Tendons connect muscle to bone and enable transmission of forces from contracting muscle to bone, resulting in joint movement. They possess the ability to adapt to changes in loading [1] and studies have shown that collagen synthesis is increased as a result of both acute exercise [2,3] and prolonged physical training [4]. The adaptation to loading can ultimately lead to increases in CSA and collagen content in chronically loaded tendons [5]. Despite this physiological ability to adapt, tendinopatic tendons represents a large and constantly growing clinical problem affecting both recreational and professional athletes as well as people involved in repetitive labour [6,7]. Years of research have unfortunately not provided much insight into the pathogenesis of chronic tendinopathy [8]. Indeed, the etiology of tendinopathy has been related to repeated micro strain below the failure threshold as an initiating stimulus for degenerative processes [9,10]. Other authors, however, have proposed that mechano-biological under-stimulation results in a degenerative cascade, through the production of a pattern of catabolic gene expression that leads ultimately to extracellular matrix degeneration [11]. Tendinopathy is characterized by activity-related pain, focal tendon tenderness, and decreased local movement in the affected area [12,13]. The general opinion is that no inflammatory cells are present in the tendinopathic tissue [14] and that tendinopathy is the result of a degenerative process with collagen disorganization, collagen fibre separation, increased cellularity, neovascularization and focal necrosis [15].

The major finding of the present study is that tendinopathy causes focal biochemical and morphological differences in the human Achilles tendon. The data obtained using TEM indicate that the structural composition of the t-area has a significantly increase in the number of smaller collagen fibrils compared to the h-area of the same tendon. This supports our hypothesis that localized structural differences are present in tendinopathic Achilles tendons, potentially as a result of an increased turnover of the tissue in an attempt to heal potentially injured tissue. This fits with previous findings where a site-specific loss of larger collagen fibrils and an increase of fibrils with a small diameter was observed in Achilles tendons after tendon rupture [26]. However the present findings differ somewhat from the results observed in another study from our laboratory [27], in which a significantly larger fibril area and a lower collagen fibril density was observed in patellar tendinopathy [27]. This discrepancy might partly be explained by the fact of the patella tendon functions more like a ligament (ligament patella) with the primary role of ensuring a fixed distance between the patella apex and the insertion of the tendon on to the tibia bone, in contrast to the Achilles tendon which functions more like a spring, providing a means of shock absorption via the connective tissue, during periods of muscle contraction and loading [28]. Since the function of the patellar and the Achilles tendon differ, it is likely that the structure of the tendons e.g. the cross-linking and length of fibrils etc. of the two tendons reflects this difference. A further explanation might be that the healthy tissue was taken from the same tendon in the present study, while Kongsgaard et al. [27] used control tissue taken from another tendon of healthy control subjects. The present design has as all study designs advantages and limitations. The advantage is that this design enables to investigate local differences in the tendon, the limitation is that no control tissue is available from tendons that never had any symptoms. Changes that occur in the whole tendon can therefore be overlooked. A previous study has shown that histological changes in the tendon were not only present at the site of rupture but also in the macroscopical normal part of the tendon, indicating that local alterations of the tissue might not necessarily be local [29]. Whether tendinopathy shows the same pattern is unclear and needs further investigation. However we believe that the present local differences between the t-area and the h-area are strong enough to justify the conclusions of the present study. To investigate if the increased number of small collagen fibrils could be due to genesis of collagen fibrils or degradation of previously much larger fibrils, the gene expression of scleraxis and tenomodulin was analysed. Both gene targets have previously been associated with tendon formation and development [30,31]. The absent difference of the expression of scleraxis and tenomodulin suggests that no fibrillogenesis took place in the t-area at this very late stage in the disease (Figure ​(Figure5).5). Furthermore a similar expression of Lysyl oxidase (LOX) in both regions of the tendon indicates that the tissue does not compensate for the localized structural changes by initiating cross-links to maintain the mechanical properties of the tendon. Previously findings showed that training increases the expression of LOX in healthy tendon tissue in rats [32]. The present data suggest that this adaptation does not take place in t-area of the tendon. Several abnormalities of the tendon structure have been investigated with histopathological analysis including fibre structure, fibre arrangement, nuclear rounding and cellularity [15]. In the present study a significant increased volume fraction of cells was observed in the tendinopathic area of the tendon using TEM. This is in line with previous animal studies of Soslowsky and colleagues [33-35], where rats ran with a velocity of 17 m/minute, 5 days/week, 1 hour/day, either uphill or downhill for a period of between 2-16 weeks. In such experiments, a decreased collagen fibre organization and increased numbers of cell nuclei were observed [36,37]. The present TEM analysis did unfortunately not allow for distinguishing between the cell types that were counted, and thus it was not possible to exclude that other cell types than just fibroblasts might have migrated into the t-area of the tendon. The significantly higher mRNA expression of both collagen I and collagen III in the t-area shows a higher collagen synthesis of the tendon. At the same time indicates the higher expression of MMP-2 and MMP-9 in the t-area an increased collagen matrix degradation. Together these findings display a higher collagen turnover in the t-area of the tendon. It has previously been shown that normal tendon tissue express matrix metalloproteinases and that a homeostatic turnover is necessary for tendon regeneration and maintenance [16]. A increased collagen turnover is usually associated with adaptation to exercise [2] or healing of the tendon [38]. It is still puzzling why the increased collagen turnover in the tendons of chronic patients like the present has not resulted in a decrease in symptoms or a healing of the tissue (symptoms range: 0.5-10 years Table ​Table4).4). However these data are confirmed in other studies also showing increased collagen turnover in tendinopathic tendons [22,24] and in tendon ruptures [39]. Alterations in proteoglycans have previously been associated with tendon pathology [40,41]. Proteoglycans and glycoproteins are essential for the maintenance of homeostasis of the ECM of the tendon and achieve this by regulating the collagen fibril assembly [42]. The upregulation of tenascin-C, and fibronectin is consistent with earlier findings [24,43]. The observed unchanged levels of versican contrasts earlier findings, where a significant down regulation of versican was observed in both tendinopathic and ruptured tendon tissue [40]. This discrepancy might lie in the medical history of the patients. The present patients had a very long history of tendon pain (range: 0.5-10 years). Thus the current biochemical situation in the tissue of these patients may have changed over time. The observed tendency to a decrease of decorin expression in the t-area of the tendon was contrary to our hypothesis. It has been previously shown that a down-regulation of decorin using anti-sense decorin injections improved ligament healing in rabbits [44]. Whether the present finding of a depressed decorin is part of the healing response of the tendons and thus beneficial for the patients is unclear. Additionally the increased expression of fibromodulin in the t-area may partly explain the observation of many thin collagen fibrils since fibromodulin participates in the matrix assembly leading to a delayed fibril formation and formation of thinner fibrils [45,46]. Treatment with injections of growth factors for tendon injuries has received much attention in recent years. Growth factors are polypeptide molecules that are decisive regulation of cell metabolism and cell proliferation and are associated with tendon healing [47-51]. Studies using local administration of bFGF [52,53], HGF [54,55] and IGF [56] have all shown beneficial effects in the healing process of tendon injuries, but not all injuries were tendinopathies though. Exogenous injections of bFGF in human patellar tendons have been shown to increase wound healing both in vitro and in vivo in patellar tendon models after surgery [52,53], and likewise, collagen type III and cell proliferation was increased after exogenous bFGF injections in patellar tendons after surgery in vivo [53]. Recently, our group showed that the cytokine IL-6 could act as a growth factor in tendon tissue [57]. Moreover, local injections of rhIL-6 have been shown to increase collagen synthesis in humans after one hour of exercise in the form of running, in healthy young men [57]. However, the issue as to whether local injections of IL-6 in tendinopathy patients may be beneficial to the healing process of a damaged tendon is still unknown. The pain that tendinopathy patients experience has previously among other been related to Substance-P, a neuro-peptide with various biological functions including pain transmission [58,59]. Since no difference in Substance P expression were observed between the two areas, the present data might indicate that other factors than substance P can be responsible for the pain in tendinopathic tendons. It is however also possible that the expression of Substance-P is increased in the whole tendon and therefore overlooked due to the previously mentioned limitations of the present design. Although the role of inflammation in tendinopathy is one that is often discussed, it has long been known that tendinopathy is primarily a degenerative condition, in which inflammatory cells in or around the lesion are absent. All markers of inflammation that were measured showed no upregulated expression in the t-area of the tendon (Table ​(Table3).3). This underlines that long-term tendinopathy is not primarily an inflammatory process, but rather an ongoing degenerative process. Although inflammation is absent in tendinopathy at this late stage, it does not rule out that an inflammation insult was present at the initiation of the tendinopathic process [60,61]. In fact, various inflammatory mediators like TNF-alpha, IL-6, IL-15, IL-18 have been shown to play a role especially in wound healing after injury [62] and in early stages of tendinopathy [23,63]. However, since inflammation can not be detected in later stages of chronic tendinopathy, the present data may indicate that the use of anti-inflammatory treatments e.g. NSAIDs are not relevant for chronic tendinopathy patients.

The pre-publication history for this paper can be accessed here: