Introduction

Tendons and ligaments are specialized fibrous tissues serving a mechanical function. The tendon links muscle to bone and allows forces from the muscle to be transmitted to the bone for movement, whereas the ligament joins bone to bone and provides stability while allowing a controlled range of movement. Although tendons and ligaments share many features, some histological and mechanical differences must be underlined [1]. Tendon fibers are mostly organized along the main axis, while fibers are more randomly oriented, with a weaving pattern, in ligaments. Total collagen content is slightly higher in tendons; the type I isoform represents up to 99 %, while it does not exceed 90 % in ligaments. Collagen type III ranges from 1 to 5 % in tendons and accounts for about 10 % in ligaments. Tendon also contain slightly less ground substances (e.g. hyaluronan, glycoproteins, proteoglycans) and elastin than ligaments. These differences in structure and extracellular matrix (ECM) composition account for the higher tensile strength and elastic modulus of tendons [2].

Injury or breakage of tendons and ligaments can lead to significant impairment in joint motion and disability [3]. Spontaneous healing produces reparative, scar-like tissue, lacking the mechanical properties of native tendon or ligament, with limited functionality [4, 5]. Each year more than 100,000 patients in the USA undergo one of the therapeutic options presently available, i.e.: 1) direct suture repair, 2) reconstruction with an auto/allograft tendon or ligament, or 3) use of prosthetic devices [6, 7].

There are disadvantages with each choice. The first, direct suturing, may not be possible and is of limited use for those injuries resulting in a gap. Additionally, the techniques for transferring or lengthening tendons can be demanding and have uncertain outcomes. As for reconstructive techniques, grafts with autologous tissue similar to the original, injured tissue eventually become lax in many patients; moreover, there is evidence that such a treatment does not alter the progression of osteoarthritis. On the other hand, allogeneic grafting of ligaments may lead to immunologic reactions preventing the adequate healing of the wound. Finally, prosthetic devices using synthetic materials without cellular components are at risk for wear and rupture [8–10].

The development of tissue-engineered substitutes is a promising new strategy to address these shortcomings [11]. The purpose of tissue engineering (TE) strategies is to provide functional tissue constructs (scaffolds with cells on board) grown in vitro which are able to develop in vivo and integrate with the host tissues. In healthy tendon and ligament tissue, the tensile properties necessary to withstand constant loading from locomotion are provided by an oriented network of collagen fibrils and fibroblasts. Mechanical cues, such as substrate stiffness and topography, play a recognized role in inducing this high level of organization [12–14]. Mechanical loading systems also induce cellular alignment and orientation of ECM; indeed, several studies have demonstrated the pivotal role of mechanical stimulation in the successful development of engineered constructs through the enhancement of cell proliferation, ECM synthesis, and the promotion of cell differentiation towards specific fibrous connective tissue lineages [15]. In vivo, mechanical loads, including tension, compression, hydrostatic pressure, and fluid shear stress, are transduced by cells to stimulate biochemical pathways and affect cellular processes [16]. It is well accepted that mechanical stimulation can be used to precondition TE constructs in vitro to mimic the biochemical and structural properties of healing tissues in vivo. The purpose of this review is to discuss recent promising results obtained with mechanical actuation devices. These systems apply a physical stimulus mimicking native loading regimens within an appropriate culturing environment to generate functional engineered tissue [17, 18].

Mechanostimulation Protocols for Tendon and Ligament Tissue Engineering

In Vitro Protocols

Critical Variables When Setting a Protocol

The most relevant variables in a mechanostimulation protocol can be summarized as follows: (i) the type of cell (e.g. tenocytes, ligament fibroblasts, stem cells (SCs)), (ii) the nature of the scaffold (e.g. synthetic, biological, structured or amorphous), (iii) the features of the strain (e.g. amplitude, static vs. cyclic, frequency, duration, continuous vs. intermittent, translational vs. rotational) and the actuator that generates the strain (e.g. linear motors, pneumatic actuators, step motors).

  1. (i)

    The type of cell

Although defining in detail which cell lineages are the most suitable for tendon and ligament TE – a topic extensively covered elsewhere (see, for example, [19]) – exceeds the purpose of this manuscript, we should point out that tenocytes, fibroblasts, and SCs have all been tested as potential candidates [19].

Tenocytes might be considered the ideal cell type, although in culture they reduce their metabolic activity and lose tenogenic markers, such as collagen type I, tenomodulin, tenascin-C and scleraxis, after serial passaging [20].

Fibroblasts isolated from tendon sheaths, as well as from ligaments, are also used. Recently, researchers have shown that dermal fibroblasts are also able to express tendon markers. Dermal fibroblasts grow more rapidly in culture than tenocytes [21], thus representing an interesting choice for tendon TE. However, only a few studies assessing whether they differentiate towards a functional tenogenic lineage have been performed to date (see, for example, [22]).

SCs are a promising resource for tendon/ligament regeneration; in fact, a number of published studies address the issue of the appropriate selection of suitable SC types (see, for example, [23]). Mesenchymal SCs (MSCs) in particular have several relevant features: they are multi-potent cells isolated from a number of tissue sources [24], they can differentiate into several lineages [25–27], and they respond to mechanical induction of phenotypic fate [28]. Due to their proliferation potential, bio-molecular production, cell-to-cell signaling, and formation of appropriate ECM, they appear to be the most promising cell source for improving the structural and biomechanical features of an injured tendon or ligament upon autologous administration [29–31]. In addition, bone marrow SCs (BMSCs) [32] and adipose-derived SCs (ADSCs) [33] have been extensively studied for tendon repair and regeneration, and some clinical trials in progress are investigating their efficacy in different tendinopathy treatments (Australian New Zealand Clinical Trials Registry, Trial Number: ACTRN12610000985088; ClinicalTrials.gov, Trial Number: NCT01687777; ClinicalTrials.gov, Trial Number: NCT01856140). However, because BMSCs and ADSCs can also differentiate into osteoblastic cells [34–36], a potential risk of ectopic bone formation should not be ignored.

The recent identification of tendon SCs (TSCs) is one of the most appealing discoveries in the field [37, 38]. Compared to BMSCs or ADSCs, TSCs feature a superior clonogenicity and proliferation capacity and express higher levels of scleraxis (a tendon-enriched specific transcription factor) and tenomodulin (a marker of adult tenocytes) [39]. Nevertheless, their low availability and inability to maintain their phenotype with time and passaging in culture currently limits their use in TE protocols [40]. A few studies have shown that induced pluripotent SCs (iPS) are also able to support the host’s endogenous healing process in tendon/ligament repair protocols [41, 42].

  1. (ii)

    The nature of the scaffold

Central to TE-based repair strategies is the choice of an appropriate biomaterial to fabricate a scaffold that supports cell growth and transmits the appropriate physical stimulus. Ideally, scaffolds should be able to harbor competent cells, support the full mechanical load initially, provide a structured environment with tissue-specific elastic properties, and grant the ability to integrate with surrounding tissue [43, 44].

Biomaterials used for scaffold fabrication include both biomimetic synthetic polymers and biological molecules. Obviously, these distinct cell-bearing matrices have different biomechanical properties — which require distinct loading protocols in a mechanically active environment. Synthetic polymer scaffolds such as polyhydroxyesters (e.g, polylactic and polyglycolic acids, their copolymers, and polycaprolactone) have the advantage of possessing reproducible mechanical and chemical properties. However, because of the risk of their releasing acidic byproducts or unnatural polyesters into the bloodstream during the degradation process, they are rarely used in clinical trials [45, 46].

Current research is exploring natural biomaterials, such as collagen derivatives, polysaccharides and silk fibroin. Collagen appears to be the favorite raw material for producing scaffolds in many physical forms (e.g. sponges, films, fibers), also featured by hierarchical three-dimensional architecture that closely imitates native tendon structure [47]. However, such scaffold conformations cannot withstand mechanical stress over time when implanted in the challenging mechanical environment of the injured tissue [48, 49].

Among polysaccharides, hyaluronan, chitosan and alginate have been used as scaffolding materials, promoting cell adhesion and tissue-specific ECM production while providing good mechanical properties. Given their crucial roles in cell signaling, one can expect that their role as scaffolding materials in going to increase in the future [49].

Silk fibroin, a natural protein, is FDA-approved for TE purposes and suitable for the manufacture of structural patterns in a range of applications, including tendon/ligament repair [50, 51]. Although contingent allergic reactions in the host have been described [52], silk-based scaffolds provide microenvironmental support and promote cell adherence, expansion, and differentiation.

Fabrication specifications are also a significant feature of any successful TE strategy because, together with the chemistry of the selected biomaterial, they contribute to the overall physical properties of a scaffold. Superficial topography and overall stiffness of the scaffold are undoubtedly critical for cell behavior, and for the ultimate performance of the tissue substitute. For example, when fibers from collagen type I are aligned within a scaffold, they mimic the ECM architecture of native tendons, inducing the implanted cells to differentiate towards the tenogenic lineage [53]. Furthermore, microgrooves and microchannels (referred to as contact guidance) at the scaffold surfaces contribute to the development of unique biomechanical properties of tendon and ligament tissue, as axial alignment of cells and their anisotropic arrangement contribute to successfully mimic the native ECM structure [54, 55]. Further information about substrate stiffness and topography is available in the literature (see, for example, [12–14].

  1. (iii)

    The features of the strain and of the actuator

Tables 1 and 2 list the features of the strain applied to induce a tendon/ligament differentiation of SCs, as reported in the literature. The amplitude of the strain, which is always cyclic, ranges between 2 and 15 %. Reported frequencies range between 0.017 and 1 Hz. The stimulus is applied for at least one hour, for up to 28 days. In most protocols the stimulus was intermittent; however, very dissimilar intervals are reported, from 10 s rest after each strain cycle to 1 h strain before 5 days of rest. Application of mono- or multidimensional strains matches the functional requirements of the tissue to be engineered. Usually, uniaxial cyclic strain is employed in tendon TE protocols, as the cells in a tendon are physiologically subjected to uniaxial loading conditions [56]. Representative modeling in Fig. 1a shows how uniform uniaxial stretching of an elastic matrix induces a homogeneous stimulation of embedded corpuscular component. On the other hand, multidimensional strains such as axial tension/torsion reproduce in vitro the complex mechanical forces applied to ligamentous tissue (e.g, anterior cruciate ligament, ACL) under normal physiological conditions. Representative modeling in Fig. 1b shows how adding a rotational strain to uniaxial stretching of an elastic matrix induces heterogeneous stimulation of the embedded cell-like component. When translational and rotational strains are applied, helically organized collagen fibers are preferentially formed in the direction of the load, especially at the periphery of the ligament [57]. Although specific markers to differentiate tendon from ligament are unavailable [58], several studies show that the expression of fibrous tissue connective markers, such as collagen types I and III, tenascin-C, and scleraxis, obtained by both mono- and multidimensional strain. These findings render it difficult to establish which approach is better to promote cell differentiation towards a specific lineage.

Table 1 Published protocols for mechanical actuation systems for the phenotype commitment of stem cell-based tendon tissue substitutes
Full size table
Table 2 Published protocols for mechanical actuation systems for the phenotype commitment of stem cell-based ligament tissue substitutes
Full size table
Fig. 1
figure 1

COMSOL Multiphysics® representative modeling of force direction and intensity (red arrows) and shear stress distribution (surface color map) affecting the same corpuscular component embedded in an elastic matrix undergoing either a uniform uniaxial 4 % stretching (panel a) or additional 45° rotational contribution to the same uniaxial stretching (panel b)

Full size image

The technology required to implement such a heterogeneous list of protocols has transformed very basic devices to more sophisticated systems that closely approach physiological conditions [11, 59, 60]. All the devices referenced in this manuscript are capable of producing at least one of these two relevant stimuli: a) an axial strain and b) a multidimensional strain.

These stimuli can be provided by linear motors, pneumatic actuators, or step motors, which are usually incorporated into custom systems. Linear motors are generally used to generate uniaxial strain, which, in addition to promoting connective fibrous tissue differentiation, can also induce muscle, cardiac or bone lineage commitment [61–64]. A pneumatic actuator such as the one used in the Flexcell tension systems (Flexcell® International Corporation, Burlington, NC, USA), generates biaxial strain, by using vacuum to deform flexible-bottom culture plates. These systems, widely used in different fields of TE, are commercially available and offer up to 33 % substrate elongation (with minimum strain resolution capability of 0.1 %) with frequency ranges from 0.01 to 5 Hz (reported in http://www.flexcellint.com/index.html). Note that similarly detailed specifications are not available for most other devices due to their custom nature. Step motors are usually used to generate rotational strain. Systems devised to impart multidimensional strains (e.g. uniaxial and rotational strains, which are used to closely mirror the physiological environment of ACL) combine more than one actuator.

As expected, the appropriate combination of cells, biomaterials/scaffolds and physical stimulation is clearly the key for a successful TE protocol, but a standard procedure to obtain tendon/ligament substitutes has yet to emerge.

Tendon Tissue Substitutes

A standard protocol for engineering tendon tissue substitutes is still lacking in the scientific literature. During this last decade, many of the proposed configurations have featured not only different cell sources, scaffolds or equipment for mechanical stimulation in vitro, but also different strain patterns in terms of amplitude, frequency, duration and on/off ratio of applied stimuli (see Table 1).

A comprehensive overview of the different published configurations follows. In 2008, Riboh et al. [65] studied different settings of strain patterns using the Flexcell Strain Unit®. They suggested that intermittent cyclic strain (4 % strain; 0.1 Hz frequency; alternating 1 h of strain and 2 h of rest) favors cellular proliferation, promotes collagen production, and maintains morphology in 4 candidate cell lines (epitenon tenocytes, sheath fibroblasts, BMSCs and ADSCs) for flexor tendon engineering. Results showed that the proposed pattern of stimulation proved to be better than continuous strain, increasing both cell number and collagen production per cell.

Two years later, researchers at Pittsburgh University [66] determined that mechanical stretching increased the proliferation of rabbit TSCs in a stretching magnitude-dependent manner. Low mechanical stretching at 4 % promoted differentiation of TSCs into tenocytes, whereas stretching at 8 % induced differentiation of some TSCs into adipogenic, chondrogenic, and osteogenic lineages. TSCs were seeded on silicone dishes and subjected to uniaxial strain (4 % or 8 % strain; 0.5 Hz; 12 h) in a custom-made device. Results also showed that cells were highly elongated and aligned along the microgrooves in the axis of stretching.

In the same year, Angelidis et al. [67] examined the effect of strain on the biomechanical properties of constructs seeded with rabbit ADSCs and flexor tendon sheath fibroblasts. Flexor tendons from the rear paws of rabbits were acellularized and seeded with autologous cells. Constructs were dynamically stimulated by uniaxial strain (1 cycle/min; alternating 1 h periods of mechanical loading and rest, for 5 days) in an Instron® Ligagen L30-40 instrument (ITW Test and Measurement Italia S.r.l., Pianezza (TO), Italy). This process achieved an ultimate tensile stress and elastic modulus comparable with those of fresh tendons, demonstrating how pretreatment together with alternative cell lines, such as ADSCs and fibroblasts, might contribute to the in vitro production of strong tendon material.

Barber et al. [68] analyzed the tensile properties of braided nanofibrous poly(L-lactic acid) scaffolds using the Bose Electroforce® 5200 BioDynamic® multi-chamber system (Bose Electroforce, Eden Prairie, MN, USA). Results showed that acellular BNFs mimicked the normal mechanical behavior of native tendon and ligament during loading. Moreover, human MSCs (hMSCs) adhered to and proliferated on the BNFs and, when a uniaxial force (10 % strain; 1 Hz; 2 h per day for 10 days) was applied, they aligned parallel to the length of the nanofibers and displayed a concomitant realignment of the actin cytoskeleton. In addition, when hMSCs were cultured on BNFs in the presence of tenogenic growth factors and stimulated with cyclic tensile strain, they differentiated into the tenogenic lineage, as demonstrated by significant up-regulation of scleraxis gene expression.

In 2011, Scott et al. [69] conducted a similar study, examining the effects of static and cyclic mechanical loading on the expression of proteins, such as scleraxis and type I collagen, which play a role in tenocyte differentiation and function. In addition to static and dynamic load, they evaluated the effects of varying mechanical parameters including (1) increasing the strain magnitude, (2) the inclusion of 10 s rest periods, and (3) increasing the number of cycles. In their study a cyclic load (10 % stretching at 0.1 Hz) was applied by a Flexcell Tissue Train® system (Flexcell® International Corporation) to a mouse mesenchymal cell line (C3H10T1/2, American Type Culture Collection) cultured in 3D collagen gel.

Two years later, Morita et al. [70] examined the effects of three types of cyclic elongation (5 %, 10 % and 15 %) on promoting the differentiation of human BMSCs (hBMSCs) into tenocytes. Cells were seeded on silicone rubber chambers and dynamically stimulated, by a simple stretching device (STB-140®, Strex Inc., Osaka, Japan), at 1 Hz for either 24 or 48 h. Results showed the expression of types I and III collagen, tenascin-C and scleraxis when cells were exposed to 10 % uniaxial cyclic stretching stimulation for 48 h. Interestingly, MSC-to-tenocyte differentiation appeared strongly associated with the cumulative elongation load of the cells as well as the cell orientation angle.

In the same year, Teh et al. [71] were the first to report the synergistic effect of mechanical stimuli with aligned topography in a 3D environment to promote tenogenic differentiation and tissue maturation in vitro. Rabbit MSCs were seeded on 3D-aligned full silk fibroin scaffold and mechanically stimulated in a custom-made apparatus which applied a 5 % translational strain and a 90° rotational strain for 12 days with a cyclic frequency of 0.1 Hz. The authors reported that while the aligned topography potentiated the expression of tenogenic markers in MSCs, such as collagen type I, tenascin-C and tenomodulin, the introduction of mechanical stimulus intensified and accelerated the tenogenic differentiation process.

Recently, Youngstrom et al. [72] proposed a novel system for in vitro tendon differentiation and TE. They suggested that a protocol applying 3 % cyclic strain at 0.33 Hz for up to an hour daily for 11 days promoted tenocytic differentiation of equine BMSCs seeded on syngeneic decellularized tendons inducing collagen type I, scleraxis and decorin gene expression. Both cyclic strain higher than 3 % (e.g. > 5 %) or no strain (0 %) produced weaker effects. They interpreted this finding to indicate that too much stretching might excessively tear some ECM components, thus rendering them inadequate for stimulating BMSC differentiation.

Ligament Tissue Substitutes

Although different cell sources, scaffolds and actuator systems have been used in several configurations, the design of a standard protocol for engineering ligament tissue substitutes is still to come (see Table 2). A comprehensive overview of the different published configurations follows.

In 2008, Zhang et al. [73] showed how mRNA expression and protein levels of collagen types I and III and of tenascin-C in rat BMSCs (rBMSCs) were up-regulated under co-culture with ligament fibroblasts or under a continuous uniaxial cyclic condition. A custom-made mechanical device was used to apply cyclic uniaxial strain (10 % strain; 1 Hz; from 3 to 36 h) to rBMSCs grown on an elastic silicone membrane pre-coated with 1 % gelatin. Their data demonstrated a differentiating potential for uniaxial strain indicating that BMSC might be useful as a cell source for ligament TE.

In 2012, Kreja et al. [74] tested the effect of a cyclic uniaxial intermittent strain protocol on hBMSCs. However, contrary to results usually reported in the literature, no effect of mechanical stimulation on the expression of ligament-marker genes was found in undifferentiated hBMSCs seeded on a polylactic (PLA) scaffold. These results suggest that higher-magnitude stretching was necessary to differentiate hBMSCs towards the tendon/ligament lineage.

In 2013, Kahn et al. [75] proposed a poly(L-lactic-co-glycolic acid)-poly(L-lactic-co-ɛ-caprolactone) (PLGA-PLCL) composite scaffold as a valid candidate for the regeneration of ACL. The scaffold’s mechanical properties were tested under a static culture condition vs 2 h per day of a simultaneous traction-torsion cyclic culture condition (10 % uniaxial strain and 90° of torsion cycles) at 0.33 Hz in a custom-made actuation system. Rat BMSCs and CRL 2703™ human fibroblasts were used to verify the biocompatibility of the scaffold and cell orientation after mechanical stimulation. Results showed that although the structure of this scaffold was suitable in terms of biocompatibility and cell orientation, it was rapidly subjected to hydrolytic and enzymatic degradation. Thus, the authors suggested reinforcing the structure to delay the degradation process.

Li et al. [76] presented another scaffold for tissue-engineered ACL reconstruction. Although they used human foreskin fibroblasts rather than SCs, this approach appears promising enough to be listed here. Silk scaffolds were fabricated with three different architectures (wired, braided and straight-fibered), and their mechanical properties were evaluated using a tensile device to carry out cyclic loading tests. The results showed that, although the cyclic loading decreased the tensile strength of the constructs, a wired structure best approximated the biomechanical tri-phasic behavior [77] of the human ACL.

Recently, Qiu et al. [78] corroborated the ability of cyclic tension to promote differentiation of MSCs into fibroblasts. Fourteen days of mechanical stimulation in a custom device caused hBMSCs, seeded on fibrous collagen-based scaffolds, to secrete more ECM, including collagen types I and III and tenascin-C, than constructs in static culture.

Overview of the Current in Vitro Protocols

Although the above-mentioned references about mechanostimulation protocols for tendon and ligament TE show unstandardized approaches, it is possible to extract information in an attempt to establish a unifying protocol design. Inasmuch as cell lineage is concerned, it seems reasonable to suggest that ADSCs are the most suitable candidate since they are more readily available than BMSCs and TSCs, adequately proliferate in vitro, and can differentiate into the desired phenotype [79]. Indeed, based on data presented in recent literature and in some human trials previously mentioned, auto-transplantation of ADSCs has given encouraging results in terms of patient safety and ability to enhance intrinsic tendon/ligament regeneration.

In order for a bioengineered construct to have adequate tensile properties, the cells need the appropriate scaffold which couples biocompatibility (particularly enhanced in natural scaffolds) and suitable physical properties (usually better in synthetic materials). In this regard, silk fibroin appears to be a promising compromise, since the wired structure of its fibers best approximates the biomechanical tri-phasic behavior of the human ACL [76]. When electrospun with aligned topography in a 3D environment, it shows biomechanical characteristics similar to native tendon and ligament tissues [71] and promotes tendon marker gene expression [80].

The features of the strain (e.g. amplitude, static vs. cyclic, frequency, duration, continuous vs. intermittent, translational vs. rotational) applied to induce the phenotype are central to this work. However, the design of a standard strain protocol is not obvious because the literature presents a truly heterogeneous landscape. Nevertheless, it appears that cyclic strain is regularly employed, and its intermittent administration is preferred as a relevant reduction in cell growth and a tendency to apoptotic death was observed when continuous stimulation is applied [65]. Intermittent strains lasting from one to 22 h stimulate cell proliferation and increase the biosynthesis of molecules expressed in tenocytes, such as collagen types I and III, tenomodulin, tenascin-C, scleraxis, and decorin.

We note that uniaxial strain is preferred, for tendon TE, whereas multidimensional loads appear appropriate for ligament TE. A 10 % deformation is usually needed to obtain a significant expression of tenocyte-marker genes in MSCs [68, 70, 74]. It is worth noting that an amplitude lower than 3 % preferentially commits MSCs towards the osteogenic lineage [74].

To induce the expression of tenocyte-marker genes, cyclic tension ranging from 0.016 to 1.0 Hz is applied. A standard value within this range has not emerged from the literature. The total duration of the entire period of strain is equally undefined. Two days were enough to increase tenocyte gene expression in BMSCs [70], but similar results were also obtained after 3 weeks of mechanical stimulation [69]. Obviously, the faster protocol is preferred when a therapeutic application is the target of the procedure.

In Vivo Approaches

The recent studies listed above highlight how tendon and ligament TE protocols continue to improve in vitro, but have yet to be translated into clinically relevant models [81, 82].

Several experimental procedures on small and large animals [83–86] were performed by various groups that tested the outcome of implanted engineered constructs. In addition, some clinical trials, based on the injection of autologous human SCs isolated from bone marrow or lipoaspirates of recruited patients, were designed to answer specific questions about the clinical efficacy of these reparative strategies in human tendon repair.

As a matter of fact, only a few reports describe engineered constructs stimulated in a mechanical actuator system and subsequently implanted in an animal model of injured tendon or ligament (see Table 3).

Table 3 In vivo approaches
Full size table

In 2006, Juncosa-Melvin et al. [87] correlated an in vitro mechanical treatment pattern with an in vivo follow-up for the first time, showing how mechanical stimulation affects the biomechanics and histology of SC–collagen sponge constructs used to repair central patellar tendon defects in rabbit. MSCs were seeded onto type I collagen sponges. After 14 days of mechanical stimulation (static condition in incubator as control), autologous tissue-engineered constructs were implanted into cell donors. Twelve weeks after surgery, repair tissues were biomechanically and histologically analyzed. Results showed that structural and mechanical properties of the stimulated tissue were significantly different from static controls. Histological analysis revealed an excellent cellular alignment with mildly increased cellularity in stimulated constructs. Moreover, the authors affirmed that no ectopic bone was observed in any occasion.

Three years later, the same group published a new study about the optimization of the mechanical stimulus in culture to obtain the in vitro biomechanical properties most useful for in vivo repair [88]. The type I collagen sponge was crosslinked by dehidrothermal (DHT) treatment in order to enhance the stiffness of the scaffold. However, their results showed that the level of DHT crosslinking of the scaffold reduced the biomechanical properties of the repair tissue; this specific treatment also appears to mask any benefits of stimulation. Moreover, altering cycling conditions with DHT crosslinking had no significant impact on in vivo repairs. Interestingly, the authors emphasized the importance of a proper mechanical preconditioning of the scaffold to better control/modulate MSCs differentiation in vitro, although this assertion needs to be verified in future studies.

Unfortunately, in the years since 2009 the literature has shown very few studies in which engineered constructs are implanted in animals after mechanical preconditioning.

In 2012, Chang’s research group at Stanford University, who had previously described the biomechanical properties of tendon constructs [67], presented a study on the mechanical optimization of tissue-engineered rabbit flexor tendons to be implanted in vivo. In this case rabbit tenocytes, rather than SC, were seeded onto decellularized rabbit rear paw flexor tendons and exposed to uniaxial cyclic strain in a custom device (1.25 N strain force per construct at a frequency of 0.017 Hz, alternating 1 h of cyclic strain with 1 h of rest for 5 days) [89]. These tendon constructs were then transplanted back to the rear paw of recipient animals and 4 weeks later they were sacrificed to examine the biomechanical characteristics of engineered tissues. Results showed that cyclic strain preconditioning had an impact on the ultimate tensile strength (UTS) and elastic modulus (EM) of tendon constructs. In addition, histology showed that cellularity was increased in the mechanically preconditioned tendons.

Recently, in 2014, Xu et al. [90] focused their attention on the potential of dynamic mechanical stimulation to facilitate the maturation of tissue-engineered tendons in vivo. In this study, mice TSCs (mTSCs) were seeded onto poly(L-lactide-co-ε-caprolactone)/collagen [P(LLA-CL)/Col] scaffolds and subjected to uniaxial intermittent strain in a custom-made mechanical traction stimulation system for 14 days. Results showed increased cell proliferation and tenogenic differentiation of mTSCs generated by mechanical stimulation. In addition, when autologous engineered constructs were implanted in nude mice, native neo-tendon tissue formation was promoted. More importantly, dynamically cultured mTSCs-P(LLA-CL)/collagen constructs were also implanted in rabbits. Results demonstrated that this heterologous implant significantly promoted the repair of injured rabbit patellar tendons with good mechanical properties by enhancing their collagen production and the expression of tendon-related proteins. Based on the ex vivo approach to engineering the extensor tendon complex developed by Wang et al. [91], Deng et al. [92] proposed a strategy to mechanically precondition a tissue construct for the in vivo repair of rabbit Achilles tendon. To this end, rabbit ADSCs were seeded onto a scaffold with an inner part composed of polyglycolic acid (PGA) unwoven fibers and an outer part, to provide mechanical strength, composed of a net knitted with PGA/PLA (polylactic acid) fibers. Cell-free scaffolds served as control group. Both constructs were cultured under dynamic stretch loading for 5 weeks with a frequency of 3 times per minute (10 s stretching and 10 s interval) and a stretch amplitude of 1/10 length of the constructs. A 2 cm defect was created on the right side of the Achilles tendon followed by the transplantation of a 3 cm cell-seeded scaffold in fifteen animals. A 3 cm cell-free scaffold was used in the control group. Dynamic culture enabled ADSCs to produce their matrix on the scaffold fibers. In addition, in contrast to the cell-free constructs, the cell-seeded ones gradually formed neo-tendon which became mature at 45 weeks, showing histological structure similar to that of native tendon.

Conclusions

Injury or breakage of tendons and ligaments is quite frequent, affecting both young and older patients, and resulting in considerable morbidity. These conditions can deteriorate into chronic pathologies, such as osteoarthritis, peritendinous adhesions and muscle atrophy. Nowadays, injuries to these tissues are treated by surgical repair and/or conservative approaches, both burdened by relevant shortcomings. Alternative strategies based on development of tissue-engineered substitutes promise to address these limitations.

A take-home message from this review of the recent literature is that mechanical actuation systems appear to be a successful tool for the application of TE principles in tendon and ligament regenerative medicine. In fact, mechanotransduction pathways have definitely proven to modulate cell-signaling cascades as well as cell functions such as proliferation and differentiation. However, most studies based on mechanostimulation protocols have yielded published results which are far from being clinically relevant. Tendon and ligament development and healing are complex processes whose most important aspects in the context of TE remain to be determined. There is a need to recognize, assess, and arrange in order of importance the criteria for certifying the quality of tissue substitutes. As an example, Breidenbach at al. [93] recently suggested (1) scleraxis-expressing cells, (2) well-organized and axially-aligned collagen fibrils having bimodal diameter distribution, and (3) a specialized tendon-to-bone insertion site as criteria of biological success in tendon TE. We here define some additional issues which merit further considereration. Too little consideration has been given to the scaffold’s isometric mechanical properties, or to the influence that the variability of its intrinsic stiffness has on tensile pre-stress. However, tensegrity can markedly affect cell shape and structure, so these features are very important pre-requisites for an immediate response to external mechanical stress. In addition, other environmental conditions, such as hypoxia [94] or the simultaneous presence of MSCs and tendon/ligament fibroblasts [95], need to be investigated, since they can further improve both structure and function of the graft. Finally, to guarantee a successful implantation, it will be necessary to study effective protocols applying adequate loading conditions in different regions of the bioartificial tissue, in order to obtain tendons and ligaments that acquire the characteristic structure at the enthesis [96]. We expect that addressing these issues will facilitate effective TE protocols, allowing the creation of functional engineered tissues capable of treating injuries without many of the undesirable side effects associated with current reconstructive options.