Tensile Forces and Nanofiber Alignment Influence Both Innervated and Non‐Innervated Skeletal Myofiber Formation in Custom Mechanobioreactors

ABSTRACT

While it is understood that muscle tissue generates contractile forces, it is less appreciated that muscle dynamically responds to applied forces during development. We previously fabricated tissue engineered muscle comprised of skeletal myocytes in co‐culture with spinal motor neurons on aligned nanofiber poly‐caprolactone scaffolding, demonstrating that innervation elicited more robust myofibers and formation of neuromuscular junctions. The current study utilized custom mechanobioreactors to apply tensile elongation to this engineered muscle platform to explore the effects of exogenous forces and scaffold topology on innervated versus non‐innervated myocytes. Nanofiber scaffold alignment played a significant role in myocyte thickness, width, and fusion under both innervated and non‐innervated conditions. A combination of tensile loading and nanofiber alignment increased myocyte fusion, suggesting these parameters work together to expedite and enhance myofiber formation and maturation. Overall, this multi‐faceted paradigm, featuring biomechanical loading, substrate topology, and innervation, mimics key features of the developmental microenvironment experienced by myocytes in vivo. These findings may facilitate more sophisticated studies on muscle development, function, and responses to trauma while also elucidating principles to support the fabrication of engineered muscle to repair major muscle defects.

Graphical Abstract and Lay Summary

The consequences of tensile stretch and nanofiber topography on skeletal myocyte maturation are described in this graphical abstract. Myoblasts plated on a nanofiber scaffold were either cultured independently or co‐currently with motor neurons. The control scaffold condition was a randomly aligned nanofiber sheet. An applied stretch force alone did not change myofiber width, length, or fusion on random nanofiber scaffolds. Myocytes increased in length on aligned nanofiber scaffolds compared to random scaffolds.   However, the combination of nanofiber scaffold alignment and applied tensile stretch increased myofiber fusion in both innervated and non‐innervated myocytes.

1. Introduction

Skeletal muscle cells, known as myocytes, develop and reside in a poly‐cellular environment where they interact with the adjacent cellular milieu while also being influenced by other external factors. Developing myocytes (myoblasts) are influenced by extrinsic cues, including the presence of motor axons [1], cell‐to‐cell interactions [2], mechanical forces [3, 4], and matrix composition and topography [5], among other factors [6, 7]. Indeed, the physical microenvironment dynamically influences cell growth and differentiation [8]. Innervation, or physical connections from axons projecting from spinal motor neurons, promotes maturation and fusion of skeletal myocytes and prevents myofiber atrophy [9, 10]. Mechanical loading, particularly associated with passive tension, leads to skeletal muscle hypertrophy resulting in increased muscle volume [11]. Substrate topography is an environmental cue that influences myocyte adhesion, growth, and anisotropy. These factors, among others, collectively provide a dynamic and complex set of inputs that shape myocyte orientation, maturation, and function; however, potentially synergistic effects between these types of extrinsic cues remain poorly understood.

The broad field of muscle tissue engineering has three major objectives: (1) to develop implantable constructs to treat major muscle loss and muscle wasting diseases [12, 13, 14, 15], (2) to create biofidelic experimental platforms for high throughput in vitro testing [16, 17, 18, 19, 20, 21, 22], and (3) to grow consumable meat substitutes [23]. Various strategies have been employed to create tissue engineered muscle, including reseeding acellular muscle [24, 25, 26], growth on planar or three‐dimensional scaffolds [27, 28, 29, 30], application of mechanical forces [4, 6, 31, 32, 33, 34], and exploring pre‐innervation and/or pre‐vascularization [10, 35, 36, 37, 38, 39]. In particular, self‐aligned myofiber growth on nanofiber scaffolds has been demonstrated [13, 39, 40, 41, 42, 43, 44, 45]. Tissue engineering strategies aim to optimize tissue construction based on scaffolding techniques and various cell sources, such as skeletal myoblasts derived from primary murine, porcine, and human cells, as well as human induced pluripotent stem cells. Although, to date, the influences of most extrinsic cues have been studied independently, potential combinational effects have not been fully explored. Indeed, potential synergistic influences of innervation, mechanical conditioning, and scaffold alignment have yet to be investigated together in the context of tissue‐engineered muscle.

As such, the goal of the current study was to investigate the effects of mechanical forces and substrate topography on development and maturation of innervated and non‐innervated myocytes. To accomplish this goal, we cultured C2C12 skeletal myoblasts—either alone or in co‐culture with primary spinal motor neurons from embryonic rats—on aligned scaffolds comprised of poly‐caprolactone (PCL) nanofibers within custom mechanobioreactors capable of applying continuous, uniaxial tensile elongation over several days in vitro (Figure 1). Nanofiber alignment was varied to be either 0° (parallel) or 90° (perpendicular) from the tensile force vector and compared to randomly aligned nanofiber controls. We uncovered that the presence of innervation and substrate topography play significant roles in skeletal myofiber expansion and maturation, with secondary effects from applied mechanical tension in this paradigm.

FIGURE 1.

Mechanical stretch of neuromuscular constructs on nanofiber scaffolds. The experimental methods and techniques are outlined in this figure. (a) A schematic illustrating a co‐culture procedure with the different alignment strategies. C2C12 mouse myoblasts were plated on random and aligned nanofiber sheets attached to a towing membrane and coated with PDL and Laminin. The myoblasts were allowed to differentiate, and subsequently, rat spinal cord‐derived motor neurons were added on top of the myocytes. The towing membrane was gradually pulled by a stepper motor at a rate of 1 mm/day to initiate continuous unidirectional “stretch growth” over 5 days. This is not to scale. (b) Components of the custom‐built mechanical bioreactors to achieve controlled mechanical stretch. (c) Nanofiber sheets are shown in the red dotted square. These images show before and after a 5‐mm stretch operation of the box in (b). Scale is shown.

2. Methods

All procedures were approved by the Institutional Animal Care and Use Committees at the University of Pennsylvania and the Corporal Michael Crescenz Veterans Affairs Medical Center and adhered to the guidelines set forth in the NIH Public Health Service Policy on Humane Care and Use of Laboratory Animals (2015).

2.1. Mechanical Properties of Electrospun PCL Nanofiber Sheets

Young's modulus is a measure of material stiffness and is the first derivative of the linear‐elastic region from a stress–strain curve. Uniaxial tension was applied (Instron, Universal Testing System 5980 Series, 2013) and tested on (1) parallel PCL nanofiber alignment, (2) perpendicular PCL nanofiber alignment, (3) random PCL nanofiber alignment, and (4) native tibialis anterior (TA) muscle. The nanofiber sheets were dimensioned at 5 mm x 0.15 mm x 10 mm (width x depth x length) and stretched at 1.2 mm/min along the length. Native tibialis anterior muscle, with a cross section of 20 mm², was stretched in the transverse direction—parallel to myofibers—for physiological relevance and only within its linear elastic limits [46, 47]. The relative force measurements were recorded on software affiliated with Instron and generated a timescale dataset. Of note, while the length of the nanofiber sheets was similar to the length of the muscle tissue, the cross‐sectional areas were 0.75 and 20 mm², respectively; this difference in area was taken into account when calculating stress. A custom‐made MATLAB code subtracted the baseline value to calibrate and calculated the Young's modulus via a linear regression approximation on the resulting stress–strain curves.

2.2. Cell Isolation and Culture Techniques

We cultured commercially available C2C12 mouse skeletal myoblasts (ATCC CRL‐1772) and primary spinal motor neurons following previously described procedures [39].

2.3. Culturing C2C12 Mouse Skeletal Myoblasts

C2C12 myoblasts were cultured and maintained in growth media composed of 20% fetal bovine serum (Sigma, F0926‐100ML), 1% penicillin–streptomycin (ThermoFisher, 15070063), 79% Dulbecco's Modified Eagle Medium/Nutrient Mixture F12 (Invitrogen, 11330032). Upon reaching 80% confluency, myoblasts were passaged with 0.25% Trypsin‐EDTA (ThermoFisher, 25200056) and plated directly onto pre‐coated PCL nanofiber sheets (described below). After growing skeletal myoblasts for 2 days in vitro (DIV) in skeletal muscle growth media, cells were given differentiation media composed of 2% normal horse serum (Fisher, 16050130), 1% pen–strep, 97% DMEM‐F12 for 5 DIV and maintained in an incubator at 5% CO₂, 37°C, and 19.2% O₂ and sufficient humidity [39, 48, 49] for 5 days.

2.4. Spinal Motor Neurons

Spinal motor neurons were dissociated from embryonic spinal cords. Before dissociation, all tissue harvest was done on frozen granite block. Spinal cords were dissected from embryonic day 16 (E16) Sprague Dawley (Charles River) rats. The mother rat was euthanized with CO₂ asphyxiation followed by decapitation. Post decapitation and tail snipping of the embryo, the spinal cord was extracted from the dorsal side and placed in HBSS [Hanks’ Balanced Salt Solution, Invitrogen 14175079]). The meninges and dorsal root ganglion were stripped away. Cords were trypsinized with 2.5% 10X Trypsin (Gibco, 15090046) in 1 mL of Leibovitz's L‐15 solution (Life Technologies Inc., 11415064), subsequently triturated in DNAse (1 mg/mL Roche Diagnostics, 10104159001) and 4% bovine serum albumin (Sigma, A4161‐5G). After a 10 min centrifugation at 280 × g, the remaining pellet was resuspended, cushioned by OptiPrep (Abcam, ab286850), and centrifuged once again at 520 × g for 15 min at 4°C. The cells were extracted from the interface and resuspended in motor neuron plating media, which was previously conditioned for 6 h in the presence of E16 spinal cord astrocytes, sterile filtered through a 0.22‐µm filter, and growth factors added afterward as described [39, 49, 50].

2.5. Cell Culture on PCL Aligned Nanofiber Sheets

Custom‐made electrospun poly‐caprolactone (PCL) aligned and random nanofiber sheets (Nanofiber Solution LLC) were cut into 5 mm x 20 mm rectangles with the nanofiber alignment, if applicable, along the sheet's length. The sheets were positioned as illustrated in Figure 1a and described in the next sub‐section. After two 30‐min UV exposures, the sheets were coated with 20 µg/mL poly‐D‐lysine (PDL, Sigma, P7405‐5MG) in sterile cell culture water overnight. Following three 5‐min washes in PBS, the sheets were coated with 20 µg/mL mouse laminin (Corning, 354232) for 2 h. Pre‐differentiated C2C12 myoblast cells were plated on the nanofiber sheets at a concentration of 2.8  × 10⁵ cells/sheet for 10 min to settle and then flooded in growth media for 48 h before being cultured in differentiation media for 5 days with regular media changes (every 2 or 3 days). At 7 days in culture, dissociated spinal motor neurons were plated on top of the skeletal myocyte layer at a concentration of 1  × 10⁵ cells/sheet, and the coculture was maintained in serum‐free co‐culture motor neuron media [50] for 7 days with regular media changes. The sheets with non‐innervated skeletal myoblasts were also kept under serum‐free co‐culture motor neuron media [50] at 7–14 DIV to preserve consistency among groups.

2.6. Mechanobioreactor Configuration and Uniaxial Stretching of Cell‐Laden Nanofiber Sheet

Nanofiber sheets were arranged in custom‐built mechanobioreactors that enable the application of mechanical tension as shown in Figure 1a and adapted from our previous work [50]. The mechanobioreactor consists of a cell expansion chamber, linear motion table, stepper motor, and controller. The system propels a towing block within the cell expansion chamber in small stepwise movements while maintaining sterility (Figure 1b). The cell expansion chamber serves as a closed tissue culture environment covered with a fluorinated ethylene propylene membrane to permit gas exchange absent water vapor loss. The linear motion table supports the cell expansion chamber and the stepper motor, and all are housed within a dehumidified incubator set to 5% CO₂ and 37°C. The towing block moves linearly in the cell expansion chamber, and in the current study, the nanofiber sheets were adhered to both the towing block on one end and the container base on the other end using a silicone adhesive (NuSil, MED‐1037). The control boxes were maintained outside of the incubators and were programmed using a LabView (National Instruments) user interface. At 9 DIV, the stepper motor was engaged to drive the towing block at a rate of 1 mm/day for 5 days, with regular serum‐free co‐culture motor neuron media changes occurring over this timeframe (Figure 1a). The total strain displacement was 25% from the original sheet length (Figure 1c) to represent a developmental growth regime. Nanofiber alignment relative to the direction of tension was recorded and served as one of the metrics evaluated for this study.

2.7. Scanning Electron Microscopy

Samples were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (EMS 15960) at a pH 7.4 for 1 h and dehydrated before taken to the SEM core facility at the Singh Center for Nanotechnology (Quanta 600 FEG ESEM) and imaged. A low‐vacuum SEM was used and allows samples to be viewed without requiring a gold sputter‐coating; this avoids a buildup of charge at the sample surface [51].

2.8. Immunocytochemistry Procedure

At 14 DIV, the samples were fixed for 35 min in 4% paraformaldehyde (EMS 15710), washed 3 times with 1× DPBS (Dulbecco's Phosphate‐Buffered Saline, Gibco 14190‐136), and permeabilized in 0.3% Triton‐X100 (Sigma 9036‐19‐5) + 4% normal horse serum (NHS, Fisher 16050130) for 60 min. All subsequent antibody solutions were prepared in 4% NHS. Motor neurons were stained using neuronal marker anti‐beta tubulin III antibody (1:250, Abcam ab18027) or neurofilament 200 (1:200, Abcam ab8135) incubated overnight in 4°C to be subsequently paired with Donkey‐anti‐rabbit 568 fluorescent antibody (1:500, Fisher A10042) incubated at room temperature for 2 h. Actin was stained with Alexafluor‐488‐conjugated phalloidin (1:400, Invitrogen A12379) and the acetyl‐choline receptors/neuromuscular junctions were stained with AlexaFluor‐647‐conjugated bungarotoxin (1:250, Invitrogen B35450), both antibodies incubated for 2 h in room temperature. All samples were stained with Hoechst‐33342 (Invitrogen‐H370) as a DNA marker. Images were acquired using a Nikon Eclipse TI A1RSI laser scanning confocal microscope.

2.9. Data Analysis and Statistical Validation

Data was processed and analyzed using GraphPad PRISM software. Data were analyzed by one‐way analysis of variance (ANOVA) testing between nanofiber alignment and stretch conditions, followed by Tukey's post‐hoc test for multiple mean comparison, while Brown–Forsythe and Welch ANOVA tests followed by Dunnett's T3 multiple comparisons tests were performed to account for variable variances between groups. Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), and p ≤ 0.001 (***). Data displayed as mean ± standard error of mean; Histograms were generated in MATLAB utilizing the entire dataset of individual cell values across at least 3 regions of interest in 5 stretch and topography groups and between 2 cell culture groups. These data were compared using the Kolmogorov–Smirnov distribution test, considering random no stretch as control group.

2.10. Data Collection Techniques

The length was measured by finding the displacement between the furthest points on a myocyte/myofiber. The width was determined by the thickest section in the myofiber. Myocyte angle orientation was measured with respect to the direction of the nanofiber alignment; if the sheet had a random nanofiber alignment, the reference angle was established as an arbitrarily assigned axis. The fusion index was gathered with the following rationale: myofibrils with 3 or more nuclei were tallied up and measured.

$$\begin{array}{ccc}& & \mathrm{Myocyte}\mathrm{Fusion}\mathrm{Index}\hfill \\ & & =\frac{\mathrm{Number}\mathrm{of}\mathrm{nuclei}\mathrm{in}\mathrm{myofibers}\mathrm{with}\mathrm{more}\mathrm{than}3\mathrm{nuclei}}{\mathrm{total}\mathrm{number}\mathrm{of}\mathrm{nuclei}\mathrm{present}\mathrm{in}\mathrm{myocytes}}\hfill \end{array}$$

The overlap of the Hoechst staining with Phalloidin confirmed the nuclei in each myocyte. The percentage of area covered by myocytes was measured using ImageJ's binary image, edge detection, and particle analysis capabilities. Separating fluorescence from the dark background amounts to a proportion of the percent area covered by cells on the nanofiber sheet.

3. Results

3.1. Mechanical Properties of PCL Nanofiber Sheets Compared to Native TA Muscle

Substrate physical properties have been shown to influence cellular growth and myocyte maturation responses [3, 52, 53, 54]. Accordingly, macro mechanical properties of the PCL nanofiber sheets and native TA muscle tissue were measured via tensile Instron testing to understand the substrate's mechanical properties relative to native muscle (Figure 2a). A stress–strain curve showed that native rat tissue had a Young's modulus of 0.0314 ± 0.027 MPa in its elastic region (Figure 2b,c), in agreement with previously reported range of values [55]. Also, it was previously shown that nanofiber scaffolds stretched along the direction of alignment had greater tensile strength (intra‐fiber forces within fibers) than fibers stretched orthogonal to the direction of alignment (inter‐fiber forces between fibers) [56, 57]. To consider how similar directionality affects the mechanical properties of the nanofiber sheets used in our study, the tensile properties of the PCL sheets were measured for orientations perpendicular, parallel, or random relative to the direction of alignment. Stress–strain curves revealed that the Young's modulus of nanofiber sheets was 0.733 ± 0.294 MPa with the alignment perpendicular to tension compared to 23.520 ± 0.733 MPa with the alignment parallel to tension (Figure 2b,c). The Young's modulus for nanofiber sheets with random alignment was measured to be 7.665 ± 0.690 MPa (Figure 2b,c). A one‐way ANOVA followed by Tukey's pairwise comparisons revealed that the Young's modulus for PCL nanofiber sheets either aligned parallel or randomly to the direction of the applied tension was significantly greater than that of native TA muscle (each p < 0.0001). In contrast, the Young's modulus for the PCL nanofiber sheets aligned perpendicular to the applied tension did not differ statistically from that of native TA muscle tissue (p > 0.05); however, given the ∼23‐fold increase in Young's modulus for the perpendicularly aligned sheets versus TA muscle, a more expansive testing regime may be warranted.

FIGURE 2.

Tensile testing offers information on stiffness and the stress–strain relationship of poly‐caprolactone nanofiber sheets (2 cm × 0.5 cm × 150 µm) compared to native tissue. (a) Representative photographs of nanofiber sheet (left image) and rat tibialis anterior (TA) muscle (right image) during mechanical testing on an Instron. (b) Bar graph showing Young's modulus defined as the change in stress over the change in strain (note this is the slope of curves in [c]) of each group, n = 3 samples per group. Data were analyzed using a one‐way ANOVA followed by Tukey's post‐hoc test for multiple comparisons. Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). Significance was compared to the native TA muscle. While random and parallel were statistically different than TA, under these experimental conditions, there was no statistical significance between perpendicular and TA. (c) Stress–strain curves for each nanofiber topographical and alignment group compared to native TA (standard deviation depicted by colored area surrounding each curve). The bottom panel is an enlarged view showing stress–strain curves for nanofiber aligned perpendicular to the direction of stretch and the native TA tissue.

3.2. Ultrastructural Analyses to Observe the Influence of Scaffold Topography on Cell Morphology

Scanning electron microscopy (SEM) was employed to observe the nanofiber surface topography and effects of applied tension. This nanoscale surface visualization technique confirmed the PCL nanofiber alignment (Figure 3) and the presence of cell bodies at the substrate's surface. In all cases, cells were grown for a total of 14 days in vitro (DIV) on the nanofiber sheets, with mechanical tension applied to the sheets for the final 5 DIV. Sheets with random nanofiber alignment displayed an unorganized fiber arrangement (acellular group, Figure 3 top panel) with cells webbed in between the pores (cellular group, Figure 3 lower panel). Nanofibers with tension applied perpendicular to their alignment were found in a wave‐like configuration and had similar cellular adhesion patterns as observed on the randomly aligned nanofiber sheets. Stretching parallel to the PCL nanofiber alignment led to homogeneous myofiber direction; therefore, a more uniform nanofiber alignment was observed both in the acellular and cellular parallel stretched conditions. Together, our SEM ultrastructural analysis suggests that the nanofiber topography influences myofiber orientation.

FIGURE 3.

The microarchitecture of nanofiber scaffolds with and without cells depicted with scanning electron microscopy. The nanoscopic topography of the scaffold sheets and the cellular adherence to the surface can be observed qualitatively. The two‐headed arrow indicates the direction of uniform tension along an axis. Acellular and cellular examples are shown, as well as the nanofiber direction on the sheet. Scale bar: 100 µm.

3.3. Combined Role of Surface Topography and Mechanical Stretch on Muscle Growth and Maturation

Nanofiber alignment and stretch directionality were manipulated to understand the effects of these extrinsic cues in influencing cell growth, orientation, and maturation. Combining topographical nanofiber alignment with tensile stretch generated from the mechanobioreactor generated a model system to measure myocyte growth relative to nanofiber orientation and direction of stretch. At 14 DIV (the final 5 of which had applied mechanical tension), monocultures of C2C12 skeletal myocytes were fixed and immunolabelled with phalloidin (actin) and Hoechst (nuclei), imaged using confocal microscopy, and structural features were then quantified (Figure 4a). With unequal standard deviations, under a Brown–Forsythe and Welch correction for a one‐way ANOVA, myocyte width did not show a statistical significance between groups but there was a trend for greater width in the aligned, non‐stretched group (p = 0.18) (Figure 4b). However, growth on aligned nanofibers resulted in longer myocytes compared to myocytes grown on randomly oriented nanofiber sheets as shown by an ordinary one‐way ANOVA (Figure 4c). Within the stretched nanofiber sheets, aligned nanofiber sheets showed a significant difference in myocyte length compared to the random nanofiber alignment (Figure 4c). Within the aligned nanofiber groups, the application of stretch did not play a significant role in myocyte length. Not surprisingly, the degree of variability in myocyte angle orientation with respect to the nanofiber direction reflected a dependency on surface topography with randomly aligned nanofibers displaying higher variability than aligned nanofibers (Figure 4d). Moreover, the myocytes cultured on aligned nanofiber sheets stretched perpendicular to the stretch direction were statistically more variable to the angle orientation of myocytes cultured on nanofiber sheets running parallel to the stretch direction, suggesting a potential role of stretch on myocyte angle orientation. The surface area covered by myocytes was significantly increased on sheets with nanofiber alignment parallel to the direction of stretch and in general trended to be greater on aligned nanofiber sheets (Figure 4e). Myocyte fusion was greatest among the myocyte monocultures on aligned nanofiber substrates and under tension, most predominantly parallel stretch to the nanofiber alignment direction compared to the randomly aligned nanofiber groups (p < 0.05) (Figure 4f).

FIGURE 4.

Structural parameters of skeletal myocytes grown on nanofiber scaffolds with and without continuous mechanical tension. (a) Representative fluorescent confocal images of samples labeled using immunocytochemistry for the presence of actin in skeletal muscle (phalloidin, green). Scale bar: 200 µm. In cultures with only C2C12 skeletal myoblasts after a total 14 days in vitro (with the final 5 DIV under continuous tension), myocytes’ physical characteristics were quantified, including (b) the average thickness or diameter of myocytes, (c) the average length of myocytes, (d) the variation of myocyte direction, (e) the percent area covered by myocytes, (f) and the myocyte fusion index. Data were analyzed using a one‐way ANOVA followed by Tukey's post‐hoc test for multiple comparisons. Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). Significance was compared to random stretch (+) and significance compared to random, no stretch (#).

In order to investigate the combined dimensional myocyte measures (i.e., length, width, and angle orientation), the population statistics across all the replicates showed distributional differences between the topographical and stretch parameters (Figure 5a). Distributional differences were determined using Kolmogorov–Smirnov tests to compare each dataset to the no stretch, random nanofiber control group. The frequency distribution of myocyte widths (Figure 5b) and lengths (Figure 5c) in the random nanofiber groups was much tighter than the aligned groups. On the contrary, angle orientation (Figure 5d) followed a reverse pattern; there was a greater spread of myocyte angle orientation in the random alignment and a tighter standard deviation in the uniformly aligned nanofibers.

FIGURE 5.

The distribution of myocyte dimensions and orientation among scaffold nanofiber direction and presence of stretch. The data shown in Figure 4a–c are represented as nested scatter plots (a), illustrating the spread and average of myocyte width, length, and angle direction from each trial. The combined populations of myocyte (b) width, (c) length, and (d) angle under tensile and surface topography conditions are shown in the remaining histograms as notated. Data were analyzed using a two‐sample Kolmogorov–Smirnov test comparing each group to the random, no stretch negative control group. Significance was taken at p ≤ 0.05 (#), p ≤ 0.01 (##), p ≤ 0.001 (###), and p ≤ 0.0001 (####).

Overall, there were significant differences in myocyte growth morphology between aligned and random nanofiber sheets. Compared to the random nanofiber groups, the myocytes grown on parallel stretched aligned nanofibers displayed greater fusion (p < 0.05), longer length (p < 0.001), and higher surface coverage (p < 0.05), indicating enhanced myocyte growth and maturation. Further, myocytes on aligned, no stretch scaffolds displayed less fusion than myocytes on aligned, parallel stretch scaffolds. Notably, the aligned, parallel stretch scaffolds produced the greatest extent of myocyte fusion, which may indicate maturation more reliably than myocyte size [58].

3.4. Combined Role of Surface Topography and Mechanical Stretch on Innervated Muscle Growth and Maturation

After showing that surface topography and mechanical stretch had morphological effects on skeletal myocytes in a monoculture, we next sought to test the effects of these parameters in myocyte‐neuronal co‐cultures. At 14 DIV (the final 5 of which had applied mechanical tension), co‐cultures of C2C12 mouse skeletal myocytes and dissociated embryonic rat spinal cord motor neurons were observed to express phenotype specific markers—phalloidin (actin) and β‐tubulin III (axons) (Figure 6a). Bungarotoxin—positive junctions were observed in this co‐culture system (Figure S1). Under a one‐way ANOVA, innervated myocytes were wider when cultured on random nanofibers than on aligned and stretched nanofiber sheets (Figure 6b). Parallel stretch, albeit only significant relative to the random oriented groups, led to thinner myofibers. Myofiber elongation was relatively consistent with the presence of motor neurons regardless of surface morphology or tensile stretch (Figure 6c). The angle spread of the innervated myocytes was greatest in the random nanofiber groups and there was no statistical difference between the three aligned nanofiber groups (Figure 6d). Provided that the distributions were unequal, a Brown–Forsythe and Welch correction for a one‐way ANOVA indicated that innervated myoblast cell density was lower on the nanofiber sheets aligned perpendicular to the direction of stretch, with notable amounts of unoccupied space between the cells (Figure 6e). However, innervated myocyte fusion was greatest when the nanofibers on the scaffolding were aligned perpendicular to the direction of tensile stretch (Figure 6f). Commensurate with changes in area covered by myocytes across the various experimental conditions, we also found trends toward increased cell density relative to the initial plating density following substrate stretch in both innervated and non‐innervated culture conditions (Figure S2).

FIGURE 6.

Structural parameters of innervated skeletal myocytes grown on nanofiber scaffolds with and without continuous mechanical tension. (a) Representative fluorescent confocal images of samples stained for the presence of actin in skeletal muscle (phalloidin, green) and the presence of neurons (β‐tubulin III, red). Scale bar: 200 µm. In a co‐culture, innervated C2C12 skeletal myoblasts behave differently than in a monoculture after 14 DIV. Physical measurements of innervated myocytes were taken, including (b) the average myocyte width, (c) the average myocyte length, (d) the angle spread of myocyte growth, (e) the percentage of area covered by myocytes, and (f) the myocyte fusion index of the formed myofibrils. Data were analyzed using a one‐way ANOVA followed by Tukey's post‐hoc test for multiple comparisons in (b) and (c) or by Browne–Forsythe and Welch corrected one‐way ANOVA followed by Dunnett's T3 multiple comparisons in (d), (e), (f). Significance was taken at p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). Significance was compared to random stretch (+) and significance compared to random, no stretch (#).

Distributional differences in the aggregated data measures were similarly determined using Kolmogorov–Smirnov tests to compare each innervated dataset to the no stretch, random nanofiber control group (Figure 7). The aligned stretched groups had a tighter distribution of innervated myocyte widths than the no stretch and random nanofiber scaffold groups (Figure 7b). Innervated myocyte length was only significantly different in the aligned, no stretch group skewing toward shorter myocytes (Figure 7c). The innervated myocyte angle orientation distribution was bimodal in the random nanofiber no‐stretch group and uniform in the random nanofiber stretch group (Figure 7d). Appropriately, the nanofiber sheets with perpendicular nanofiber alignment to the stretch direction had a wider distribution of innervated myocyte angle orientation than the sheets with parallel nanofiber alignment to the stretch direction. The nanofiber sheets with uniform alignment but with no stretch showed a wide unimodal angle distribution, therefore differing from the random nanofiber controls. Not surprisingly, neurons and neurites were also found to be preferentially orientated based on nanofiber alignment (Figure S3).

FIGURE 7.

The distribution of innervated myocyte dimensions and orientation among scaffold nanofiber direction and presence of stretch. The data shown in Figure 5a–c are represented as nested scatter plots (a), showing the spread and average of innervated myocyte width, length, and angle direction. The combined populations of innervated myocyte (b) width, (c) length, and (d) angle under tensile and surface topography conditions are shown in the remaining histograms as notated. A two‐sample Kolmogorov–Smirnov test compares each group to the random, no stretch control group. Significance was taken at p ≤ 0.05 (#), p ≤ 0.01 (##), p ≤ 0.001 (###), and p ≤ 0.0001 (####).

4. Discussion

The microenvironment of developing skeletal muscle is a complex niche providing multi‐faceted extrinsic cues that likely elicit combinatorial effects on cell growth and maturation. These extrinsic cues individually affect different mechanisms for muscle growth. However, we postulated that skeletal muscle likely responds differently to concurrently applied cues; moreover, these responses may vary based on developmental timing. Our group has demonstrated that pre‐innervated tissue engineered muscle exhibited more robust myofiber and neuromuscular junction formation, among other hallmark features of maturation [39]. Likewise, tensile loading and mechanical stretch have been shown to enhance myofiber formation on electrospun nanofibers [6]. Previous studies have also revealed that minute physical stresses developed on skeletal myocytes, such as those induced by exercise, increases the fiber width and bulkiness [59].

To better understand these phenomena, the current study assessed for the first time the combined influence of mechanical loading and substrate topography with or without innervation (e.g., structural and soluble effects of motor neurons/axons) on the maturation of C2C12 skeletal myoblasts. We demonstrated that skeletal myocytes morphologically mature with the presence of (1) aligned nanofiber scaffolding and (2) motor neurons compared to their mono‐cultured counterparts. In non‐innervated cultures, we found that tensile stretch reduced myofiber density on scaffolds with random nanofiber alignment, while there were no other notable differences in myofiber morphology between the stretched and non‐stretched groups. We also found that uniformly aligned nanofiber topography, regardless of tensile stretch, induced morphological indicators of skeletal myocyte maturation without motor neurons. Notably, this aligned nanofiber topography may be more biologically relevant than the randomly oriented counterparts. In innervated cultures, we found that the combination of nanofiber alignment and tensile stretch increased skeletal myocyte fusion. Our results establish an in vitro paradigm that may be useful for further parameterized investigation of multiple factors simultaneously influencing skeletal muscle development and maturation.

Stress–strain curves, as shown in Figure 2, indicated that the PCL nanofiber sheets behave elastically at small displacements. Aligned nanofibers are anisotropic, as are the naturally occurring fasciculation and striations found in myofiber. Not surprisingly, we found that the mechanical properties of PCL nanofiber sheets varied greatly based on directionality of nanofiber alignment relative to the applied tension field; in particular, the Young's modulus for nanofiber sheets with fibers oriented parallel, randomly, and perpendicular to the direction of applied tension were increased relative to native TA muscle by 749‐fold, 244‐fold, and 23‐fold, respectively. Explaining this finding, in the parallel direction, the tensile force is determined by the intra‐fiber forces within the PCL nanofiber sheet, whereas in the perpendicular direction, the tensile force is determined by inter‐fiber forces across the PCL nanofiber sheet. PCL nanofibers will therefore have a greater Young's modulus, or stiffness, when stretched along the fiber direction due to the intra‐fiber forces than when stretched perpendicular to fiber direction due to inter‐fiber forces. This is supported by the results in Figure 2b,c and is in accordance with previously reported studies [56, 57]. As the mechanical environment (Young's modulus) was different across the experimental groups and tended to vary compared to that of native tissue, it is unclear if the findings from stiffer PCL nanofibers can be related to actual cellular behaviors in native tissues that are significantly softer. Indeed, additional studies are necessary to better understand how different stiffnesses may play a role in translating the observations on PCL nanofibers to the actual cellular behavior in soft tissues. However, of note, any potential PCL degradation and consequent mechanical property alterations over days in culture may make the PCL nanofiber sheets more similar to native muscle and therefore should be considered in future studies. Also, the sample lengths for the PCL nanofibers in the bioreactors (2 cm) was longer than the length used for mechanical testing using the Instron (1 cm); while Young's Modulus is a fundamental material property and should not be affected by these differential lengths, this could be empirically shown in future studies. Overall, further scaffold optimization to better match the mechanical properties of mature and/or developing muscle may lead to more representative growth and maturation of innervated skeletal muscle in vitro.

PCL nanofiber scaffolding supported C2C12 cell survival, demonstrated in Figure 3, most readily on the aligned nanofibers compared to random nanofibers. The topography of aligned nanofibers enables myofibers to extend along their natural axes whereas the random nanofibers disrupt that extension. SEM images confirmed the microarchitecture of the nanofiber sheets. The presence of aligned scaffolding had an overwhelming morphological and directional effect on skeletal myofiber multinucleation and formation as seen in Figures 4 and 6. Nanofiber scaffold alignment provided an intrinsic framework for muscle cells to fuse and multinucleate along, taking a more dominant role on myofiber configuration during in vitro growth [39]. A random nanofiber alignment led to unaligned myofibers which formed clusters and did not readily fuse into longer multinucleated fibers. The random surface architecture alone had an attenuating effect on myocyte growth as noted by all the metrics in the non‐innervated groups (Figure 4). Innervation appeared to override the topographical setbacks in the random aligned groups, particularly demonstrated in innervated myocyte length and width (Figure 6). However, although the cells might be longer and thicker, they did not multinucleate and fuse as readily. Mature myofibers typically extend between 200 and 1000 µm [54, 60], however the myofiber length in this study only extended less than 150 µm, possibly indicating limited growth [3]. Similarly, the width of the myofibers did not reach a conventional matured size, regardless of the stretch or surface topographic parameters (Figure 4b). Additional immunostaining that targets proteins in mature myofibers, such as myosin heavy chain, α‐actinin, titin, myomesin, and underlying transcriptomic changes via RT‐qPCR as well as functional measurements would offer a more comprehensive investigation of myotube characterization and maturating to be pursued in future studies.

The distribution of the myocyte length, width, and angle showed the diversity and range of myocyte growth (Figures 5 and 7). All groups were compared to the random, no stretch condition. Notably, the collective measurements of myocytes grown in all the aligned groups were significantly different from the random no stretch group across all measured quantities, indicating that scaffold alignment plays a role in muscle growth (Figure 5). Innervated myocytes on aligned nanofiber surfaces also behaved differently than those grown on random nanofiber surfaces (Figure 7). As expected, nanofiber alignment influenced myocyte angle orientation in both innervated and non‐innervated myocyte groups. Interestingly, the distribution of innervated myocyte angle orientation was much wider in the aligned, no stretch group compared to the other aligned groups (Figure 7d) and in comparison to the non‐innervated group (Figure 5d). Myoblasts ultimately self‐aligned and mature myofibers lengthened and bundled together, forming striations along a uniform direction. The aligned nanofiber sheets guided and promoted this myocyte alignment. The nanofiber scaffold alignment changed the distribution of myocyte morphology and orientation (Figure 5). Intuitively, the random nanofiber surface oriented the cells in scattered directions whereas the aligned nanofiber surface oriented the cells more homogeneously in a uniform direction. Innervation played an important role in myocyte morphology in conjunction with nanofiber scaffold alignment (Figure 7). Our findings suggest that under the mechanical parameters used in this study, continuous tensile stretch played a less significant role in the distribution of morphological measurements.

Skeletal myofibers experience physical loading via applied external forces and other extrinsic cues to maintain structure and to guide their formation in vivo, motivating the applied forces used in in vitro models. Tensile strain and stiffness of longitudinal skeletal muscle varies [61] with different animal types such as pig [46, 62], rabbit [63], and rat [47]. Myofibers are unique in that they are force generating cells; however, as a key component of their growth and maturation, they are also subjected to external loads. For instance, myofibers undergo cyclic tensile loading—with or without muscle lengthening—via activation of antagonistic muscle groups [64, 65]. During embryogenesis through adolescence, myofibers are subjected to continuous gradual tensile forces due to bone growth (e.g., limb and spinal column lengthening). It is believed that these different loading regimes are each important to promote myofiber growth [31]. Applying tension or a low impact stretch, whether continuous or cyclic, to an engineered muscle scaffolding may lead to thicker myofibers as the microenvironment may replicate facets of development and function, reflecting post‐natal muscle growth along relatively fast‐growing bone [58]. Uniaxial stretch, or tensile loading, is known to have an aligning and higher maturation effect on skeletal myocytes in vitro [3, 31, 32, 34, 66, 67, 68]. While the current study was limited in that we only applied uniaxial static stretching to skeletal myocytes, it would be worthwhile for future studies to employ multi‐axial dynamic loading representative of actuation in vivo.

A goal of the current study was to demonstrate a biomechanical loading paradigm that affected muscle growth. We expected that muscle cells would grow more when exposed to stretch and be able to tolerate and adapt to such indirect forces. The strain applied in these studies has both a developmental precedent and is within the range of strain regimes evaluated in similar studies, which have typically reported that stretch stimulates proliferation and facilitates myogenesis [34, 69, 70, 71, 72]. Indeed, during development nerves are “stretch‐grown” at rates at or exceeding 1 mm/day, and we have used similar strain regimes to replicate so‐called axonal “stretch‐growth” in our custom mechanobioreactor systems [50, 73, 74, 75]. On the contrary, the results of this study suggest that uniaxial tensile stretch at 25% elongation has a minimizing effect on skeletal muscle width when cultured on aligned nanofiber sheets and no additional effect on myocyte length on both innervated and non‐innervated myocytes (Figures 4b,c and 6b,c). However, myocyte fusion increases on stretched aligned nanofiber sheets in both innervated and non‐innervated conditions (Figures 4f and 6f). Some sources indicate that stretch at specified cyclic frequencies and conditions may have inhibitory or downregulating effects [33, 69, 76, 77]. It has been shown that at 20% elongation over 24 h, cyclic mechanical stretch is a down‐regulatory signal for C2C12 myoblast differentiation [78]. While myoblast determination protein 1 (MyoD) and myocyte nuclear factor (MNF‐α) were both down‐regulated in the stretch condition [78], these factors are typically upregulated upon differentiation, particularly in the presence of chronic nerve stimulation or innervation [79].

Innervation plays a central role during postnatal skeletal muscle development. Spinal motor neurons activate skeletal muscle and enable both coarse and fine motor control. Muscle fibers differentiate independent of innervation at the first embryonic stages of development [1, 80]. However, bulked muscle formation requires nervous system input for proper development and functionality [1, 81]. Adding motor neurons to differentiated myocytes in culture allows neuronal trophic factors to guide muscle maturation [1], recapitulating developmental neuromuscular interaction. The effects of innervation in this experiment were indicated by the presence of bungarotoxin‐positive clusters in our cultures, indicating the presence of nascent neuromuscular junctions (Figure S1).

The microenvironment for cellular growth also provides critical extrinsic cues to cell maturation, elongation, and overall health. Surface geometries [39, 41, 82], small molecules [7, 12, 83], and chemical coatings [84, 85] are only a few examples of substrates that influence cellular processes. Cell adhesion and planar motility depend on the presence of chemical substrates, which provide ligands for receptor attachment and extracellular matrix for the larger cell attachment [86]. Mature skeletal myofibers align with the support of anisotropic forces that can be expressed through micropatterning of aligned grooved surfaces. The combined overall effects of cellular adhesives on electrospun micropatterned scaffolds lead to greater myocyte maturation and individual cell elongation [28, 82, 87].

Topographical and spatial factors can play a role in catalyzing metabolic pathways, can serve as a microarchitecture for cell adhesion, migration, and elongation, and can change the protein or phenotypic expression of a cell population. Nanofiber alignment facilitates anisotropic forces to enable myofiber self‐assembly and uniform alignment. Previous literature has shown muscle cells align orthogonal to the direction of stretch when plated on flexible‐bottomed culture plates coated with collagen‐I [65]. Myocytes conform into a collective direction based on a nanofiber alignment. Mimicking the mechanical and topographical cues present in endogenous myofibers in vivo can recapitulate features of the native microtissue environment. Our study exhibited that under non‐stretch conditions, aligned nanofibers facilitated an increased myocyte width and length in monocultures but did not affect these parameters in motor neuron/skeletal myocyte co‐cultures (Figures 4b,c and 6b,c). However, alignment supplemented with a stretch condition led to greater myocyte fusion in both innervated and non‐innervated skeletal myocyte cultures (Figures 4f and 6f). These results reinforce that topographical factors and stretch can collectively contribute to skeletal myocyte growth and fusion.

5. Conclusion

We investigated the combined effects of mechanical forces from tensile stretch and scaffold topography on the morphological maturation of both innervated and non‐innervated C2C12 murine skeletal myoblasts. Aligned nanofiber scaffolding, resembling the morphology of in vivo musculoskeletal environments, played a significant role in skeletal myocyte thickness, width, and area coverage. We found that surface topography and tensile forces were two isolated environmental factors that play a role in myocyte growth in vitro, in both non‐innervated and innervated myocytes. Our results suggest that topographical guidance had a greater effect on myocyte growth and maturation compared to tensile loading, at least for the particular loading paradigm used in this study. Also, we found that innervation had a critical effect on skeletal myocyte growth and maturation, in conjunction with the effects of scaffold topography and mechanical tension applied to the scaffold. The combination of applying tensile loading and aligned nanofiber topography to innervated muscle mimics key features of the in vivo developmental microenvironment and thus may build toward a functionalized regenerative strategy for repair of major muscle defects. Further investigation may advance this platform to assess the effects of other extrinsic parameters on muscle growth and maturation.

Author Contributions

Melanie C. Hilman, Suradip Das, D. Kacy Cullen conceived of the approach. Melanie C. Hilman and Suradip Das designed experiments with input from Foteini Mourkioti and D. Kacy Cullen. Melanie C. Hilman carried out experiments and analyzed data. Melanie C. Hilman wrote the first draft of the manuscript, with input from Suradip Das, Foteini Mourkioti, and D. Kacy Cullen.

Ethics Statement

These studies were done in accordance with approvals from the Institutional Animal Care and Use Committee (IACUC) at the University of Pennsylvania and the CMC VA Medical Center in Philadelphia.

Conflicts of Interest

D.K.C. is a co‐founder of Axonova Medical, Inc., and Innervace, Inc., which are University of Pennsylvania spin‐out companies focused on translation of advanced regenerative therapies to treat nervous system disorders. US Patent Application 17/291,869 and PCT International Patent Application PCT/US2019/060585 titled “Engineering of Innervated Tissue and Modulation of Peripheral Organ Activity” (D.K.C. and S.D.) are related to the techniques described in this paper. The other authors declare that they have no competing interests.

Supporting information

Supporting file 1: biot70047‐sup‐0001‐FigureS1‐S3.docx

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.