of ASC-Exos on tenocytes was achieved mainly by suppressing the GCs-induced transcription of ROS and MMPs, and increasing the transcription of TIMPs and GCs-suppressed tenocytic matrix molecules (including decorin and biglycan) and the type I/III collagen ratio. The analysis of the rat model of chronic rotator cuff tear further showed that the detriments on both the histological and mechanical properties of the injured-intact rotator cuff induced by a single GCs injection at 1 week were restored by an additional ASC-Exos injection.
Subacromial GCs injections are commonly performed to treat multiple shoulder pathologic conditions to reduce pain and improve range of motion [3][4][5]8]. However, a major concern limiting the further use of such GCs injection is the adverse effects that accelerate tendon degeneration [2,6,7,15,18]. Nowadays, it is commonly accepted that subacromial GCs injections lead to a signi cant decrease in the mechanical properties of both injured and normal rotator cuff tendons [2,6,15]. Inspired by the bioactive effect of platelet-rich plasma (PRP), an autologous blood plasma concentrate that contains various growth factors, Jo et al. investigated the effect of concomitant treatment with PRP and GCs on tenocytes in a degenerative condition induced by IL-1β [7]. Their results showed that PRP produced a synergic antiin ammatory effect with GCs and antagonized their adverse effects [7]. In vitro evidence from other studies has also showed that some bioactive factors have a cytoprotective effect on tenocytes against GCs-induced cellular senescence and apoptosis [27][28][29]. However, these previous studies only evaluated the anti-in ammatory effect with or without cytoprotective effects in vitro rather than in vivo.
Macrophages play a key role during the in ammatory in ltration process after initial rotator cuff injury to cause shoulder pain, degeneration, and dysfunction [11][12][13][14]30]. During the early in ammatory in ltration stage, macrophages cause damage to normal tissues when phagocytizing and clearing necrotic tissues [11][12][13]30]; during the in ammatory regression stage, residual lipid-laden macrophages differentiate into adipocytes, resulting in fatty accumulation in rotator cuff muscles [31]. Furthermore, multiple pro-in ammatory cytokines secreted by macrophages start the in ammatory in ltration cascade that leads to the secondary damage [11][12][13]30]. Thus, the main therapeutic objective to improve impaired shoulder function is to inhibit macrophage proliferation, migration, and secretion of pro-in ammatory cytokines. In the present study, a single GCs injection signi cantly decreased rat raw cell proliferation, migration, secretion of pro-in ammatory cytokines including TNF-α, IL-1α, and IL-1β, and increased the secretion of the anti-in ammatory cytokine IL-4. Additional treatment with ASC-Exos plus GCs further signi cantly decreased the secretion of pro-in ammatory cytokines and increased secretion of antiin ammatory cytokines including IL-4 and IL-10, suggesting a stronger synergic anti-in ammatory effect.
However, this synergic anti-in ammatory effect was not achieved by further decreasing the proliferation and migration of rat raw cells, as the proliferation and migration intensity in the GCs + ASC-Exos group were not different to the control group. We considered that this synergic anti-in ammatory phenomenon might be achieved by the ASC-Exos-driven M2 phenotypic transformation of macrophages. Recent studies from Chen et al. found that a medium conditioned with human bone marrow-derived stem cells, which could be regarded as exosomes from bone marrow-derived stem cells, had a signi cant impact on macrophage polarization by inhibiting M1 phenotype polarization and promoting M2 phenotype polarization [12]. Zhang et al. also reported that MSC-Exos enhance the survival of cartilage grafts via the phenotypic transformation of macrophages from M1 to M2 [23]. The results from these previous studies indicate that the synergic anti-in ammatory effect obtained after additional treatment with ASC-Exos plus GCs in the present study might be achieved by regulating the phenotypic transformation of macrophages.
Further studies are needed to investigate this phenomenon.
Additional treatment with ASC-Exos plus GCs not only exerted a stronger synergic anti-in ammatory effect than GCs treatment alone, but also restored the GCs-induced detrimental effects on tenocytes. The generation of ROS is reportedly a sensitive indicator of GCs-induced mitochondrial detriment, resulting in decreased cell proliferation and migration, and increased cell senescence and apoptosis [18]. In the present study, additional treatment with ACS-Exos signi cantly reduced the GCs-induced ROS transcription, decreasing the senescence and apoptosis, and increasing the proliferation and migration of tenocytes. MMPs are responsible for the degradation of the extracellular matrix, while this process is antagonized by TIMPs during tendon development, morphogenesis, and normal remodeling [26,32,33].
Although the roles of MMPs and TIMPs may be either pathological or physiological, the present results suggest that additional treatment with ASC-Exos somewhat decreased the GCs-induced MMP level and increased the TIMP level, showing the possible role of ASC-Exos in maintaining the metabolic homeostasis of tendons [26,32,33]. Increased amount of type ΙΙΙ collagen is regarded as a repair response to tissue injury and formation of wound bed granulation tissue, while successful remodeling is characterized by the replacement of type III collagen with type I collagen in the healing of torn rotator cuff tendons [26,32]. The present results show that GCs treatment signi cantly decreased the type I/III collagen transcription ratio, which might result in a poorly organized tendon, while additional treatment with ASC-Exos was able to restore that adverse effect. Decorin and biglycan are also major components of leucine-rich proteoglycan in tendons [22]. Additional treatment with ASC-Exos signi cantly increased the GCs-suppressed transcription of decorin and biglycan, further indicating that ASC-Exos had a positive effect on tenocytic differentiation.
Our hypothesis was further supported by in vivo studies that showed that an additional injection of ASC-Exos signi cantly improved both the histological and mechanical properties of GCs-induced detriments on rotator cuff tendons. Consistent with our in vitro studies, the disorganization, collagen degenerative changes from type I to type III, fatty in ltration, and compromised tendon mechanical properties induced by a GCs injection were restored by an additional injection of ASC-Exos. The potential bene ts underlying the contribution of ASC-Exos to synergic anti-in ammatory effects and antagonizing the detrimental effects of GCs might be attributable to stem cell-derived exosome-regulated immunity and growth effects [20,22].
Published data concerning the protection against GCs-induced tendon damage are limited. Poulsen et al.
investigated GCs-induced detriments of tenocytes and found that vitamin C and insulin antagonize the GCs-induced inhibition of the ERK and Akt signaling pathways as well as the GCs-induced increased level of FOXO, restoring normal cell proliferation and collagen synthesis in GCs-treated tenocytes [18]. These previous results identify the signaling pathways underlying the GCs-induced detrimental effects on tenocytes, and the authors suggested potential strategies against the adverse effects of GCs.
The present study had the following limitations. First, although we have shown that ASC-Exos exerted a synergic anti-in ammatory effect with GCs and override their detrimental effects on the rotator cuff, but we did not perform further investigations to elaborate this phenomenon and reveal underlying mechanism, which is worth investigating in our future studies. Second, we used normal rather than degenerative tenocytes in the in vitro study, and the results might differ under different conditions. Third, we only investigated the effect of a single GCs injection with or without ASC-Exos on the rotator cuff tendons. As previous studies have showed that the detrimental effect of a single GCs injection on tendons is transient, both the histological and mechanical properties of the GCs-treated tendon may return to normal levels by 3 weeks [2]. Thus, the effect on the rotator cuff of repeated GCs injections with or without ASC-Exos is worth investigating in the future. Fourth, the anti-in ammatory and analgesic effects are the main areas of interest regarding GCs injections in clinical practice. However, it is di cult to evaluate changes in pain levels after GCs injections in an animal model, and this important aspect was lacking in the present study. However, the present ndings have substantial implications for clinical application. As GCs injections lead to a signi cant decrease in the mechanical properties of tendons, repeated GCs injection is avoided in clinical practice. The present study found that concomitant treatment with GCs and ASC-Exos not only exerted a stronger synergic anti-in ammatory effect, but also restored the GCs-induced detrimental effects on tendons. This might be a possible strategy for repeated GCs injections in the future.

Conclusion
The combination of GCs and ASC-Exos exerted a stronger anti-in ammatory effect than GCs alone and overrode the detrimental effects of GCs on the rotator cuff.        Quantitative analysis of (A) fatty in ltration intensity, (B) type I collagen staining intensity, (C) type III collagen staining intensity, and (D) semi-quantitative analysis of injured supraspinatus in the saline treated group (Control), GCs treated group (GCs), and GCs + ASC-Exos treated group (GCs + ASC-Exos).
Data are expressed as mean ± SD. * indicates signi cant differences, P<0.05. Quantitative analysis of (A) type I collagen staining intensity, (B) type III collagen staining intensity, and (C) semi-quantitative analysis of intact infraspinatus tendon in the saline treated group (Control), GCs treated group (GCs), and GCs + ASC-Exos treated group (GCs + ASC-Exos). Data are expressed as mean ± SD. * indicates signi cant differences, P<0.05. Figure 12