Cell and Biologic-Based Treatment of Flexor Tendon Injuries

Abstract

The two primary factors leading to poor clinical results after intrasynovial tendon repair are adhesion formation within the digital sheath and repair-site elongation and rupture. As the outcomes following modern tendon multi-strand repair and controlled rehabilitation techniques are often unsatisfactory, alternative approaches, such as the application of growth factors and mesenchymal stem cells (MSCs), have become increasingly attractive treatment options. Successful biological therapies require carefully controlled spatiotemporal delivery of cells, growth factors, and biocompatible scaffold matrices in order to simultaneously (1) promote matrix synthesis at the tendon repair site leading to increased biomechanical strength and stiffness and (2) suppress matrix synthesis along the tendon surface and synovial sheath preventing adhesion formation. This review summarizes recent cell and biologic-based experimental treatments for flexor tendon injury, with an emphasis on large animal translational studies.

Introduction

Among the most common and challenging hand injuries, intrasynovial flexor tendon transections have motivated over five decades of research designed to improve primary operative and rehabilitation techniques.^1–10 Finger lacerations are the most common upper extremity injury encountered in the emergency room, with an incidence of 221 per 100,000 person-years or 1 in 452 people per year,¹¹ mostly caused by glass or knives.¹² Even small lacerations < 2 cm presenting to the emergency room often cause deep tendon injuries (~60% of cases).¹² Major repair technique advances by Kessler,⁹ then Pennington,¹⁰ and then Winters and Gelberman⁴ have changed Zone II intrasynovial flexor digitorum profundus (FDP) tendon treatment from an inoperable “no man’s land”⁸ to a common surgical procedure. Following several decades of repair^{3,4,9,10,13–22} and rehabilitation^23–25 improvements, we have reached a plateau in Zone II flexor tendon repair outcomes with current methods. Clincial outcomes remain highly variable, necessitating alternative approaches.^3,26,27

The two primary factors leading to poor results are adhesion formation within the digital sheath and repair-site elongation and rupture. Adhesions severe enough to limit range of motion occur in up to 40% of flexor tendon repairs.²⁸ While adhesions can be decreased with passive motion rehabilitation,^6,29 they still occur frequently, even with closely controlled techniques.^25,30 Experimental studies report repair-site elongation and gap formation preventing satisfactory healing in up to 48% of canine FDP tendons undergoing state-of-the-art operative repairs. In a clinically relevant, controlled canine repair model, repair site gap formation during the first six postoperative weeks did not correlate with formation of intrasynovial adhesions or loss of digital motion.³¹ In clinical settings, surgeons pursue a balance between repair and rehabilitation approaches promoting tendon strength and digital excursion.³² Flexor tendon repair complications are attributed to a slow accrual of repair-site strength and stiffness and to an increase in gliding resistance within the digital sheath during the first few weeks following tendon suture.^31–39 The healing of paucicellular, hypovascular intrasynovial tendon appears to be limited by the relatively low levels of collagen synthesis and remodeling during the early stages of healing.^40,41

Recent approaches in the canine model seek to increase time-zero strength, enabling better coaptation of tendon stumps, by increasing interaction between the suture and tendon tissue. Adhesive-coatings on sutures increase the interaction and distribute load transfer over a longer length of suture. Mechanically optimized adhesive coatings have potential to improve repair strength by several fold.⁴² Experimental crosslinking agents coating sutures, including 1-ethyl-3-(3-dimethylaminopropyl) carbo-diimide hydrochloride (EDC) and cyanoacrylate, also increase suture-tendon interactions and crosslink the tendon tissue immediately adjacent to the suture.^43,44 These mechanical approaches offer an opportunity to improve repair strength, but do not inherently decrease adhesions or enhance the healing process.

Therefore, we look to biological approaches, such as the application of growth factors and mesenchymal stem cells (MSCs), for the next generation of approaches to improve tendon and ligament repair.^{37,39,45–48} The goal of recent studies has been to: (1) promote matrix synthesis at the tendon repair site leading to increased biomechanical strength and stiffness and (2) suppress matrix synthesis along the tendon surface and synovial sheath preventing adhesion formation.^31,33,35,36 Biological approaches to augment repair have the potential to advance both of these goals. This review summarizes recent cell and biologic-based experimental treatments for flexor tendon injury, with an emphasis on large animal translational studies.

Flexor tendon natural healing response

Similar to healing paradigms in other tissues, intrasynovial flexor tendons follow three successive, overlapping stages of healing: acute inflammation (days 0–7 post injury), proliferation (days 3–14), and remodeling (days 10+).^32,35,49 The commonly injured region of the flexor tendon is intrasynovial, defined as Zone II by Kleinert and Verdan.⁵⁰ The tendon lies within a synovium-lined fibro-osseous sheath that extends from the distal aspect of the palm to the distal aspect of the A4 pulley. Intrasynovial flexor tendons are paucicellular⁵¹ and hypovascular,^52,53 with limited blood supply delivered by long and short vinculae originating from the digital arteries and supplying the tendon segmentally.⁴¹ In addition, the tendon receives nutrients and lubrication from the synovial fluid produced by the tendon sheath.^3,32 As healing intrasynovial tendon has few intrinsic cells and has limited vascularization, there is little intrinsic healing from tendon fibroblasts until delayed time points. At early time points, cell proliferation and matrix synthesis are dominated by cells that migrate to the injury site (Figure 1).^32,33,35 As a result, zone II flexor tendon injuries have substantially poorer healing outcomes following operative repair than do tendon injuries to extrasynovial flexor tendons.^3,8,15

Figure 1.

Representative histologic sections of healthy and repaired canine flexor tendons 1, 3, and 9 days post-operatively. The sections were stained with H&E and viewed under bright field for cell identification. An overview of a representative section from each time point is shown to the left (4x objective, 2mm scale bar). High magnification images (20x objective, 200mm scale bar) of the section outlined in blue are shown to the right. Inflammatory cells are seen infiltrating the repair site via the tendon surface.

Acute inflammation in the first several days after tendon injury attracts circulating inflammatory cells to the injured tendon.^35,54,55 This inflammatory infiltration is dominated by polymorphonuclear cells during the first day, especially in the fibrin clot that forms at the repair site, followed by a transition to monocytes and macrophages by the third day.⁴⁹ Activated macrophages exhibit two major phenotypes: M1 and M2. The M1 macrophages, prevalent during acute inflammation,^56,57 promote extracellular matrix deposition (scar) and inflammation,^55,58 bridging the transected tendon ends but also leading to adhesions. Following acute inflammation, the proliferative phase of healing ensues. In addition to M1 macrophages,⁵⁵ there is an increase in the number of fibroblast-like cells synthesizing extracellular matrix at the proliferative phase.⁴⁹ Most of the fibroblast-like cells are likely derived from epitenon cells⁴⁹ and resident tendon fibroblasts.⁵⁹ Morphologic studies of repaired canine tendons at 7 days after tendon transection and repair show that regions with well coapted collagen fibers had a stronger endotendon response compared to those where the gap only had a few fibrinous strands serving as a scaffold for epitenon cell migration.³⁵ New blood vessels emerge at the surface of canine tendons 9 days following suture.⁴⁹ By 14 days, repaired canine tendon stumps show spontaneous neo-vascularization.³⁵ The final phase, remodeling, lasts many weeks to months, during which M1 macrophages subside and M2 macrophages appear. M2 macrophages suppress inflammation, promote matrix deposition, and facilitate tissue remodeling.^55,56,60 Reorganization of the granulation tissue at the repair site leads to improved tendon strength.

Animal models

The most commonly used animal models for studying flexor tendon repair and tendon rehabilitation^18,61 are the canine, mouse, horse,^62–64 rabbit,⁶⁵ and chicken.^66–69 The canine model for Zone II FDP tendon laceration and repair has been extensively used since 1962.^1,70 Canine flexor tendons are similar to human flexor tendons in both anatomy and function,^61,71 as well as in response to tendon injury, repair, and rehabilitation.^3,24 The canine FDP tendon size is approximately one half the size of a human FDP tendon. Approximate size match enables surgeons both to perform surgical repairs identical to those performed clinically and to achieve similar time-zero mechanical strength to that seen in humans.^72,73 The canine Zone II FDP tendon repair surgical model allows direct testing of surgical modifications and biological approaches before performing clinical trials in humans.^{24,25,43,48,74–76}

Several groups are currently investigating murine models for flexor tendon repair.^{46,54,59,77–80} These models offer high genetic versatility and low cost, enabling in vivo studies of the healing response, biology of adhesion formation,^54,59,79 and effects of biological interventions.⁴⁶ However, the models and hypotheses tested need to be considered carefully due to anatomic and technical challenges that limit clinical relevance. Specifically, the small size of the tendon requires a simpler surgical technique using 8-0 caliber or smaller suture. Furthermore, to prevent repair rupture, all murine models to date require either partial laceration, which modifies the healing process, or proximal unloading. Wong and colleagues perform a partial laceration in Zone II in the murine digit.⁵⁴ Other groups opted to fully or partially lacerate the extrasynovial Zone III tendon and perform proximal transection to protect the repair,^59,78,80 leading to large scar formation between tendon ends.⁷⁹ Finally, rehabilitation postoperatively cannot be controlled due to the small size of the animal. Despite these limitations, the availability of transgenic mouse models opens up possibilities for mechanistic basic science experiments, including cell lineage tracing, gene deletion, and cell ablation.

Biologic treatments

A number of recent reports have indicated that biological approaches, such as the application of growth factors and mesenchymal stem cells (MSCs), have the potential to improve tendon and ligament repair.^{37,39,45,47,48,81} By introducing cells into the paucicellular intrasynovial flexor tendon milieu and inducing a developmental paradigm between the repaired tendon ends, biological approaches attempt to accelerate healing and regenerate normal tissue. Multipotent MSCs from a variety of adult tissues have an excellent capacity to differentiate into the relevant tissue-specific phenotype and to provide potent immunosuppressive and anti-inflammatory effects.^82,83 However, MSC delivery has been ineffective in improving the strength and stiffness following the repair of intrasynovial tendons in vivo.⁴⁸ Similarly, likely due to the paucity of tendon fibroblasts in the region of repair, growth factor application in isolation has been unsuccessful in stimulating enhanced tensile properties following tendon suture, although some improvements have been achieved in digital range of motion.^37,39,84 This has led to more recent focus on combinations of growth factors, cells, and specialized delivery approaches to improve flexor tendon repair.

Delivery of biofactors

Several biofactor delivery approaches have been investigated to improve healing after flexor tendon suture. The simplest delivery method, systemic drug delivery, has not been widely adopted clinically due to low bioavailability at the tendon and concern of side effects. Oral nonsteroidal anti-inflammatory drugs (NSAIDs, e.g., ibuprofen) have been used, with varying results, to limit adhesions experimentally and clinically.^85–88 Local bolus delivery of cells⁸⁹ or growth factors^90,91 by simple injection has yielded limited results, since few cells graft to host tissue without a supporting scaffold and the delivered growth factor is rapidly cleared from the repair site⁹². Recent studies have shown that biological interventions require controlled spatiotemporal delivery to the repair site to improve tendon healing.^39,93–96

In order to effectively deliver cells and growth factors to the repair site, two major tissue engineering paradigms have been investigated using a variety of scaffold biomaterials. Approaches typically either interpose cell- and/or growth factor-seeded scaffolds between the repaired tendon stumps^97–99 or deliver scaffolds on the surface of the repaired tendon.^48,100 Interposition delivers factors directly to the injury site where they are needed for repair, but scaffolds may form a barrier between the tendon stumps that is detrimental for healing.¹⁰¹ Our group has explored scaffold delivery in a longitudinal slit made within the canine flexor tendon, enabling factor delivery to the injured site while retaining tendon stump coaptation. However, the slit was found to have injurious mechanical effects that must be overcome before improvement in healing can be achieved.^81,84,102 Alternatively, scaffolds placed on the surface of tendon adjacent to the repair site deliver factors to the general vicinity but rely on biofactor diffusion or migration to impact the repair itself. Furthermore, scaffolds wrapped around the tendon may induce adhesions or cause excessive bulking that limits tendon gliding within the fibro-osseous sheath. To minimize adhesion formation, lubricating biomaterials such as lubricin and hyaluronic acid^{46,48,103–107} and anti-mitotic drugs such as 5-fluorouracil^33,108 have been successfully utilized in animal models. While these materials improve tendon gliding, some studies have shown deleterious effects on repair strength.^48,103,109 Other materials including silicone, polyethylene, and cellophane have been used clinically as an artificial sheath to reduce adhesions, but have not gained widespread acceptance in the United States.^2,3,110,111

Biomaterial selection is crucial to the function of tissue engineered scaffolds. Fibrin delivery systems with heparin-bound growth factors have enabled sustained drug delivery during healing,^{39,76,81,93,102,112,113} as have some microsphere-based approaches.^114–116 The scaffold backbone is also essential for promoting stem cell integration and differentiation. Scaffold mechanical properties and fiber diameter influence cell activity and differentiation.¹¹⁷ Synthetic polymer approaches using electrospun polylactic co-glycolic acid (PLGA) nanofiber scaffolds have provided a strong, fibrous backbone and delivered viable cells and growth factors to the repair site (Figure 2). However, these scaffolds release acidic byproducts that increase the proinflammatory cytokine IL-1β and negatively impact healing.^81,102 Naturally occurring polymers, including collagen- and fibrin/heparin-based delivery systems, have been explored for their enhanced biocompatibility. Future delivery approaches should be biocompatible, appropriate for cell seeding, able to provide sustained growth factor delivery, and have appropriate surgical handling characteristics for implantation into the relatively dense tendon tissue.

Figure 2.

A representative PLGA-fibrin scaffold with 11 alternating layers of aligned electrospun PLGA nanofiber mats separated by fibrin containing adipose-derived MSCs. (A) Micrograph showing the scaffold in vitro; the PLGA was labeled with FITC (green), the fibrin was labeled with Alexa Fluor 546 (red) and the adipose-derived MSC nuclei were labeled with Hoechst 33258 (blue) (scale bar = 200 μm). (B) Micrograph showing the scaffold in vivo 9 days after implantation in a canine flexor tendon repair (scale bar = 100 μm). (C) A schematic of the layered scaffold is shown.

Growth factor treatments

The growth factors bone morphogenic protein (BMP) 12, BMP13, and BMP14, (a.k.a., GDF7, GDF6, and GDF5, respectively) which are expressed in developing tendons and ligaments, have been shown to have the greatest potential for improving tendon healing.^45,118–122 These BMPs act by inducing tenogenesis in stem cells in vitro via Smad 1/5/8 phosphorylation.^{45,118,121,123} BMP12 effectively increased the expression of the tendon markers scleraxis and tenomodulin in canine adipose-derived mesenchymal stromal cells (ASCs) in vitro at both mRNA and protein levels.⁴⁵ Consistent with these results, BMP12 induced scleraxis promoter driven-GFP and tenomodulin expression in mouse ASCs. BMP12 administration concurrently reduced expression of the bone marker osteocalcin, but not the osteogenic transcription factor runx-2. There was a mild increase in the expression of the cartilage matrix gene aggrecan, though still to considerably lower levels than those detected in tendon fibroblasts. BMP14 had similar but less potent effects.⁴⁵ However, these factors alone, without concurrent cell delivery, have not been sufficient to improve repair strength. Hayashi et al. interposed collagen gels with BMP14 without cells between cut ends of canine FDP tendon under in vitro tissue culture conditions, but this did not significantly change ultimate healing strength or stiffness compared to repaired controls.⁴⁷ Similarly, adenoviral-mediated gene transfer of human BMP13 did not improve healing in a rat rotator cuff repair model.¹²⁴

Several other growth factor approaches have attempted to promote cell proliferation and matrix synthesis in order to improve flexor tendon healing. Exogenous basic fibroblast growth factor (bFGF) was shown to accelerate wound closure and promote fibroblast proliferation and matrix synthesis in vitro in rat patellar tendon fibroblasts and in canine tendon fibroblasts.^76,125,126 In an in vivo chicken flexor tendon injury and repair model, adeno-associated virus-2 carrying bFGF was directly injected into injured tendons. The bFGF treated chicken flexor tendons had significantly higher ultimate strength at 2, 4, and 8 weeks, fewer ruptures, and similar adhesion scores.⁶⁹ These promising in vitro and chicken in vivo studies motivated testing the effects of sustained bFGF delivery to injured flexor tendons in the clinically relevant canine large animal model. Sustained delivery of biologically active bFGF was achieved by incorporating bFGF into a fibrin delivery system. The bFGF release profile was tuned by bFGF dosage and heparin concentration, where increasing concentrations of heparin significantly slowed release. At a 1:1000 growth factor to heparin ratio, 37% of the loaded bFGF was released within the first 2 days and 71% was released within the first 10 days by passive release in vitro (i.e., mediated by diffusion and degradation of the fibrin carrier). Active bFGF release (i.e., mediated by canine tendon fibroblast cells) also achieved a sustained delivery profile over at least 10 days. The released bFGF stimulated increased cell number, increased gene expression of the extracellular matrix lubricating proteins lubricin and hyaluronic acid synthase 2, increased expression of the degradation proteins matrix metalloproteinase 1 and 13, and decreased expression of collagen I and III.⁷⁶ However, at 21 days in vivo in the canine flexor tendon injury and repair model, sustained bFGF delivery not only accelerated the cell-proliferation phase of tendon healing, but also promoted neovascularization and inflammation in the earliest stages following the suturing of the tendon. Despite a substantial biologic response, the administration of basic fibroblast growth factor failed to produce improvements in either the mechanical or functional properties of the repair. Rather, increased cellular activity resulted in peritendinous scar formation and diminished range of motion (Figure 3).⁸⁴

Figure 3.

There were more adhesions (arrows) between the flexor tendon surface and its surrounding sheath in the bFGF-treated tendons (top panel) than there were in the non-bFGF-treated control (CTL) tendons (bottom panel). Arrows indicate adhesions proximal to the repair site in a tendon treated with 1000 ng bFGF. The paired control tendon did not have any apparent adhesions.

Similarly, exogenous platelet derived growth factor-BB (PDGF-BB) was shown to stimulate canine flexor tendon fibroblast proliferation, decrease collagen I and III gene expression, and increase lubricin, hyaluronic acid synthase-2, and matrix metalloproteinase 1 and 13 gene expression in vitro.^76,93,126 Sustained delivery of PDGF-BB using a fibrin/heparin based delivery system in canine flexor tendon repairs in vivo consistently improved range of motion, collagen remodeling, and cell proliferation at 14, 21, and 42 days following suture. However, PDGF-BB delivery did not increase repair strength at any timepoint.^37,39,40,81

Transforming growth factor β (TGFβ) has also produced mixed effects on flexor tendon healing: while TGFβ isoforms are important during flexor tendon healing to stimulate collagen production, they also lead to fibrosis and scar formation.^38,127–131 While some studies used TGFβ with the goal of improving strength,¹¹⁸ many have attempted to inhibit TGFβ in order to reduce adhesions, albeit at the cost of reducing repair strength.^132–136 This decrease in repair strength precludes the use of TGFβ inhibitors as a standalone therapy. Successful TGFβ-modulating treatments would require controlled spatiotemporal activity to enhance collagen production in the healing tendon itself, but decrease scar and adhesion formation at the tendon surface.

A promising recent growth factor approach is based on connective tissue growth factor (CTGF), which has been shown to induce MSC differentiation into tendon fibroblasts and/or chondrocytes.^137–139 Similar to BMP12 and positively promoting BMP12 effects, in vitro CTGF effectively increased the expression of the tenocyte lineage markers scleraxis and tenomodulin, as well as the fibroblast proteins collagen I and tenascin-C.¹³⁸ During rat rotator cuff healing, CTGF is highly expressed in the tendon midsubstance and at the tendon-to-bone insertion for several weeks following injury.¹⁴⁰ In chicken flexor tendons, CTGF is relatively highly expressed in normal tendons and throughout healing.¹⁴¹ CTGF and cell combination studies are described below.

Cell treatments

As noted above, early enthusiasm for cell therapy, based on patellar tendon^89,142 and Achilles tendon¹⁴³ results, has been largely unsuccessful in rotator cuff¹⁴⁴ and flexor tendon animal models. In two studies using a canine in vitro tissue culture model, interposition of a multilayered collagen patch seeded with bone marrow-derived MSCs into the repair site did not improve flexor tendon healing mechanics compared with control repairs without interposed patches.^47,145 MSC implantation in vivo in rabbits decreased adhesions but did not improve biomechanical properties 3 or 8 weeks after surgery.¹⁴⁶ Racehorses that received direct injection of bone marrow-derived MSCs during superficial digital flexor tendon repair had reduced re-injury rates compared with historical controls,^147–149 however, the equine superficial digital flexor tendon has substantially different functional, structural, and material properties from human FDP tendon.^63,150

Cell-growth factor combination treatments

Though cells and growth factors in isolation have not markedly improved flexor tendon healing, combination therapies offer greater potential to improve outcomes. While interposition of bone marrow-derived MSCs only or BMP14 only did not improve repair mechanics in an in vitro canine flexor tendon tissue culture model, the combination of MSCs with BMP14 or platelet-rich plasma on collagen patches improved strength and stiffness.^47,145 This approach, combined with surface lubricin for in vivo canine flexor tendon repairs to decrease adhesions, unfortunately resulted in substantially worse repair strength 42 days after repair.⁴⁸ Similarly, application of adipose-derived MSCs in combination with BMP12 in an in vivo canine Zone II flexor tendon repair using PLGA and fibrin scaffolds led to increased total collagen compared to repairs with acellular scaffolds, but did not improve tensile properties at 28 days after surgery compared to the acellular group. The delivery method used in these studies was a critical component driving the outcomes: the PLGA-fibrin scaffolds had a deleterious effect that may have counteracted any beneficial effects from the MSCs and/or BMP12.¹⁰² A previous study delivering the same PLGA-fibrin scaffolds containing MSCs and PDGF-BB demonstrated retained cell viability after 10 days, but also mild inflammatory reactions, possibly due to the PLGA scaffold (Figure 2).⁸¹

Connective tissue growth factor and cell combination approaches have not been thoroughly evaluated in flexor tendon in vivo, but CTGF-based approaches show promise in other tendon repair scenarios. Tendon-derived CD146+ stem cells cultured with CTGF promoted tenogenic differentiation in vitro.¹⁵¹ Tendon-derived stem cell sheets stimulated with CTGF promoted improved anterior cruciate ligament graft healing and biomechanics in vivo in rats, including improved osteointegration.¹⁵² Similarly, cell sheets with CTGF and ascorbic acid enhanced biomechanical and histology-based outcomes at 8 weeks in an in vivo rat patellar tendon repair model. Further studies introducing CTGF and/or BMP growth factors with cells and biocompatible matrices will be important for defining the next generation of therapies for flexor tendon repair.

Conclusions

Intrasynovial flexor tendons are notoriously challenging to repair, with highly variable clinical outcomes due to the competing requirements for repair strength to avoid rupture and minimal adhesion formation to maintain adequate range of motion. Advances in rehabilitation protocols have led to significant improvements in tendon healing and gliding outcomes, but post-operative complications remain. Biological approaches have potential to simultaneously tackle both of these problems. Successful biological therapies will require carefully controlled spatiotemporal delivery of cells, growth factors, and biocompatible scaffold matrices in order to simultaneously promote matrix synthesis at the tendon repair site to improve strength and suppress matrix synthesis along the tendon surface to prevent adhesion formation.