A simple and scalable 3D printing methodology for generating aligned and extended human and murine skeletal muscle tissues

As basis for a controlled approach for directing the self-assembly of myotubes from myoblasts, we formulated two sets of gelatin inks, one of physiological and one of sub-physiological stiffness, see figures 2(a) and (b). For the first, we employed high-bloom (HB) gelatin at 7.5% (w/v) with and without 0.25% (w/v) xanthan gum (XG) as rheology modifier. These inks undergo sol-gel transition near room temperature at ∼24 °C, and so, to avoid clogging, we printed these using a slightly heated reservoir set at 32 °C, where they behave as shear-thinning liquids. For soft domains we employed, low-bloom (LB) gelatin at 4% (w/v) with and without 0.25% (w/v) XG. These inks must be cooled slightly to gel, and can therefore be extruded using room temperature reservoir. For the lower concentration, LB gelatin inks, XG prove highly beneficial for increasing viscosity and obtaining well-defined prints. Thus, when printing line-patterns of 4% LB gelatin onto 7.5% HB gelatin, the apparent line width decreases from ∼225 to ∼175 µm when adding 0.25% XG. Parallel lines can consequently be spaced more closely before they start to merge, see figure 2(c). Notably, the height of these LB gelatin lines are in the order of ∼20 µm, see figure 2(d). Hence, they do not constitute an impassable barrier. Following enzymatic crosslinking using microbial transglutaminase (mTG), the 7.5% HB gelatin gels had an apparent Young's modulus of ∼15 kPa, in the range of the physiological stiffness of skeletal muscle [11], while the stiffness of both 4% LB gelatin gels were well below 1 kPa, see figure 2(e).

Zoom In Zoom Out Reset image size

We next compared the degree of alignment of murine C2C12 myotubes generated from myoblasts on isotropic and patterned substrates. On uniform, planar substrates composed of either 7.5% HB gelatin or 4% LB gelatin + 0.25% XG, the degree of alignment matched that found in conventional tissue culture polystyrene (TC-PS) dishes. However, when introducing simple line patterns spaced by 400 µm onto printed 7.5% HB gelatin substrates, the degree of myotube alignment increased dramatically, see figure 2(f). Notably, while all line-patterned substrates displayed an increased alignment, we observed the highest numerical degree of alignment for line-patterns of 4% LB gelatin + 0.25% XG on 7.5% HB gelatin where the mechanical contrast between the regions is largest.

We next sought to identify the minimal spatial density of cues required for generating predictable myotubes architectures, when applying gelatin inks of contrasting stiffness. Using a conventional non-tapered 70 µm (ID) metal nozzle, we printed 4% LB gelatin + 0.25% XG lines at increasing spacing on printed 7.5% HB gelatin substrates and evaluated the degree of alignment of C2C12 myotubes developed on the substrates, see figure 3(a). For a spacing of 400 and 300 µm we observed a similarly high degree of alignment, which slightly decreased when the spacing was increased to 600 µm. When increasing the nominal line spacing to 800 and 1200 µm we observed a notable decrease in the degree of alignment, see figure 3(b). Interestingly, the primary angle of orientation appeared fixed at ∼36° from that of the printed lines for the aligned samples, and only changed, as global alignment was lost see figures 3(c) and 4(a). This angle of orientation was consistent across all samples, batches and all areas of the samples. To explore the consistency these largely self-organized architectures, we printed concentric circles of 4% LB gelatin + 0.25% XG on 7.5% HB gelatin, see figure 4(b). After myoblast seeding and maturation these pattern gave rise to a spiraling architecture of myotubes with almost perfect regularity, and locally consistent with a shift of ∼36° between the orientation of the LB gelatin lines and the myotubes orientation. This illustrates that within some design limitations, our method allows for creation of non-linear architectures with high degree of predictability.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Achieving a high degree of myotube alignment is an important aspect of replicating the native in vivo architecture of skeletal muscle tissue. Still, it does not necessarily result in myotubes of improved physiological relevance, as reflected in the length, sarcomerogenesis and contractility. To investigate whether the increased organization observed on the printed substrates also had a positive influence on myotube length, we did an unbiased quantification of myotube length for C2C12 myotubes matured for 1 week on TC-PS, plain gelatin or gelatin line patterns composed of 4% LB gelatin + 0.25% XG on 7.5% HB gelatin, see figures 5(a) and (b). Doing so, we observed almost a doubling in mean myotube length from 476 ± 89 to 888 ± 164 µm, when comparing TC-PS to line-patterned gelatin, while on plain gelatin the mean myotube length was found to be 564 ± 68 µm, see figure 5(b). Similarly, while >30% of the myotubes on TC-PS, and >27.5% of the myotubes on plain gelatin had a length of 300 µm or less, only ∼10% of the myotubes developed on patterned gelatin substrates fell in this range, see figure 5(a). The average myotube width appeared smaller on both gel types as compared to TC-PS, though the differences was not statistically significant, see figure 5(c). The degree of myotube fusion as indicated by the fusion index and coverage of sarcomeric contractile proteins, were both the largest at ∼50% for the line-pattern substrates, though this could be artifact of incomplete staining and differences from plain gelatin and TC-PS were not statistically significant, see figures 5(d) and (e). In addition to an impact on myotube length, the compliant mechanics of gelatin hydrogel substrates enables longer culture times than rigid TC-PS, where we observed typically delamination after 1 week of differentiation. Increasing the maturation time to 3 weeks enables myotubes with a more mature contractile apparatus, as indicated by more clearly defined sarcomeres at a higher density, see figures 5(f) and (g). To verify the contractile function of the myotubes we applied a commercial electrical pacing system based on carbon electrode inserts. Electrical pacing can accelerate the maturation of myotube tissues, and the electrically induced contraction further serves as an indication of myotube fusion and sarcomerogenesis [5, 12, 28]. We found that pacing induced immediate contraction of the myotubes and verified that the myotubes could be paced for at least 1 week, see supplementary videos 1–3 (available online at stacks.iop.org/BMM/17/045013/mmedia). Lastly, to demonstrate that the printing procedure can be applied on a range of substrates we generated patterned gel coatings onto commercial multi-electrode systems for recording electrical signaling in tissue cultures. Doing so, we observed that the contraction-associated depolarization of myotubes was highly synchronized, further validating myotube fusion, see supplementary figure 1.

Zoom In Zoom Out Reset image size

While murine C2C12 myotubes can be useful as low-cost models, engineered human tissues naturally has more relevance for biomedical research. We therefore investigated whether the patterned gelatin would also affect the development of human myotubes. We thus seeded and matured primary human myoblasts into myotubes on three types of substrates: printed, line-patterned gelatin, isotropic gelatin and TC-PS substrates, see figure 6. Mirroring our findings for C2C12 myotubes, we observed a significantly larger degree of macroscopic orientation on the patterned substrates for human myotubes. We further wondered whether as similar self-organization would take place for other classes of engineered muscle tissues, not least cardiac. However, when seeding mouse primary neonatal ventricular myocytes at 150-, 200-, and 300.000 cells cm⁻² on the patterned substrates with and without Matrigel coating, we did not observe any indications of macroscopic organization, see supplementary figure 3.

Zoom In Zoom Out Reset image size

HB gelatin (G2500, Sigma-Aldrich), LB gelatin (48723, Sigma-Aldrich) and XG (G1253, Sigma-Aldrich) were used for ink preparation. All inks were prepared in sterile conditions. For a stock solution, 1% w/v XG was dissolved in PBS (D8537, Sigma-Aldrich) at 80 °C and stirred sporadically until the mixture is homogeneous. Stock solution was kept at 4 °C maximum 3 days before printing. Stock was diluted 1:3 with DMEM (D5796, Sigma-Aldrich) for a final 0.25% XG concentration before gelatin addition. All inks were prepared by freshly dissolving gelatin in DMEM (with/without XG) at 45 °C for 45–60 min. LB gelatin was dissolved at a 4% w/v ratio, HB gelatin was dissolved at a 7.5% w/v ratio. A cross-linking solution composed of 10 U ml⁻¹ mTG (ACTIVA® TI transglutaminase, 100 U g⁻¹, 1002) solution was prepared in DMEM. The mTG solution was sterile-filtered and kept cooled on ice until usage.

A RegenHU 3D Discovery bioprinter was used for printing patterned gelatin substrates under a biosafety cabinet. Once the inks were dissolved at 45 °C, they were loaded on the cartridges and equilibrated approximately 1 h before the print. A 7.5% HB gelatin (with/without XG) was printed from a temperature-controlled cartridge set to 32 °C using volumetric dispensing. A 4% LB gelatin (with/without XG) was printed with a cartridge at RT (20 °C–23°C) using pneumatic dispensing. Unless otherwise noted, metal nozzles with a nominal inner diameter of 70 µm (Cellink Swe) were used. Six-well tissue culture plates were used as printing substrates and further cell culture. In general, an unstructured, homogeneous, layer of gelatin of ∼50 µm thickness was first printed using a meander-pattern with a line-spacing of 200 µm and a nominal nozzle height of 50 µm. On top of this flat layer, line-patterns of a second gelatin was printed at a nominal nozzle height of 70 µm. After printing, the plates were kept on ice for 15 min to ensure complete gelation. Chilled mTG cross-linking solution was added on the prints (1 ml well⁻¹) and the plates were kept on ice for an additional 15 min before 1 h incubation at 37 °C. After incubation, the prints were washed with PBS. Plates were kept at 4 °C in PBS maximum 3 days until the cell seeding.

The rheology of each ink was analyzed using a Discovery Hybrid Rheometer (TA instruments, DE, USA) equipped with a Peltier plate thermal controller and a plate geometry with a diameter of 40 mm and a fixed gap of 1 mm. All inks were freshly prepared in PBS before measuring. Amplitude sweep experiments were conducted to calculate the Young's modulus. The measurements for 4% LB gelatin ± 0.25% XG, and 7.5% HB gelatin ± 0.25% XG, were performed at 25 °C and 37 °C, respectively. For this, stock solutions of the inks with the following concentrations were prepared and mixed 1:1 with a 20 U ml⁻¹ mTG solution in PBS: 8% LB gelatin, 8% LB gelatin + 0.5% XG, 15% HB gelatin, 15% HB gelatin + 0.5% XG. The inks were mixed with mTG solution by pipetting up and down three times and depositing the mixed gel on the peltier plate of the rheometer. The plate geometry was lowered to 1 mm and the ink kept at 15 °C for 10 min to allow for physical gelation of the gelatin ink. Afterwareds, the gel was heated to 37 °C to allow chemical cross-linking for 1 h. This procedure was repeated three times for each ink composition. The storage modulus G' and loss modulus G'' were recorded as a function of the oscillation strain (0.1%–200%) at a fixed frequency of 1 Hz at 25 °C and 37 °C. The Young's modulus E was calculated solely for the samples recorded at 37 °C as follows: E = 2G' (1 + ν).

With G' being the storage modulus in the linear viscoelastic regime and ν being the Poisson's ratio. Here, we assumed a perfectly incompressible material with a Poisson's ratio of 0.5. Flow sweep experiments were conducted at the respective printing temperature, 4% LB gelatin ± 0.25% XG at 25 °C, and 7.5% HB gelatin ± 0.25% XG at 32 °C. The inks were first equilibrated at the respective temperature for 30 s prior to measuring. The viscosity of the inks was recorded as a function of the shear rate from 0.1 to 200 s⁻¹. Gelation curves were recorded from 45 °C to 15 °C and from 15 °C to 45 °C with a soak time of 30 s at a rate of 1 °C min⁻¹. The strain and frequency were fixed to 1% and 1 Hz, respectively. G' and G'' were recorded as a function of temperature, the gelation point was determined by plotting tan(δ) as a function of temperature.

C2C12 myoblasts were cultured using standard sterile technique and incubator (37 °C, 100% humidity, 5% CO₂). Prior to seeding on gel substrates, myoblasts were maintained in a tissue culture flask in C2C12 growth medium; DMEM (D5796, Sigma-Aldrich) with 10% fetal bovine serum (S1810, Biowest) and 1% Pen/Strep (P0718, Sigma-Aldrich). The culture was passaged at ∼80% confluence. The C2C12 cells were seeded on hydrogel prints at a density of 10 000 cells cm⁻². The cultures were allowed to proliferate in C2C12 growth medium reaching ∼90% confluence before initiation of myotube formation. The culture was rinsed with PBS (D8537, Sigma-Aldrich) and C2C12 differentiation medium formulated as DMEM with 2% horse serum (H1270, Sigma-Aldrich) and 1% P/S was added. C2C12 differentiation medium was changed three times per week, unless cells were electrically stimulated.

hSkMs (CC-2580, Lonza) were cultured following the manufacturer's instructions. Cells were maintained in skeletal muscle growth medium formulated as SkGM™-2 BulletKit™ Medium (CC-3245, Lonza). The culture was passaged upon reaching ∼60% confluence. The hSkMs were seeded on freshly prepared hydrogels at a density of 10 000 cells cm⁻². The cultures were allowed to proliferate in skeletal muscle growth medium reaching 70% confluency before initiating myotube formation. Once sufficiently confluent, the culture was rinsed with PBS (D8537, Sigma-Aldrich) and switched to skeletal muscle differentiation medium formulated as DMEM:F12 (D6421, Sigma-Aldrich) with 2% horse serum (H1270, Sigma-Aldrich) and 1% P/S (P0718, Sigma-Aldrich). Differentiation medium was changed three times per week.

Following myotube formation and further maturation (differentiation day 14/pacing day 0), C2C12 cells were paced using a MyoPacer Field stimulator and a six-well C-dish carbon electrodes (IonOptix). The pulse generator was set for a constant 1 Hz frequency with 2 ms pulse duration with 11.5 V field potential. The cells were stimulated for 7 days (until differentiation day 21/pacing day 7). The medium was changed daily during the stimulation period. The pacing videos were recorded after daily medium change. Only for video recording purposes at pacing day 0 and pacing day 7 the frequency was set also with 0.5 and 2 Hz for shortly (max. 20 s) with the same pulse duration and field potential.

C2C12 myotubes were fixed in 4% paraformaldehyde in PBS, rinsed and permeabilized with 0.1% Triton-X in PBS. The fixed samples were incubated in 1% BSA in PBS blocking solution with 1:400 sarcomeric α-actinin monoclonal antibody (MA1-22863, Invitrogen) at room temperature for 4 h on a shaker. 1:200 Alexa Fluor Plus 488 conjugated goat anti-mouse IgG secondary antibody (A32723, Invitrogen) was applied together with 1:400 Alexa Fluor™ 647 conjugated phalloidin (A22287, Invitrogen) for F-actin staining in blocking solution overnight at 4 °C on a shaker. 4ʹ,6-diamidino-2-phenylindole, dihydrochloride (DAPI) (62247, Thermo Scientific) was used as nuclear stain. Stained samples were then washed with PBS. The samples were imaged either in liquid mounting medium (50% Glycerol in PBS) or with ProLong™ Gold Antifade (P10144, Molecular Probes). Fluorescent images were acquired with Nikon Eclipse Ti2 microscope and NIS-Elements software. The bright field images were acquired with Zeiss Observer Z1 microscope with a mounted Zeiss AxioCam.

Myotube alignment analysis were run on ImageJ plugin Orientation J [31] based on F-actin staining at differentiation day 7. The false color images were obtained by using the analysis tool color-survey where the hue and the saturation corresponds to the orientation angle and the coherency, respectively. The data obtained from the distribution tool was used to calculate the alignment score. Briefly, for each image, angular fraction of distribution was plotted. The angle with maximum fraction of distribution corresponds to the principal orientation angle. The total fraction of distribution within the frame of ±15° around the principal orientation angle corresponds to the alignment score.

Myotube lengths were calculated similar to Jensen et al using bright field images [10]. In ImageJ, nine random counting windows (275 × 275 mm² each) were drawn on each image in a fixed pattern (1775 mm separation in a 3 × 3 positioning, with the central counting window being in the center of the image). The length of all the myotubes passing through the random counting windows were measured manually.

A 0.4 mm² of areas were imaged for alpha-actinin, phalloidin and DAPI staining. For each condition, three replicates were imaged at ten different areas in total. Among ten image per condition, three images were randomly selected for further analyses. Contractile protein expression: Actinin positive area was selected on alpha-actinin stained images by Huang's fuzzy thresholding method on ImageJ. Selected area over total area was calculated as actinin positive area percentage (n = 3 per condition). Myotube width: The average myotube width was calculated on phalloidin stained images (>20 myotubes per condition) on ImageJ by measuring membrane to membrane distance.

Fusion index: number of nuclei were counted in alpha-actinin and DAPI composite images (>1200 nuclei per condition). The fusion index was calculated as a percentage of the number of nuclei located in the alpha-actinin stained myotubes (with more than two nuclei per myotube) over the total number of nuclei in the field of view.

Statistical analysis was conducted using Origin Pro 2021. The data were analyzed for normality using Saphiro–Wilk test, and analyzed using one-way ANOVA followed by Tukey's post hoc tests for multiple comparison, p-values <0.05 were regarded as statistically different.