Tendon Regeneration and Scar Formation: The Concept of Scarless Healing

Abstract

Tendon healing is characterized by the formation of fibrovascular scar tissue, as tendon has very little intrinsic regenerative capacity. This creates a substantial clinical challenge in the setting of large, chronic tears seen clinically. Interest in regenerative healing seen in amphibians and certain strains of mice has arisen in response to the biological behavior of tendon tissue. Bone is also a model of tissue regeneration as healing bone will achieve the mechanical and histologic characteristics of the original tissue. The ultimate goal of the study of genes and mechanisms that contribute to true tissue regeneration is to ultimately attempt to manipulate the expression of those genes and activate these mechanisms in the setting of tendon injury and repair. Clearly, further research is needed to bring this to the forefront, however, study of scarless healing has potential to have meaningful application to tendon healing.

Tendon injuries are some of the most common orthopedic problems, and account for substantial pain, disability, and time off of work. While many tendon injuries are acute, a very large number are chronic, degenerative conditions. Repair in either case results in the formation of fibrovascular scar that never attains the gross, histological, or mechanical characteristics of normal tendon. Especially in the shoulder, failure to heal is a common and complex problem. The field of tendon research is recently emerging as a central focus, but is generally lagging behind that of other musculoskeletal tissues and general medical research. The precise mechanisms of degeneration, tearing, and subsequent healing are not known. The concept of “scarless healing” has recently become of interest to physicians and researchers in application to tendon healing, as tendon has very little regenerative capacity of its own.

Current Biologic Strategies

Currently, biologic strategies employed to enhance healing have largely been market driven and have not achieved tendon regeneration. While some improvements have been documented in healing rates and function in some studies, there is no scientific evidence of regeneration of normal tendon tissue.

Extracellular Matrix (ECM) Patches for Rotator Cuff Repair

The concept of an ECM patch is to reinforce the strength of a repair by offloading the tendon repair. Potentially, the patch could provide a scaffold for new growth and differentiation, and in the future, these may be a delivery vehicle for cells and growth factors. Most available materials are made of dermis, intestinal submucosa, and synthetic materials. Very few prospective, randomized trials exist in the literature. One utilized a product that induced an aseptic inflammatory response which precluded its future use in this application.¹ The other investigated a dermal product which suggested better scores and improved healing in the treated group.² Another series, also using a dermal product, showed some favorable results with some second look histology findings, but there was no comparison group.³ Overall, in spite of some early enthusiasm for this treatment strategy, there is not enough evidence to support routine use, and no strong evidence, even in animal studies, that these facilitate tendon regeneration or scarless healing.

Platelet Rich Plasma (PRP)

PRP is a sample of autologous blood with a concentration of platelets 3–4 times greater than baseline. Its use is compelling, as platelets contain alpha granules which contain a multitude of growth factors that have a role in cell recruitment, proliferation, and angiogenesis. There are several different preparations commercially available. Several studies have evaluated the role of PRP in tendinopathies and repair scenarios. PRP may well have a positive role in the treatment of tendinopathies such lateral epicondylitis and Achilles tendinitis. However, PRP has not demonstrated a strong benefit in the setting of rotator cuff repair.^4–7 Two randomized trials showed no benefit,^6,8 and another showed a small improvement in tendon integrity but no clinical difference compared with untreated repairs.⁹

One of the barriers to usefulness of PRP may be the timing of delivery of the growth factors. Once the PRP is activated, the alpha granules release the factors within an hour. Most of these growth factors have a half life of only a few minutes.¹⁰ Most studies of tendon healing show a spike in cell proliferation, growth factor production, and extracellular matrix production from 7 to 14 days after injury and repair.¹¹ Delivering the growth factors at the appropriate peak in metabolic activity is likely critical for success. Therefore, sustained delivery with the use of platelet rich fibrin matrix is compelling,^{4,6,8,12–14} as this product contains the platelets in an absorbable fibrin matrix, providing a slower, more controlled release of factors as the fibrin matrix absorbs. Nevertheless, further research is necessary to maximize the use PRP, and even still, there is no evidence that delivery of growth factors alone induces tendon regeneration or scarless healing.

Animal Models for Studying Scarless Healing

In mammals, the early-gestation fetus has the ability to heal skin wounds by tissue regeneration without scar formation. It has long been thought that the absence of an inflammatory response during early gestation is an important factor that allows tissue regeneration (“scarless healing”). However, once a functioning immune system develops in later gestation, the inflammatory response to wounding leads to healing via a scar-mediated pathway. Although there are other factors that govern the response to wounding in the fetus and lead to scarless healing, the inflammatory response appears to play an important role.

Several animal models have been used to study scarless healing in vertebrates. The most obvious way to study fetal wound healing is to use the developing fetus as a model. Intra-uterine surgery in pregnant ewes has been used to study fetal wound healing. For example, a partial tenotomy was performed in the lateral extensor fetal tendons of fetal sheep at 80–85 days of gestation.¹⁵ Analogous tenotomies were created in the maternal limbs. One week after the tendon injury, there was a gap with granulation tissue and inflammatory cells in the adult tendon wounds, while regenerative healing was found in the fetal tendons, where there was reconstitution of collagen architecture. There was upregulation of TGF-b expression in adult tendon wounds, while there was minimal expression in the fetal tendons. The sheep model has also been used to evaluate the role of wound size on fetal tendon healing.¹⁶ A tendon wound was made in pregnant ewes at 69–77 days of gestation. The authors reported that small tendon wounds showed a regenerative healing phenotype with orderly deposition of collagen fibers, while large tendon wounds healed via scar formation based on histologic and molecular criteria.

The sheep model has also been used to examine whether regenerative healing is intrinsic to the fetal tissue itself or the result of its environment. Injured fetal sheep tendon that was transplanted into an adult environment was found to retain a regenerative healing pattern after injury.¹⁷ The authors concluded that scarless repair ability is intrinsic to fetal tendon itself, suggesting that the adult environment is not an impediment to scarless repair. Insight into the mechanism(s) for this response was gained by the finding that injured adult tendons demonstrated elevated levels of TGF-β1, bFGF, and CD44 at the wound site, whereas the fetal specimens showed little or no such changes.

Mouse models have been used to study scarless healing. The Murphy Roths large (MRL) mouse possesses the ability to heal via tissue regeneration in the post-natal setting. The PU-1 null mouse is genetically incapable of mounting a standard inflammatory response following injury because it lacks macrophages and functioning neutrophils.¹⁸ These mice are able to repair skin wounds in a scar-free manner with a similar time course to wild-type mice. The study of the cytokine profile and cell kinetics at the healing wound site in these animal models can provide further insight into the basic cellular and molecular mechanism(s) of scarless healing.

Insight into the process of scarless healing can also be gained from a diversity of other animal models. Axolotls are salamanders with unsurpassed healing and regenerative capacities.¹⁹ Skin wounds in the axolototl heal without fibrosis, with absence of neutrophils and minimal TGF-β expression. Fetal mouse skin xenografts transplanted onto the chorioallantoic membrane of fertilized chicken eggs provide another useful model for the study of fetal wound healing.²⁰ A heart infarction model has been used to study cardiac regeneration in zebrafish.²¹ In this model fibrotic scar tissue initially forms following the infarction, but later is progressively eliminated by cell apoptosis and becomes replaced with a new myocardium, resulting in scarless regeneration. Tissue remodeling in the infarct zone is associated with accumulation of vimentin-positive fibroblasts and with expression of the extra-cellular matrix protein tenascin-C.

Investigative Techniques and Methods of Evaluation

Histologic Techniques

A fundamental consideration is the lack of rigorously defined histologic or molecular criteria to define “scar.” The currently accepted way to define scar in tendon is based on histologic techniques to assess tendon micro-structure. Routine histologic analysis under light microscopy is used to examine collagen fibrils and tendon architecture. Specialized stains, such as Safranin-O and toluidine blue, can be used to evaluate areas of chondroid metaplasia that is sometimes seen in scar. Polarized light microscopy exploits the birefringence of collagen to allow evaluation of collagen organization, with scar generally lacking the organization seen in normal tendon. Collagen birefringence is measured as brightness under polarized light microscopy and is generally proportional to collagen fibril orientation and fibril thickness, with scar tissue generally having less birefringence. Picrosirius red staining can be used to enhance collagen birefringence.

Other more advanced imaging techniques can provide quantitative measurements of collagen organization. Circularly polarized light microscopy is a more advanced method that allows quantification of collagen organization.²² Second harmonic imaging microscopy and multiphoton microscopy are other techniques that provide information about tendon matrix. Computerized image analysis techniques can be used to measure angular deviation of collagen fibers as another method to quantify matrix organization.²³

Electron microscopy (EM) provides information about matrix structure at the ultrastructural level. Transmission EM (TEM) is used to measure collagen fibril diameters, while scanning EM can produce three-dimensional images of the surface of tendon. A transmission electron microscope can magnify a sample up to 1,000,000 times, while scanning electron microscope can magnify a sample up to 200,000 times. SEM also provides a 3-dimensional image while TEM provides a 2-dimensional picture.

Transcriptional (Gene Expression) and Translational (Protein) Analysis

While the imaging modalities above are used to evaluate matrix microstructure, further insight is gained from analysis of matrix composition. Immunohistochemical staining may be used to identify specific matrix proteins and their spatial arrangement. ELISA and Western blot is used to quantify levels of specific matrix proteins in a tissue extract. Analysis of gene expression using standard techniques such as in situ hybridization, PCR, and microarrays provides further insight into characterizing tissue as normal tendon versus scar tissue. Transcriptome data from micro-array experiments show differences between wounds that heal by scar versus those that heal by tissue regeneration.¹⁶ One of the challenges for studying scarless healing is that there are no reliable and universally acceptable markers to identify regeneration of neotendon. Most markers are expressed in scar and adult healing, so discriminating between development of new tendon and scar formation is difficult.

Imaging Evaluation

Imaging techniques are being developed that allow non-invasive evaluation of healing tendon, permitting the ability to carry out serial imaging over the course of an experiment. Magnetic resonance imaging (MRI) is the optimal modality due to its superior soft tissue contrast; however, the high field strength (7.0 Tesla) required for high resolution imaging of small animal specimens requires a small bore that can only accommodate small specimens ex-vivo. Quantitative MRI techniques that hold promise for evaluation of tendon microstructure include diffusion tensor imaging and ultrashort TE (T2*).²⁴

Ultrasound is a readily available modality that can image superficial tendons. Speckle tracking is a validated technique that allows estimation of intratendinous strain.²⁵ This technique would be suitable for evaluation of superficial tendons, such as Achilles or patellar tendon, in larger animals. Nuclear medicine techniques such as positron emission tomography (PET) may be used to identify specific metabolic processes, such as inflammation or osteoblastic activity. Development of other radio-labelled substrates may allow PET to be used to track other metabolic processes, such as collagen synthesis or remodeling activity.

Biomechanical Testing

The functional properties of tendon are evaluated with biomechanical testing. Structural properties are measured as load-to-failure and deformation during tensile testing. Material properties include measures of tissue stress, strain, and elastic modulus. These measures can be done on the whole tendon or on a smaller, representative section of the tendon. Techniques are now available to measure mechanical properties of isolated collagen fiber bundles and even individual collagen fibrils.

Regenerative Healing

It is common knowledge that amphibians display a spectacular ability to regenerate damaged or lost tissues and organs to restore the normal anatomic state both architecturally and functionally (epimorphic regeneration). This is generally not seen in mammals²⁶ although there are many tissues that can regenerate based on stem cell populations found in and supported by their specific niches including skin, hair follicles, intestine, liver, and bone marrow. However, there is a clear inability to fully re-grow organs and appendages.

Perhaps the earliest evidence of mammalian epimorphic regeneration came from observations of rabbit ears in which holes made in the ear pinnae completely closed and cartilage regrew.^27,28 Other studies examining deer antlers showed a similar degree of regenerative ability with the formation of a blastema and then differentiation into mature tissue.²⁹ In 1998, it was first reported that a strain of adult mouse, the MRL mouse, could also fully close ear holes with the replacement of cartilage and without scarring.³⁰ Originally thought to be an outcome of a mutation in the fas gene which leads to uncontrolled proliferation of peripheral immune cells of the MRL/lpr mouse, (used for many years as a model of lupus erythematosis), it soon became clear that MRL/MpJ mice not bearing the lpr mutation also displayed full ear hole closure. Interestingly, the MRL/MpJ does retain a mild form of autoimmunity.

Besides ear hole closure and elastic cartilage regeneration, there have been many tissues shown to regenerate in the MRL mouse, and among the tendon-related tissues are digit,^31,32 spinal cord intervertebral discs,^33,34 and knee joint articular cartilage.^35–37 Discussions at the “New Frontiers in Tendon Research” meeting (Mt. Sinai Medical Center, NYC, September 2014) revealed that there are multiple laboratories that are using the MRL mouse to examine enhanced tendon healing, and in fact, over 80% recovery of tendon function has been observed.³⁸

There are key events in epimorphic regenerative response. Both amphibians and the MRL mouse share a similar pattern and trajectory of tissue responses to injury. Rapid and complete coverage and re-epithelialization of skin wounds can take place within 12–24 h in amphibians and within 24–48 h in the MRL mouse, much earlier than in other mouse strains. Another key event is the formation of the regeneration blastema, a self-organizing group of progenitor cells which expand and form a copy of the original structure.³⁹ In amphibians, blastema cells initially accumulate by cell migration for approximately the first 7 days, followed by a proliferative phase through the second week after injury.⁴⁰ The cells in the blastema express progenitor cell markers through a process of de-differentiation. After the proliferative phase, cells re-differentiate and form a mature structure. These events have a parallel in the MRL mouse (ms submitted).

Mechanisms of Mammalian Regenerative Ability

In defining the mechanisms involved in MRL regeneration, it was discovered that the gene CDKN1a or p21^cip1waf1 was not expressed or poorly expressed as protein in cells from the ear pinnae in either unwounded or post-wounded animals. Examination of CDKN1a knockout mice showed an ear hole closure response similar to that seen in the MRL mouse.⁴¹ This suggested that cell cycle control was potentially important; it may be no surprise that cell division and regeneration are functionally related. However, there are multiple possibilities for its function though these have not yet been addressed.

There are now other examples of transgenic mutants in which otherwise non-regenerating mice can close ear holes. These point to related or to other potential mechanisms for this healing response. Two examples show that the over-expression of angiogenic molecules lead to ear hole closure. This includes angiopoietin-1 and angiopoietin-related growth factor or AGF.^42,43 ENU mutagenesis has produced a mouse with a mutant TGFbR that can heal ear holes⁴⁴ and may relate to previous results showing aberrant TGFb expression⁴⁵ and ear hole closure. Most recently, over-expression of the microRNA binding protein Lin28 was shown to enhance ear hole closure.⁴⁶

Applying Concepts of Scarless Fetal Wound Healing to Enhanced Regenerative Healing in Adults

As discussed above, there is evidence that scarless fetal wound healing should inform the study of adult regeneration. One line of research suggests that it is at the embryonic stage where the inflammatory response first arises that converts scarless fetal wound healing to scarred adult-type healing. However, there are studies that suggest that this may not be the case. It has been shown that there are limitations to limb regeneration in fetal mice post-amputation which is highly dependent on the stage of development.^47,48

The Role of Inflammation

The suggestion from scarless fetal wound healing studies is that inflammation is a problem for regenerative responses. However, there are multiple examples where a pro-inflammatory response enhances regeneration. Inflammation plays a positive role in axonal regeneration and neural protection.^49,50 Specifically, Leukotriene B4 and lipoxin A4⁵¹ have been shown to regulate stem cell proliferation and differentiation and eicosanoids in general have been shown to positively contribute to regenerative healing⁵² (28); IL-4 can enhance liver regeneration.⁵³

A striking example showing the effect of inflammation comes from two strains of mice, AIRmax and AIRmin, which were selectively bred for maximum and minimum acute inflammation, respectively.⁵⁴ These mice show different neutrophil and macrophage migration phenotypes. Furthermore, these two mice show differences in ear hole closure with the highly inflammatory mice (AIRmax) showing complete closure, and AIRmin mice showing a negative response. There was also a significant genetic linkage of ear hole closure, neutrophil infiltration, and the AIRmax allele of the gene Slc11a1 or Nramp, a protein involved in endosomal ion transport in macrophages and neutrophils.⁵⁵ This may play a role in macrophage-enhanced wound healing responses.^56–59

Considering that the MRL/MpJ mouse retains (subdued) autoimmune features suggested that inflammation might also play a pro-regenerative role here. This was confirmed by treating MRL mice with meloxicam, a COX-2 inhibitor, and finding that this drug could block ear hole closure.⁶⁰ Analysis of MRL inflammatory cell types showed that macrophages, neutrophils, and mast cells were increased in the MRL ear pinna pre- and post- injury. Pathway analysis of an ear hole injury gene expression library comparing MRL and C57Bl/6 tissue showed that the major pathways included inflammation, remodeling, and metabolism. Molecules such as MPO and mast cell protease 7 (MCP7)/tryptase alpha are highly up-regulated in the MRL mouse. MCP7 is a potent inflammatory mediator that affects clot formation, fibrinogen/integrin interactions, and is pro-angiogenic.⁶¹ On the other hand, there are many molecules that are reduced in the MRL, such as CTGF (connective tissue growth factor) and Loxl4 (lysyl oxidase-like molecule) both of which are involved in scar formation. One molecule, the mast cell high affinity Ig receptor, Fcer1g, is a key element in inflammation and is 57 fold reduced in the MRL. This should result in reduced inflammatory cell-mediated tissue destruction.⁶²

In conclusion, the MRL mouse should be a useful animal model for tendon regeneration research as there is now a body of cellular and molecular targets for intervention studies. In particular, the use of NSAIDS and other anti-inflammatory drugs post-injury should possibly be reconsidered.

Fracture Healing as a Model of Regeneration and Scarless Healing

Developmental Aspects of Fracture Healing/Recapitulation of Endochondral Bone Formation

Fracture healing and bone repair are unique post-natal processes that recapitulate many of the ontological events that take place during endochondral bone formation, processes that are observed during the embryological development of the skeleton.^63–67 It is generally believed that the recapitulation of these ontological processes makes fracture healing one of the few post-natal processes that is truly regenerative. In contrast to most injury responses that lead to fibrotic scarring and incomplete restoration of tissue structure, fracture healing restores both the structure and cell composition of the original bone tissue and in this regard it is a true regenerative process. The appearance of the endochondral process of bone formation during fracture healing is very similar to the cellular processes used during embryological formation and the growth of mammalian skeletal tissues. Understanding how fracture repair differs from other types of soft tissue repair associated with fibrotic scar will help elucidate the processes that make it a regenerative process.

Over the time course of fracture healing, multiple cellular lineages give rise to cartilage, bone, vascular, and hematopoietic tissues. Cells are all recruited and contribute to the regeneration of the injured skeletal organ.⁶³ Many of the genes that are preferentially expressed in embryonic stem cells and the morphogenetic pathways that are active during embryonic skeletal development are detected in fracture callus. They direct these cells to undergo appropriate differentiation and coordinate their activities in an appropriate temporal and spatial manner such that a regenerative process occurs.⁶⁸ It would be instructive to understand the temporal and spatial signals that recruit tendon and ligament stem cells, as well as the morphogenetic signals that direct embryonic tendon and ligament development in order to compare these processes to those that take place during post-natal repair.

Role of Skeletogenic Stem Cells during Fracture Repair

The recruitment of mesenchymal stem cells (MSCs) and skeletal progenitor cells is a central feature of fracture repair. Growing evidence indicates that these cells are derived from multiple tissue niches within post-natal tissues including tendon sheaths, adipose tissue, and satellite cells of the muscle.^69–72 The highly vascularized periosteum, a bilayer membrane lining the outer cortical surface of bone has been shown to provide a large number of the pluripotent cells that are required for successful bone repair.^65,73–75 Although MSCs in vitro direct adherent marrow stromal cell populations to differentiate into chondrocytes, adipose or bone cells, current literature suggests that the periosteal response provides the majority of the osteochondral progenitor cells needed to form the bridging callus of a fracture.^76,77 Finally, a population of alpha Smooth Muscle Actin (SMA) positive cells was identified within the periosteum and muscle that in response to fracture, proliferate and contribute to callus formation.^78,79 Collectively, these studies suggest that there are several different tissue niches that harbor populations of skeletogenic stem cells that potentially contribute to the repair of tendons or ligaments after injury.

Soft tissue trauma can induce ectopic or heterotopic bone formation in which MSCs are recruited within muscle tissues or the entheses of joints. Subsequently these cells will undergo condensation, chondrogenesis, and endochondral ossification. The inflammatory response may initiate this process although the role inflammation in the initiation phase is not understood.^80,81 HO is associated with polytrauma involving brain or spinal cord⁸² and HO often occurs around joint capsules of specific bones.⁸³ In a rat model of head trauma, fractures healed faster and serum contained factors that promoted mesenchymal stem cells to undergo chondrogenic differentiation in vitro.⁸⁴ The combination of a traumatic brain injury and a femoral fracture results in HO in the thigh muscle of approximately 54% of patients.⁸² The thigh muscle is the most common site of HO with head injury; however, HO develops in 20% of forearm fractures in the presence of head injury.⁸² Other orthopedic interventions such as total hip replacements, burns, oncology, and internal fixation of acetabular and elbow fractures can lead to HO, albeit at lower frequency.^82,85These results suggest therefore that stem cells are present within muscle tissues and around ligaments and tendons of entheses. In particular, the occurrence of HO around joints and at the entheses indicate that these stem cells are inhibited in some manner from undergoing progression to a cartilage or bone pheno-type under normal conditions but in response to injury show misdirected development. Thus, it may be inferred that the regeneration of tendon and ligaments may not be impeded by the lack of stem cells capable of forming these tissues, but rather may be limited by the appropriate set of temporal spatial signals that direct their progressive differentiation into the appropriate types of terminally differentiated cells found within these tissues.

In one recent study, murine stem cells grown in heterotopic locations were able to form fibrocartilaginous and fibrous entheses that had tenogenic characteristics and still maintain an osteo-/chondrogenic capacity.⁸⁶ Multiple tissue types concurrently develop to regenerate damaged tissue in fracture repair. Consistent with this, a new concept is emerging that biphasic constructs such as the bone and cartilage seen in joints, also regenerate from multiple cell lines rather than engraftment of a single cell type, thereby producing a more stable and long lasting repair.⁸⁷ Other studies have similarly shown that the maintenance of the different cell types found at osteotendon/ligament junctions in conjunction with the appropriate application of mechanical loading that is observed within these tissues may be important in achieving the successful repair of these tissues at the osteotendon/ligament junctions.^88,89 The development of multi-phasic tissue engineered scaffolds or hybrid scaffold cell constructs to repair damaged tendons and ligaments is an emerging approach to the repair of these sites of injury.⁹⁰

Tissue Remodeling During Fracture Healing as an Integral Part of Tissue Regeneration

Fracture healing involves the combination of both anabolic and catabolic processes. Fracture healing starts with an initial anabolic phase that is characterized by de novo recruitment and differentiation of stem cells that form skeletal and vascular tissues followed by a prolonged period of remodeling. As cartilage tissue development progresses, cells that will form the nascent blood vessels are recruited from within the surrounding muscle sheath to form new feeding blood vessels for the developing bone.^91,92 As chondrocyte differentiation progresses the cartilage extracellular matrix undergoes mineralization and this period terminates with the apoptosis of the chondrocytes.^93,94 During the initial catabolic period, cartilage resorption occurs, primary bone formation is initiated and there is a continuation in primary angiogenesis as the nascent bone tissues are formed. Subsequently, secondary bone remodeling begins in which the first mineralized matrix produced during primary bone formation is resorbed by osteoclasts followed by a prolonged period of coupled remodeling. In this final period of coupled remodeling, the marrow space and hematopoietic tissues are re-established and the regeneration of the original structural features of the injured skeletal organ is achieved. This final period of the catabolic phase is also accompanied by extensive vascular remodeling in which the higher levels of vascular flow that were seen in the anabolic period return to their pre-injury levels.^95,96

The exact nature of the catabolic processes and tissue remodeling that occur during soft tissue repair and the mechanism of tissue remodeling that are essential to appropriate regeneration of the injured tendons and ligaments are only partially understood. Recent studies have shown that the attenuation of early phases of initial tissue remodeling by MMP9 leads to fewer adhesions and scar tissue development in a murine model of flexor digitorum longus (FDL) tendon injury. It was speculated that the continued presence of MMP9 promotes the uncontrolled synthesis of fibrotic tissue.^97,98 In contrast, this enzyme is needed in fracture healing to facilitate the normal progression of cartilage resorption and the progression of angiogenesis that is associated with the bone tissue development.⁷⁴ MMP9 was also shown to mediate indirect effects on skeletal stem cell differentiation by regulating both the inflammatory response and the distribution of the types of inflammatory cells that migrate into the injured area. The presence of inflammatory cells and the expression of MMP9 were shown to lead the local periosteal cell differentiation toward chondrogenic or osteogenic linages in fracture healing models.⁹⁹ Thus, MMP9 not only has direct effects on the accumulation of extracellular matrix, but also has effects on mediating inflammatory response and local stem cell differentiation.

MMP2 and MMP14 are also seen during fracture healing¹⁰⁰ and are sequentially expressed as cartilage is resorbed. They also appear to be expressed late in tendon healing during the later, active remodeling, and reorganization phase of healing, around 21 days. This suggests that Mmp-2 and Mmp-14 were involved in the transition from fibroblastic granulation tissue to a more organized collagen structure.⁹⁸

Conclusion

The mechanisms of tendon degeneration and healing are not well understood. Tendon healing is characterized by fibrovascular scar that never attains the mechanical properties of normal tendon, leading to a relatively high failure rate after repair, especially in the setting of rotator cuff repair. The basic mechanisms of tendon healing, including the growth factor expression and cell line recruitment are poorly understood. Given that tendon has so little regenerative capacity, the concept of scarless healing is very compelling. If mechanisms of true regenerative, scarless healing can be elucidated and applied to tendon tissue and its attachment sites, many hope that would be an answer to the current challenge of poor healing. Substantial research is necessary before this will be achieved.