The tendon is composed of longitudinally arranged collagen fibers and a scattered distribution of spindle-shaped tendon cells. Its primary function is to transmit the force generated by contraction of the muscles and to drive the movement of the bones (1). It is hypothesized that when the tendon is overused for a long time, bears a large load or is stretched repeatedly, many pathological changes will occur in the tendon including cell and extracellular matrix (ECM) lesions, increased proteoglycan and damage to the collagen structure (2), which leads to tendon injury. In addition, other factors may also lead to tendon injury, including age, incorrectly performed exercise, previous injury, weight and medication (3). In recent years (4), injured athletes have accounted for the majority of injured people. Numerous athletes suffer from chronic tendinopathy due to overwork, especially those who play basketball, football and volleyball, as well as those who perform in the high jump. The injured areas include the Achilles and patellar tendon, rotator cuff and the tendons around the elbow joint, all of which result in inconvenience to the lives of those affected.

Many effective treatment methods (Fig. 1) have been employed in a number of clinical practices. Patients with tendon injury initially engage in conservative treatment, including eccentric training, shock wave therapy and injection therapy. However, in 10% of patients, conservative treatment has no significant effect (5). For example, patients with partial tears of the supraspinatus tendon and shoulder pain lasting >3 months were treated with mesenchymal stem cell (MSC) injections at the lesion site. However, final treatment effect in terms of pain, shoulder function and tear size was smaller and less significant than the clinical effect of MSC injection compared with the control treatment (6). However, there is potential of MSCs in tissue regeneration. Therefore, surgical treatment, including open and arthroscopic surgery, is required for the small percentage of patients who do not do well with conservative treatment. Although the aforementioned treatment approaches have achieved good results in the treatment of tendon injury, they still cannot fully restore the composition, structure, and mechanical properties of the injured tendon (5).

In recent years, studies have shown that MSCs are a promising treatment method. Many clinical trials have demonstrated that MSCs have good therapeutic effects (7-9). Firstly, MSCs differentiate into targeted cell types, and under specific induction conditions in vivo or in vitro, they can differentiate into tendon cells to stimulate tendon tissue regeneration. Secondly, they have a paracrine effect and can secrete cytokines and growth factors into neighboring cells, thereby promoting vascularization and cell proliferation in damaged tissue and helping to repair the damaged area (10). MSCs also have immunomodulatory properties and decrease the inflammatory response of damaged tissue (11). Additionally, MSCs have a wide range of sources; they can be isolated and prepared from bone marrow, adipose tissue, placenta, umbilical cord and other tissues. Overall, MSCs have numerous advantages over other conservative treatments such as NSAIDs, low level lasers, including strong multiplication capability, safety, economy and efficiency. The present review summarized the mechanisms, progress, and challenges of MSCs in the treatment of tendon injury based on published literature and provides support for future clinical practice and research.

Cell-based tissue regeneration therapies are attractive and well-explored therapeutic approaches, especially in the application of tendon repair (12,13). The most discussed cell-based therapies include MSCs (from sources such as bone marrow, adipose, umbilical cord) and tendon, embryonic and induced pluripotent SCs (iPSCs) (14-17). Tables I and II summarize the properties of MSCs and other cellular therapies in tendon repair and their characteristics.

MSCs are widely available, relatively simple to obtain and can be injected directly or following processing, purification and amplification. Several studies have shown that MSCs in the tendon are actively involved in the tendon repair process (15,18). They migrate to the injury site following tendon injury and secrete growth factors and other soluble cytokines that induce cell proliferation and regulate signaling, in addition to enhancing the tendon-forming properties of tendon stem/progenitor cells (TSPCs), thereby promoting tendon repair (19). Studies have showed that the efficiency of differentiation of MSCs into tendon cells could be better improved by injecting growth factors such as bone morphogenetic protein-12 (BMP-12), growth/differentiation factor-5 (GDF-5) and TGF-β compared to treatment with MSCs alone (20,21). There are numerous studies on the use of exogenous MSCs of different sources in tendon repair following injury through intravenous or topical wound injection, bioengineered scaffolds and gels (22-24). For example, Smith et al (25) injected bone marrow MSCs (BMSCs) into racehorses with tendon injury; after 6 months of treatment, the tendons of racehorses showed enhanced biomechanics, morphology and normalization of extracellular matrix (ECM) component of the tendon (25). In particular, adipose-derived MSCs (ADSCs) show advantages over other sources of BMSC in terms of decreased donor morbidity and avoiding ethical concerns. Therefore, it has a high value in the treatment of tendon injuries (26).

The tendon also contains a small population of resident cells that maintain homeostasis of tendon growth and repair (27). Similar to other SCs, these TSPCs have the capacity to undergo self-renewal and multidirectional differentiation. Numerous studies have exploited this feature to promote the self-proliferation of TSPCs and induce differentiation to tendon cells by injecting growth factors such as TGF-β and basic fibroblast growth factor (bFGF) in vivo or creating hypoxic states in vitro. Subsequently, TSPCs upregulate IL-10 and TIMP-3 through the JNK/STAT signaling pathway, thus playing a regulatory role in inflammation and tendon remodeling (16,28). However, alterations in the tendon microenvironment following injury may lead to misdifferentiation of TSPCs to chondrocytes, osteoblasts and adipocytes, resulting in failure of tendon healing (29). There are many triggers that lead to misdifferentiation, including age-associated cellular aging, mechanical stretch stimulation >8% and some inflammatory factors such as prostaglandin E2 (PGE2) (30,31). Understanding the factors that induce TSPC (mis)differentiation may facilitate use of TSPCs in the treatment of tendon injuries.

Embryonic SCs (ESCs) are isolated from early embryos (before protointestinal embryonic stage) that have properties of unlimited proliferation, self-renewal and multidirectional differentiation (32). ESCs can be induced to differentiate into almost all cell types, both in vitro and in vivo. Therefore, they have potential in regenerative medicine (17). It has been shown that human ESCs can be induced to differentiate into tendon-like cells by the addition of exogenous BMP-12, GDF-7 and BMP-13 (33). It has also been shown that tendon injury sites treated with ESCs recover better and collagen fibers can be restored to a more normal linear fiber pattern compared to other cellular therapies (33). However, there are concerns regarding the use of ESCs. First, ESC isolation destroys the embryo, which may be considered a violation of bioethics. Despite the potential use of ESCs in both basic research and clinical applications, research on ESCs and their applications is prohibited in some countries, such as the United States and some European countries where religious organizations are prevalent (14). ESCs can theoretically be induced into various types of somatic cells for tissue regeneration. Therefore, there is a risk of teratoma formation following application of ESCs for treatment (34). Teratomas consist of three embryonic germ layers, which are due to residual undifferentiated cells in the transplanted population. Therefore, it is necessary to remove residual undifferentiated stem cells from ESCs before application (35). Similar to ESCs, iPSCs) can be prepared from the patient's own somatic cells, thus avoiding immune rejection (36). Compared with ESCs, iPSCs can be obtained from more convenient sources, such as fibroblasts and hepatocytes, and do not involve ethical concerns. In horses, the application of iPSCs promotes tendon tissue regeneration and significantly decreases the frequency of re-injury (37). However, since iPSCs have the ability of multidirectional differentiation, there is also a risk of teratoma formation (38). In addition, the time and cost required to prepare iPSCs may prevent them from becoming a therapeutic option.

Tendons are dense tissues that connect muscle and bones. Their unique composition and structure give them appropriate mechanical properties (39). Therefore, understanding of the association between the composition, structure and mechanical properties of normal tendons can help to prevent tendon injury and select the most appropriate method for treatment.

In terms of ultrastructure, tendons are hierarchical structures with a regular arrangement of collagen fibers (40). This hierarchical structure provides ideal load-bearing and tensile force transmission properties (41). The smallest structural unit of the tendon is the fibril, with a diameter of 20-500 nm, consisting of rod-shaped collagen molecules (42). Electron microscopy in the absence of load shows fibrils become 'crimped'. This is thought to be due to a non-linear change in the strain-stress curve caused by small tensile forces at low strains. Studies have shown that this change can be effective for cushioning and shock absorption in tendons (43,44). The fibrils are cross-linked to form a stable structure, referred to as a collagen fiber. Multiple collagen fibers reassemble to form the tendon fascicle, which is the largest structural unit of the tendon. The tendon fascicle is a tubular-like structure 150-500 µm in diameter, aligned parallel to the long axis of the tendon. Each fascicle is surrounded by connective tissue called the endotenon, thus forming a complete tendon structural unit (45). The tendon is covered with a layer of connective tissue attached to the endotenon called the epitenon. The epitenon effectively reduces friction between the tendon and adjacent tissue (46). In addition, there are nerves and blood and lymphatic vessels on the endotenon and epitenon, which serve a key role in development of the tendon. A layer of loose connective tissue, called paratenon, surrounds the tendon away from the joint area. During joint movement, the paratenon facilitates the smooth gliding of the tendon and completion of the movement (47). Normal tendons consist of collagen, water, proteoglycans, glycoproteins, cells and other components (48). Type I collagen is the primary component of tendons and is also the main element in connective tissue responsible for transmitting force (49). Type III collagen is less abundant in normal tendons (50). Due to its rapid cross-linking properties, type III collagen increases rapidly following tendon injury, allowing for rapid repair in the injured area and tendon healing (51). In addition to containing high amounts of type I and type III collagen, collagen fibers expressed at lower levels, including types V, VI, XI, XII and XIV, serve key roles in regulating the biological properties of fibers in terms of diameter, number and density (52-54). Tendons also contain a large amount of water; studies have shown that the higher the water content, the lower the stiffness of tendons (55,56).

In addition to collagen, non-collagenous glycoproteins, proteoglycans, elastin, and other components of the ECM also play important roles in the growth and development of tendons (57). Non-collagenous glycoproteins mediate signaling between TSPCs and muscle, thereby promoting maturation of the tendon-muscle junction and maintaining tendon stability (58). Proteoglycan is an important component of extracellular matrix (59); it mainly regulates the diameter of linear and lateral fibers during the later stage of tendon development and cooperates with growth factors to regulate cell proliferation, thus promoting collagen production (60,61). Glycoproteins are part of the ECM; cartilage oligomeric matrix protein (COMP) is the most abundant glycoprotein in tendons. The amount of COMP is positively associated with tensile stress and stiffness (62). Tenascin-C is glycoprotein expressed at low levels that is primarily present in locations where the tendon is subjected to higher load (63). Tenascin-C maintains mechanical properties of the ECM by interacting with collagen fibers (64). Different numbers of elastic fibers are found between tendon fascicles and around tenocytes; multiple elastic fibers surround groups of tenocytes and travel longitudinally along tendon, whereas fibers present between fascicles form a loose mesh-like organization oriented in the transverse direction (65); these play important roles in maintaining tendon strength and conferring elasticity to tissue such as ligaments, aorta and skin (66). Thus, elastin fibers may contribute to recovery of fibers to their original form after stretching (46).

Tendon cells are primarily divided into two categories. Tendon cells present between collagen fibers, known as tenocytes, are the main cells in the tendon tissue. These produce ECM components such as collagen, fibronectin and proteoglycans, and therefore serve an important role in maintaining tendon homeostasis (67). Furthermore, tenocytes increase expression levels of the junctional and stress fiber components when exposed to tensile load. Thus, in tendinopathy, loading with an appropriate tensile load promotes recovery of mechanical properties of the damaged tendon (68). Another group of cells located in the interfascicular matrix (IFM) is called interfascicular cells (69). Interfascicular cells are round in shape and are more densely distributed compared with tenocytes. They are also more metabolically active because of faster protein turnover in the IFM (47). Marr showed that CD146+ cells are interfascicular cells (70). Following injury to a rat Achilles tendon, CD146+ cells migrate toward the injury on days 4 and 7 and to fill the wound on day 21 after the injury (70). Studies have demonstrated that CD146+ cells promote cell proliferation through mTORC2 signaling and thus serve a role in tendon repair (71,72). CD146+ cells also bind Laminin α4 and may play a role in the resolution of inflammation but the exact mechanism of action is currently unknown (73). The presence of interfascicular cells may be key for maintaining tendon function but their exact mechanism of action requires further study. In addition to the aforementioned cells, other cells are present in tendons, such as chondrocytes, synovial cells and tendon SCs (46). Among them, the recently identified tendon SCs have good ability to maintain homeostasis of tendons and promote repair of tendinopathy (74). Tendon SCs have the ability to self-renew and differentiate into tendon cells (75). Their differentiation into tendon cells is promoted at low levels of mechanical stretch (4%) and produces collagen, thereby promoting remodeling of the tendon ECM. However, at high mechanical stretch (8%), tendon SCs are induced to differentiate into non-tendon cells, such as adipocytes, chondrocytes and osteocytes, resulting in histopathological features of lipid deposition, proteoglycan accumulation and calcification in the tendon, thus leading to the development of tendinopathy (76).

Because of the unique layered structure and composition, tendons have characteristic biomechanical properties such as high mechanical strength and viscoelasticity (46). The unique mechanical properties of tendons are reflected in a stress-strain curve composed of four regions. In the initial area when the tendon is stretched <2%, the curled fiber is straightened. When the degree of stretch is 2-4%, it is referred to as the linear area. As the tendon is stretched, the stiffness of fibers increases rapidly and they distort in a linear manner. The slope of this region is called the Young's modulus of the tendon and is used to express the stiffness of the tendon. When a tendon is stretched >4%, microscopic tears occur in the fibers. Tendons undergo significant tissue damage when subjected to >8% strain and continued stretching leads to tendon rupture (46). Tendons have several other mechanical properties, including non-linearity, viscoelasticity, and heterogeneity (77). Viscoelasticity may result from the interaction between collagen, water and proteoglycan (78). Viscoelasticity is important for load transfer in tendons. Tendons are more prone to deformation at low strain rates and are less effective in transferring loads. At high strain rates, tendons are less prone to deformation with higher stiffness and are more effective at transmitting larger loads (79). Thus, the unique mechanical properties of tendons are key to their function of carrying and transmitting loads. The high levels and regular arrangement of collagen create high tensile strength required to provide efficient load transfer. However, it is not clear how small changes in the structure and composition of the tendon lead to changes in mechanical properties.

MSCs are considered to be 'medicinal cell factories' that are capable of secreting a range of bioactive molecules, either in the form of soluble biofactors or through MSC-exosomes (Exos), which have functions in immunomodulation, anti-apoptotic activity, and promoting synthesis of ECM components, such as collagen (80). MSCs have been used in regenerative medicine since their discovery in the late 1960s (80). MSCs are isolated from a number of tissues, including adipose, muscle, tendon, synovial sac, dental pulp, skin, lung, placenta and umbilical cord (11). Furthermore, MSCs self-renew and differentiate to produce specialized cell types such as chondrocytes, muscle cells, and skin cells (82). MSCs derived from BM and adipose are most frequently used in the treatments of tendon injury as they exhibit self-renewal, multidirectional differentiation, and paracrine functions. And they can also lead to tendon healing and functional improvement through minimally invasive treatment approaches (83). The effect of MSCs from different sources in tendon healing is summarized in Table III.

MSCs were originally isolated from BM (84). BMSCs are usually obtained from the iliac crest by minimally invasive puncture and isolated by density centrifugation (85). Chong et al (86) identified two roles of BMSCs in tendon healing. BMSCs secrete growth factors and promote tendon healing by differentiating into tenocytes and participating in collagen synthesis and remodeling. However, activity of alkaline phosphatase is increased after treatment of tendon injury using BMSCs, which leads to the formation of ectopic bone (84,87) and impedes tendon healing.

In recent years, more attention has turned to ADSCs because they are easier to isolate compared to other sources of MSCs. And because BMSCs need to be collected using a trocar to drill through the iliac crest, while ADSCs can be collected using only minimally invasive liposuction techniques, this is easier and less painful for the patient. In addition, over time, the donor site providing BMSCs is prone to pain and stiffness, whereas ADSCs have a much lower incidence (15,88,89). Specifically, adipose tissue is easily aspirated from abdominal subcutaneous adipose; collected tissue is passed through specific systems including forming, granulating, cutting, purification, centrifugation, nitrification, absorption of antimicrobial or antitoxin arrangements, cleansing, partition, and lyophilization to obtain ADSCs (90,91). Furthermore, ADSCs can be obtained from autologous or allogeneic sources, but ADSCs isolated from autologous adipose tissue are the best candidates for the treatment of tendon injuries because they do not induce immune rejection after application in injured tendons (92). Studies have detected transcription factors associated with hypoxia, such as hypoxia-inducible factor-1 (HIF-1), in models of ruptured Achilles tendon that show ectopic ossification; therefore, hypoxia may be associated with formation of cartilage in tendons (93,94). Adding ADSCs in the early stage of tendon healing can reverse or prevent hypoxia by inhibiting inflammation and promoting formation of new blood vessels to inhibit occurrence of heterotopic ossification (95). Consequently, ADSCs have advantages over BMSCs in the treatment of tendon injury. In addition, compared with MSCs from other sources, ADSCs enhance tenogenic properties of tendon resident cells, increase the ratio of collagen I/III and promote repair of ECM (15). These characteristics make ADSCs promising in tendon healing.

MSCs secrete biologically active soluble factors (cytokines, growth factors, chemokines, MSC-Exos) to accelerate healing of tendons (96,97). MSC-Exos are extracellular vesicles (EVs) containing complex RNAs and proteins that target cells via endocytosis, membrane fusion or receptor-ligand interactions and are key paracrine factors for MSCs. MSC-Exos serve important roles in immune regulation, apoptosis and tissue regeneration in numerous types of disease such as osteoarthritis (98). In recent years, there have been numerous studies on the use of MSC-Exos in tendon repair (99-101). MSC-Exos serve roles in tendon repair by regulating biological factors and activating signaling pathways (102). For example, MSC-Exos inhibit the inflammatory response, primarily by promoting AMPK signaling (103). The angiogenesis phase, which is important for tendon healing, also involves MSC-Exos, which secrete growth factors such as vascular endothelial growth factor (VEGF), which is associated with angiogenesis (104). In addition, MSC-Exos play a key role in subsequent cell proliferation and collagen synthesis. MSC-Exos promote proliferation and differentiation of tenocytes by activating the SMAD signaling pathway and increasing the ratio of type I/III collagen genes, thereby promoting type I collagen synthesis (105,106). Other growth factors secreted by MSCs that serve important roles in tendon repair include insulin-like growth factor-I, TGF-β, platelet-derived growth factor (PDGF) and bFGF. These growth factors promote tendon repair by participating in intercellular messaging, as well as signaling pathways during the three phases of tendon healing: inflammation, proliferation, and remodeling (10,107). The biologically active soluble factors secreted by MSCs and their effects on the molecular structure of the tendon are presented in Fig. 3. In addition, mechanical stimulation induces MSCs to differentiate into tenocytes, thereby accelerating the repair of tendon tissue (108,109). For example, in Zhu's (110) experiments, a mechanical stretching force of 10% was applied to MSCs and stimulated for 6 h with 30 cycles per minute. However, in Kasper's (111) experiments, he applied a mechanical load of 10 kPa to MSCs at a frequency of 1 Hz for 72 h of stimulation.

Tendon injury is common in orthopedics. After tendon injury, the tendon shows a local inflammatory response, hypocellularity, lack of collagen and blood vessels and increase levels of proteoglycans (218). The injured tendon exhibits discontinuous and disorganized tendon fibers. Since tendon is a tissue with low cellular content and poor blood supply, the tendon has a limited ability to heal (219,220). The process of tendon healing is divided into three overlapping phases: Inflammatory, proliferative and remodeling phase. During the remodeling phase, scars often form (116). Scars have different biomechanical properties compared with natural tendons, including decreased strength, increased stiffness and greater tendency to form adhesions. Reconstructed tendon tissue exhibits poorer biochemical and mechanical properties compared with uninjured tendon (221). This leads to dysfunction in the limb and makes the tendon more prone to re-rupture (11). With the development of SC research, MSCs have attracted attention (222). MSCs have high proliferation and self-renewal capacity (223). MSCs differentiate into target cells and directly promote tissue regeneration; MSCs also secrete biological factors and EVs, thus indirectly affecting tissue repair (220). MSCs can be obtained from numerous types of tissue and MSCs from different sources show different characteristics, indicating potential advantages and disadvantages of each type of MSC for specific clinical applications (224). Among sources, the most commonly used are BM and adipose tissue. ADSCs are more readily available compared with BM, yield more abundant MSCs after isolation and also decrease donor site morbidity (225). This is because BMSCs need to be collected using a trocar to drill through the iliac crest, while ADSCs can be collected using only minimally invasive liposuction techniques, this is easier and less painful for the patient. In addition, over time, the donor site providing BMSCs is prone to pain and stiffness, whereas ADSCs have a much lower incidence (88). In addition to application in tendon injury, MSCs can also be used to treat fractures, osteoarthritis and other disease (226). MSCs have shown satisfactory results in tissue engineering and regenerative medicine, not only in promoting tissue regeneration, but also in restoring the tissue to its original biomechanical function to the greatest extent possible (227).

The roles and mechanisms of MSCs may involve promoting angiogenesis, cell proliferation and differentiation and collagen formation and decreasing inflammation during tendon repair (129,228). MSCs participate in localized anti-inflammatory response in the early stage following tendon damage. Anti-inflammatory factors and MSCs-EVs are hypothesized to be intercellular messengers in immune regulation (229). They interact with various types of immune cell, including T and B lymphocytes, natural killer cells, macrophages, neutrophils and monocytes (230). MSCs promote angiogenesis mainly by releasing VEGF and Exos (157). During the remodeling and collagen production phase, in vitro experiment indicated MSCs enhance their ability to differentiate into tenocytes by co-culture with growth factors and TSPCs, thus promoting tenocyte proliferation and differentiation, as well as collagen fiber production and ECM remodeling (198,217).

The present review summarized the functions and mechanisms of MSCs in tendon repair but there are still some issues with the clinical application of MSCs that need to be addressed. To the best of our knowledge is no consensus on practical considerations regarding the source, dose, administration technique and timing of MSC usage (231). Although the commonly used sources of MSCs are adipose and BM tissue, there are still debates on the applications of both; for example, whether ectopic bone can develop after application of BMSCs for treatment and whether the application of ADSCs involves ethical issues. MSCs isolated from young living sources survive longer, secrete more Exos and have a broader differentiation capacity than MSCs isolated from older tissue (232). Therefore, MSCs isolated from embryonic sources may be promising therapeutic tools for tendon repair and regeneration (233). The current mode of administration is direct injection of MSCs at the site of injury, but a study (234) has shown that intravenous MSCs can promote better interaction with the immune system and initiate systemic anti-inflammatory effects. However, due to small sample, more research is needed on intravenous MSC administration (235). Although study have shown that MSCs show satisfactory therapeutic effects when co-cultured with growth factors or TSPCs, few studies have compared the effects of culture conditions (236,237). Therefore, randomized controlled trials are required (238). In addition, current treatments to promote tendon repair lack standardization so treatment results may differ. Therefore, future studies should investigate the effects of clinical treatment with MSCs alone to develop standardized treatment modalities, which may lead to more uniform outcomes.

In conclusion, MSCs are a promising cell therapy to promote tendon healing and understanding of the functions and mechanisms of MSCs in tendon healing can help improve its efficiency. However, further studies are required to maximize the therapeutic value of MSCs.