Pre-Innervated Tissue Engineered Muscle Promotes a Pro-Regenerative Microenvironment Following Volumetric Muscle Loss

Volumetric Muscle Loss (VML) is defined as traumatic or surgical loss of skeletal muscle tissue beyond the inherent regenerative capacity of the body, generally leading to a severe functional deficit. Autologous muscle grafts remain the prevalent method of treatment whereas recent muscle repair techniques using biomaterials and tissue engineering are still at a nascent stage and have multiple challenges to address to ensure functional recovery of the injured muscle. Indeed, appropriate somato-motor innervations remain one of the biggest challenges for both autologous muscle grafts as well as tissue engineered muscle constructs. We aim to address this challenge by developing Pre-Innervated Tissue Engineered Muscle comprised of long aligned networks of spinal motor neurons and skeletal myocytes. Here, we developed methodology to biofabricate long fibrils of pre-innervated tissue engineered muscle using a co-culture of myocytes and motor neurons on aligned nanofibrous scaffolds. Motor neurons lead to enhanced differentiation and maturation of skeletal myocytes in vitro. These pre-innervated tissue engineered muscle constructs when implanted in vivo in a rat VML model significantly increase satellite cell migration, micro-vessel formation, and neuromuscular junction density in the host muscle near the injury area at an acute time point as compared to non-pre-innervated myocyte constructs and nanofiber scaffolds alone. These pro-regenerative effects can potentially lead to enhanced functional neuromuscular regeneration following VML, thereby improving the levels of functional recovery following these devastating injuries.


Introduction
Innervation plays a crucial role in the development, maturation and functioning of different muscles in the body and yet remains largely unexplored in tissue engineering studies related to cardiac, smooth or skeletal muscle. There are limited reports showing the importance of reinnervation in a transplanted heart or in a tissue engineered urinary bladder in order to achieve complete functional recovery [1][2][3] . Tissue engineered constructs typically lack preformed neural networks and depend on host-induced innervation to integrate with the native nerve supply.
Although the concept of "pre-innervation" (akin to pre-vascularization), wherein appropriate neuronal populations are cultured along with relevant muscle cells has been established for a long time in vitro 4-7 , it is yet to be studied in vivo in models of muscle injury. The present study is focused towards exploring the effect of pre-innervation in skeletal muscle tissue engineering using a severe muscle trauma model like Volumetric Muscle Loss (VML).
VML is defined as traumatic or surgical loss of a large mass of skeletal muscle tissue beyond the inherent regenerative capacity of the body, generally leading to a severe functional deficit 8 .
Due to a significant loss of nerve and vascular supply accompanied by inflammation driven fibrosis the self-repair process is not adequate to generate sufficient new muscle in time to prevent a chronic scar 9,10 . Although VML is widespread among the civilian population, military The present study was focused on exploring the role of pre-innervation on myocytes in vitro and host neuromuscular environment in vivo following implantation. a) For in vitro studies, our overarching hypothesis were that innervation would augment skeletal myocyte fusion, maturation and formation of Neuromuscular Junctions (NMJs). b) Volumetric Muscle Loss (VML) is defined as frank loss of muscle volume that is accompanied by chronic motor axotomy leading to denervation of the injured area. We used a standardized rat model of VML where >20% of the Tibialis Anterior (TA) muscle volume was resected to create a defect leading to potential damage to intramuscular branches of the host nerve and loss of motor end plates (or NMJs) near the injury area. For in vivo studies, our overarching hypothesis were that implantation of pre-innervated constructs would enhance Acetylcholine Receptor (AchR) clustering and promote innervation of AchRs (mature NMJs) near the implant site at acute time point. personnel in particular are more prone to such damage due to combat related musculoskeletal injuries 11 . According to a recent study, 65% of soldiers who retired due to various injuries reported a muscle condition, while 92% of these cases included VML 12 .
Free functional muscle transfer (FFMT) is the preferred procedure to treat VML, which entails transfer of donor muscle along with nerve and blood vessels from another part of the body to the injury site to facilitate re-innervation and re-vascularization of the graft region 9,11,13 . Although FFMT remains the gold standard, its success is limited by donor site morbidity, long operative time and prolonged de-innervation of motor end plates in the donor muscle 14 . As an alternative approach, tissue engineered skeletal muscle constructs have been fabricated using scaffold based as well as scaffold-less technologies. Decellularized extracellular matrix (dECM) remain the most prevalent scaffold material used for muscle reconstruction following VML injuries 15,16 .
Although dECM-based scaffolds recapitulate the native ECM composition, their efficacy is challenged by their fast resorption rates in the body and failure to provide topographical guidance to regenerating myofibers thereby leading to random fiber recruitment and subsequent scar tissue formation 17,18 . The ECM secreted by skeletal myofibers is arranged in a network of micronano fibrils which are aligned along the myofibers. Aligned nanofiber-based scaffolds accurately replicate the topographical cues of native ECM architecture and has been reported to promote aligned myogenesis, cell migration, survival and angiogenesis 19 .
Persistent inflammation, fibrosis, revascularization and reinnervation remain the major impediments to complete recovery of contractile function following major skeletal muscle trauma like VML [20][21][22][23][24] . Hence, in addition to topographical guidance, tissue engineered scaffolds must be myo-conductive, promote angiogenesis through either pre-vascularization or incorporation of vasculogenic accelerant, have immunomodulatory effect as well as facilitate reinnervation to overcome the pathophysiology of VML and achieve functional restoration. Direct administration of anti-inflammatory agents can reduce fibrosis but has been shown to hinder muscle regeneration 25,26 . Acellular scaffolds cannot address such multidimensional challenges and reinforces the necessity of incorporating multiple cell types in a scaffold 18 . Pre-vascularized constructs fabricated by coculture of myoblasts and endothelial cells on aligned nanofibrous scaffolds have been found to promote organized myofiber regeneration and vascular integration in murine VML model 27 . Although angiogenic and immunomodulatory strategies have been explored in VML repair, lack of innervation in an engineered muscle remains one of the major impediments to its success as a functional muscle replacement 28 .
In the absence of neural In the present study, we report the development of a novel pre-innervated tissue engineered muscle construct for application in VML repair (Fig 1). Aligned nanofiber sheets were used to coculture skeletal myocytes and spinal motor neurons and explore the effect of innervation on maturation of myocytes in vitro (Fig 1a). The bioscaffolds were further implanted in athymic rat model of VML and evaluated for cell survival, satellite cell proliferation, microvasculature and neuromuscular junction density (NMJs) at an acute time point (Fig 1b). This is the first report studying the acute effect of pre-innervated constructs on the regenerative micro-environment of injured muscle following severe muscle trauma.

Pre-Innervation promotes myocyte fusion and formation of NMJs in vitro
Mouse skeletal myoblast cell line C2C12 were cultured on aligned polycaprolactone (PCL) nanofiber scaffolds and allowed to differentiate. Differentiated myofibers were found to align along the direction of nanofibers as observed by staining for F-actin (Fig 2a). Similarly, spinal motor neurons cultured on the nanofibers exhibited axons aligning along the nanofiber orientation (Fig 2b). Subsequently, both motor neurons and myocytes were cocultured on the nanofiber scaffolds. Motor neuron-myocyte coculture led to formation of thick intertwined nervemuscle bundles aligned along the nanofibers (Fig 2c-c′′). Within 7 days of coculture on the nanofiber scaffolds, NMJs were observed by colabelling for presynaptic marker Synaptophysin and Bungarotoxin mediated identification of post synaptic Acetylcholine Receptors (AchR) (Fig   3a-a′). Motor neurons were also found to promote myocyte maturation and fusion in vitro leading to significantly higher myocyte fusion index (MFI) as compared to myocyte only cultures (Fig 3bc). Taken together, these data clearly demonstrates that innervation not only leads to NMJs in vitro but also facilitates myocyte maturation.

Bioscaffold implantation in athymic rat model of VML
The tibialis anterior (TA) muscle of athymic rats was exposed and a 10 mm×7 mm×3 mm (length × width × depth) segment of the muscle was excised corresponding to ~20% of gross muscle weight to create a VML model (Fig 4a-b). The animals were randomized into the following repair  (Fig 4c). At terminal time point of 7 days post implant, the nanofiber sheets were visible upon TA exposure and appeared intact (Fig 4d). Further, the graft area in the No Repair group appeared to be recessed and atrophied as compared to the Repair groups (Fig 4e-f).

Evaluation of acute cell survival in bioscaffolds upon implantation in VML model
All animals were sacrificed after 7 days and the whole anterior muscle compartment of the hind limbs were fixed in paraformaldehyde. The muscles were cryopreserved, embedded in OCT, sectioned and stained. Immunohistochemical analysis of cross sections of the injury/repair sites were performed to detect implanted cells on the nanofiber sheets and assess overall muscle health. We found that the nanofiber sheets were intact after 7 days in all the animals. Myocytes positive for Phalloidin (F-actin) were observed in both MN-MYO and MYO groups ( Fig. 5a-b).
Animals implanted with SHEETS only did not show significant Phalloidin+ cells within the implant region while the NO REPAIR group was left with a gap that was eventually found to be filled with and motor axons within the implanted nanofiber sheets (Fig. 6). These results indicate the survival of implanted motor neurons and myocytes at acute time point following a VML repair.

Pre-Innervated constructs promote satellite cell migration near injury area
Satellite cells are resident myogenic precursor cells essential for muscle regeneration 30 .
Activation and mobilization of satellite cells to the sites of injury is a major contributor to the regenerative capability of skeletal muscle 31 . Satellite cell migration near the injury area was observed across all groups by staining with Pax7 (Fig 7a-d). Pax7 + nuclei located on the periphery of Skeletal Muscle Actin + myofibers and colabelling with pan-nuclear marker DAPI were identified as satellite cells (Fig 7e).

Pre-Innervated constructs lead to increased microvasculature near injury area
Revascularization of the injured area/implant is critical for survival of implanted cells and integration with host vascular system. Vascularization near the injury area was evaluated by staining tissue sections with endothelial cell specific marker-CD31 and Smooth Muscle Actin (SMA) (Fig 8). CD31 + /SMA + structures with a visible lumen and cross-sectional area greater to the injury site while other groups had more punctate CD31 + cells (Fig. 8). The presence of significantly higher microvasculature around the injury area in MN-MYO group suggests that innervated tissue engineered muscle constructs can potentially augment revascularization following VML repair.

Innervated constructs
Acetylcholine Receptor (AchR) clusters have major implications in formation and maintenance of motor end plates during muscle development as well as regeneration 32,33 . Indeed, bungarotoxin staining, a known marker of nAchR a7 receptors 34 , showed the presence of pretzel-shaped AchR clusters around the injury area across all groups (Fig 9a-e). A count of AchR cluster near the injury area revealed that the MN-MYO group had significantly more AchR clusters than the other groups (Fig. 9f). inflammation, denervation and atrophy 35 . Mature NMJs were identified as pretzel shaped structures which were colabelled with presynaptic marker Synaptophysin and postsynaptic AchR marker (Bungarotoxin) (Fig 10 a-e′). The percentage of AchR clusters near the injury area which were positive for Synaptophysin was calculated to quantify the amount of mature NMJs (Fig   10f). Pre-innervated constructs (MN-MYO) were found to have significantly higher percentage of mature NMJs as compared to other groups indicating the potential role of pre-innervation in augmenting formation of mature NMJs following implantation in VML model (Fig 10f).

Discussion
Severe musculoskeletal trauma like VML is accompanied by progressive motor axotomy over several weeks, leading to denervation of the injured muscle thereby severely limiting functional recovery 36 . Hence, appropriate somato-motor innervations remain one of the biggest challenges in fabricating a fully functional muscle. Apart from augmenting re-innervation process, tissue engineering strategies need to provide accurate cellular alignment and enable bulk muscle repair. The present study is the first report on using synthetic polymer based aligned nanofiber scaffolds in a rat VML model. We have used commercially available nanofiber sheets made of polycaprolactone (PCL)-which is an FDA approved slowly degrading, bioresorbable polymer 38 .
These aligned PCL nanofiber sheets were used as scaffolds for 3D motor neuron-myocyte coculture. We studied the effect of motor neurons on myocytes in vitro and observed that motor neurons cultured on pre-differentiated skeletal myocytes led to formation of mature NMJs and promoted fusion and bundling of myocytes to form multinucleate myofibers (Fig 3). This is in agreement with previous report which describes that enhanced fusion and maturation of myocytes are only observed when the myocytes are allowed to fully differentiate before introduction of the motor neurons and the coculture is maintained subsequently in serum-free conditions 5 .
To evaluate the in vivo potential of pre-innervated constructs as a reconstructive approach to VML, we used a standardized model of VML in the rat TA muscle 39 (Fig 4). The aligned nanofiber sheets were highly porous (80% porosity) and our method of stacking three layers of sheets allowed exchange of nutrients and oxygen through blood perfusion thereby facilitating survival of the implanted motor neurons and myocytes. Subsequent immunohistological analysis of transverse and longitudinal sections of the muscle allowed visualization of multiple layers nanofiber sheets and confirmed the presence of long thick bundles of skeletal myocytes and motor axons on the nanofiber sheets confirming acute survival of the implanted cells (Fig 5-6). In order to achieve a comprehensive understanding of the acute effects of innervation on the regenerative milieu of an injured muscle, we proceeded to investigate the density of satellite cells, microvasculature, AchR clusters and mature NMJs near the injury area.
The robust regenerative capacity of skeletal muscles can be largely attributed to the resident myogenic precursor cells called muscle satellite cells 30 . These satellite cells lie quiescent in between the basal lamina and sarcolemma and gets activated within a few days after an injury.
Activated satellite cells then differentiate to form myoblasts which fuse together to form new skeletal muscle fiber 30 . Satellite cells can be reliably identified by paired box transcription factor Pax-7 which is expressed in both quiescent and activated stages 46 . Although pre-vascularized tissue engineered constructs have been shown to promote satellite cell activation upon implantation in a mild muscle injury model, the effects of pre-innervation on host satellite cell population are yet to be addressed 47 . Separate studies indicate that various neurotrophic factors like NGF and BDNF play a critical role in modulating satellite cell response within an injured muscle 48,49 . For instance, exogenous treatment with BDNF was enough to recover the regenerative capacity of satellite cells in BDNF-deficient mice after skeletal muscle injury 48 .
Spinal motor neurons used in the present study for fabrication of pre-innervated constructs have been shown to secrete BDNF 50 . This can potentially explain the presence of significantly more Pax-7+ satellite cells near the injury area in MN-MYO group having pre-innrevated constructs comprising of motor neurons and myocytes (Fig 7). However, unlike Czajka et al's report showing satellite cell migration within a pre-vascularized tissue engineered construct within 3 days of implantation, we did not observe any Pax-7+ satellite cells within our constructs by 7 days 47 (Fig 7). This is likely due to the difference in models; VML presents a very different pathophysiology than does a mild incision injury. It is also possible that the inherent hydrophobic nature of the PCL nanofibers used here could have restricted host satellite cell infiltration 51,52 .
Tissue engineering strategies for VML repair demands bulk muscle reconstruction. Inadequate re-vascularization remains one of the major challenges to engineer thick skeletal muscle limiting nutrient exchange and survival of implanted cells 23 . Pre-vascularized constructs comprising of preformed vascular networks have been shown to promote microvasculature, vascular perfusion of the graft and inosculation with host vascular system thereby significantly improving muscle regeneration following VML 23,27,47 . Neurotrophic factors like NGF, BDNF, GDNF, NT-3 have been reported to enhance angiogenesis in different tissues like skin, heart and cartilage through receptor mediated activation or recruitment of proangiogenic precursor cells [53][54][55] . Spinal motor neurons secrete BDNF whereas astrocytes can express a range of neurotrophic factors 50,56 . It is to be noted that although we strive to obtain a pure motor neuron population, we have detected minimal glial cells in our motor neuron cultures (data not shown). We have observed that the pre-innervated constructs used in this study lead to significant increase in microvasculature near the injury area following implantation in a VML model (Fig 8). Although the molecular mechanisms of how pre-innervated constructs promote vascularization is the scope of future studies, it is reasonable to postulate, that neurotrophic factors from our spinal cord derived cell population (comprising of motor neurons and glia) triggered this increased microvasculature.
Interestingly, despite such increased microvasculature near the injury area we did not find any evidence of endothelial cells within the implanted nanofiber sheets. Aside from the inherent hydrophobicity of PCL, the absence of endothelial cell infiltration within the graft can also indicate that the evaluation time of 7 days was too early to observe vascular integration of the construct.
Muscle nicotinic AchRs are pentameric structures that are dispersed along the basal membrane of myofibers (extra-junctional) during fetal stage and progressively redistribute to form localized (junctional) pretzel shaped clusters on adult muscle 57 . AchR clustering plays a pivotal role in skeletal muscle function and regeneration through formation of stable motor end plates thereby effecting functional restoration following severe musculoskeletal trauma. Early physical rehabilitation involving exercise has been shown to benefit patients with VML in recovering muscle force and range of motions 58 . Similarly, rehabilitative exercise in conjunction with bioengineered constructs augments functional restoration in murine models of VML by promoting formation of AchR clusters and mature NMJs 59-62 . One of the key findings of the present study is that pre-innervated constructs augments AchR clustering and NMJ formation.
This is reflected in our results which show MN-MYO group have significantly more AchR receptor clusters and mature NMJs within 7 days of implanting in a VML model (Fig 9-10). Clustering of AchRs in skeletal muscles is mainly controlled by motor innervation through secretion of neural agrin by the motor neurons 33,63 . This may be a reason behind increased density of AchR clusters near the injury area following implantation of pre-innervated constructs. In a denervated muscle following injury, AchR clusters can again disperse to form extra-junctional immature receptors.
This often leads to an increase in overall count of Bungarotoxin+ structures in an injured denervated muscle 64 . Hence, in order to investigate if the observed increase in AchR clusters was only an injury effect, we looked for innervated AchR clusters that would indicate formation of mature NMJs and preservation of the motor end plate. The MN-MYO group was found to have significantly higher percentage of innervated AchR clusters as compared to other groups (Fig   10). This confirms that pre-innervated constructs promote formation of mature NMJs around the injury area at acute time points which can potentially have a significant effect in augmenting functional restoration at more chronic time points.
Although our study demonstrates the potential of pre-innervated constructs in promoting a regenerative environment following VML, several limitations exist. First, this study was conducted in athymic rats lacking an intact immune system and was terminated at an acute time In summary, the present study is the first to explore the implications of pre-innervation on the regenerative microenvironment in a rat VML model at an acute time point. This is also the first report on the use of synthetic polymer derived aligned nanofiber scaffolds as an implant in a VML model. Our results indicate that pre-innervation promotes myocyte maturation in vitro, satellite cell migration and vascularization in the injury area as well as facilitates formation of mature NMJs thereby providing a favorable regenerative microenvironment for neuromuscular regeneration following VML. We believe that these findings in skeletal muscle injury model would stimulate further research into developing pre-innervated tissue engineered constructs for application in smooth muscle as well as cardiac tissue engineering. In future work, these nervemuscle constructs may also be fabricated using cells derived from adult human stem cell sources (e.g., iPSCs), thereby making them translational as an autologous, personalized bioengineered construct. These pro-regenerative effects can potentially lead to enhanced functional neuromuscular regeneration following VML, thereby improving the levels of functional recovery following these devastating injuries.

Isolation and Culture of Rat Spinal Motor Neurons
Motor neurons were harvested from the spinal cord of E16 Sprague Dawley rat embryos following previously described procedure 65,66 . All harvest procedures prior to dissociation were conducted on ice. Briefly, spinal cords were extracted from the pups and digested with 2.5% 10X trypsin diluted in 1mL L-15 for 15 mins at 37 o C. The digested tissue was triturated multiple times with DNAse (1mg/mL) and 4% BSA and centrifuged at 280g for 10minutes to pool all the cell suspension. Subsequently, the cell suspension was subjected to Optiprep mediated density gradient centrifugation at 520g for 15 minutes to separate the motor neuron population.
Following centrifugation, the supernatant was discarded, and cells were resuspended in motor neuron plating media consisting of glial conditioned media. Glial conditioned media was made as described earlier 65  At least three 2mm 2 area was considered per sample for counting MFI and the average was plotted for each sample (Fig 3c).  ). Prior to implantation, the sheets were washed thoroughly with PBS to remove any leftover media. The fascia, connective tissue, and skin were closed in layers with 8-0 prolene, 6-0 prolene, or staples, respectively. At the conclusion of the surgery, the area was cleaned with alcohol and animals were given a subcutaneous injection of sustained-release meloxicam (4mg/kg). Animals were placed on heating pads until recovered and returned to home cages.
For quantitative measurement of satellite cell, micro-vessel, AchR cluster and mature NMJ density, an area of 5mm 2 (5mm long and 1mm wide) was chosen at 100µm from injury/implant site towards the host muscle and defined as the injury area. At least 3 cross-sections each separated by 300µm was considered for counting and average density was plotted in the graph and compared across groups.

Statistical Analysis
All quantifications reported in this study were performed by personnel blinded about the treatment groups. All statistical analysis was performed using GraphPad PRISM software. For comparison between two groups only (Fig 3c), an unpaired two-tailed Student's t-test with Welch's correction was used. For comparison between multiple groups, a one-way analysis of variance (ANOVA) was performed with post hoc Tukey's adjustment with 95% Confidence Interval (Fig 7f, 8f, 9f, 10f). Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). All graphs were made in GraphPad PRISM and display mean ± standard error of mean (SEM).

Data Availability
Data supporting the conclusions of this paper are available from the corresponding author upon reasonable request.