Background

As a metabolic disease, diabetes mellitus (DM) is associated with many diseases and complications, such as retinopathy [1], nephropathy [2], osteoporosis [3] and impaired wound healing [4]. Currently, the influences of DM on the musculoskeletal system have been noted. The negative impacts of DM on the tendons may contribute to tendinitis [5], tendon ruptures [6], adhesive capsulitis [7] and even the diabetic foot [8].

Epidemiological studies revealed higher prevalence of Dupuytren’s disease, carpal tunnel syndrome and shoulder adhesive capsulitis in both type I and type II diabetic patients with poor glycaemic control [9, 10]. Patients with insulin-dependent diabetes were significantly associated with chronic rotator cuff tendinitis and limited range of motion (ROM) in the shoulders [5].

Recently, some studies reported the tendon alternations in patients with diabetes following imaging examinations, such as ultrasound, computed tomography (CT) and magnetic resonance imaging (MRI) [11]. In this review, we summarized the current evidence on the alternations of tendons from both clinical patients and animal models of DM. The clinical manifestation, imaging examination, biomechanical properties, histopathological features and cellular and molecular changes in tendons of patients with diabetes were summarized. The aim of this review was to explore the links between the DM and the underlying cellular and molecular mechanisms of diabetic tendon alternations.

Clinical manifestation of tendon pathologies in diabetic patients

Patients with DM with poor glycaemic control often suffer from chronic pain, limited range of motion (ROM) of the joints and have a higher risk of tendon tears [10]. The prevalence of connective tissue diseases, such as Dupuytren’s disease, trigger finger, carpal tunnel syndrome, rotator cuff tears and shoulder adhesive capsulitis (frozen shoulder), were increased in both patients with type 1 and type 2 diabetes [9, 12, 13]. These diabetic patients with stiff shoulder have poor prognosis of nonoperative treatment [14].

Symptomatic rotator cuff tears and acute Achilles tendon ruptures are common in patients with DM [6, 15]. After surgical repair, these patients showed a limited ROM and a higher incidence of re-tears [16, 17]. For the ruptured Achilles tendons, patients with DM have higher rate of postoperative infection and poorer tendon healing [6, 18]. Clinical research revealed that patients with diabetes with poor glycaemic control easily develop severe lower limb infection and even diabetic foot, and are at a higher risk of lower extremity amputation [8]. There is increasing evidence that reduced ankle joint ROM, static stiffness of the Achilles tendon and the diabetic neuropathy may be responsible for the development of diabetic foot [19–21].

However, not all the patients with DM are symptomatic, some of them are asymptomatic, even with an increased thickness or structural abnormalities in supraspinatus and biceps tendons and Achilles tendons found with imaging exams [11]. It is suggested that the insensitive neuropathy may reduce or even block the transmission of pain signal, thus leading to asymptomatic conditions in some patients with DM [21].

Imaging findings

Many imaging techniques, such as ultrasonography, MRI and CT, have been used in investigating the changes of tendons and ligaments in patients with diabetes. Currently, there is no longitudinal study to examine the thickness and the structural abnormalities in the tendons of patients with diabetes.

In a clinical study, Giacomozzi et al. [21] discovered a tendency for increased thickness of the Achilles tendon in patients with DM compared to healthy subjects, although there was no statistical difference. However, in diabetic patients with neuropathy, the thickness of the tendons was increased significantly. In another study, Akturk et al. [22] found that the average thickness of the Achilles tendon was increased in patients with type 2 DM by ultrasound, but the significance was only among the female groups. Batista et al. [8] discovered disorganization of the Achilles tendon fibers as well as calcification in some asymptomatic individuals with diabetes by ultrasound, but they were not able to demonstrate any statistical relationship between the structural abnormality with age, length of time of diagnosis, or diabetes management, and no data about the location of disorganization and calcifications in the Achilles tendon were provided. In addition, the anterior-posterior thickness of the Achilles tendon was greater in the control subjects than in the diabetic patients, but still lacked a statistical analysis. The latest study showed the degenerative features and calcifications localized at the enthesis of the Achilles tendons were significantly increased in the asymptomatic patients with diabetes [12]. Papanas et al. [23] found a significantly greater volume of the Achilles tendon in patients with DM than controls in both male and female groups, while the thickness of the Achilles tendon showed no difference.

Recently, a study by Abate et al. [24] revealed the thickness of the supraspinatus tendon and the biceps tendon in asymptomatic elderly patients with diabetes, which were greater than that in controls by sonographic evaluation. Besides, they also found more sonographic appearances of degenerative features in the rotator cuff and biceps in DM subjects. Another study also discovered that the quadriceps tendons had more buckling in diabetic patients with MRI examination [25].

All these studies presented abnormal changes of the thickness and structure of collagen fibers in diabetic tendons, although the association between diabetes mellitus and tendon alterations could not be sustained due to methodological drawbacks [11]. However, all the studies reported the increasing trend of thickening tendons in diabetic patients. Overuse and increasing BMI in DM subjects are considered as the major pathogenic factors for the increased tendon thickness and reduced ROM of joints [12].

Biomechanical changes

A few reports focused on the biomechanical properties of tendons in DM patients. In the DM animal models, most of the studies indicated DM having negative impacts on the mechanical properties of the tendons [26–32]. Most of the tendons were isolated from the diabetic animal models induced by streptozotocin (STZ) or other methods. The main findings are summarized in Table 1.

Table 1 Characteristics of studies in biomechanical properties of diabetic tendons
Full size table

In the STZ induced diabetic rat model, a decreased Young’s modulus which was calculated as stress divided by strain was found in both the diabetic Achilles and patellar tendons, as well as the accumulation of advanced glycation end products (AGEs) in tendon extracellular matrix (ECM) [26, 27]. In the study by De Oliveira et al. [27], besides the decreased elastic modulus, specific strain patterns were also seen in the subjects with diabetes. Lehner et al. [32] also discovered the significantly decreased maximum tensile load of Achilles tendon in the STZ treated animals. All these data indicated a tendency of impairing biomechanical properties in the diabetic tendons. De Oliveira et al. [29] compared the sedentary and aerobic physically trained diabetic rats with the matched controls, the elastic modulus of the sedentary diabetic group was decreased when compared to the trained diabetic subjects, meanwhile the deformation at maximum force and energy/tendon area was significantly decreased.

In the acute rotator cuff repair diabetic rat model, the supraspinatus tendon-bone complex of animals with diabetes demonstrated a significantly reduced mean load-to-failure and the stiffness of constructs [30]. Ahmed et al. [31] discovered that the injured tendons of the diabetic group exhibited 35 % lower stiffness compared to the control tendons. In the intact diabetic tendons, the stiffness was 19 % lower than that of the non-diabetic controls. In the mice with type 2 diabetes, the mice fed with high fat diet showed decreased maximum force of uninjured flexor digitorum longus (FDL) tendons compared with lean diets at 24 weeks. What’s more, impaired biomechanical properties of FDL tendon in high fat diet mice at day 28 post injury were also observed, and the subsequent histological examination of the injury site showed a smaller area and lower cell content which showed a defective repair in the mice with type 2 diabetes [33].

More evidence suggested that the micro-structure of the tendons was altered in the diabetic patients. The latest study by Yoshimasa et al. [34] revealed the accumulation of the pentosidine (a well-characterized AGEs) in the diabetic tendons. They also found the ultimate stress of the Achilles tendon decreased as the amount of pentosidine deposition increased in the diabetic rats simultaneously, and suggested that accumulation of AGEs might deteriorate the biomechanical properties of the tendons in patients with diabetes.

In the diabetic subjects, a markedly increased accumulation of AGEs in the tendon was noted, forming covalent cross-links within the collagen fibers and interacting with the receptor for advanced glycation end products (RAGE) on the cell surface, which may lead to deterioration of the biomechanical properties of the diabetic tendons [26, 35].

On the contrary, the glycated Achilles tendons of rabbits in vitro exhibited significantly increased mechanical properties [28]. It has also been shown that the hyperglycemia could reduce proteoglycan levels in the tendon cells in high glucose condition exposure to AGEs. Decreased proteoglycan levels could increase intracellular water and cellular edema, and these alterations may lead to the increased stiffness in diabetic tendons [36].

All these studies indicated a tendency of impairing biomechanical properties in the diabetic tendons. These reduced biomechanical properties may lead to the easy tendon rupture and ROM of the affected joints limited in patients with DM.

Histopathological changes

Gross observation and histological features

The diabetic tendons appeared fragile, atrophic and demonstrated a yellowish discoloration compared to the more robust, healthy tissue observed in the healthy control subjects [26, 30]. In the STZ induced diabetic rat model, the thickness of the Achilles tendon increased [37]. In the injured diabetic tendons, cross-sectional area of the healing site decreased compared to the control group [30].

Pathological changes were seen in the diabetic tendons. In contrast to the tight, parallel, bundled appearance in the healthy Achilles tendons, some collagen fibres in the diabetic tendons were separated and lost their parallel orientation, unequal and irregular with a decrease in fibre diameter and density of collagen. In addition, there were some chondrocyte-like cells within the Achilles tendon in the subjects with diabetes and the micro-tears were observed [38]. We have also successfully built an STZ-induced DM rat model, and similar alternations were also observed in diabetic patellar tendons with H&E staining (Fig. 1d and e). Besides the disordered collagen fibers in diabetic tendons, we also found the tendon cells in diabetic tendons surrounding the micro tear sites were rounding changed (Fig. 1e). In a clinical study with ultrasound examination, the calcifications localized at the enthesis of the Achilles tendons were found in asymptomatic patients with diabetes [8]. What’s more, the significant increase of fibrocartilage metaplasia and granulation tissue hyperplasia were reported in patients with diabetes with stenosing flexor tenosynovitis compared to those subjects without diabetes [39]. Local hypoxia, which is induced by rupture of collagen fibers, and disturbance of microcirculation, may contribute to calcifications [24]. Some studies suggested that calcifications might be attributed to the erroneous differentiation of tendon-derived stem cells (TDSCs) towards to osteogenesis [40]. In our study, we found that the expression of osteopontin (OPN) was significantly increased in diabetic tendons when compared to the healthy tendons with immunohistochemical staining (IHC) (Fig. 1c and f).

Fig. 1
figure 1

Normal structure of patellar tendon from healthy SD rats (H & E staining, A×200; B ×400). Altered structure of patellar tendon from STZ induced diabetic rats (H&E staining, D ×200; E ×400). Negative expression of OPN in healthy tendon (IHC staining, C×400). Positive expression of OPN in diabetic tendon (IHC staining, F×400). (unpublished data)

Full size image

In the STZ-induced DM rats, the increased type I collagen was observed with a disorganized manner. This may attribute to the increased cellularity in diabetic tendons, and the cells with excessive accumulation of collagen fibers and other extracellular matrix components may lead to tissue fibrosis resulting in the thickening of diabetic tendons [37, 41]. In the tendon-healing model, less organized collagen fibers were found at the tendon-bone interface in the subjects with diabetes [30]. The tendons in the rats with diabetes had lower regenerating activity, with a smaller transverse area, poorly organized collagen fibers, and decreased vascularity [37]. Most of the collagen fibers were of yellowish red colour and arranged irregularly, the collagen III-like fibers were less at the rupture site of the diabetic animals while the highest density of collagen III-like fibers were found in the controls [31]. The decreased transverse area as well as collagen III production might hamper the tendon repair.

Vasculopathy and neuropathy

It was reported that an adequate circulation and neuronal supply are beneficial for collagen synthesis at the injured site of the tendons [42]. However, reduced neovascularization in the degenerated tendons was found in type 2 diabetes patients [43], which is consistent with the findings in the injured diabetic rat tendons [41]. The reduced numbers of macrophages in the diabetic tendons impaired angiogenesis and tendon healing since macrophages mainly take part in the inflammatory process are also important in initiating angiogenesis [41]. Down-regulated expression of peroxisome proliferator activated receptor gamma (PPAR-γ) has been reported, which can reduce the vessel and nerve ingrowth involved in the pathogenesis of diabetic tendinopathy [38]. The insensitive neuropathy in the diabetic subjects could lead to excessive use of tendons, and might be a possible reason for those patients who were clinically asymptomatic and only had thickening and disorganized changes in tendons.

One study reported the vascular hyperplasia in human patellar tendinosis, and the increased prevalence of mast cells may be associated with these changes [37, 44]. In the STZ induced rat model, the average number of blood vessels per field was statistically higher in the diabetic group, and the blood vessels were often present in the tendons of the diabetic group. The cross-sectional area of the blood vessels was significantly larger and the VEGF expression was significantly higher in the diabetic rats [37]. The higher expression of VEGF in diabetic tendons may be related to the chronic degeneration of tendons, hypoxia and increased mechanical load, and the accumulation of AGEs may also be involved [45–47].

Inflammatory responses

In the diabetic patients with stenosing flexor tenosynovitis, the presence of a small number of inflammatory cells was found in the hyperplastic granulation tissue [39]. In the diabetic rats, a higher accumulation of nitrite/nitrate (NOx) was also found indicating chronic inflammatory status in the diabetic tendons [37]. However, in the injured diabetic tendons, the inflammatory factors were reduced compared to the normal controls, which resulted in a reduced numbers of PMNs and macrophages, and aborted tendon healing [41]. A transitory inflammatory response is necessary for the normal progression of tissue healing by releasing cytokines that regulate chemotaxis and fibroblast proliferation and extracellular matrix synthesis and remodeling [48]. Reduced inflammatory response in the injured diabetic tendons might account for the impaired normal tendon healing capacity in DM condition. Chronic inflammatory status was found in the DM condition, and this might be one of the proper reasons for the tendons' alternations in DM [37, 39]. Many studies also demonstrated that inflammatory enzymes, cytokines and reactive oxygen species (ROS) contribute to cell death and many diseases [49]. In the diabetic tendon, the chronic inflammatory condition is also harmful to tendon repair. Although there was still no report that decreasing the chronic inflammatory status could help to regain the normal inflammatory response in tendon, all these findings suggested that regaining the normal inflammatory response might be a way for DM tendinopathy treatment.

One study reported that the inflammation could promote mineralization of human bone marrow stem cells [50]. Although accumulation of calcium was found in both human and animal models, however, up to now, no studies reported the connection between inflammatory response and TDSCs differentiation.

AGEs

The excess of AGEs deposited in ECM is an important pathological change of DM, and it correlated with the histological and biomechanical changes of the diabetic tendons [26, 30]. The key characteristic of reactive AGEs is the formation of covalent cross-links among the collagen, which could be generated via the enzymatically driven and the non-enzymatic glycation or oxidation-induced pathways [10]. Once formed, the cross-links within the collagen fibrils could dysfunction the ECM of diabetic tendons and deteriorate the biomechanical properties [10]. Recently, one study reported the catalyst Fe2+ could accelerate the formation of AGEs in type I collagen, and lead to the glycosylation of collagen and other matrix proteins in diabetic tendons [51]. AGEs interact with RAGE on the cell surface and activate several critical molecular pathways including pro-oxidant events and up-regulation of inflammatory mediators; the intracellular accumulation of AGEs could lead to negative effects, such as inhibiting cell growth, destruction of nitric oxide and promoting apoptosis [52–54]. AGEs is also involved in the scar formation and cause of tendon cell death [36]. The tendon cells treated with high glucose concentrations showed the decreased proteoglycans and increased transforming growth factor β1 (TGF-β1) levels in an AGE-independent manner, and high TGF-β levels is reported to associate with scar formation in tendon and tenocyte death [55].

With histological analysis, tendons in subjects with DM showed degenerative features, and the inflammatory responses were accelerated with the accumulations of AGEs in the ECM, which deteriorated tendon biomechanical properties as well as the functions of tendon cells. However, the differences in the findings in vasculopathy may be attributed to the different duration of DM.

Cytological and molecular changes

Tendon cells

Tenocytes are the basic functional unit of tendon and secrete ECM maintaining tendon function [56]. In the diabetic tendons, the density of mast cells were four-fold higher than the controls, and mast cells play an important role in soft tissue remodeling through releasing fibroblastic factors and tryptase [37, 44]. The effects of hyperglycemia on the proliferation of different cell types, such as human periodontal ligament fibroblasts, notochordal cells, tenocytes and peritoneal fibroblasts, are divers [57–60]. It was reported that the proliferation of tenocytes were not affected by glucose concentrations up to 25 mM in vitro [61]. Recent studies reported that the proliferations of rat notochordal cells were significantly decreased in high glucose concentration while the autophagy and apoptosis were increased via high glucose-induced oxidative stress [59, 60]. And the senescence of young annulus fibrosus cells was also accelerated in such conditions [62]. The expression of sox9 and scleraxis (Scx) of primary human tenocytes were up-regulated and the collagen synthesis was increased when they were treated with low extracellular glucose [63]. There were also reports that the cell proliferation of tendon cells in diabetic rats was reduced [41], with reduced collagen III expression, and decreased collagen I (Col 1a1) expression in the injured diabetic tendons [31].

Tendon stem/progenitor cells (TSPCs)

Besides the tenocytes in the normal tendon tissues, there is a small population of TSPCs present in tendons. TSPCs have stem cell characteristics, including clonogenicity, self-renewing capacity and the multi-differentiation potential in vitro and in vivo [64]. TSPCs could promote tendon repair and regeneration, and might play important roles in maintaining tendon homeostasis [43, 65]. Mechanical loading could promote BMP-2 expression in TDSCs, which could induce osteo-chongdrogenic and adipogenic differentiation of TDSCs in vitro. Our recent work found that the properties TDSCs isolated from patellar tendons in the collagenase-induced (CI) tendinopathy rat model were altered with higher osteo-chondrogenic differentiation potential, lower tenogenic differentiation potential and lower proliferative capacity, when compared to healthy TDSCs [40]. These findings suggested that erroneous differentiation of TDSCs might play a role in the pathogenesis of tendinopathy.

There are a few studies which investigated the properties of TSPCs in diabetic tendons. Recently, we isolated TDSCs from the STZ-induced diabetic rat patellar tendons, and found that the mRNA expression of Scx and Col1a1 were significantly lower in the diabetic TDSCs than that of the healthy TDSCs. High glucose could suppress the mRNA expression of Scx, tenomodulin (Tnmd) and Col1a1 of healthy TDSCs in vitro. One study reported that there were insulin expressing cells in diabetic tendon and these cells express tendon progenitor markers such as Scx and nestin [32]. The brain and thymus have been reported as the extra-pancreatic insulin producing tissues upon glucose stimulation [66]. Our unpublished data has also found that the mRNA expression of insulin in diabetic TDSCs was significantly lower than that in the healthy TDSCs. Taken together, TSPCs might participate in the process of the pathological alternations in diabetic tendons and might be a therapeutic target for diabetic tendinopathy [67].

Metalloproteinases (MMPs)

MMPs play important roles in ECM network remodeling [68]. Previous studies revealed the altered MMPs levels during tendon healing [69, 70]. In a diabetic rat model, both diabetic and healthy tendons exhibit strong expression of MMP-13 and MMP-3 in tenocytes. Meanwhile, in the tendon healing model, MMP-13 expression was upregulated and MMP-3 expression decreased [31]. Both MMP-3 and MMP-13 participate in collagen degradation; MMP-3 could activate other MMPs and is also involved in ECM remodeling [61]. Up-regulated mRNA expression of MMP-9 and MMP-13 in tendon cells treated with high glucose concentration was reported, as well as the enzymatic activity of MMP-9; these results indicated that high glucose concentration might induce collagen degradation and lead to a negative impact on the tendon structure [61].

Conclusions

Many studies in both patients with diabetes and animal models suggested a strong association of diabetes mellitus and alternations in tendon quality (Table 2). Excessive accumulation of AGEs, the altered inflammatory responses, neovascularization and insensitive neuropathy may account for the altered tendon biomechanical properties, cytological and molecular changes in the diabetic tendons. Formations of cross-links among collagens, increased tenocyte apoptosis, dysregulation of TDSCs differentiation, and altered expression of MMPs may be the underlying mechanisms of pathological alternations in diabetic tendons. Our latest findings speculated that the erroneous differentiation of TDSCs may be attributed to the pathological changes in diabetic tendons.

Table 2 Alternations in diabetic tendons
Full size table