Engineering interfacial tissues: The myotendinous junction

The myotendinous junction (MTJ) is the interface connecting skeletal muscle and tendon tissues. This specialized region represents the bridge that facilitates the transmission of contractile forces from muscle to tendon, and ultimately the skeletal system for the creation of movement. MTJs are, therefore, subject to high stress concentrations, rendering them susceptible to severe, life-altering injuries. Despite the scarcity of knowledge obtained from MTJ formation during embryogenesis, several attempts have been made to engineer this complex interfacial tissue. These attempts, however, fail to achieve the level of maturity and mechanical complexity required for in vivo transplantation. This review summarizes the strategies taken to engineer the MTJ, with an emphasis on how transitioning from static to mechanically inducive dynamic cultures may assist in achieving myotendinous maturity.


INTRODUCTION
Musculoskeletal conditions continue to plague our community with approximately 1.71 Â 10 9 people affected worldwide, making them the leading contributor to disability globally. 1Although only being responsible for 2% of total hospital discharges, musculoskeletal conditions reported 5.4% of the total hospital costs for children and adolescents aged 20 years or younger, correlating to a $7.6 billion economic burden.In fact, in 2011, there was an estimated $213 billion annual cost of direct treatment and lost wages in the United States alone. 2Arising from various diseases, injuries, and myopathies, musculoskeletal conditions are largely incurable or irreparable.Trauma to muscles and tendon alone contributes 20.6% of all work-related musculoskeletal injuries in Australia (Safe Work Australia, 2016), in which the MTJ represents the primary site of injury at 58.7%. 3 Current surgical methods, such as suturing, allografts, and xenografts, result in residual scar tissue that disrupts the fragile biomechanics of the muscle-tendon interface, and hence do not achieve the therapeutic rehabilitation required for patients to return to pre-injury activities.Although the interdisciplinary field of tissue engineering has previously encompassed interfacial tissues, this domain remains relatively unexplored, particularly regarding the MTJ.Groups that have targeted the MTJ have observed interface-specific markers indicating MTJ development; however, current efforts have failed to recapitulate the maturity and mechanical complexity of native MTJs.Herein, we review the finite studies targeted at MTJ regeneration, revealing a common trend that may be extrapolated to other interfacial tissues.In particular, this review focuses on the mechanosensitive traits of skeletal muscle and tendon constructs in three-dimensional (3D) culture, in which the prospect of maturing MTJ constructs under uniaxial strain has been explored.
To rationally design a MTJ engineering strategy, we should consider the process by which it is originally formed.Herein, we begin by describing the development and mechanical characteristics of MTJ constituents in utero, followed by the biomechanics of fully developed MTJs in humans.Subsequently, through summarizing the finite studies targeted at MTJ engineering, the lack of dynamic culture methods has been identified as a limitation.Finally, this gap has been linked back to the established literature on individual tissue responses to uniaxial strain in dynamic cultures, revealing a targeted future perspective aimed at recapitulating the native environment more appropriately to be identified.

MTJ CHARACTERISTICS Embryonic development
MTJ formation is unequivocally reliant on the interactions between developing tissues during embryogenesis.Originating from the myotome, myogenesis relies on a complex, yet not fully elucidated signaling network comprised of molecules, such as Wingless-related integration site proteins (Wnts), sonic hedgehog (Shh), and bone morphogenetic proteins (BMPs). 4These molecules, especially Wnt1 and Wnt3a, heavily influence the expression of myogenic regulatory factors (MRFs) responsible for myogenic lineage progression and differentiation. 4,5Contrarily, tenogenesis originates from the undifferentiated mesenchymal cells within the lateral plate mesoderm. 6Initially, Shh signals derived from the notochord induce paired box transcription factors (Pax) 1 and 9, which regulate mesenchymal differentiation.This process ultimately commits SRY-box transcription factor 9 (Sox-9)-expressing chondrogenic mesenchymal cells to their tendinous fate. 7Post mesenchymal differentiation, tendon progenitors express various transcription factors, including Scleraxis (Scx), Mohawk (Mkx), early growth response 1 (Egr1), and early growth response 2 (Egr2). 8J formation begins as tenocytes attach to skeletal muscle cells (myocytes).The chemical and mechanical signaling between these tissues during embryonic development is poorly understood; however, electron microscopy has been widely incorporated to analyze the origins of the MTJ.The earliest morphological modification observed at the MTJ is the formation of close associations between myogenic cells and tendon fibroblasts. 9Here, a dual identity can be seen as fibroblasts transdifferentiate by switching on a myogenic program, allowing fusion into myofibers. 10Simultaneously, extracellular material accumulates at the surface of muscle cells, representing the first appearance of the basement membrane. 9,11Myofibril production subsequently increases, allowing subsarcolemmal densities to appear at the intracellular surface of the MTJ. 9 Finally, associations between myofibril thin filaments and subsarcolemmal densities occur, resulting in membrane folding, later depicted as invaginations. 11o date, there has been limited reported data regarding the mechanical properties of embryonic skeletal muscle, particularly in mammals.Although myotendinous development is a generic process in vertebrates, Fig. 1 displays not only the variability between species but also the rapid formation in utero as evident in the increased stiffness and ultimate tensile strength (UTS).A study by McBride et al. focused on individual tendons adjacent to the femur and tibiotarsus of fertilized white leghorn eggs.At post-fertilization day (PFD) 14, a Young's modulus (E), UTS, and strain at failure (SAF) were determined to be 0.216 6 0.060 MPa, 2.052 6 1.112 MPa, and 12.77% 6 1.91%, respectively.At PFD 17, these values had increased to 1.02 6 0.27 MPa, 21.411 6 2.634 MPa, and 29.83 6 5.33, indicating not only the rapid pace of embryonic tendon development but also the disparity between embryonic and adult tendon mechanical properties. 12This disparity may be seen by Nakagaki et al.'s elastic modulus of caged, 8-month-old calcaneal tendon of chickens (210.51 6 46.01 MPa). 13It has been determined from Marturano et al.  that the elastic modulus of the embryonic tendon increased nonlinearly as a function of embryonic stage at both the nano-and microscale. 14A further in-depth review of embryonic development may be found here. 15

Composition
For the MTJ to develop into an integrated mechanical unit that can support and transmit great forces, many connective complexes must be considered. 26These connections must not only allow high tensile stresses to be endured but also act as the junction in which intracellular myofilament contractions transmit force to extracellular proteins found in the tendon. 27In total, five milestones of MTJ formation have been reported in the literature, including 26,[28][29][30][31] (i) Actin microfilaments extend from the last Z-line of sarcomeres and terminate in the ridge-like protrusions.(ii) Actin-binding proteins cross bind highly aligned actin filaments together.(iii) Actin filament bundles insert into the sarcolemma plasma membrane via an electron-dense subsarcolemmal layer.The anchorage of these actin filaments may be perpendicular or oblique.(iv) Transmembrane proteins that link cytoskeletal elements to basal lamina components.Basil lamina is seemingly thicker in these regions.
(v) Proteins that link the basement membrane to collagen fibril-rich matrix in such a way that the collagen fibrils are parallel to the muscle myofilaments.
Two major and distinct transmembrane linkage systems have been described at the MTJ, which rely on the dystrophin-associated glycoproteins complex (DGC) and the binding and signaling protein "a7b1" integrin.Both systems constitute a structural link between cytoplasmic actin and tendinous extracellular matrix proteins.These links heavily rely on laminin 211, which is prevalent in both complexes, and represents the unique isoform that is found in adult human MTJs. 26,27

Ultrastructure
The adult MTJ contains a complex morphology to combat stress concentrations and avoid injury.Advancements in scanning techniques, such as transmission and scanning electron microscopy (TEM and SEM, respectively), and focused ion beam (FIB) have steered scientists away from the idea of a simplistic planar divisional interface between skeletal muscles and tendons, revealing a more complex, interwoven design.Here, tendinous extracellular matrix (ECM) folds and protrudes into invaginations of the muscle cell membrane (sarcolemma) as represented in Fig. 2. In fact, while studying the human MTJ in 3D, Knudsen et al. more accurately described the connection between tissues as ridge-like protrusions 30 that serve two primary mechanical purposes.First, the interwoven nature of the MTJ greatly increases the contact area between skeletal muscles and tendons, therefore significantly reducing stress concentrations. 322D TEM images taken by Noonan et al. revealed a 10-20-fold increase in surface area compared to a smooth planar transition; 33 however, recent 3D modeling by Knudsen et al. indicates an even greater increase. 30Second, the intertwining design positions the muscle cell membrane at acute angles relative to the applied force, compelling the sarcolemma to be predominately exposed to shear forces. 34Shear forces are optimal as cell membranes display greater strength of adhesion while subject to contractile shear stresses as opposed to tension loading. 35Thus, the ability of the MTJ membranes to transmit force is accentuated through invaginations which result in greater load-bearing capabilities.The ultrastructure of the MTJ may be seen below in Fig. 2.

Biomechanics
Generally, collagen fibrils within the tendon are crimped until the onset of strain, resulting in an initial toe region of a stress-strain curve up until 2%-2.5% strain, 36 in which tendonous strains primarily reside under physiological conditions. 37Once stretched beyond 4%, plastic deformation occurs, representing mild injury.If stretched over 10%, however, complete rupture will occur. 38,39The tendon is also described as a viscoelastic tissue, meaning compliance at low strain rates and resilience at high strain rates.This was elegantly displayed by Wren et al., who increased the rate of strain from 1 to 10 mm/s during tensile loading of the human Achilles tendon (AT). 40Here, a greater strain rate resulted in increased E, UTS, and SAF from 816 6 218 to 822 6 211 MPa, 71 6 17 to 86 6 24 MPa, and 7.5% 6 1.1% to 9.9% 6 1.9%, respectively.
To date, insufficient data have been acquired regarding the in vivo biomechanics of the MTJ due to the difficulty in attaining results during physical activity.Therefore, postmortem destructive tests are the primary source for mechanical properties of the MTJ.Several studies 27,41 claim that the mechanical properties of the MTJ are that as displayed in Table I.
Upon further investigation, many of these references are obtained from studies conducted on not only different tissues but also different species.In fact, analysis of the original data referenced in these reviews provides greater insight into the variability of myotendinous tissue (see Table II).Here, Ã represents the corresponding value presumed to be identified in summary Table I.
Comparing the original data to the referenced values (Table I), although being within similar range, this table appears to be an oversimplification of MTJ properties as it does not reveal the extent to which biomechanics vary upon location or species.Considering postmortem mechanical properties of the human MTJ constituents, Loren and Lieber studied tendons of the human wrist, observing E (at maximal tetanic tension), UTS, and SAF of 438.   s expected, there is a greater abundance of studies conducted on animals as opposed to humans; however, the study by Pollock et al. revealed that the elastic properties of tendons do not vary significantly from animals of different body mass. 550]67 Focusing on living human mechanical properties, Maganaris et al. stimulated the human tibialis anterior (TA) muscle using conductive aluminum pads, enabling ultrasonography and magnetic resonance imaging of the connecting AT. 68 By stimulating isometric loads, a maximum stress, strain, and stiffness of 25 MPa, 2.5%, and 1.2 GPa were found.Conversely, Kongsgaard et al. used a calf raising apparatus to study the AT revealing a max stress, strain, and modulus of 29 6 3MPa, 4.2% 6 1.1%, and 2.0 6 0.4 GPa, respectively. 69urthermore, the Achillies MTJ averaged a proximal displacement of 7.1 6 0.9 mm.A summary of living human mechanical properties of MTJ constituents under load can be seen below in Table III.
Apart from stimuli type and analysis method, simulation protocol, subject age, and sex, the definition of free tendon plays an important role in comparable outcomes.For example, previous methods have estimated Achilles tendon mechanical properties from measurements of distal medial gastrocnemius MTJ displacement in relation to an external marker. 69This method, however, does not account for elongation or contraction of the tendon-aponeurosis structure and, therefore, the results vary.It is concluded that heterogeneous analysis methods and tissue types produce inconsistent outcomes while considering the biomechanics of myotendinous tissues.

Injuries and current interventions
Despite the adaptable and complex inner workings of the MTJ, injuries are still prevalent. 86Although most tears occur near the MTJ, as opposed to through it (indirect tears), 78 the MRI reveals the  [79][80][81] Tears within close proximity to the MTJ generally occur between the cell membrane and lamina densa of the basement membrane. 35hese injuries mainly arise from a passively overstressed tissue, or fast eccentric contractions, where the skeletal muscle contracts while lengthening. 82In fact, eccentric forces greater than 20% of the average maximum isometric force are sufficient to induce rupture. 29n skeletal muscles, it is widely accepted that either disrupted sarcomeres, or alternatively, damage to the excitation-contraction coupling system represents the immediate signs of injury. 83Furthermore, sarcomeres may be disturbed as a lack of homogeneity results in uneven energy absorption. 84Most commonly, muscles that are exposed to these contractions include, but are not limited to, the hamstring, rectus femoris, hip adductor, and calf muscles. 82Although a universal grading system is yet to be agreed upon, a classification system compromising four grades are identified by location (distal, middle, and proximal), severity, and symptoms.These grades are as follows: [85][86][87][88][89][90] • Grade 0: edema or fluid adjacent to an intact tendon/aponeurosis/epimysium without myofibril detachment.• Grade 1: myofibril detachment without tendon/aponeurosis/epimysium change.
Grade 1 and 2 tears may be treated with non-steroidal antiinflammatory drugs (NSAIDS), protection, rest, ice, compression, and elevation (PRICE) protocol, 91 and physiotherapy where a full recovery is expected, whereas grade 3 tears require surgical intervention. 92urrently, suturing of allografts and xenografts are the only available clinical option.A review of current suturing approaches for the MTJ may be found here. 93Although this generally allows for the reattachment of the torn muscle-tendon unit, significant scar tissue results in compromised biomechanics and increased likelihood of reinjury. 93In cases where the interface is completely severed, suture may not allow the possibility to re-join the MTJ.Since most myotendinous injuries are untreated, surgical outcomes are not well documented. 92,94urthermore, if surgery is delayed, muscular atrophy may make it impossible to re-attach.Although experimental pharmacologic agents are still being explored, 95 suture-based approaches remain the predominant surgical intervention for repair of a grade 3 MTJ injury.Suturebased approaches remain the predominant surgical intervention for repair of a grade 3 MTJ injury.

TISSUE ENGINEERING
Tissue engineering (TE) approaches provide great hope to assist the natural repair of musculoskeletal diseases.Recently, a trend has been developing around the hypothesis that tissue-engineered scaffolds that mimic embryonic developmental forces may dictate the differentiation of embedded cells down a specific lineage more effectively.These approaches, however, remain undeveloped and insufficient due to the limited understanding of tissue interactions during embryogenesis.Extracellular forces that contribute to embryonic development may, however, be applied to tissue-engineered scaffolds in hopes of recreating developmental-specific microenvironments.For example, Kardon et al. revealed that in the avian hind limb, the initial morphogenetic events, formation of tendon primordia, and initial differentiation of myogenic precursors occur autonomously with respect to one another. 96This effect was further iterated by Kieny et al. who studied the development of an embryonic chick wing.Here, they concluded that tendons start to develop autonomously from the muscle bulks, but for their maintenance and further development they require connexion to a muscle belly. 97Although being imperative to interfacial tissue development, in utero relationships remain heavily unexplored, therefore, tissue engineering advancements are predominantly governed by the understanding of individual tissues as opposed to the relationship between the two.Thus, the pillars of tissue engineering (cells, growth factors, and scaffolds) must be adapted to house complementary environments for both skeletal muscles and tendons prior to in vivo transplantation.Few attempts have been made to engineer the MTJ, some of which utilize scaffolds to provide initial mechanical support and disparity between skeletal muscles and tendons.Not only do scaffolds provide mechanical support but they also pre-define tissue geometry and replicate the ECM, thus allowing for provisional cellular adherence, migration, 98 proliferation, and differentiation. 99To achieve desired outcomes, the following five parameters should be considered and manipulated in such a way that provokes intended biological responses: 100 1. Biodegradability 2. Biocompatibility 3. Engineered scaffold architecture 4. Surface properties 5. Mechanical properties Further increasing the complexity, these factors are all timedependent, making the ideal scaffold challenging to fabricate.Scaffolds for individual tissues will not be reviewed here, as they have been elsewhere. 101Instead, heterogeneous structures able to mimic dissimilar tissue properties will be analyzed.

Heterogeneous scaffolds for interfacial tissue engineering
Although planar homogeneous scaffolds are most often reported in the literature, they are incapable of producing spatially altered mechanical properties. 102Despite this, architectural adaptations, such as the transition from linear to crimped fibers, may be introduced to make tissue-specific scaffolds.For example, Hochleitner et al. used melt electro-writing below the critical translation speed to obtain sinusoidal polymer fibers. 103Under tensile loading, these fibers, although homogeneous, exhibited a toe region resembling the uncrimping of collagen fibrils in tendinous tissue.Producing a similar structure, Wu et al. fabricated a tendon-bioinspired wavy scaffold via electrospinning, displaying a tensile toe region and UTS consistent with that of the native tendon.Expectedly, however, the tensile modulus within this toe region fell below that of the native tendon, indicating optimization is required. 104In general, electrospun fibers maintain high surface area-volume ratios that mimic the natural ECM nanostructure.One disadvantage of electrospun fibers, however, are their ability to maintain sufficient mechanical properties, particularly for tendons, and, therefore, will only be discussed in hybrid scaffold situations.For example, Sahoo et al. electrospun poly (lactic-co-glycolic acid) (PLGA) nanofibers onto the surface of a knitted PLGA scaffold, 105 resulting in increased proliferation, cellular function, failure load, elastic stiffness, and toe stiffness, whereas cell attachment remained comparable.
Heterogeneous scaffolds are a promising approach for mimicking the mechanical impedance seen at interfacial tissues.Melt electrowriting (MEW)-controlled deposition style presents a unique approach to achieving these mechanical and architectural variances.As practically all transplantable tissues require a 3D geometry, a cylindrical collector is often used for target tissues, such as heart valves, tendon, ligament, and skeletal muscle.For example, Saidy et al. printed polycaprolactone (PCL) onto a 22 mm diameter cylinder for heart valve tissue-engineered scaffolds, 106 creating discrete and controlled architectural changes as diamond-shaped pores turned rectangular [Fig.3(b)].Although not having undergone appropriate tensile testing, compression testing revealed that, as expected, variations in tubular morphology had a direct impact on mechanical properties.
Steering away from biomechanics, changes in pore sizes have been shown to dictate cell attachment, alignment, and growth rate.For example, Xie et al's Melt electro-wrote high-resolution PCL scaffolds with four distinct zones segmented by pore size. 107Here, the proliferation rate of bone marrow-derived stem cells (BMSCs) and human umbilical vein endothelial cells (HUVECs) in small pores was determined to be three times faster than in larger pores, with changes in fiber diameter dictating the spread of cells, attachment, and alignment [Fig.3

(e)]. Similarly, Nguyen et al. seeded NIH-3T3 cells on MEW
PCL scaffolds [Fig.3(f)], confirming that cell growth is considerably influenced by lattice structure grid size, enabling the control of spatially attached cell populations. 108gineering the myotendinous junction To date, advances in tissue engineering the MTJ rely extensively on the knowledge gained from individual tissue types.Engineering strategies can be segregated into scaffold-based and scaffold-free approaches, depending on their intended function.Scaffold-based approaches provide greater promise in mimicking the translational stiffness between skeletal muscles and tendons.Here, individual scaffolds are engineered for skeletal muscles and tendons, where an overlap defines the interface (Fig. 4).
For whole composite structure produced an E and UTS of 7.339 6 2.131 and 0.5058 6 0.2130 MPa, respectively, and could last up to 100 cycles during cyclic loading.Subsequent embedding of C2C12 and NIH/3T3 cells into evenly distributed bovine collagen I provided no reduction in cell viability and accommodated attachment, survival, and differentiation of myoblasts into myotubes.These traits, however, did not progress into evidence of cell re-organization at the interface.
Similarly, Merceron et al. also used spatial deposition of synthetic polymers with a 10% overlap to create an interfacial region. 111Here, bioprinting was utilized to deposit polyurethane (PU) with C2C12 cells (representing muscle), and PCL with NIH/3T3 cells (representing tendon), in which the deposition of cells had a negligible effect on viability.Mechanical characterization of the composite structure displayed a yield strain of 300%, which was expectedly driven by the superior elasticity of PU.Elastic modulus of PU, PCL, and the interface was 0.39 6 0.005, 46.67 6 2.67, and 1.03 6 0.14 MPa, respectively, which, compared to Ladd et al., were mechanically inferior.On the muscle side, C2C12 cells expressed both desmin and myosin heavy chain (MHC), aligned along a unilateral plane, and began to show multinucleation, indicating their differentiation into myotubes.On the tendon side, cells began to secrete collagen I, producing a distinct interfacial region (Fig. 5).Focusing on the junction, upregulation of MTJassociated genes, such as pax, talin 1 (tln1), vinculin, integrin b1, laminin a1, and laminin a2, were prevalent.
Yet another example of bioprinting's efficacy in creating spatially deposited co-cultures was described by Laternser et al. 112 Here, a dumbbell-shaped construct was deposited in between two posts, where primary human skeletal muscle-derived cells were printed between five successive layers of gelatin methacrylate (GelMA), and tenocytes between a GelMA/polyethylene glycol dimethacrylate (PEGDMA)based ink.Interestingly, tenocytes were printed around posts, whereas skeletal muscle-derived cells were printed in between, with a gap of 3 mm to create a clear border, maintaining a >95% cell viability.Myoblasts formed aligned areas and were able to attach to the tendinous tissue; however, after 1-2 days, the structure tore at one of the Another form of mechanical stimulation for engineering the MTJ was investigated by Kim et al., who used a combination of biochemical and biophysical cues. 113One culture analyzed cell-laden human adipose-derived stem cells (hASC) in bioinks including collagen, muscle-derived ECM (mECM), and tendon-derived ECM (tECM)), in which flow-induced alignment of cells in bioinks was achieved by controlling (i) the collagen concentration in the core and (ii) the flow rates in the core and sheath.Using the idealized flow rates for optimized cellular alignment, 14 days post culture, paxillin, integrin, fibronectin, and talin were found in the MTJ region, suggesting MTJ formation.Furthermore, upregulated expression of integrin b1and MTJassociated genes [Pax, Tln1, thrombospondin 1 (Thbs1), collagen type 1 (Col1a1), laminin alpha 1 (Lama1), myosin heavy chain 2 (Myh2), and Scx] was identified, leading to the conclusion that MTJ formation can be clearly affected by the physical interfacing shape between muscle and tendon cells.
Scaffold-free approaches hold promise in circumventing issues associated with scaffold fabrication, degradation, and reduced biologically active volume.Larkin et al. evaluated the co-culture of skeletal muscle satellite cells and fibroblasts obtained from pregnant Fischer 344 rats based on their ability to create MTJs. 114One week of monolayer culture enabled the muscle and tendon cells to roll up into 3-D constructs around tendon tails, in which a highly aligned interface of collagen fibrils and myotubes formed.Furthermore, although less concentrated than found in adult samples, both paxillin and talin were localized at the interface.Proceeding to tensile testing, an average E of 37.2 6 10.3 kPa (for muscle construct) was revealed, being comparable to one quarter the passive stiffness reported for young adult rat soleus muscle.Conclusively, Larkin et al. successfully engineered constructs resembling neonatal MTJs using a cell approach, greatly expanding the potential to control the phenotype of skeletal muscles in culture. 85ontinuing from this, Kostrominova et al. adopted the same method to engineer self-organizing tendon (SOT) constructs. 115riefly, once myocytes fused to form spontaneously contracting multinucleated myotubes, SOT constructs were pinned onto the muscle cell monolayer, allowing the construct to roll up around the anchors.At the early stages of culture, MTJ-like structures were represented by sub-sarcolemma densities, which later developed into well-defined folding of the plasma membrane.These plasma membranes were surrounded by type I collagen with the characteristic striation pattern.
Several other studies regarding engineered MTJ constructs have been summarized in Table IV and may be visualized in Fig. 5 below.
Currently, the task of engineering interfacial tissues, especially that of the MTJ, reveals the lack of understanding in several other domains.For example, knowledge about how tissue-specific cells interact with the microenvironment and ECM of opposing tissues is scarce.Gaffney et al. decellularized the porcine AT and gastrocnemius muscle to form tissue-specific prehydrogel digests.Seeding of C2C12 myoblasts and tendon fibroblasts in their opposing ECM hydrogel revealed a 50%-70% increase in paxillin expression, and collagen XXII to a lesser extent. 122Most recently, Gaffney et al's continuation of this research represents the only mechanically stimulated study focused on engineering a myotendinous construct. 121Here, C2C12 cells were seeded in either type 1 collagen or tendon-derived extracellular matrix (tECM) and conditioned in a customed designed bioreactor.Cyclic tensile exposure of 10% strain, for 10 800 cycles at 1 Hz (3 h per day) resulted in the upregulation of myosin heavy chains (Myh1, Myh2, and Myh4), Pax, and type XXII collagen.Interestingly, tECM consistently produced greater upregulation of these factors compared to that of type 1 collagen, possibly due to the presence of tendon-specific ECM factors.Furthermore, transmission of uniaxial loads from scaffolds to cells is heavily influenced by hydrogel stiffness, as is actinmyosin striation.In fact, hydrogel mechanical properties have become increasingly influential in dynamic cultures and must, therefore, be considered an imperative aspect of interfacial tissue engineering.As previously stated, different hydrogels are commonly used in MTJ constructs; however, evolving to dynamic cultures involving uniaxial strain introduces more variables, such as tensile modulus, flexibility, and fatigue.Engler et al. found an optimal gel modulus of 12 kPa for skeletal muscle constructs, 48 within the range of native skeletal muscle.Decellularized MTJ (D-MTJ) has only been reported once in the literature by Zhao et al., who seeded muscle satellite cells into D-MTJ derived from porcine AT MTJ.This study failed to investigate the coculture of tenocyte and myoblasts; however, it successfully characterized the mechanical properties of decellularized MTJ. 62Considering human cells, Tsuchiya et al. cocultured semitendinosus and gracilis tendon-muscle tissues, revealing three main findings: (i) humanderived tenocytes do not enhance proliferation of myoblasts, (ii) human-derived tenocytes enhance myotube formation, and (iii) human-derived myoblasts and tenocytes release factors important for myotube formation. 123n essence, the foundation of myotendinous engineering is fragile due to the limited knowledge of cellular interactions in their opposing ECMs.Therefore, it is difficult to identify the exact mechanisms responsible for the expression of interfacial characteristics, especially since this is not fully understood during in vivo development.The majority of data indicating the dependency of skeletal muscles and tendons during development has arisen from NULL studies, which strategically eliminate inputs, such as muscle contraction, to investigate the effect on tendon development.

Trends and limitations
As discussed, all techniques used to engineer the MTJ have employed heterogeneous structures in combination with cocultured skeletal muscle and tendon cells.Here, the junction is represented in a slight overlap of hydrogels, cells, and scaffolds (where relevant).Although the exact mechanisms behind MTJ development are for the most part unknown, tissue engineers commonly analyze MTJ-specific gene and protein expression to determine the success of their study, such as collagen XXII.Furthermore, the existence of invaginations directly correlates with mechanical competency and is, therefore, commonly seen as an indicator of MTJ maturation.Although there has been some investigation into the response of tendon and SM cells in their opposing dECMs, this information is scarce, making it increasingly difficult to predict in vitro outcomes and create a targeted, standardized approach.The trend of further maturing MTJ constructs via mechanical stimuli has been identified as future steps by many authors; 41,111,124 however, few have investigated the effect of mechanotransduction on engineered MTJ constructs. 121The transition to dynamic cultures must, however, be related back to the vast body of knowledge gained from individual tissues, such as skeletal muscles and tendons, respectively.As the progression from 2D to 3D cultures often produces conflicting results, it is important to only refer to 3D cultures for implantable tissue constructs.

RECAPITULATING THE BIOPHYSICAL ENVIRONMENT Bioreactors
To facilitate the transition to dynamic cultures, mechanically inductive bioreactors are required.Hermetically sealed, contaminantfree bioreactors provide an environment that can mimic complex and dynamic in vivo environments. 1256][127] Pre-conditioning is of great importance as loss of cellular function is currently one of the main limitations in tissue engineering. 128Thus, to replicate in vivo interactions, bioreactors aim to control physico-chemical stimuli parameters to elicit specific cellular responses.Several bioreactor categories have been developed since the early 1980s, including rocking bed, batch stirred tank, rotating wall vessels, perfusion, and isolated expansion automated systems. 126,127,129Considering the myotendinous junction, stretch bioreactors are favorable; however, only few are commercially available.General bioreactor considerations revolve around the following five domains: 130 1. Physical design/material selection 2. Mass transfer (nutrient) 3. Mechanical stimulation 4. Electrical stimulation 5. Feedback control system Here, only two imperative domains will be described, including mass transfer and mechanical stimulation, as they have the most relevance to MTJ engineering.

Mass transfer
In vivo, cells are generally within 100 lm from capillary networks that provide gas and nutrient supply and waste disposal. 131Bioreactors must, therefore, emulate the rich and extensive vascular network of the human body via the circulation of nutrient-rich culture media.Three forms of mass transfer may occur to accommodate this in a bioreactor, including (i) convection, (ii) diffusion, and (iii) perfusion.Mass transfer in static cultures solely rely on diffusion to achieve adequate mass transfer, which, due to the diffusive penetration threshold of $100-200 lm, commonly results in heterogeneous cell distributions. 132Lack of adequate perfusion is commonly visible in tissue constructs by the  126 As practically all transplantable tissues will exceed these dimensions and lack vasculature, mass transfer drastically restricts scalability. 125,133Conversely, dynamic cultures further incorporate convection via continuous media flow, and perfusion, via mass flow throughout the scaffold itself.Dynamic environments are most commonly facilitated by pumps and rotary motion; however, they may also be generated with heat exchangers, humidifiers, bubble traps, and oxygenators.
Suitable mass transfer, thus, allows culture media constituents, such as nutrients, glucose, proteins, vitamins, oxygen, and pH, to be homogeneously delivered to engineered tissue.The concentration of these constituents, however, varies with tissue growth and type and must continuously be replenished.This may be achieved via peristaltic pump systems that induce the constant supply and circulation of fresh medium.In essence, the addition of fluid flow within bioreactors drastically increases the complexity of analysis and design considerations.For example, pumping flow rate may be set to induce laminar or turbulent flow, which is further influenced by a bioreactor's architecture. 134Ideally, mass transfer from supplying arteries and arterioles of the perimysium at the myotendinous junction must be emulated to avoid necrosis and enable reliable tissue growth.

Mechanical stimulation
Regarding the maturation of myotendinous structures, the implementation of mechanical forces in vitro appears to increase tissue maturity.To date, there is convincing evidence that physical forces are imperative for tissue development during embryogenesis. 126In fact, a study by Sungsoo et al. displayed that stress-induced signal transduction is at least 40 times faster than growth factor-induced signal transduction while considering the human airway smooth muscle (HASM) cells. 135Forces assumed to contribute to this consist of hydrodynamic 125 and hydrostatic pressure, fluid dynamics, mechanical stresses and strains, and electrical cues. 133To replicate one aspect of these, a dimension-based mechanical stimulation is often embedded into bioreactors. 136n three dimensional tissue culture, forces may be applied via three mechanisms, including (i) fluid flow, (ii) mechanical vibration, and (iii) mechanical stimulation. 137,138Focusing on mechanical stimulation, forces may be applied via uniaxial or biaxial tension, compression, and/or torsion depending on the native tissue environment that is being emulated.Tensile strain is the most commonly applied biophysical stimulus used to mimic the native muscle environment.While each individual muscle belly shortens during movement, the contraction of individual myofibers against two ends that are tethered to the rigid bone induces tensile strains in the tissues as they are contracting. 139In reported stretch bioreactors, one end is often clamped into a stationary position, whereas the other is clamped to an oscillation source, commonly connected to a stepper motor. 140Here, strain concentrations increase in magnitude approaching the clamps, leading to altered cellular responses which may be able to mimic interfacial tissues, such as the MTJ.
Studies demonstrate that, dependent on the exposure routine, mechanical input can (i) stimulate ECM production, 141 (ii) improve cell/tissue organization, 136 (iii) direct cell differentiation 142 and alignment, 143 and (iv) enhance targeted tissue functions. 144For example, by analyzing cells more susceptible to stretching, such as fibroblasts, cellular alignment along the axis of strain can be observed after just three hours of exposure. 145This is expected as fixation of a scaffold between two anchor points creates predictable lines of isometric strain which cells sense and respond to accordingly.This mechanically mediated internal tension is sufficient in promoting cellular alignment along the principal axis of strain, 140,146,147 and as previously described, was used by both Laternser et al. 112 and Gaffney and Laternser 121 to engineer the MTJ.
Considering uniaxial strain exposure, little is known about which specific mechanical force(s) or regimes of application (magnitude, frequency, and duty cycle) induce specific cellular responses. 131Brown et al. showed that tissue engineering programs should be customized for axial vs limb tendons while studying Scx-GFP mice cells under a cyclic uniaxial strain. 148Significant consideration must, therefore, be given to determine the appropriate application of biophysical stimuli, such as static vs cyclic, continuous vs intermittent, low vs high amplitude, among others.For example, a frequency of 1 Hz is similar to the natural stride frequency, mimicking the strain cycle of muscles in locomotion. 139Referring back to embryonic cues however, Kodama et al. found that muscle-driven movements began at E14 in mice, creating a timepoint benchmark for the transition from static culture to mechanically inducive dynamic cultures. 24These findings have further implications to the cues dictating tendon development, suggesting tendon progenitor cells (TPCs) begin to differentiate as a function of their particular anatomical microenvironments prior to E14.This was further consolidated by Lipp et al., who found muscular contraction to begin at E13.5. 149dhering to this embryonic phenomenon, Foolen et al. applied a cyclic stretch after a static culture to study the subtle balance between contact guidance and stress concentrations in 3D. 147Here, cells in the core displayed alignment along the principal axis of the strain; however, cells became increasingly oblique toward the exterior due to decreasing contact guidance toward the exterior of the construct.Similar results can be seen from Boerboom et al. 150 and Rubbens et al. 151 Conversely, Chen et al. restrained cell deformation between two posts for 24 h before applying a cyclic uniaxial strain, resulting in homogeneously aligned cells from surface to core. 146It has been suggested that the earliest forces experienced by MTJ constituents is via the elongation of the connecting bone, 16 in which the increasing cellmediated tension between bioreactor posts may have mimicked this function and enhanced homogeneous cellular alignment. 146yclic loading, however, produces inconsistent results on signaling and myogenic marker expression, making it difficult to determine optimal cyclic parameters, whereas static loading largely produces consistent results.Although the exact science behind mechanotransduction is yet to be elucidated, it has been hypothesized that the strain effect on myogenesis may occur by signaling intracellular transcription factors.This further iterates that the expression of myogenic markers is a direct indicator of the progression of immature constructs to functional implants.There have been many studies conducted on the effect of mechanical strain on skeletal muscles and tendons individually; however, there have not yet been many studies specifically investigating the effects of mechanical strain at the MTJ (see Tables V and VI).
Overall, the mechanical stimulation of skeletal muscle cells promotes a higher degree of alignment in the direction of the strain, 136 thicker and longer myotubes, 136 greater sarcomeric patterning, 136 promoted myoblast fusion, 181
Increased cross striations in MPC constructs.
enhanced proliferation, 183 unidirectional nose to tail orientation, 189 generation of contractile force, 180 elongated and aligned morphology, 188 higher level of multinucleated cells, 190 promoted hCPC migration, 192 and upregulation of scleraxis, collagen I, decorin, and tenascin C 193 while maintaining comparable cell survival. 198Similarly, engineered tendon constructs promoted alignment, 143,161,167 ECM production, 157,161,179 maturation, 160,169,178 and expression of tenogenic markers. 143,175,178As depicted above, a uniaxial strain has a profound effect on various cellular activities.Further studies are reviewed by Somers et al. 139 To induce this strain, several strain-inducing bioreactors have been commercialized and are available from a number of companies (Fig. 6) including Strex/Strexcell, Electroforce, Flexcell, Biodynamic, Tissue growth technologies, Ebers, Cellscale, and Con Whitley scientific.The vast majority of strain bioreactors have been custom designed and are primarily research focused, such as those seen in Table VII.Lab-specific bioreactors commonly incorporate a 3Dprinted frame, transparent Perspex lid, stepper motor linear actuator, and a wide range of clamps to constrain the engineered tissue, similar to commercialized devices.199][200][201] Given this large range of materials, however, various cleaning protocols and surface treatments are mentioned.For example, to improve biocompatibility and hydrophobicity, sylgard-184 has previously been coated on the bottom of a resin-based bioreactor chamber. 202Similarly, basic features, such as gas vents, are common throughout, where only the most intricate designs incorporate mechanical feedback (via in line or bending beam load cells) and automated media exchange systems.Finally, there are a finite number of designs that incorporate electrical stimuli with a uniaxial strain; 158,186,200,[203][204][205] however, they will not be summarized here as they are not the focus of this review.

CONCLUSION
It is widely hypothesized that engineered tissues that mimic embryonic (as opposed to adult) mechanical properties will produce greater biological responses.Although producing promising results, the paucity of knowledge surrounding MTJ's embryonic development, particularly that of skeletal muscle, makes outcomes difficult to quantify.Currently, attempts to engineer the MTJ rely on heterogeneous scaffolds aimed to provide the appropriate environment for skeletal muscle and tendon tissues simultaneously.In doing so, the interfacial region may be analyzed for interface-specific genes, proteins, and architectural characteristics.To date, although favorable results have been achieved, the mechanical discrepancy and architecture complexity between skeletal muscles and tendons is yet to be mimicked.To overcome this issue, many groups identify the possibility to mature myotendinous constructs in a dynamic, strain-inducive bioreactor as a means of tissue maturation; 41,111,124 however, few have advanced to this stage. 121The translation to dynamic cultures must be reliant on the established body of knowledge known about individual tissues, such as skeletal muscles and tendons separately.Here, the significant progress of engineering tissues may be merged to develop targeted strategies focused on interfacial regions such as the MTJ.Future studies are advised to direct their attention to the translation from static to strain-inducive dynamic bioreactors, where a cyclic regime following an initial static strain period shows the greatest results.

FIG. 4 .FIG. 5 .
FIG.4.Common structural design of engineered myotendinous junction.Here, cells and scaffolds are spatially deposited, in which the 10% interfacial region signifies the main area for analysis of specific MTJ markers.Created with BioRender.com.

TABLE I .
Reported mechanical properties of the myotendinous junction and its constituents.

TABLE II .
Analysis of references cited by several reviews elucidating the mechanical properties of adult human MTJ.
Ã Indicates a possible origin source for the abovementioned summary table (TableI).

TABLE III .
Mechanical properties of living, in vivo adult human myotendinous tissues.

TABLE IV .
Current biomaterial and biomaterial-free approaches to engineering the myotendinous junction in three-dimensions.

TABLE V .
3D tendon constructs under uniaxial strain.

TABLE VI .
3D skeletal muscle constructs under uniaxial strain.

TABLE VII .
Custom-made bioreactors capable of providing uniaxial strain on three-dimensional tissue-engineered constructs.indicates transparent at top and bottom.indicates transparent at top and bottom.