Decellularized skeletal muscles display neurotrophic effects in three‐dimensional organotypic cultures

Abstract Skeletal muscle decellularization allows the generation of natural scaffolds that retain the extracellular matrix (ECM) mechanical integrity, biological activity, and three‐dimensional (3D) architecture of the native tissue. Recent reports showed that in vivo implantation of decellularized muscles supports muscle regeneration in volumetric muscle loss models, including nervous system and neuromuscular junctional homing. Since the nervous system plays pivotal roles during skeletal muscle regeneration and in tissue homeostasis, support of reinnervation is a crucial aspect to be considered. However, the effect of decellularized muscles on reinnervation and on neuronal axon growth has been poorly investigated. Here, we characterized residual protein composition of decellularized muscles by mass spectrometry and we show that scaffolds preserve structural proteins of the ECM of both skeletal muscle and peripheral nervous system. To investigate whether decellularized scaffolds could per se attract neural axons, organotypic sections of spinal cord were cultured three dimensionally in vitro, in presence or in absence of decellularized muscles. We found that neural axons extended from the spinal cord are attracted by the decellularized muscles and penetrate inside the scaffolds upon 3D coculture. These results demonstrate that decellularized scaffolds possess intrinsic neurotrophic properties, supporting their potential use for the treatment of clinical cases where extensive functional regeneration of the muscle is required.

including nervous system and neuromuscular junctional homing. Since the nervous system plays pivotal roles during skeletal muscle regeneration and in tissue homeostasis, support of reinnervation is a crucial aspect to be considered. However, the effect of decellularized muscles on reinnervation and on neuronal axon growth has been poorly investigated. Here, we characterized residual protein composition of decellularized muscles by mass spectrometry and we show that scaffolds preserve structural proteins of the ECM of both skeletal muscle and peripheral nervous system. To investigate whether decellularized scaffolds could per se attract neural axons, organotypic sections of spinal cord were cultured three dimensionally in vitro, in presence or in absence of decellularized muscles. We found that neural axons extended from the spinal cord are attracted by the decellularized muscles and penetrate inside the scaffolds upon 3D coculture. These results demonstrate that decellularized scaffolds possess intrinsic neurotrophic properties, supporting their potential use for the treatment of clinical cases where extensive functional regeneration of the muscle is required. of diseased tissues. 1 Synthetic scaffolds have the advantage over natural biomaterials in that their structure, topography, and mechanical properties can be finely tuned to design an optimal environment for a particular biological application. 1 Despite incredible improvements have been achieved in biomaterial manufacturing, many challenges remain in preparing scaffolds that recapitulate in vitro, the complexity of the tissue microenvironment. The peculiar combination of the tissue-specific extracellular matrix (ECM) biochemistry, biomechanics, and three-dimensional (3D) organization cannot be fully reproduced in the lab. [2][3][4] Therefore, there has been increasing interest in using naturally derived ECM itself, as decellularized (decell) tissues or whole organs, where such complexity can instead be preserved. [5][6][7] The decellularization process removes cellular and nuclear content of the native tissue, while retaining ECM mechanical integrity, biological activity, and 3D architecture. 6 Decell scaffolds are highly biocompatible and show absence of rejection after allogeneic or xenogeneic transplantation. 8 Altogether, these properties make them an important and promising alternative biomaterial for the treatment of clinical cases as traumatic injuries, surgical ablations, and congenital malformations. 9,10 Indeed they have already been obtained from different organs and used for regenerative medicine strategies in animal models, as well as in clinical trials. 9,10 In particular, decell muscles have been shown to promote muscle regeneration in volumetric muscle loss models. 7 We recently developed perfusion methods for the generation of skeletal muscle scaffolds which retain 3D structural organization of the tissue, as well as ECM components and growth factors. Decell muscles were used as xenograft to promote tissue regeneration in a murine model of volumetric muscle loss, which also allowed innervation and regeneration of the neuromuscular junctions. 11 In agreement with this, other studies have demonstrated that, when implanted in vivo, decell muscles are not only able to restore muscle mass, but also trigger the regeneration of the nervous system with overall functional recovery. 12,13 Finally, it was recently reported that decell muscles guide nerve regrowth in a diaphragmatic hernia mouse model. 14 During in vivo peripheral nerve regeneration, injured axons are able to elongate into the distal nerve stump if they find a permissive substrate. This is mainly provided by trophic support from Schwann cells, connective cells, and ECM. Eventually, regenerating axons will be mostly able to reach the distal target organs and reinnervate them, thus allowing for the recovery of lost functions. 15,16 The ECM is an essential player required for the formation of axonal tracts as well as for the maturation and function of synapses in the peripheral nervous system. [16][17][18][19][20] As extensively demonstrated by in vitro studies and tissue engineering strategies, axonal regrowth and guidance are enhanced by ECM components, including collagen IV or laminin. [16][17][18][19][20] In agreement with this, decell scaffolds prepared from nervous system-derived tissues (including decell nerves) have been proved to support long-distance axon regeneration in peripheral nerve injury in animal models [21][22][23][24] as well as in patients. [25][26][27] Other decell tissues used for nerve repair in in vivo models include scaffolds derived from small intestinal submucosa, 28 amniotic tissue grafts, 29,30 and umbilical cord. 31 Few studies also reported the ability of implanted decell muscles to repair nerve injury in vivo 27,32-35 .
The evidence that decell muscles promote myogenesis have been observed both in in vivo and in vitro models, indicating that these scaffolds directly preserve biological activity able to guide myogenic cells toward the generation of myofibers. 12,36 On the contrary, innervation of decell muscles has only been observed in in vivo models. 11,12,14 Therefore, it remains unclear whether the neurotrophic properties of the scaffolds observed in vivo could be the result of the overall regenerative response to the implant, or if decell muscle could directly promote axon invasion.
The implementation of the nervous system is essential for skeletal muscle tissue functionality, which is a necessary feature for the future use of decell muscles in clinical application and 3D in vitro modeling. The purpose of this study was, therefore, to investigate the direct ability of decell scaffolds to promote axonal sprouting and invasion in vitro. To do so, we developed a 3D coculture system of organotypic spinal cord slides (oSpC) and decell muscles. This experimental approach allows the study of the neurotrophic effect of the scaffolds by excluding the influence of other cellular and/or systemic components that instead exist in vivo.

| Decellularized muscle preparation
Rats (250-350 g) were used as a source of muscle for decellularization.
The leg was dissected from the rest of the body by splitting the pelvis at the pubic symphysis and the sacroiliac joint. We performed decellularization as previously described. 11 Briefly, a 24 G cannula was inserted into the abdominal iliac artery and advanced distally to allow perfusion condition at a flow rate of 1 mL/min. Limbs were perfused with 0.25% SDS (Sigma) for 72 hours and washed in deionized water for 48 hours. After decellularization, the muscles of interest were dissected, treated with 137Cesium irradiator (IBL 437C), and preserved at 4 C, in phosphate buffered saline (PBS, Gibco) with 1% Penicillin/Streptomycin (P/S, Gibco).

| Proteomic sample preprocessing
The decellularized matrix was freeze-dried for 72 hours (Labconco

| Proteomic liquid chromatography-tandem mass spectrometry analysis
Protein identification by liquid chromatography tandem mass spectrometry (LC-MS/MS) was performed using Thermo Fusion Mass Spectrometer with Thermo Easy-nLC1000 Liquid Chromatography.
Ninety minutes of LC-MS gradients were generated by mixing buffer A (0.1% formic acid in water) with buffer B (0.1% formic acid in 80% acetonitrile (ACN) in water) by different proportions. Using nanospray ionization (NSI) as the ion source and Orbitrap as the detector, the mass scan Rang was at 300 to 1800 m/z, and the resolution was set to 120 K. The MS/MS was isolated by Quadrupole and detected by Ion trap, whose resolution was set to 60 K. The activation type was higher-energy collisional dissociation (HCD).

| Proteomic bioinformatics analysis
Peak list files were searched against UniProt Mus musculus reference proteome by Thermo Proteome Discoverer 2.2, due to the high similarity but more complete annotation of this proteome respect to that of Rattus norvegicus ( Figure S1). Searches were performed using a 10 ppm precursor ion tolerance for total protein level profiling. The product ion tolerance was set to 0.02 Da in mascot TMT6 quantification searches. TMT6 modification (229.163 Da) and carbamidomethyl on cysteine (+57.021 Da) were set as static modifications. The oxidation of methionine residues (+15.995 Da) was set as a variable modification. Peptide-spectrum matches were adjusted to a 1% and then assembled further to a final protein-level false discovery rate of 1%.
Proteins not identified in all three replicates or identified with a q-value >0.05 were filtered out. Protein localization was annotated according to the following gene sets: matrisome (structural and associated), 37   Hoechst 33342 (ThermoFisher, H1399) was used to stain nuclei.

| Imaging acquisition and analysis
Samples were analyzed with the following microscopes: epifluorescence Olympus BX60; fluorescence stereomicroscope Leica MZ16F, equipped with Canon EOS1000D camera; confocal Leica TCS SP5 microscope; confocal ZEISS LSM 800 microscope; wide-field motorized stage Leica DM6B; modular multiphoton microscope (Bergano-II, Thorlabs) coupled with two synchronized pulsed laser beams (excitation 800 nm). ImageJ software was used for image processing, contrast and intensity level adjustment, and 3D reconstruction. For directionality analysis, Directionality ImageJ plugin was used to analyze bright field images of the region between the oSpC and the scaffolds or the axons sprouting from the oSpC in absence of scaffolds. The plugin measured the amount of structures in a given direction every 2 (from −90 to +90 ), where scaffolds were placed at 0 axis. Images with completely isotropic content are expected to give a flat histogram, whereas images in which there is a preferred orientation are expected to give a histogram with a peak at that orientation. The quantification was expressed as the mean of 4 to 6 independent biological replicates.

| Statistical analysis
All the analyses were performed by using GraphPad Prism 6 software.
Plotted data were expressed as mean ± SEM. We determined statistical significance by unequal variance Student's t test or one-way ANOVA (analysis of variance) and Tukey's multiple comparison test or Kruskal-Wallis and Dunn's multiple comparison test. A P value of less than .05 was considered statistically significant.

| RESULTS
Based on the cell instructive cue exerted by decell muscles upon implantation in volumetric muscle loss models 11 and on the role of ECM during innervation, [16][17][18][19][20] we first characterized the residual protein composition of decell scaffolds by mass spectrometry. After applying stringent filtering criteria, we identified 2081 proteins (Table S1). At the protein level, our results indicate that decell muscles are not only composed by structural ECM proteins, but these includes also numerous other associated proteins ( Figure 1A). In detail, we identified 72 ECM structural proteins (including collagens, laminins, fibronectin, nidogen-1 and -2, and proteoglycans) and 46 ECMassociated proteins ( Figure 1A and Table S2). Within the latter category, multiple proteases and other ECM remodeling enzymes were included. Among these, we identified cathepsins, a disintegrin, ADAM family metalloproteases (Adam10 and Adamtsl4), and protease inhibitors (as Serpineb1a, Serpineb6, Serpine2, Serpinf1, Serpinh1). We already demonstrated that decell muscles preserve single anucleated myofibers (that could also be isolated) and sarcolemmal proteins such as dystroglycans. 11 Here, we confirmed these findings, identifying cytoskeleton proteins (including myosins, actins, and desmin) and sarcolemma proteins (GO-CC: 0042383), including dysferlin (Dysf) and aquaporins (Aqp1, Aqp4 and Dag1) with its interacting partners (such as Lama2, Dmd, and Cav3; Figure 1A and Table S1). Direct associations between mitochondria and the cytoskeleton exist in myofibers. 41 In agreement with this, proteins known to be involved in mitochondria motility along microtubules such as dyneins, dynactins, and kinesins were preserved ( Figure 1A and Table S1). We also revealed proteins composing extracellular vesicles in decell muscles ( Figure 1A and Table S1). Interestingly, the identification of proteins involved in the ECM remodeling or composing extracellular vesicles strengthen the concept that decell muscle retains biological cues typical of the native tissue.
To determine the tissue-specificity of protein content preserved in decell scaffolds, we compared our identified proteins to the curated protein composition of different tissues ( Figure S1C,D). When we selected the proteins specific for each tissue, skeletal muscle was prevalent with approximately 10-fold increase in the number of proteins among the tissues under consideration ( Figure 1B). Moreover, we found that decell scaffolds preserved specific ECM proteins known to play a role in nerve regeneration and neurite outgrowth, as collagens (I, IV, VI), laminins (α2-, α4-, β1-, and γ1-chains), and fibronectin. 15,17,42 These proteins are known to be directly interacting with each other and form a well-connected network ( Figure 1C).
Moreover, ECM components of nerves, such as myelin constituents (including myelin basic protein, Mbp, myelin P2 protein, Pmp2, and myelin proteolipid protein, Plp1), and specific ECM proteins of the synaptic basal lamina of neuromuscular junctions (including nidogen-2, Nid2; laminin α4-, α5and α2-chains) were also identified (Table S1 and S2). Altogether, these data support the hypothesis that decellularized muscles retain both muscular and neuronal tissuespecific ECM components, including proteins that have been shown to drive nerve regeneration and promote neurite outgrowth.
To test whether decell muscles could have an intrinsic neurotrophic effect, scaffolds were cocultured with oSpCs in a 3D environment. The use of this culture system had the aim to retain 3D F I G U R E 1 Proteomic analysis of decellularized skeletal muscle. A, Classification of the 2081 identified proteins based on their localization. Edge thickness is proportional to the number of proteins in common between the two linked categories; circle radius is proportional to the number of proteins in that category. B, Number of proteins identified in our data that are annotated exclusively to the indicated tissue. Reference data of human proteins per tissue are described in Section 2. C, Protein-protein interaction network, including only ECM structural proteins having at least one reported neighbor. Collagens, laminins, and fibronectin are highlighted in blue, green, and orange, respectively. Edge thickness is proportional to confidence of interaction. ECM, extracellular matrix organization, multiple cell composition, and cell-ECM interaction of neural cells within organotypic spinal cords, 43,44 as well as the specific skeletal muscle environment provided by the decell scaffolds.
This strategy should allow to better mimic the in vivo innervation process, excluding the contribution of muscle regenerative and systemic responses (ie, inflammation) to the innervation process. In particular, we investigated the ability of decell scaffolds to sustain neural projection sprouting within its 3D environment by culturing oSpC in close proximity to the scaffolds or to attract neural axons when cocultured at a distance from each other.
To reach this aim, we first characterized 3D oSpC culture within When used in vivo to repair a resected muscle, decell scaffolds were permissive to innervation along the entire length of the implants. Importantly, this included the median region located far from the host tissue. 11 Therefore, we also hypothesized that the scaffolds could possess direct neuroattractant properties. To investigate this, oSpC were cocultured with decell muscles by seeding them at a distance from each other and embedding them in Matrigel droplets to allow a gel-mediated physical connection (Figure 2A). Axonal spouting was monitored during the culture period at 4, 7, and 14 days after seeding ( Figures 4D-F and S5). We first quantified axon directionality, comparing oSpC culture performed in Matrigel droplets (a) in the absence of decellularized muscle, (b) in the presence of decellularized scaffolds, or (c) in the presence of inert scaffold. Four days after seeding, no F I G U R E 3 3D coculture of oSpC onto decellularized muscles. A, Representative stereomicroscope live imaging of Calcein (green) incorporation from oSpC cultured onto decellularized muscle at 14 days after seeding. Scale bar = 1 mm. BF, bright field. B, Two-photon live imaging of SC-derived neural projection incorporating Calcein (green) at 14 days after seeding onto decellularized muscles. Scale bar = 200 μm. C, z-stack two-photon image of whole mount oSpC cultured onto decellularized muscle immunostained for Tuj1 (green) and laminin (red) at 14 days after seeding.  (Figure 4D,E). Conversely, marked orientation of axons was revealed in oSpC cocultured with decell muscles, with projections sprouting toward the decell muscles ( Figure 4D,E). Based on these results, we also evaluated axon length 4 days after seeding. The presence of decell scaffolds did not influence significantly the length of neural projections, when compared to oSpC cultured in Matrigel or in the presence of inert scaffold F I G U R E 4 Evaluation of multicellular composition and axonal attractant effect of decellularized muscles on oSpC culture. A-C, z-stack images showing immunostaining for ISL1/2 (green) and laminin (red; A), nkx6.1 (green) and laminin (red; B), or S100B (green; C) of cross-sections performed in the middle region of the oSpC/decell muscle at 14 days after seeding. Nuclei were stained with Hoechst (blue). Scale bars = 50 μm.
Dashed lines indicate the interface between scaffolds and oSpCs. D, Representative bright field images of oSpC cultured into Matrigel, cocultured with inert scaffolds or cocultured with decell muscles at 4 days after seeding. The insets show axon projections within the Matrigel droplet. Scale bars = 1 mm (left panel) and 100 μm (right panel). E, Quantification of neuronal projection directionality in oSpC section cultured in presence of Matrigel (green), of inert scaffold (red) or decell muscle (black) at 4 days after seeding. Data are shown as mean ± SEM of four independent replicates; multiple comparison one-way ANOVA (analysis of variance) was used. ***P < .01 among all the experimental groups. F, Two-photon live imaging of oSpC-derived neural projection incorporating Calcein (green) that run within the Matrigel toward decell muscle identified with SHG (gray) at 14 days after seeding. The image shows neural projections sprouted from the central body of oSpC, which instead is out from the optical field. Scale bar = 200 μm. SHG, second-harmonic generation ( Figure S4). Moreover, the preferential organization of neural projections observed during the first days of culture was not appreciated anymore at longer time points, due to incremental sprouting in all the directions ( Figure S5). Notably, projections of viable cells were present within the decell scaffold at 14 days after seeding, while nuclei were almost completely retained in the body of the oSpC. This was confirmed via two-photon microscopy imaging of Calcein incorporating oSpC-derived projections and by second-harmonic generation imaging that revealed decell ECM ( Figure 4F). These results indicate that decell scaffolds promote neural attraction during the early stages of the 3D coculture. Accordingly, projections sprout toward the scaffold also when oSpC and decell muscles are positioned at a distance from each other.

| DISCUSSION
The perfusion process of decellularization allows for the maintenance of the native skeletal muscle complexity, 7,45 not only composed of myofibers, but also constituted by other structures such as peripheral nerves. Here, we show that upon decellularization, the skeletal muscle scaffold proteomic composition is much more complex than a mere network of structural ECM proteins associated with the myofibers.
Organotypic cultures of neural slides have been shown to represent a middle ground between dissociated cells and in vivo studies, and have been instrumental in enhancing our understanding of axon guidance. 43,44 The use of our 3D coculture system allowed to retain cell-cell and cell-matrix interactions and to investigate the direct neuroattractant effects of decell muscles, excluding other players that could instead operate in vivo during the regeneration of the implants. In particular, our model sustained the preservation of neurons expressing Isl1/2 and Nkx6.1, which are both markers required for motor neuron specification, 46-48 as well as neuronal cells expressing S100B, which is localized in many neural cell-types, including astrocytes. 49 It is known that the ECM can either promote or inhibit the elongation of neurites and modulate axon growth. [16][17][18][19][20] For example, laminin is an adhesive component of the ECM secreted by Schwann cells.
Laminin-rich basal lamina scaffolds have shown pro-regenerative capability following nerve injury, promoting axonal outgrowth in vitro 15,16 . Moreover, studies aiming at using decell tissues to repair peripheral nerve injuries suggested that the basal lamina of implanted decell scaffolds (that includes collagen type IV, fibronectin, and laminins) could mimic the endoneurial tube, thus promoting in vivo nerve outgrowth and nerve regeneration. 27 Our proteomic analysis of decell muscles confirmed the preservation of ECM proteins involved in axon growth (such as collagens, laminins, and fibronectin), as well as components of the peripheral nerves such as myelin constituents.
Therefore, we can speculate that the observed ability of axons to sprout over and within the scaffolds could be the result of a physical guidance exerted by the structural and molecular composition of the ECM preserved in decell muscle.
Furthermore, the ability of decell muscles to attract neuronal axons strongly supports the hypothesis that the scaffolds have an intrinsic neurotrophic nature. The identification of the mechanism underpinning axon attraction from decell muscles remains an intriguing aspect that needs further investigation. However, the preservation of the signaling components in decell muscles strongly suggests that the scaffold could serve as a reservoir of neuro-attractant molecules.
Indeed, decell muscles preserved not only structural proteins of the ECM, but also extracellular vesicles and proteins involved in the ECM remodeling. Extracellular vesicles comprise a heterogeneous population of membrane vesicles with particular lipid, protein, and nucleic acid composition that are considered as an additional mechanism for intercellular communication, including the regulation of signal transduction and cell adhesion. 50 Interestingly, together with the presence of such signaling components, we also demonstrated that decell scaffolds preserve chemokines, such as vascular endothelial growth factor (VEGF) and insulin-like growth factor (IGF-1), abundance of which was estimated to be approximately 39 pg/mg wet decell tissue and 44 pg/mg wet decell tissue, respectively. 11 Both IGF-1 and VEGF are neurotrophic factors that have been shown to promote peripheral nerve regeneration, axonal targeting and outgrowth, and to be protective in both in vitro and in vivo models of neuronal degeneration. [51][52][53] A number of studies have demonstrated that soluble factors released from the ECM and its degradation products themselves are capable of recruiting both neural differentiated cells and progenitors to the site of remodeling, 54 as well as Schwann cells from products derived from decellularized small intestinal submucosa. 55 It is, therefore, not unconceivable that signaling molecules could be released from the decell muscles during the time in culture, creating the chemotactic gradient, within the 3D culture system that promoted axonal sprouting toward the scaffolds.

| CONCLUSION
Our study demonstrated that decell muscles obtained preserving the native tissue environment have direct neurotrophic properties. This strongly suggests that our model could represent a powerful tool to investigate in vitro axon sprouting and guidance within a complex native-like skeletal muscle 3D environment.