The use of MSCs is a promising and outstanding therapeutic efficacy in regenerative medicine. The stem cells have been successfully derived from multiple tissues, including bone marrow, adipose tissue, tendon, umbilical cord blood, bursa, and urine. In this review, we summarize the animal and clinical studies of these stem cells for rotator cuff injuries. The therapeutic effects of MSCs for rotator cuff injuries are summarized in Figure 2.

MSCs are thought to mediate therapeutic functions in a paracrine manner in addition to their multipotent differentiation capacity and direct intercellular interactions. In the example of a rotator cuff injury model, conditioned medium (CM) of human bone marrow-derived stem cells promotes rotator cuff healing by increasing histologic score, bone mineral density and biomechanical tensile after surgery (Reiner et al., 2017; He et al., 2021). To search for the therapeutically active components, Bruno and colleagues successfully fractioned CM of MSCs by ultracentrifugation and discovered the therapeutic vesicular structures (Bruno et al., 2009). Collectively, these nano-sized particles with a lipid bilayer, naturally released by cells, are called extracellular vesicles (EVs) (Théry et al., 2018). They can be categorized into exosomes, microvesicles, and apoptotic bodies according to their cellular origins. Exosomes are the smallest vesicle types (40–120 nm), and they originate from the inward budding of late multivesicular endosomes (MVEs) and are released upon fusion of MVEs with the plasma membrane. Meanwhile, microvesicles are formed by budding from the plasma membrane, and their size can vary from 50 nm to 1 μm. Unlike exosomes and microvesicles, which are released by all cells, apoptotic bodies are vesicles (50 nm–5 μm) produced by cells undergoing apoptosis. EVs have become an attractive approach in regenerative medicine because they exert biological activities like those of stem cells and overcome the shortages of cell-based therapy, such as cell expansion, low survival rate, and potential immunological rejection (Keshtkar et al., 2018; Woo et al., 2020). It is believed that MSCs cultured in chemically defined serum-free media may be more suitable for the manufacture of EVs. The methods of separation and concentration may vary depending on the size of EVs and the purpose of end-use. According to a worldwide ISEV survey in 2015, ultracentrifugation was the most widely used primary EV separation and concentration technique (Gardiner et al., 2016). A variety of techniques have also been developed to achieve better specificity of separation, such as density gradients and size-exclusion chromatography (Théry et al., 2018). However, there are currently no accepted metrics for assessing the purity or degree of purity in EV preparation (Reiner et al., 2017). Although the mechanism of EVs is not fully understood, it is believed that EVs secreted by stem cells can promote tissue repair and regeneration by inducing the proliferation of cells, promoting angiogenesis, modulating the inflammation process, and affecting cell apoptosis (He et al., 2021). The therapeutic effects of MSC-EVs for rotator cuff injuries are summarized in Figure 2 and Table 3.

Notably, the most important impact of EVs on tissue regeneration is their immunomodulatory properties at both humoral and cellular levels. They can reduce injury-induced inflammation by dampening but not inhibiting complement activation through CD59, and they involve the promotion of anti-inflammatory and pro-regenerative (M2) macrophages over pro-inflammatory M1 macrophages and concomitantly enhance the expression of anti-inflammatory cytokines such as IL-10 instead of pro-inflammatory cytokines such as IL-1β and TNF-α (Toh et al., 2018). Additionally, EVs contain a large amount of biological information, including biologically mRNAs, miRNAs, and lncRNAs, which are important for modulating the signaling of the endogenous and exogenous cells of the injured site (Forsberg et al., 2020). Because of this advantage, engineered EVs are also regarded as candidate cargo to realize gene therapy for injuries. Additionally, the different resources of human MSC-derived EVs can exhibit distinct characteristics that reveal their potential applications in different fields. Bone marrow MSC-derived EVs have shown superior regeneration ability, and adipose tissue MSC-derived EVs have played a significant role in immune regulation, whereas umbilical cord MSC-derived EVs are prominent in tissue damage repair (He et al., 2021).

BMSC-derived EVs (BMSC-EVs) are widely used in the musculoskeletal regeneration field. Recently, one study solely injected BMSC-EVs into the vein to promote rotator cuff repair (Huang et al., 2020). The results showed that BMSC-EVs promoted angiogenesis around the tendon–bone interface, histologic histological appearance, and biomechanical strength, and they inhibited the secretion of pro-inflammatory factors in rat rotator cuff reconstruction. They believed that the mechanism by which BMSC-EVs achieve the healing process may be through the proliferation, migration, and angiogenic tube formation of human umbilical vein endothelial cells (HUVECs) by regulating the angiogenic signaling pathway, inhibiting the polarization of M1 macrophages, and also inhibiting the secretion of pro-inflammatory factors by M1 macrophages (Huang et al., 2020). A previous study reported BMSC-EVs can also suppress inflammation by increasing the expression of anti-inflammatory mediators IL-10 and IL-4 at an early phase of healing (Shi et al., 2019). Moreover, local delivery of BMSC-EVs can promote tendon regeneration by facilitating the proliferation, migration, and tenogenesis differentiation of endogenous TPSCs (Shi et al., 2019; Yu et al., 2020).

ADSC-derived EVs (ADSC-EVs) have regeneration and immunomodulation capacities (Chen et al., 2021; Wang et al., 2021). To investigate the effect of ADSC-EVs on chronic rotator cuff tears, Wang C. et al. (2020) established a bilateral rotator cuff chronic tear model and injected isolated ADSC-EVs at the injury site of the supraspinatus muscle at the time of repair. This demonstrated that the local injection of ADSC-EVs inhibited fatty infiltration, regenerated fibrocartilage, and increased biomechanical strength. Similarly, Fu et al. (2021) reported that rotator cuff injected with ADSC-EV-loaded hydrogel exhibited aligned collagen fiber and muscle bundles and enhanced mechanical properties. A variety of mechanisms may contribute to ADSC-EVs in rotator cuff repair. TPSCs regulate the proliferation, migration, and tenogenic differentiation of TPSCs. Studies have reported that ADSC-EVs can upregulate the expression of the tenogenesis genes TNMD, TNC, and Scx in vivo (Liu H. et al., 2021; Fu et al., 2021). Smad signaling pathways play vital roles in regulating stem cell activity. This effect may be related to activated SMAD2/3 and SMAD1/5/9 signaling pathways, which play vital roles in regulating stem cell activity (Liu H. et al., 2021). In addition, ADSC-EVs regulate the early inflammatory response in rotator cuff healing by decreasing the M1 macrophage, enhancing the M2 macrophage, and reducing the secretion of pro-inflammatory cytokines, such as IL-1β, IL-6, IL-8, and MMP-9 (Liu H. et al., 2021). In a human supraspinatus explant experiment, ADSC-EVs maintained homeostasis of the impaired tendon by increasing expression of COL1A1, COL3A1, and an elevated type I/III ratio and by decreasing expression of MMP-9 and MMP-13 (Zhang et al., 2021). The underlying mechanism might be enhancing AMPK signaling to suppress Wnt/β-catenin activity or NF-κB pathway (Ma et al., 2019; Zhang et al., 2021).

Few research observations of TPSC-derived EVs (TPSC-EVs) exist on rotator cuff repair. Recently, a study reported that TPSC-EVs suppressed inflammation and apoptosis at 1 week after surgery; the tendon exhibited a more continuous and regular arrangement and a larger collagen fiber diameter in the TPSC-EV-treated group compared to the non-TPSC-EV-treated group at two and eight weeks after surgery (Zhang et al., 2020a). The augment of TPSC-EVs can be partially explained by promoting the proliferation and migration of tenocytes in a dose-dependent manner in an in vitro study, which was related to the activation of the PI3K/AKT and MAPK/ERK1/2 signaling pathways (Zhang et al., 2020a). Nevertheless, the detailed mechanism of TPSC-EVs is still poorly understood and needs further investigation.

Many studies have reported the potential of human umbilical cord mesenchymal stem cell-derived EVs (HUMSC-EVs) in tendon repair. Kim et al. (2022) demonstrated that the HUMSC-EVs laden injectable collagen could effectively promote bone-to-tendon healing via collagen maturation in the bone–tendon interface and prevent fatty degeneration of the rotator muscle at 4 weeks after rotator cuff repair. In a rat Achilles tendon injury model, treatment with HUMSC-EVs improved the histological structure, enhanced tendon-specific matrix components, and optimized biomechanical properties of the Achilles tendon, which was related to the overexpression of miR-29a-3p regulated by PTEN/mTOR/TGF-β1 signaling (Yao et al., 2021).

An increasing number of studies have utilized gene therapy to enhance and expand the therapeutic effectiveness of stem cells in tendon repair. In comparison to conventional stem cell therapy, modified stem cells or their EVs yield more production of a gene in the local injured site, with greater biological activity and lower immune response. There are two main methods of transferring the gene to target cells: viral and non-viral vector methods. Viral vectors have been widely used in gene therapy due to their high efficiency in gene delivery into the cells they infect. Commonly, recombinant viruses include adenovirus, lentivirus, retrovirus, and adeno-associated virus in tendon repair. However, the use of viral vectors may also meet the challenges of high-cost expenses and safety issues. Thus, non-viral vectors, such as plasmids, increase the interest of researchers in gene delivery because of their safety, simple manufacture, and lower immunogenicity.

Concerning rotator cuff regeneration, many studies have focused on facilitating the tenogenic differentiation of stem cells to promote rotator cuff repair. Gulotta et al. (2011a) showed that BMSCs transfected to overexpress Scx promoted the formation of fibrocartilage at the tendon insertion and improved biomechanical strength at 4 weeks for rats who underwent unilateral detachment and repair of the supraspinatus tendon. Owing to the tear of enthesis (Zones 3 and 4) in rotator cuff injuries, it was meaningful to transfer osteogenic and chondrogenic genes to enhance tendon–bone healing. It was demonstrated that, when transfected to overexpress developmental genes, membrane type 1 matrix metalloproteinase (MT1-MMP), which is thought to direct the process of ossification, promotes the formation of fibrocartilage at the tendon insertion and improves biomechanical strength (Gulotta et al., 2010). However, ADSC transduced with the osteogenic factor bone morphogenetic protein 2 (BMP-2) led to impaired healing by losing bone mass and decreasing biomechanical properties (Lipner et al., 2015).

In terms of chondrogenic genes, the genetically modified BMSCs with overexpressing BMP-13 showed no improvement in either biomechanical parameters or histological appearance in acute rotator cuff repair (Gulotta et al., 2011b). Additionally, gene-modified stem cells can inhibit inflammation during the healing process. Cheng et al. (2014) discovered that the silent TNF-α stimulated gene/protein 6 (TSG-6) of TPSCs reduced biomechanical strength, indicating that TPSCs might promote rotator cuff healing through regulating anti-inflammatory response by TSG-6 signaling. Schnabel et al. (2009) explored using the insulin-like growth factor-I (IGF-1) gene enhanced BMSCs significantly improved tendon histological scores and reduced ECM degradation in collagenase-induced bilateral tendinitis lesions, but the benefit of IGF-1 gene enhancement was not obvious compared to untreated BMSCs.

EVs contain miRNAs from donor cells that can be transferred to recipient cells, thereby promoting the expression of specific proteins. Therefore, an increasing number of studies pay attention to utilizing engineered EVs to transfer genes in musculoskeletal disorders such as osteoporosis (Yang et al., 2020) and osteoarthritis (Tao et al., 2017). Several studies have demonstrated the promising results of engineered EVs in tendon repair and regeneration. Early tendon remodeling plays a vital role in tendon regeneration. Yao et al. (2021) revealed that miRNA-29a-3p loaded HUMSC-EVs reduced the area of the lesion and improved histological scores in a tendinopathy model. The underlying pathway was reducing transcript levels of collagen III via the PTEN/mTOR/TGF-β1 signaling pathway (Millar et al., 2015; Watts et al., 2017).

Another strategy of gene therapy for tendon healing is to inhibit the fibrous process of the tendon and surrounding tissues. The EVs derived from antagonists targeting miR-21a-3p treatment of HUMSC, which expressed low levels of miR-21a-3p, expanded the inhibition of tendon adhesion by manipulating p65 activity, suggesting that delivering low-abundance miR-21a-3p may inhibit tendon adhesion. Furthermore, some miRNAs participate in tenogenic differentiation and prevent chondro-osteogenic differentiation, including miR-124, miR-135a, miR-140, and miR-337-3p (Chen et al., 2015; Wang et al., 2016; Geng et al., 2020; Liu Y. J. et al., 2021). However, the effectiveness of the delivery of these genes by EVs has not been confirmed.

To date, various natural and synthetic materials have been developed to promote stem cells in rotator cuff repair and regeneration. In our review, natural polymers are most widely used in stem cell therapies for rotator cuff injuries (Gulotta et al., 2010; Gulotta et al., 2011a; Degen et al., 2016; Liu et al., 2019). They have the advantages of non-toxicity, biocompatibility, and biodegradation, as well as cell proliferation and cell adhesion. Nevertheless, it is difficult to modify their physical and chemical properties, which remains a potential immunogenicity problem (Garg et al., 2012). Commonly used natural materials include ECM-derived biomaterials, hyaluronic acid (HA), chondroitin sulfate (CS), and fibrin. As the main component of the ECM, ECM-based biomaterials provide a biomimetic environment suited for tissue remodeling. For instance, Liu et al. (2019) reported a novel biomaterial that uses engineered tendon–fibrocartilage–bone composite (TFBC) augmentation with BMSCs to form a “sandwich” structure that can enhance rotator cuff healing in terms of anatomic structure, collagen organization, and biomechanical strength. Decellularized matrices have been explored for their regenerative effects on tendon repair; however, tissue resources should be considered. While the tendon-derived decellularized matrix promoted the tendinous phenotype in TSPCs and inhibited their osteogenesis, the dermal skin-derived collagen matrix had no apparent effect on TSPC differentiation (Yin et al., 2013). Hyaluronic acid (HA) is an anionic, non-sulfated glycosaminoglycan that is distributed in the intercellular matrix of most connective tissues. It has been reported that HA decreases the cell proliferation and expression level of procollagen α1 (III) mRNA of tendon-derived fibroblasts (Yamada et al., 2007). Moreover, CS is a natural polymer and a major ECM component that has the ability to reduce inflammation by diminishing NF-κB activation and nuclear translocation (Vallières and Du Souich, 2010). Human mesenchymal stem cells cultured within the decellularized amniotic matrix wrapped around the collagen-chondroitin sulfate scaffold could maintain metabolic activity and down-regulate the pro-inflammatory cytokines (Hortensius et al., 2018). Fibrin is formed following the cleavage of fibrinogen and thrombin and can be processed into hydrogels or fibrous scaffolds. Tissue-engineered construction based on fibrin hydrogel has better extracellular matrix organization and biomechanical properties compared to collagen-based hydrogels (Breidenbach et al., 2015; Thangarajah et al., 2018).

Synthetic materials are also used extensively in tendon regeneration since their molecular weight, hydrophobicity, and degradation speed can be easily modified, and due to their low cost of fabrication (Ruiz-Alonso et al., 2021). Notably, it is possible to achieve similar mechanical properties with tendon tissue and good structural integrity, which are important in the regeneration of tendon repair (Pina et al., 2019). Compared to natural biomaterials, synthetic materials have a low risk of disease transmission because they are not obtained from biological organisms or tissues. For this reason, they have a high risk of immune response (Won et al., 2014). Since synthetic materials are hydrophobic in nature, they may also cause poor cell adherence, low proliferation rates, and altered phenotypes of stem cells (Theisen et al., 2010). Numerous synthetic materials are used for tendon tissue repairs, such as poly-ε-caprolactone (PCL), poly (lactic acid) (PLA), poly (glycolic acid) (PGA), poly (ethylene glycol) (PEG), and poly (lactic-co-glycolic acid) (PLGA). These synthetic polymers can be electrospun into nano- and microfibrous scaffolds, which mimic aligned collagen fibers in tendon tissue and promote tenogenic differentiation (Vuornos et al., 2016; Laranjeira et al., 2017; Calejo et al., 2019; El Khatib et al., 2020). The rate of degradation determines its usage. For example, polymers with a low degradation rate, such as PCL, are suitable for building longer-term tendon scaffolds (Laranjeira et al., 2017; Calejo et al., 2019), while polymers with faster degradation rates are less suitable since they may increase the inflammation response, including PLA, PGA, and PLGA (Yokoya et al., 2008; Vuornos et al., 2016; Chen et al., 2019; Chen P. et al., 2020; Araque-Monrós et al., 2020; El Khatib et al., 2020). According to this characteristic, a PEG-based hydrogel system with a range of degradation rates can control the timing of MSC delivery to the target site of tendinopathy (Qiu et al., 2011).

The topographical and mechanical properties of biomaterials impact the proliferation and tenogenic differentiation of stem cells, including fiber diameter, pore size, alignment, surface roughness, and matrix stiffness. Fiber with a large diameter promotes the expression of tenogenic genes, such as Scx, in stem cells. This suggests that large-diameter fibers (e.g., >2 μm) may be more suitable for MSC differentiation into tendon lineage than small-diameter fibers (Cardwell et al., 2014). Generally, the pore size of the scaffold plays an important role in migration ability (Zheng et al., 2017); a larger pore size of PLGA scaffolds significantly enhances the migration of BMSCs in vitro (Dai et al., 2018). Fiber alignment provides tissue-specific biomechanical cues to resident cells in the native tendon. Studies have indicated that scaffolds with aligned fibers enhance cell infiltration, ECM deposition, collagen alignment, and tendon-related gene expression of stem cells when compared to nonaligned fibers (Orr et al., 2015; Zheng et al., 2017). In the massive rotator cuff repair model, scaffolds with aligned fibers exhibit more conspicuous native microstructures, better alignment, and better mechanical properties at 12 weeks post-implantation (Zheng et al., 2017). Due to the mechanical microenvironment of the tendon, matrix stiffness impacts stem cells during tendon repair. The proliferation of TPSCs increases and more stress fibers form with increasing matrix stiffness. Furthermore, the differentiation of TPSCs into tenogenic lineages is inhibited on stiff hydrogel with reduced expression of tendon-specific genes THBS4, TNMD, and SCX by regulating FAK and ERK1/2 pathways (Liu et al., 2018).

Concerning rotator cuff repair, the scaffold is an effective tool for transmitting mechanical stimulation to delivered cells; thus, the mechanical environment provided by biomaterials should be considered in cell delivery. The mechanical stimulation of stem cells is vital in tendon tissue repair and has been shown to influence the differentiation and proliferation of stem cells (Wang H.-N. et al., 2020). The magnitude of stretching could lead to different cell fates. Studies indicated that 4% stretching promoted the differentiation of TPSCs into tenocytes with increased gene expression of COL1A1; 8% stretching, however, promoted the differentiation of TPSCs into non-tenocytes, including adipocytes, chondrocytes, and osteocytes, aside from differentiation into tenocytes, as evidenced by higher expression levels of genes such as PPARγ, COL2A1, Sox9, and Runx2 in vitro (Wang H.-N. et al., 2020).

Furthermore, mechanical stimulation of BMSCs significantly increased the expression of tenogenic genes and anti-inflammatory cytokines (Ciardulli et al., 2020). In addition, biomaterials containing magnetic elements have been developed to mechanically stimulate stem cells in tendon regeneration. Human adipose stem cells cultured on the magneto-mechanical actuation scaffold increased the expression of tendon-related genes Scx and Tnmd when compared to static culture and steered the mechanosensitive YAP/TAZ signaling pathway. This magneto-mechanical stimulation also modulated the inflammation response by upregulating the expression of anti-inflammatory cytokines IL-4 and IL-10 while reducing the expression of pro-inflammatory cytokines COX-2 and IL-6 (Tomás et al., 2019).

To achieve the different requirements of delivery, “smart” delivery systems, such as stimulation-responsive hydrogel, can provide possibilities for precise treatment for different stages of healing (Bawa et al., 2009; Yun et al., 2015). These hydrogels respond to visible or UV light and release drugs for tissue regeneration. A gelatin methacryloyl hydrogel loaded with TPSC-EVs was placed in the Achilles tendon defect to promote tendon healing. By using a 405 nm blue light source at a distance, the carrier is converted to the gel state by irradiation for 10–20 s. After delivering TPSC-EVs, tendon repair is promoted by suppressing inflammation and apoptosis and regulating ECM balance (Zhang et al., 2020b). Additionally, matrix metalloproteinase 2 (MMP-2) is a valuable endogenous trigger for responsive release systems, achieving localized and on-demand drug delivery. MMP-2 is a member of the zinc endopeptidase family and has the ability to cleave ECM components, which is upregulated in the peritendinous area where the adhesion tissue forms after tendon injury. Cai et al. (2021) have designed an innovative anti-adhesion electrospun nanofiber scaffold system for the on-demand and unidirectional release of polyplexes to inhibit fibroblast proliferation and collagen deposition by gene therapy. The MMP-2 degradable hydrogel is fabricated by crosslinking allyl glycidyl ether (AGE) modified carboxymethyl chitosan (CMCS-AGE) and the MMP-2 substrate peptide CPLGLAGC (MMP-2 sp).