The Effects of Type II Diabetes Mellitus on Tendon Homeostasis and Healing

Abstract

Over 300,000 tendon repairs are performed annually in the United States to repair damage to tendons as a result of either acute trauma or chronic tendinopathy. Individuals with type II diabetes mellitus (T2DM) are four times more likely to experience tendinopathy, and up to five times more likely to experience a tendon tear or rupture than non-diabetics. As nearly 10% of the US population is diabetic, with an additional 33% pre-diabetic, this is a particularly problematic health care challenge. Tendon healing in general is challenging and often unsatisfactory due to the formation of mechanically inferior scar-tissue rather than regeneration of native tendon structure. In T2DM tendons, there is evidence of an amplified scar tissue response, which may be associated with the increased the risk of rupture or impaired restoration of range of motion. Despite the dramatic effect of T2DM on tendon function and outcomes following injury, there are few therapies available to promote improved healing in these patients. Several recent studies have enhanced our understanding of the pro-inflammatory environment of T2DM healing and have assessed potential treatment approaches to mitigate pathological progression in pre-clinical models of diabetic tendinopathy. This review discusses the current state of knowledge of diabetic tendon healing from molecular to mechanical disruptions and identifies promising approaches and critical knowledge gaps as the field moves toward identification of novel therapeutic strategies to maintain or restore tendon function in diabetic patients.

INTRODUCTION

The dramatic increase in type two diabetes mellitus (T2DM) as part of the obesity epidemic is one of the most pressing health challenges facing the United States. As of 2015, more than 23 million people (nearly 10% of the U.S. population), were diagnosed as having diabetes.¹ It is estimated that over 7 million cases remain undiagnosed.¹ An additional 84 million people are considered pre-diabetic, a condition that if left untreated often progresses to T2DM within five years.¹ For those diagnosed with diabetes, the health care costs are staggering; in 2012, the average annual medical cost for a diabetic patient was over $13,000, with approximately $7,900 of the cost attributed directly to diabetes care and management.² Combined, the estimated annual health care burden of diabetes in the US is over $240 billion, a number that is predicted to rise precipitously in coming years.² By 2060, the number of adults with a diagnosis of diabetes is expected to triple to over 60 million, or nearly 18% of the U.S. population.³

The etiology of T2DM is multifactorial. Environmental, lifestyle, and genetic factors have all been shown to play important roles in the pathogenesis and progression of T2DM.⁴ Increased adiposity (abdominal fat in particular) is thought to be one of the major underlying causes of T2DM through its contribution to aberrant endocrine function and elevated secretion of pro-inflammatory cytokines.⁵ T2DM is characterized by systemic metabolic dysfunction including dyslipidemia and elevated plasma glucose levels (hyperglycemia). Insulin sensitivity in both hepatic and peripheral tissues is reduced, leading to decreased insulin receptor signaling and impaired glucose uptake. Combined, these alterations in systemic homeostasis wreak havoc on organ function. The effects of T2DM on the cardiovascular system, for example, are well characterized; patients with T2DM have decreased coronary artery diameter, increased presence of atherosclerotic plaques, and higher atheroma volume compared to non-diabetic patients.⁶ Subsequently, heart disease and stroke are the primary causes of death and disability among T2DM patients.¹

There is also a clear link between the metabolic dysfunction found in T2DM and musculoskeletal pathologies. T2DM accelerates the progression and severity of osteoarthritis^{7, 8} and increases fracture risk.^{9, 10} In addition to its effects on cartilage and bone, T2DM impairs both tendon homeostasis and repair following acute injury. Though the impact of T2DM on tendon function is a major complication of the disease, it remains poorly understood. In this review, we will cover recent advances toward our understanding of how T2DM affects tendon homeostasis both clinically and at the cellular level, as well as how T2DM complicates an already challenging healing landscape in the case of tendon injuries.

CLINICAL BURDENS, MANIFESTATIONS AND COMPLICATIONS OF T2DM ON TENDON HOMEOSTASIS

The effects of T2DM on tendon homeostasis in the absence of acute injury have been largely overlooked in the basic science and pre-clinical literature, which may be attributed to an insufficient appreciation for the magnitude of the clinical problem. However, chronic pathological changes to tendon structure represent a major clinical and social burden as patients experience decreased mobility and quality of life.^11–13. Importantly, Zakaria et al., found that T2DM patients were at a significantly greater risk of a tendon rupture requiring hospitalization, than non-diabetics.¹⁴ In general, these patients have increased incidence of disruptions in tendon homeostasis known broadly as diabetic tendinopathy, which can include functional changes such as impaired range of motion ^15–17 and structural abnormalities including loss of collagen organization (Figure 1),¹⁸ or thickening,^{16, 19, 20}, and calcification,²¹ all of which may increase the risk of tendon rupture. The precise incidence of diabetic tendinopathy is difficult to quantify and large differences in the relative prevalence are observed between studies, as outlined in the sections examining effects in specific tendons below. These discrepancies are likely due to differences in the size of the study patient populations, patient demographics and the specific tendon examined. However, the overwhelming majority of data demonstrates a clear increase in incidence of tendinopathy in T2DM, relative to non-diabetic patients.^{18, 19, 22–24} Furthermore, there is clear evidence that the burden of tendinopathy increases as a function of T2DM disease duration.^{18, 25–28} The silent, asymptomatic development of pathological changes to tendon structure, as well as the high proportion of un-diagnosed individuals with T2DM likely contribute to the advanced stage of diabetic tendinopathy that presents clinically. Because T2DM is a multifactorial disease, it is difficult to parse the relative contributions of a single factor to diabetic tendinopathy. However, elevated serum hemoglobin A1C (HbA1c) levels are strongly associated with tendinopathy development. More specifically, HbA1c >7% has been identified as an independent risk factor for tendinopathy.²⁹

Figure 1.

Decreased collagen matrix organization is observed in tendons from diabetic patients. Alcian Blue/ Hematoxylin/ Orange G (ABHOG) staining of tendon samples from non-diabetic and T2DM human flexor digitorum longus tendons. Images are representative of 5 T2DM patients and 3 non-diabetic patients. Scale bars = 50 microns.

Importantly, while this review is focused on T2DM, it is important to consider the potential confounding effects of obesity. Given that ~90% of T2DM patients are over-weight or obese,³⁰ it is likely that many of these pathological changes in the tendon are due at least in part to obesity. Interestingly, Titchener et al. demonstrated that the BMI of patients with rotator cuff (RC) pathology was significantly higher than control patients, however, patients with and without RC pathology in this study had BMIs in the overweight range.³¹ A systematic review by Gaida et al. found that 50% (n=14) of studies demonstrated a positive relationship between adiposity and tendon injury, while 46% (n=13) demonstrated no association, and one study (3.5%) reported a negative association in a subpopulation of study subjects.³² Taken together, these studies demonstrate that there may be a link between obesity and tendinopathy, but it is unclear to what extent obesity versus diabetes may drive these changes. Perhaps the strongest evidence delineating the contributions of obesity and T2DM come from assessing the impact of type I diabetes mellitus (T1DM) on tendon function. Both T1DM and T2DM have diminished insulin signaling, due to destruction of insulin producing beta cells in T1DM, and loss of insulin sensitivity due to chronically elevated insulin levels in T2DM. Therefore, comparing diabetic tendinopathy phenotypes between T1DM and T2DM may help inform which phenotypes in obese/T2DM patients are due to the altered insulin receptor signaling relative to obesity. Diabetic hand syndrome occurs in a similar proportion (up to ~50%) of T1DM and T2DM patients.²⁶ However, Al-Matubsi et al. identified lower rates of tendon-related hand pathologies in T1DM vs T2DM, including limited joint mobility,²⁵ while Rechardt et al. showed that chronic RC tendinitis was much more prevalent in T1DM men than T2DM or non-diabetics.³³ While there is some evidence of consistent disease manifestation between T1DM and T2DM suggesting a potential role for altered insulin receptor signaling in tendon pathology, there are clearly unique features of each disease that contribute to tendon pathogenesis.

While T2DM drives pathological changes in many different tendons, there is some evidence that not all tendons are equally susceptible to diabetic tendinopathy. For example, Lu et al. demonstrated that 54% of T2DM patients in their study experienced limited joint mobility in the hand (vs. 7% in the non-diabetic population),²² while Thomas et al. reported a 27.5% incidence of shoulder disorders in T2DM patients (vs. 5% in non-diabetics),²⁴ and Batista et al. reported nearly a 9-fold increase in altered Achilles tendon organization, via ultrasound imaging, relative to non-diabetics.¹⁸

In terms of the efficacy of tendinopathy treatments, there is strong evidence to suggest that T2DM modifies the responsiveness to treatment. Specifically, diabetic patients have a decreased response to corticosteroid treatment for trigger finger,^{34, 35} a higher proportion of these patients failed to progress following treatment which necessitated surgical intervention, and surgical outcomes were not successful in 13% of the T2DM patients.³⁵ As for the effectiveness of disease modifying treatments for T2DM, Abate et al. suggest that exercise in elderly T2DM patients has important benefits but can also increase the incidence of Achilles tendinopathy.³⁶

Below we explore the frequency and pathogenesis of diabetic tendinopathy in the three most commonly affected anatomical regions: the hand, rotator cuff (RC) and foot/Achilles tendon (AT).

Diabetic Hand Syndrome

DHS encompasses several discrete pathologies including limited joint mobility (LJM)/diabetic cheiroarthropathy, Dupuytren’s disease (DD), and flexor tenosynovitis/ trigger finger (FTS). Regardless of which pathology a patient is affected by, DHS is typically characterized by a positive prayer sign phenotype,¹⁵ with patients unable to touch their fingers and palms together. Importantly, a population based study by Chen et al., which included over 1.2M patients, calculated the incidence density of DHS in diabetic patients as 117.7 patients/ per 10,000 person-years, compared to 80.7/10,000 person-years in non-diabetics.²³

Diabetic cheiroarthropathy is the most common aspect of DHS with a prevalence ranging from 20–54% in T2DM patients.^{22, 25, 27} LJM is characterized by stiffness in the hands resulting from flexion contractures of the digits, which includes the flexor tendon, synovial sheath and subcutaneous tissues.²⁸ Dupuytren’s disease, which is characterized by fibrosis of the palmar fascia leading to flexion contractures of the digit, has a prevalence between 14–63%.^{25, 37–41} FTS is the third most common component of DHS with an incidence of 11–20%^{25, 27, 37, 42} and results in ‘locking’ of the finger during flexion. While the first, third and fourth digit are most commonly affected, the likelihood of function in multiple fingers being impacted is more common in diabetics.²⁸ In addition, multiple of these conditions can occur in a single patient, further compounding functional deficits. For example, Al-Matubsi et al. found that 68.4% of T2DM patients had both LJM and FTS.²⁵ Carpal tunnel syndrome (CTS), which results in compression of the median nerve by the transverse carpal ligament,²⁸ is often included in discussions of the diabetic hand, even though it is considered a compression neuropathy. Nonetheless, it is one of the most prevalent pathological changes in the diabetic hand, with CTS observed in 14–60% of diabetic patients.^{25, 27, 37, 43–45} Cumulatively, these pathologic changes impact hand function resulting in weakened grip, lower dexterity and decreased ability to perform fine finger motions.⁴⁶

Rotator Cuff Tendinopathy

T2DM is associated with an increase in the prevalence of general shoulder pathology, and was recently reviewed very comprehensively by Lee et al.⁴⁷ In terms of the magnitude of the clinical burden, Thomas et al. reported that shoulder disorders occur in 27.5% of T2DM patients relative to 5% of non-diabetics.²⁴ These pathologic changes include increased shoulder pain,³³ rotator cuff tendinopathy,¹⁹ and rotator cuff tendinitis in diabetic males.³³ More specifically, Lin et al. identified a greater than 2-fold increase in RC disorders,⁴⁸ and T2DM RCs become thickened,^{16, 19, 20} have decreased range of motion,¹⁶ and there is increased prevalence of calcified tendinitis in diabetic patients.²¹

Alterations in the diabetic foot and Achilles tendons

Changes in the structure and function of the tendons of the feet can have profound effects on daily life and may lead to altered gait and loading, which can increase the risk of diabetic foot ulcers. D’Ambrogi et al. examined changes in foot function, specifically the AT-plantar fascia-metatarsophalangeal joint complex. Significant AT and palmar fascia thickening were observed in diabetic patients, while joint mobility was significantly decreased, and loading was altered.¹⁷ Cumulatively, these changes can alter gait and loading in T2DM patients. Related to this, foot ulcers are thought be related to increased passive stiffness of the muscle-tendon unit. Using ultrasound, Batista et al. demonstrated a significant increase in tendon fiber disorganization prevalence (89% of T2DM vs. 10% of non-diabetic controls) of the Achilles tendon.¹⁸ Interestingly, Abate et al. also showed that T2DM patients that were asymptomatic for AT pathology had an increased incidence of ultrasound abnormalities (T2DM: 25.7% vs. 11.7% of non-diabetics).⁴⁹ This suggests that there is likely a large population of diabetic patients with degenerative tendon changes that have not reached the threshold for symptoms, thus we are only beginning to appreciate the true magnitude of diabetes-related tendon pathology. In addition to changes within the midsubstance of the tendon, there are also notable changes to the enthesis, and Ursini et al. showed that nearly 25% of T2DM patients that were asymptomatic for AT pathology had entheseal thickening (vs. 8.7% of controls), while 74.4% had enthesophytes (vs. 57.5% of non-diabetics).⁵⁰ Consistent with these non-invasive imaging and functional studies, we demonstrate that human diabetic flexor digitorum longus (FDL) tendons demonstrate marked structural changes including decreased compactness of the collagen fibers and decreased cellularity relative to tendon from non-diabetic patients (Fig. 1).

MOLECULAR MECHANISMS OF DISRUPTED HOMEOSTASIS IN T2DM TENDONS

Though the clinical effects of T2DM on tendon homeostasis have been documented for years, the molecular mechanisms underlying these complications remain poorly understood. This lack of information is due in part to the difficulty of obtaining relevant human clinical samples of diabetic tendon for study. As previously discussed, T2DM tendon-related complaints do not typically present clinically until there is an acute injury or severe co-morbidity. Though important in documenting the longitudinal effects of T2DM on overall tendon structure (including decreased collagen fibril diameter, increased fibril packing, changes in fibril morphology, and fibrillar disorganization⁵¹), samples obtained from these patients shed no light on the pathological changes to the tendon tissue that precede acute injury. To elucidate how the systemic effects of T2DM can manifest as impaired tendon function, investigators have turned to a variety of animal and in vitro models of T2DM.

Altered tendon homeostasis in animal models of T2DM

There are a number of well-characterized T2DM rodent models available that display various aspects of the human disease (Table 1), though no one model recapitulates T2DM entirely.⁵² As with human cases, the majority of diabetic mouse models are obese with varying levels of severity. Male C57BL/6J mice are commonly used to model diet-induced obesity (DIO) and associated T2DM as they become obese and develop decreased glucose tolerance with moderate insulin resistance when fed a high-fat diet (HFD).^{53, 54} The DIO mouse model also recapitulates hallmark complications of human T2DM including impaired wound healing⁵⁵ and peripheral neuropathy.⁵⁶ Using this model, the FDL tendons of mice fed a HFD for up to 48 weeks exhibit functional deficits including decreased range of motion and maximum load in addition to decreased collagen fibril density.^{57, 58} Importantly, these structural alterations persisted even after normal metabolic function was restored by a returning these mice to a low fat diet.⁵⁷ This suggests that tendon-specific intervention, rather that general systemic T2DM management, is necessary to mitigate the negative effects of diabetes on tendon function. Like mice, some rat strains become diabetic when placed on a HFD and exhibit similarly altered tendon mechanical properties. Gonzalez et al. found that tail tendons from the Zucker diabetic Sprague-Dawley rat display changes in collagen fiber morphology and stiffness that translated to decreased strain to failure of the fibrils.⁵⁹

Table 1.

Animal models of the effects of T2DM on tendon homeostasis

Model	Tendon	Alterations compared to non-diabetic controls	Ref
Zucker diabetic Sprague-Dawley rats	Tail	Increased nanoscale stiffness	59
		Decreased microscale strain to failure
db/db mice	Supraspinatus, Achilles, and patellar	Reduced CSA and transition strain	65
		Decreased stiffness at the insertion site
		No change in GAG or collagen content
		Decreased collagen content in midsubstance of patellar tendon only
db/db mice	Achilles	Increased diameter	66
		Decreased mechanical properties (max load, stiffness, tensile stress, elastic modulus)
		Mild ECM disorganization near calcaneal insertion
ob/ob mice	Achilles	Increased vascularity at the insertion	64
		ECM disorganization
C57BL6/J mouse DIO	FDL	Decreased max load	57
		Decreased functional properties (decreased range of motion, increased gliding resistance)
		Decreased collagen fibril density

CSA= cross sectional area, GAG= glycosaminoglycan

In addition to the DIO model, monogenic mouse strains are also commonly used to study the pathogenesis of obesity-related T2DM. Mice carrying mutations in the leptin (ob/ob) or leptin receptor (db/db) gene are hyperphagic, quickly becoming morbidly obese even on a regular chow diet, and subsequently develop decreased glucose tolerance, hyperinsulinemia, and increased insulin resistance.⁶⁰ Both the ob/ob and db/db models also exhibit some degree of pancreatic islet dysfunction^{61, 62} and liver steatosis⁶³ and therefore represent more severe T2DM than the DIO model. Despite their popularity in other fields, there have been only a few studies on how these genetic models of T2DM affect tendon homeostasis. In ob/ob mice, Ji et al. reported collagen fiber disorganization in the AT, along with increased vascularity and evidence of spontaneous rupture at the calcaneal insertion.⁶⁴ AT from db/db mice have altered mechanical properties compared to non-diabetic controls,^{65, 66} including decreased stiffness and elastic modulus, with mild degenerative changes reported near the insertion in some cases.⁶⁶ As seen in human cases of diabetic tendinopathy, there also appears to be a tendon-specific effect of T2DM in the db/db model; a study by Connizzo et al. found that while the Achilles, patellar, and supraspinatus tendons of db/db mice exhibit decreased stiffness and cross-sectional area, collagen content was decreased only in the patellar tendon.⁶⁵

Increased accumulation of advanced glycation end products (AGEs) in the collagenous tissues of diabetic patients is thought to one of the main drivers of tissue dysfunction.⁶⁷ AGEs form via a non-enzymatic reaction between sugars and the free amino groups of proteins and lipids, creating cross-links that alter the biomechanical properties of the tissue matrix. An early study by Kent et al. found that culturing rabbit tendons ex vivo in high glucose increased the degree of collagen glycosylation present and resulted in the formation of stable collagen cross-links, demonstrating that a hyperglycemic microenvironment in T2DM tendon can lead to changes in tendon structure.⁶⁸ More recent studies, however, have found that the presence of AGEs in tendon is insufficient to induce tissue-level impairments in mechanical properties, suggesting an alternative mechanism for the damaging effects of AGEs on tendon function.^{69, 70}

Effects of hyperglycemia on tendon cell behavior in vitro

A number of studies have evaluated the effects of culturing tendon cell in high glucose-containing medium to determine potential cell-mediated mechanisms for the mechanical deficits seen in T2DM tendons (Table 2). Lin et al. found that culturing rat patellar tendon cells in high (15mM or 25mM) glucose for up to 48 hours lead to decreases in the expression of collagen type I (Col1) and the tendon cell markers scleraxis (Scx) and tenomodulin (Tnmd), along with increased apoptosis and decreased proliferation compared to cells grown in low glucose (5.5mM).⁷¹ In similar studies, Ueda et al. and Tsai et al. found that culturing rat AT cells in high glucose (33mM and 25mM, respectively) for up to 72 hours elicited increased expression of catabolic enzymes (matrix metalloproteinases (Mmp) 2⁷², 9, and 13⁷³) and the pro-inflammatory cytokine interleukin-6.⁷² In contrast to these acute exposures studies, Wu et al. reported that culturing rat AT cells in high (25mM) glucose-containing medium for two weeks did not alter the expression of Scx, Tnmd, or Col1, but did decrease the expression of other tendon-related genes including mohawk, biglycan, and transforming growth factor β−1.⁷⁴ Collectively, these studies suggest that hyperglycemia may impair tendon cell homeostasis and alter the expression of pro-inflammatory/pro-fibrotic mediators. Interestingly, Poulsen et al. found that extracellular glucose concentration can affect tendon cell behavior in response to oxidative stress; human hamstring tendon cells cultured in low (5mM) glucose exhibited increased expression of Col1 and Scx whereas culture in high glucose lead to cell death via apoptosis following acute exposure to oxidative stress (100μM H₂O₂ for 18 hours).⁷⁵ As part of a mechanically active tissue, tendon cells are regularly exposed to oxidative stress. Therefore, in addition to affecting the baseline activity of tendon cells, the presence of T2DM-associated hyperglycemia may impact the ability of tendon cells to properly respond to normal mechanical stimulation.

Table 2.

In vitro studies on the effects of high glucose on tendon cell function

Cell source	Species	Culture conditions	Effects of culture in high glucose compared to low	Ref
Patellar tendon	Rat	High (15 or 25mM) or low (5.5mM) glucose DMEM, 24 or 48 hours	Decreased gene expression of Scx, Col1 (24 and 48 hours)	71
			Decreased expression of TNMD, COL1 protein (24 and 48 hours)
			Increased apoptosis (24 and 48 hours)
			Decreased cell proliferation (up to 5 days)
Achilles tendon	Rat	High (25mM) or low (5.5mM) glucose DMEM, 7 or 14 days	No change in gene expression of Scx, Col1, Tnmd	74
			No change in cell proliferation
			Decreased gene expression of Mkx, Egr1, Tgfb1, Bgn at 14 days
			Inactivation of AMPK signaling
Achilles tendon	Rat	High (12mM, 25mM) or low (6mM) glucose DMEM, 24 hours	No change in cell proliferation	73
			Increased gene expression of Mmp9, Mmp13
			Increased MMP-9 activity
Achilles tendon	Rat	High (33mM) or control (12mM) glucose DMEM, 48 or 72 hours	Increased ROS accumulation at 48 hours	72
			Increased gene expression of Nox1, Mmp2, Timp1 (48 and 72 hours)
			Increased gene expression of Col3, Timp2, Il-6 at 72 hours
			Decreased gene expression of Col1 at 72 hours
			Decreased cell proliferation (48 and 72 hours)

Mkx = mohawk homeobox, Egr1= early growth response 1, Bgn= biglycan, ROS= reactive oxygen species, Nox1= NADPH oxidase 1, Timp1/2= tissue inhibitor of metalloproteinases 1/2, Il-6= interleukin-6, DMEM= Dulbecco’s modified Eagle medium

While there is still more work to be done to clarify the short- versus long-term effects of hyperglycemia and the specific pathways/mechanisms involved, these in vitro studies clearly demonstrate that elevated extracellular glucose concentration can affect tendon cell behavior in ways that may lead to cell-mediated damage to the extracellular matrix. In addition, although hyperglycemia is certainly involved in the pathogenesis of T2DM tissue dysfunction, the characteristic dyslipidemia seen in T2DM is also an important contributor to many T2DM complications. Despite this, no studies have specifically evaluated the effects of lipoprotein abnormalities on tendon cell function. In a mouse model of hyperlipidemia, the patellar tendons of apolipoprotein E-deficient mice exhibit decreased elastic modulus compared to wild type mice, indicating that prolonged exposure to elevated cholesterol levels can impact tendon function.⁷⁶ Whether these changes are due to alterations in the tendon structure or pathologic behavior of tendon cells is unknown, but further investigation is clearly warranted.

DIABETIC TENDON HEALING

In addition to altering tendon homeostasis and baseline function, T2DM dramatically impairs the healing response following tendon injury and surgical repair. While physiological tendon healing can often result in unsatisfactory outcomes,^{77, 78} the addition of T2DM dramatically impairs the healing process and exacerbates the natural inclination of the tendon to heal via fibrosis. This is particularly important given that T2DM increases the risk of tendon tear or rupture up to five-fold, relative to non-diabetics.⁷⁹ In terms of healing outcomes in specific tendons, the most abundant clinical data exists in the rotator cuff, with clear evidence of diminished healing and increased risk of repair failure (greater than 2-fold).^{80, 81} There is also some evidence that limitations in healing are particularly pronounced during the early phases of healing. For example, Clement et al. found that, while improvements in pain and function were observed in T2DM patients at 6-months post-op, the extent of improvements were markedly reduced relative to non-diabetic patients.⁸² However, Hsu et al. identified no difference in outcomes long-term (>24 months).⁸³

Despite the well-known complications that arise during diabetic tendon healing, there are few animal studies that examine the mechanisms and cell types involved in the abnormal healing response (Table 3). Similar to what is seen in clinical cases of T2DM tendon injury, T2DM rodent models exhibit impaired tendon healing.^{54, 58} More specifically, using the DIO mouse model, David et al. reported increased extracellular matrix disorganization, limited tenocyte migration, and impaired biomechanical properties following a biopsy punch injury to the FDL mid-substance.⁵⁸ We recently reported that FDL tendons from DIO T2DM mice heal with an amplified fibrotic response characterized by increased expression of Col1, collagen type 3 (Col3), Mmp9, and Mmp2, impaired range of motion, and increased gliding resistance following transection and repair, compared to mice fed a low fat diet.⁵⁴ Importantly, we also observed alterations in macrophage polarization including increased and prolonged markers of M2 macrophage activity, suggesting that aberrant activity of these cells may be critically involved in fibrotic T2DM tendon healing. In a model of Achilles transection without repair in non-obese T2DM rats (the Goto-Kakizaki rat), Ahmed et al. reported decreased vascularity, callus area, collagen organization, and stiffness in diabetic injured tendons compared to non-diabetic.^{84, 85} Interestingly, these deficits in healing were associated with decreased expression of Col1, Col3, and Mmp3.⁸⁴ Whether this difference in the effect of T2DM on the expression of collagen and catabolic enzymes between the rat and mouse models is due to differences in the injury type (repair vs. non-repair), tendon, or species remains to be clarified.

Table 3.

Animal models of T2DM tendon injury and repair

Model	Injury type	Tendon	Outcomes compared to non-diabetic controls	Ref
C57BL6/J mouse DIO	Punch injury	FDL	ECM disorganization	58
			Limited tenocyte migration into granulation tissue
			Decreased mechanical properties (max force, work to max force, stiffness)
C57BL6/J mouse DIO	Transection and repair	FDL	Increased gene expression of Col1, Col3, Mmp2, Mmp9	54
			Increased protein expression of IL-1RA, TNFα, F4/80
			Altered macrophage polarization (increased/prolonged M2 activity)
			Decreased mechanical properties (max force, work to max force, max load)
			Decreased functional properties (decreased range of motion, increased gliding resistance)
Goto-Kakizaki rats	Transection	Achilles	ECM disorganization	84
			Decreased gene expression of Col1, Col3, Mmp-3
			Decreased stiffness
			Decreased callus area
Goto-Kakizaki rats	Transection	Achilles	Decreased gene/protein expression of Tβ–4 and Vegf	85

IL-1RA= interleukin-1 receptor antagonist, TNFα= tumor necrosis factor alpha, Tβ–4= thymosin beta-4, Vegf= vascular endothelial growth factor

CHALLENGES AND OPPORTUNITIES FOR FUTURE RESEARCH

The magnitude of the effects of T2DM on tendon function and healing are only beginning to be understood due in part to a lack of comprehensive, well-powered studies on the incidence and prevalence of T2DM tendon complications. Importantly, there is a pressing need for large cohort studies that prospectively collect potentially confounding variables and incorporate multivariable models in their subsequent statistical analyses. For example, inclusion of obese, non-diabetic, and non-obese diabetic patients may help delineate the relative contribution of obesity vs. T2DM to this pathology. The recent adoption and widespread implementation of electronic medical records will facilitate these more detailed studies in the future, including those aimed at determining the extent of T2DM tendon complications, the relative contributions of co-morbidities, identification of additional risk factors for tendon disorders, and the degree to which different tendons are affected by T2DM. In terms of improving the quality of care for T2DM patients, one challenge currently facing the field is the fact that many cases of diabetic tendinopathy remain subclinical until the point of acute injury or rupture, leading to increased severity of the injury and decreasing the quality of the healing response. The development and greater implementation of quantitative screening tools or non-invasive methods of evaluating tendon function (e.g. ultrasound) will allow tendon function to be evaluated in at-risk populations (i.e. T2DM patients) prior to the development of tendon pathologies and to monitor the effectiveness of interventions aimed at halting the progression of established diabetic tendinopathies. Given the rapidly rising number of T2DM patients, the development of effective therapeutics to prevent and treat tendon disorders in this population is critical. To this end, there is still an enormous amount of work to be done on characterizing pre-clinical models of T2DM tendinopathy and tendon healing to better understand the cell types and signaling pathways involved in these processes and to ascertain points for intervention. As previously discussed, in vitro studies have shown that exposure to high glucose can affect tendon cell behavior, but the degree to which these specific alterations may contribute to diabetic tendinopathy or impaired healing in vivo is still unknown. Moreover, while animal models of T2DM show some similarity to human cases, the lack of early stage clinical samples makes determining whether animal models truly recapitulate the pathogenesis of T2DM tendinopathy difficult. There is also a scarcity of information regarding which tendon changes seen in the various T2DM rodent models are consistent between models, and which may be attributed to specific models or tendons. More thorough characterization of both the available in vitro and in vivo models of T2DM is needed before the field can progress to the identification of effective therapeutic targets. However, the increasing use of sophisticated genetic models to better define the cellular and molecular basis of tendon healing and tendinopathy has tremendous potential to both identify aberrant cell/ molecular processes in the context of T2DM, and to accelerate discovery of promising therapeutic targets to improve tendon function in T2DM and non-diabetic patients.