Tendons are specialized tissues that connect muscle to bone and transmit the forces generated by muscle to bone, resulting in joint movement. Tendon injuries are common and affect a substantial portion of recreational and professional athletes and those in many occupations involving repetitive work [16, 37, 60, 79, 102]. Tendinopathy (often called tendinitis or tendinosis) is the most common tendon disorder [86, 99]. It is characterized by activity-related pain, focal tendon tenderness, and decreased strength and movement in the affected area. The histological features of tendinopathy are further described in the current study. Tendinopathy can occur in almost any tendon. Common examples include plantar fasciitis, Achilles tendinitis, patellar tendinitis, tennis elbow, golfer’s elbow, and supraspinatus tendinitis. Tendinopathy is poorly understood and has many described remedies with very little evidence to support their efficacy. One of the reasons there are very few, if any, good treatments for tendinopathy is lack of knowledge regarding its pathogenesis.

We summarize recent cellular and molecular findings in tendinopathy to identify potential preventative and treatment strategies and specific areas needing further investigation.

We performed a systematic review of peer-reviewed, original English language papers published on the etiology, histopathology and molecular biology/pathology of tendinopathy using Ovid MEDLINE and PubMed database from 1950 to November 2007. Keywords used in the search were: tendinopathy; pathogenesis; tendon cells; extracellular matrix; proteoglycan; metalloproteinases. Subheadings used in the search were: etiology; pathology; chemistry; metabolism. This analysis revealed 441 papers (432 papers from 1997–2007), of which 86 met our selection criteria. From this analysis we came to the following conclusions:

The histopathologic changes in tendinopathy are well-established (Table 1) (Fig. 1) [101]. Normal tendon is brilliant white in color and has a firm fibroelastic texture. In contrast, tendinopathic tendon is grey or brown, and is soft, thin, and fragile [57]. Microscopically, the collagen bundles are disorganized, there is increased ground substance, and the nuclei are darkly stained and round and found in increased numbers. This contrasts to the well organized parallel collagen bundles found in normal tendon with spindle-shaped tenocyte nuclei arranged in parallel alignment [46, 54, 68, 89]. At the electron-microscopic level the collagen fibers in tendinopathic tendon are angulated, vary in diameter and orientation, and have bubbles. There are changes consistent with hypoxia, including lipid vacuoles, enlarged lysosomes, and degranulated endoplasmic retinaculum [54]. Rarely have inflammatory cells been identified in tendinopathic tendon [12, 46, 54, 89]. Tendinopathic tendon often has infiltrations of vascular and small blood vessels [4, 46, 68, 85], an upregulation of vascular endothelial growth factor (VEGF) [80, 84], and ingrowths of small nerves [4, 64, 93]. These pathologic changes are consistent with “degeneration” and attempts of “regeneration”.

Table 1  Comparison of normal and tendinopathic tendon by macroscopic, light and electron microscopic and ultrasound findings [3, 4, 17, 28, 46, 54, 64, 68, 85, 89]

Findings	Macroscopic	Grey-scale or color and power doppler ultrasound	Light microscopic	Light microscopic
Normal tendon	• Brilliant white • Fibroelastic firm texture	• Parallel hyperechoic or bright white line • Regular uniform fiber structure	• Organized parallel collagen bundles • Spindle shape tenocyte nuclei • Nuclei parallel alignment	• Densely packed collagen fibers • Uniform in diameter and orientation of collagen fibers
Tendinopathy tendon	• Grey or brown • Tissue is thin, fragile and disorganized • Loose texture	• Localized widening of the tendon • Local hypoechoic areas • Irregular fiber structure • Neovascularization correlated with tendon changes	• Disorganized collagen bundle • Increased ground substance consisting of proteoglycans and glycosaminoglycans (GAG) • Large mucoid patches and vacuoles between fibers • Round with darker-staining tenocyte nuclei • Markedly increase number of nuclei with loss of parallel alignment • Increase of vascular and nerve ingrowths	• Angulation, bubble formation of collagen fibers • Variations in the diameters and orientation of collagen fibers • Hypoxic changes in tenocytes (lipid vacuoles, enlarged lysosomes and degranulated endoplasmic retinaculum)

Fig. 1A–B  The normal tendon and tendinopathy tendon is shown. (A) Normal tendon has parallel, longitudinal architecture with scattered elongated tenocytes. (B) Tendinopathy tendon shows disorganized collagen architecture, high cellularity of rounded tenocytes. (This figure was published in Soslowsky LJ, Thomopoulos S, Tun S, Flanagan CL, Keefer CC, Mastaw J, Carpenter JE. Neer Award 1999. Overuse activity injures the supraspinatus tendon in an animal model: a histologic and biomechanical study. J Shoulder Elbow Surg. 2000;9:79–84, © Elsevier 2000, with permission.)

---

Immunohistochemical studies demonstrate the number of substance P (SP)-positive nerve fibers are higher in painful tendinopathic samples compared to normal tendon samples [64, 93]. Microdialysis demonstrate higher neurotransmitter glutamate levels in chronic painful Achilles and patellar tendinosis compared with pain-free normal control tendons [2]. Microarray studies also report an increased mRNA of glutamate in healing rat Achilles tendon [76] and tendinopathic supraspinatus in rats [75]. The presence of neovascularization and innervation, and the increase of neurotransmitter in tendinopathy may be part of the reason tendinopathy patients often have chronic pain [1, 3, 82]. Danielson et al. [32, 33] recently reported the tenocytes in tendinosis patellar tendon exhibited more immunoreactions for adrenergic receptors and catecholamine. These findings are of relevance as studies have demonstrated stimulation of adrenergic receptors can lead to cell proliferation and/or cell degeneration and apoptosis [6, 23].

Fibroblast-like cells are the major cell type in tendons, and have been histologically classified as elongated tenocytes or ovoid tenoblasts [25, 53]. These cells are important for maintenance of healthy tendon as they can proliferate, produce collagen and maintain the appropriate extracellular matrix [19, 53]. Ovoid tenoblasts – often described as an immature or activated form of tenocytes – have a higher proliferation index and apoptosis index than those of elongated tenocytes [25]. There is evidence that changes occur in the cells (tenocytes appear rounder, proliferate, become necrotic or apoptotic, and have an increased expression of local insulin-like growth factor-1 (IGF-1)) before the overt development of tendinopathy [8, 28, 96]. Other cell types, such as synovial-like cells, smooth muscle cell and endothelial cells, can be found in the endotenon and epitenon of tendon.

Transformation of tendon cells towards a fibrochondrogenic phenotype has been observed in torn rotator cuff tendons [46], and Archambault et al. [7] reported tendinopathic rat supraspinatus tendon had increased expression of cartilage genes such as col2a1, aggrecan, and sox9. Tenocytes from the site of tendinopathy also produce abnormal amounts of collagen III, commonly associated with wound healing, even when the repetitive motion is no longer present [69].

Using a combination of cell markers and flow cytometry analysis, Bi and colleagues [18] identified a unique cell population termed tendon stem/progenitor cells (TSPCs) in human and mouse tendons with several universal stem cell characteristics including clonogenicity (the ability to form clones), multipotency (multidifferentiation potential towards osteogenesis, adipogenesis, and chondrogenesis) and self-renewal capacity (higher doubling capacity than bone marrow stromal cells [BMSCs] from same sources). Moreover, the isolated human or mouse TSPCs could regenerate tendon-like tissues with either HA/TCP or Matrigel in vitro, whereas mouse dermal fibroblasts transplanted with Matrigel did not form any tissue. When transplanted with HA/TCP onto the surface of mouse calvaria, human TSPCs formed condensed collagen fibers inserted into the bone which were similar to Sharpey’s fibers [18]. Terminal differentiation of single cells selected from a group of equivalent precursors may be random, or may be regulated by external environment/signals. Some specific microenvironments/additional signals are essential for tenocyte differentiation and proliferation [94, 114]. A recent report suggests an extracellular matrix (ECM)-rich niche, organized in part by biglycan (Bgn) and fibromodulin (Fmod), controls the self-renewal and differentiation of TSPCs [18]. Depletion of Bgn and Fmod in mice leads to decreased expression of the tendon marker scleraxis (Scx) and of Type I collagen, and increased sensitivity to BMP2 in TSPCs when compared to cells from wild-type mice [18]. BMP12 acts as signaling molecules during embryonic tendon/ligament formation in animal experiments [67, 112, 114] and can stimulate patellar tendon fibroblasts proliferation in humans [40]. Expression of Scx, a member of the basic helix-loop-helix (bHLH) superfamily of transcription factors in the progenitors and cells of all tendon tissues [94], is crucial for differentiation of all force-transmitting and intermuscular tendons in Scx^−/− mice [78] and for activation of the COL1a1 gene in mouse tendon fibroblasts in an in vitro study [63]. These studies indicate the genes that coordinate with the matrix regeneration are also important for regenerating TSPC’s.

The extracellular matrix (ECM) is a complex structural entity surrounding and supporting cells. The ECM is composed of three major classes of biomolecules: structural proteins (collagen and elastin), specialized proteins (eg. fibrillin and fibronectin), and proteoglycans [21]. The interaction between tendon cells and EMC is bidirectional: alteration of EMC may be initiated by tendon cells [8, 28], and changes of EMC microenvironment may also lead to cell proliferation, migration, apoptosis, and morphogenesis [108]. The maintenance of the tendon matrix has important consequences for the ability of the tendon to resist mechanical forces and to repair response to injury [58]. Some authors suggest an imbalance in the synthesis and degradation of ECM leads to structural deterioration and degeneration of the tendon [9, 20, 48].

Proteoglycans are protein/polysaccharide complexes in the ECM that trap water and affect the viscoelastic properties of the tissue, helping the tissue resist compressive forces [115]. Proteoglycans consist of a protein core with attached glycosaminoglycans (GAGs). Tendon contains a wide variety of proteoglycans, including the large aggregating proteoglycans and a variety of small leucine-rich proteoglycans (SLRPs) [116]. Proteoglycans and their constituent GAGs can influence many important physiological processes in tendon, including ion transport, water retention, the diffusion of nutrients, mediating cell-matrix interactions, resistance of compression and sequestration of growth factors and enzymes in the matrix [43, 116]. Animal studies demonstrate biglycan may serve both a structural [109] and a signaling role [77] in developing tendon and biglycan and collagen VI are coexpressed in tendon development [61]. It is possible the presence of tendon fibrocartilage proteoglycan/glycosaminoglycan in normal supraspinatus is part of a normal functional adaptation to mechanical forces in tendon [91].

Studies from gene-knockout mice demonstrate targeted deletion of decorin, fibromodulin, or lumican or deletion of both lumican and fibromodulin cause abnormal collagen fibril and fibril bundle morphology [24, 31, 72], indicating these proteoglycans play an important role in regulation of collagen fibril formation and maturation. Although increases of aggrecan and biglycan in tendons have been implicated in early tendon healing process [104], the changes of other type of proteoglycans in tendinopathy are different to those in normal and healing tendons (Table 2). Very little is known about the changes in noncollagen glycoproteins in tendinopathy. Fibronectin and tenascin-C are key factors in the tendon repairing process by promoting fibroblast migration, and adhesion of fibroblasts to fibrin [41, 106]. In addition to the genomic study [73] showing variants of the tenascin-C gene are associated with Achilles tendon injury, a persistent increase in expression of fibronectin and tenascin-C has been reported in tendinopathy (Table 2) and may contribute to the pathogenic matrix remodeling in tendinopathy [90, 105].

Metalloproteases (MMPs) are a large family of enzymes that degrade all tendon matrix components, and these enzymes and their inhibitors play a major role in the degradation of matrix during development, adaptation, and repair [87, 108]. In addition to their important role in normal physiological events in tendon homeostasis and repair [108], these enzymes may be key effectors of the pathological processes in tendon disease [26].

A disintegrin and metalloproteinase (ADAM) is a transmembrane protein that contains a disintegrin and metalloprotease domain and, therefore, it potentially has both cell adhesion and protease activities [83]. A novel family of extracellular proteases, a disintegrin-like and metalloprotease with thrombospondin motifs (ADAMTS), apparently plays an important role in proteoglycan turnover in tendon [50, 103]. Nineteen ADAMTSs have been identified so far, but many of them remain to be fully characterized [50]. ADAMTS-2, -3, and -14 function as key regulators of collagen fibril assembly [27]. ADAMTS-1 and -4 are capable of cleaving certain matrix proteoglycans such as versican, brevican, and aggrecan [81, 113]. ADAMTS-4 also cleaves nonproteoglycan ECM components such as fibromodulin and decorin [55]. However, there is very little knowledge about the roles of ADAMs and ADAMTS played in tendinopathy (Table 3).

The activities of MMPs are normally tightly controlled in vivo, with regulation at the levels of transcription, translation, activation, and inhibition [87]. The activities of MMPs are inhibited by a family of tissue inhibitors, the tissue inhibitors of metalloproteinases (TIMPs). This family contains four human gene products, namely TIMP1, TIMP2, TIMP3, and TIMP4 [71]. All TIMP members inhibit MMP members to varying degrees [13], however, only TIMP-3 is a potent inhibitor of members of ADAM and ADAMTS [5, 45, 56]. The activity of MMPs is inhibited reversibly by TIMPs in a noncovalent fashion in a 1:1 stoichiometry [22]. Decreases in TIMP-2, -3, and -4 are consistently found in tendinopathy (Table 3). MMP inhibitors prevent the activation of MMP-13 and inhibit pericellular matrix degeneration and the loss of material properties in stress-deprived tendons in vitro [9].

The data suggest the balance between metalloproteinases and their inhibitors is likely essential in the maintenance of tendon ECM homeostasis, and an imbalance may result in uncontrolled tendon damage.

The aim of this review was to identify recent advances in the understanding of tendinopathy, particularly from a cell and molecular biology perspective. There has been much new information and there are many gaps in our understanding of the pathogenesis of tendinopathy. These gaps are most glaring from a timeline perspective, or the sequence of events. The research uses small windows to look into a complex event that changes in location and time.

Our current hypothesis is that tendinopathy is induced when tendon cells experience a large volume of repetitive load (Fig. 2). Tendons of certain anatomical locations are more susceptible as are individuals who are older, heavier and male in those genetically predisposed to tendinopathy. There is debate as to whether a loss of integrity of the matrix, or the cells in the tendon matrix, initiate the changes in tendinopathy. Given the high incidence about (70%) of tendinopathy in the shoulders of “elite” asymptomatic swimmers [98] we favor the later. High doses of cyclical strain induce genes for two major pathways (Fig. 2): (1) oxidative stress – apoptosis; and (2) cartilage-like genes.

Fig. 2  A conceptual diagram shows our current hypothesis on the pathogenesis of tendinopathy.

---

The interaction between these two pathways is undetermined. Once tendinopathy is initiated the tendon cells become rounded and apoptotic and produce a matrix that contains less Type I collagen and is more cartilaginous and “immature” in nature. Once the normal cell – matrix complex is disrupted, “relative” stress deprivation is induced [8, 9] and metalloproteinase matrix destruction is initiated. The enthesis becomes more painful, more vascular and mechanically inferior to normal tendon. The histologic data suggest attempts at repair with vascular and neuronal infiltration occur and that if the matrix is not adequately repaired the adrenergic responses associated with this process may lead to persistent pain and/or complex regional pain syndrome. The new information on tendon stem cells is important as it implies these cells, if induced and directed to the correct location, can reverse the degenerative process.