Perfusable Biohybrid Designs for Bioprinted Skeletal Muscle Tissue

Engineered, centimeter‐scale skeletal muscle tissue (SMT) can mimic muscle pathophysiology to study development, disease, regeneration, drug response, and motion. Macroscale SMT requires perfusable channels to guarantee cell survival, and support elements to enable mechanical cell stimulation and uniaxial myofiber formation. Here, stable biohybrid designs of centimeter‐scale SMT are realized via extrusion‐based bioprinting of an optimized polymeric blend based on gelatin methacryloyl and sodium alginate, which can be accurately coprinted with other inks. A perfusable microchannel network is designed to functionally integrate with perfusable anchors for insertion into a maturation culture template. The results demonstrate that i) coprinted synthetic structures display highly coherent interfaces with the living tissue, ii) perfusable designs preserve cells from hypoxia all over the scaffold volume, iii) constructs can undergo passive mechanical tension during matrix remodeling, and iv) the constructs can be used to study the distribution of drugs. Extrusion‐based multimaterial bioprinting with the inks and design realizes in vitro matured biohybrid SMT for biomedical applications.


Introduction
Skeletal muscle tissue (SMT) is the largest organ by mass in the human body, and it is crucial for locomotion, posture, respiration, physiology, and energy homeostasis. [1,2] Genetic or acquired diseases, as well as damage and aging can render this organ dysfunctional and cause significant impairment in human quality of life and health. [3] In the last three decades, several 3D SMT culture systems have been developed to mimic the native muscle microenvironment and understand skeletal muscle function, DOI: 10.1002/adhm.202300151 plasticity, and disease. [2,4] This understanding not only improved cell culture models for biomedical research [5,6] and engineered tissue grafts for medical implantation, [1,7] but also enabled controllable biomachines with dynamic abilities (i.e., biohybrid robots). [8][9][10] Through the years, SMT was engineered in vitro via approaches that aim at replicating the ultrastructure of native muscle tissue, which is composed of highly cooriented myofibers. Topographical, chemical, and electrical cues can align adjacent myoblasts to generate uniaxially oriented myotubes. [11][12][13][14] These local cues deliver stimuli to cells when these cells grow on surfaces, in planar geometries (e.g., muscular thin films), or in microtissues. [15][16][17][18] However, they cannot be easily applied to block-like 3D SMT constructs over the centimeter scale, as the stimuli would hardly reach the cells residing in the internal core of the constructs, leading to inhomogeneous cell stimulation. In contrast to local cues, applying mechanical stress to a scaffold filled with myoblasts is an effective way to form myotubes and control the orientation of the forming fibers all over a scaffold's volume. [19] To convey stress and strain to cells, one can use biohybrid tissue configurations that interface the biological tissue with synthetic elements. Biohybrid SMT constructs contain synthetic structures, such as tendon-mimicking anchors and skeletons, to counteract forces occurring during muscle tissue maturation. These forces arise from the shrinkage of the hydrogels that are used as cell scaffolding materials and the remodeling action of cells on the scaffolding materials. Biohybrid designs simplify the mechanical stimulation of muscle progenitor cells to form intact co-oriented assemblies of fibers in cm-scaled 3D models.
Despite their utility in engineering millimeter-scale SMT, [8,19,20] biohybrid designs have only been marginally investigated and adapted for the optimal development of SMT constructs at the centimeter scale. [21,22] The two essential challenges in scaling up SMT designs are related to i) the perfusion and ii) the structural stability of the final constructs. First, biohybrid designs should allow for the efficient distribution of nutrients and oxygen to guarantee cell survival. Second, biohybrid designs should feature coherent interfaces between the muscle tissue and the synthetic elements; these interfaces increase the stability of constructs during in vitro culture when www.advancedsciencenews.com www.advhealthmat.de tissue remodeling and maturation takes place, and they reduce force dispersion that would otherwise occur during the muscle contraction if there is a mismatch in adherence between different materials. A suitable design and a fabrication strategy are needed to enable perfusion and structural stability within a construct.
Bioprinting technologies can position biomaterials, cells, and bioactive factors in a single construct to achieve architectures that mimic native tissues. [23] SMT constructs (mm 3 -cm 3 scale) have been fabricated via extrusion-based 3D bioprinting and used as pathophysiological models, implantable grafts, and drug screening platforms. [23,24] For example, Kim et al. coprinted a cell-laden bioink and a sacrificial gelatin-based ink to generate an SMT construct of 1 cm 3 in size with channels (≈400-500 μm in diameter) throughout and two supporting anchors at each end. [22] Tissue maturation and longevity of their SMT were hindered because the anchor's design obstructed the fluid exchange between the channels and the environment, and no mechanical stressing was applied to the anchors.
Here, we formulated bioinks and applied them to extrusionbased bioprinting to fabricate our biohybrid SMT design that i) models the functional perfusion of native muscle structure with a high-resolution distribution of vessels, which preserves cell viability within the cm-scale tissue; ii) supports crucial mechanical tensioning for uniaxial fiber growth; and iii) allows for controlled fiber formation and maturation. To realize this maturable muscle biohybrid structure, we optimized the 3D construct design for a stable, yet perfusable, interface of tissue with anchoring structures. Then, we formulated suitable synthetic inks and a bioink, and characterized them for their rheological properties and printability to select the most suitable components for the realization of the biohybrid. We achieved i) perfusable longitudinal microchannels of 200 μm diameters that were finely distributed in the tissue and supported not only cell survival but also alignment homogeneously all over the cm-scale construct, and ii) stable incorporation of synthetic anchors to undergo effective tissue maturation under mechanical tension. Then, we characterized the hypoxic cell responses and biological development of our tissue in vitro, and correlated it with data on the perfusion and stability of the biohybrid assembly, as well as the distribution of drug models with different molecular weights.

3D Bioprinting of Muscle Constructs with Structural Biomimicry
We used multimaterial extrusion-based 3D bioprinting to realize perfusable, biohybrid SMT ( Figure 1A). To confer mechanical tension for muscle maturation, the constructs were fixed during the differentiation phase to agar beds at the bottom of culture wells. [25] Pillars were passed through the anchor holes of the SMT and pierced into the agar ( Figure 1A). The SMT consisted of a multilayered tissue structure with a nominal size of 3.7 × 2.4 × 15 mm and was filled with internal perfusable microchannels with a 200 μm diameter ( Figure 1B). To generate the channels within the constructs, a sacrificial ink composed of Pluronic F-127 was simultaneously printed with the bioink and then removed after printing by exposing the constructs to a temperature below its gelation point. In addition to the channels, the tissue was coprinted with integrated anchoring structures composed of a mixture of polyethylene glycol diacrylate (PEGDA) and Pluronic F-127, which served to fix the construct to the culture templates used for tissue maturation.
The design of the biohybrid SMT constructs was optimized for high print fidelity via iterative prototyping ( Figure S1, Supporting Information) resulting in a multilayered, triple-material assembly with seven deposition layers ( Figure 1C). This assembly was based on a central rectangular block fused with identical anchoring structures at the extremities. The central block (serving to generate the living tissue) contained lines of two inks, namely, the cell-laden bioink (blue) and the sacrificial ink (white). These two inks were interspersed with each other and ran along the longitudinal axis of the central structure. To create perfusable channels, the lines of the sacrificial ink continued beyond the central block and interfaced directly with the external environment. The anchors were designed with an inner perfusable interface to allow for perfusion of the central block's microchannels left upon removal of the sacrificial ink ( Figure 1D). To stabilize the perfusable interface, the anchor ink (orange) was patterned to interweave with the bioink. The anchor ink also formed a central hole that served as the insertion point to stabilize the pillars in the maturation template ( Figure 1D).
After bioprinting and photocrosslinking ( Figure 1E,F), we kept the constructs for four days in a growth medium to boost cell proliferation. We then mounted the constructs on a maturation template and cultured them in a differentiation medium for eleven days to form myofibers ( Figure 1F,G). Constructs lacking channels (−Ch/+Anc/+Tens), anchors (+Ch/−Anc/−Tens), or both (−Ch/−Anc/−Tens) were used as control groups and cultured with the same cell culture protocol ( Figure 1H). We also coprinted a design with channels and anchors and then let those constructs in culture without pillar-mediated fixation (+Ch/+Anc/−Tens) to see the effect of no mechanical tension during maturation (Figure 1H).

Material Selection and Fabrication
To fabricate SMT constructs with cointegrated anchor structures, we developed a bioink with suitable mechanical properties for bioprinting, muscle tissue development, and stable coassembly with synthetic structures. We optimized the bioink for viscoelastic behavior and printability by testing different formulations with variable relative fractions of gelatin methacrylate (GelMA) and sodium alginate (NaAlg). The fractions varied in the ranges of 4-8% and 3-11% for GelMA and NaAlg, respectively (Table S1 and Figures S2-S10, Supporting Information). The formulation based on 8% GelMA and 7% NaAlg displayed a linear viscosity profile over a predefined frequency range (Figure 2A) and a solgel transition depending on the applied shear stress ( Figure 2B). Both behaviors facilitate the selection of the optimum printing parameters. Moreover, the bioink was sufficiently soft ( Figure 2B) to support the printing and culture of myoblasts, and could be printed with an accuracy of 90% ( Figure 2C).
The line width of the bioink and sacrificial ink was ≈450 and 200 μm, respectively, and coprinting them allowed us to create architectures that mimicked the target design with high structural fidelity (Figures S11-S16, Supporting Information). In particular, we observed i) parallel microchannels running along the longitudinal axis of the constructs, which were open at the extremities and contiguous with the external space (Figure 2D-F; Figure  S13, Supporting Information); ii) homogeneous size of the microchannel lumen (average size: 191.7 ± 17.8 μm) and interchannel distance across the construct's volume (Figures S14 and S15, Supporting Information); iii) compact and homogeneous internal matrix and intact construct borders ( Figure S16, Supporting Information).
We generated the bioink by combining the GelMA-NaAlg blend with skeletal muscle cells (25 × 10 6 cell mL −1 ) and then coprinted the bioink with the sacrificial ink and the synthetic ink for the anchors to create biohybrid SMT constructs. Coprinting sacrificial ink with the cell-laden bioink resulted in channels with a diameter of ≈200 μm (211 ± 27.3 μm), which retained their diameter after the removal of the sacrificial ink ( Figure 2D,E; Figure  S13, Supporting Information). Four channels per layer were observed in the longitudinal cross-sections ( Figure 2F,G, Supporting Information) demonstrating the morphology of the realized structures matched the desired design in its entirety. The biohybrid construct maintained its structural integrity after washing out of the sacrificial ink ( Figure 2H) and the bioink and the adjacent anchor structures interfaced with each other ( Figure 2I).
The postprinting cell viability was assessed via Live/Dead staining (Figure 3A), which showed that the cells retained high viability (>90%) in all types of bioprinted designs. This result suggests that the presence of the inks for the channels and anchors, as well as the related bioprinting protocols, did not substantially alter the cell viability. To assess cell viability over time, the whole constructs were analyzed via Live/Dead staining at days 3, 6, 12, and 15 ( Figure 3B; Figures S17 and S18, Supporting Information). High cell viability was found in all construct designs and conditions. However, in the designs with channels (+Ch/+Anc/+Tens; +Ch/−Anc/−Tens; +Ch/+Anc/−Tens), we detected lower amounts of red (dead) cells as compared to the conditions without channels (−Ch/+Anc/+Tens; −Ch/−Anc/−Tens). Moreover, imaging showed the effect of mechanical tension during the differentiation phase on the morphological change of cells toward myotube formation. The constructs under mechanical tension (+Ch/+Anc/+Tens; −Ch/+Anc/+Tens) displayed strong cell alignment on days 3 and 6, and long and viable myotubes at different tissue depths (100 and 500 μm) on day 15 ( Figure 3B; Figures S17 and S18, Supporting Information). In the absence of mechanical stress (+Ch/−Anc/−Tens; −Ch/−Anc/−Tens; +Ch/+Anc/−Tens), the cells aligned only to a limited extent at the early maturation stage (day 6), and no evident viable myotubes were visible in the later maturation stage (day 15).
The morphological evolution of both cells and full 3D constructs complicated the imaging-based quantification of cell viability, which is why we monitored cell proliferation by quantifying cell metabolic activity with a Resazurin assay ( Figure 3C). The constructs that were fixed to the agar template during culture (+Ch/+Anc/+Tens; −Ch/+Anc/+Tens) showed a biphasic metabolic curve that reflected the culture conditions: during the first days of culture, the resazurin reduction increased, which confirmed the proliferation of cells within the scaffold. After day 5, the rate of cell proliferation decreased, likely due to the initiation of the differentiation and the density-dependent growth inhibition. The metabolic activity of muscle tissue developed in the absence of mechanical stimulation started to decrease at a later stage (after day 8). Even though the channeled constructs were expected to have fewer cells than the ones with no channels (≈3.75 × 10 5 less), their reduction in resazurin levels was always greater than those of constructs with no channels. This observation suggests that a perfusable mesoarchitecture can strongly affect cell viability and behavior.

Perfusion of the Scaffolds
To understand the distribution of oxygenation gradients and their effects on cell viability, we assessed the constructs for their perfusability and cell response to hypoxic gradients. Constructs without channels displayed high cell death distributed over a wide area from the scaffold border ( Figure 4A). By contrast, constructs with channels had fewer dead cells, suggesting that the presence of channels supports cell viability.
Coronal histological sections of the cell-laden constructs (Figure 4B) showed that at the end of the tissue maturation process, the size of the channels' lumen decreased due to tension application and SMT remodeling (from ≈200 to 50 μm). However, the channels were neither collapsed nor obstructed. We verified the functionality of the channels to transport fluid and their contribution to cell viability by subjecting the printed constructs to a perfusion assay with a staining solution based on cell culture medium containing a red dye. Shortly after bioprinting, the constructs were placed in a tilted position (angle 30°) on a glass substrate ( Figure 4C) and the staining solution was repeatedly added to one side of the construct with exposed channels. The staining solution entered the void spaces of the constructs, revealing the channel structures, and perfused the hydrogel material in a few minutes ( Figure 4C). By measuring over time the stained fraction of the construct from optical images, we estimated the perfusion kinetics, which revealed more than threefold faster staining of the channeled tissues as compared to the control groups with no channels seven layers (top) and a longitudinal cut above the fourth layer (bottom). The cell-laden bioink, the sacrificial ink (Pluronic F-127), and the anchor ink (PEGDA-Pluronic F127) are shown in blue, white, and orange, respectively. D) Magnified visualization of the perfusable anchor structure with hole. E) Manufacturing process steps: conversion of design into G-code; selection and optimization of constituent materials; mixing with cells and formation of the bioink; extrusion-based bioprinting; crosslinking the polymeric matrix; and culture of SMT constructs under mechanical tension using a maturation template. F) Culture conditions for muscle tissue maturation, including culture in growth medium for four days first, and then in differentiation medium for eleven days after mounting on the maturation template. G) In the first phase of culture (left), the SMT constructs floated in the cell culture media to favor cell proliferation. In the second phase (right), the constructs were fixed through their anchors to maturation templates to receive mechanical stimulation via passive tension. H) Target SMT construct and control groups used for the study (Ch = channels; Anc = anchors; Tens = tension). Scale bars: 1 cm.

Figure 2.
Material characterization and printing performance. A) Variation of the viscosity ( ) of the ink composed of 8% GelMA and 7% NaAlg over the shear rate (˙). B) Rheological behavior of the GelMA-NaAlg formulation before and after crosslinking. The storage (G′) and loss (G″) moduli of the inks are plotted against the shear rate (˙), and shown in black and cyan, respectively. Squares and circles refer to formulations analyzed before or after crosslinking, respectively. C) Printing accuracy of the optimized bioink (8% GelMA and 7% NaAlg) as compared to other formulations with variable fractions of GelMA (green) or NaAlg (orange). D) Histological section stained with hematoxylin and eosin (H&E) of a cell-laden construct with channels showing the axial section of channels before and after wash out of the sacrificial ink. E) H&E histology of SMT within the constructs after printing, as shown in the axial and coronal view (top and bottom, respectively). Red arrows and dashed lines indicate the lumen of channels. Scale bar: 150 and 100 μm (top and bottom, respectively). The dispersion of the channels within the matrix was observed via microscopic imaging of channeled constructs printed with no cells (deparaffinized section, F) and with cells (H&E staining, G). In (F), one channel differed in size possibly due to compression during the microtome sectioning. Red dashed lines indicate the external construct border. Scale bar: 200 μm. H) Optical picture of an anchored construct prototype in cell culture media, showing the stable assembly of tissue and anchors, and I) the detail of an anchor structure in a multimaterial construct prototype. Scale bars: 5 mm.
( Figure 4D; Figure S19, Supporting Information). The presence of the anchors did not affect the perfusion kinetics, suggesting that the lumen of the channels was open to the external environment and the anchors did not obstruct the lumen ( Figure 4D; Figure S19, Supporting Information). At the endpoint of the maturation process (day 15), the channeled constructs were still perfusable with faster kinetics as compared to a solid muscle without channels. These kinetics were unaffected by the presence or absence of tensioning during maturation ( Figure 4E,F).
Channeled mature biohybrid constructs were fully perfused by the staining solution after only 30 min of staining exposure, whereas constructs matured in the absence of channels were fully stained after 10 h of staining exposure ( Figure 4F).
We quantified through a colorimetric reaction the content of HIF-1ɑ transcription factor as a major regulator of hypoxiaresponse intracellular pathways ( Figure 4G,H). [26] The absence of a perfusable network within the scaffolds led to a higher amount of functional HIF-1ɑ than in the constructs with chan- nels. In the constructs with channels developed under anchorage and tension (+Ch/+Anc/+Tens), the HIF-1ɑ levels were 3.5-fold lower than in the equivalent but not perfusable system (−Ch/+Anc/+Tens). Confocal imaging revealed colocalization of the HIF-1ɑ and nuclei signal within the nonperfusable constructs, but not in the constructs with channels ( Figure 4H), indicating that the absence of the channels correlates with the onset of hypoxic responses in the cells.

Tissue-Anchor Interface and Construct Stability
Despite the SMTs deformation during maturation, the anchors retained their shape and could be penetrated by pillars for mounting to the maturation template ( Figure 5A). The SMT and the anchor materials formed an intertwined pattern, which consolidated over time to generate a coherent interface between the two phases. The interface pattern was externally visible also at the end of the maturation process ( Figure 5A, bottom). Histological analysis of the internal structure of the interface showed interpenetrating areas of the SMT and synthetic ink ( Figure 5B, blue and red arrows). Void areas in the histological sections re-vealed the scheme of perfusable channels at the intersection between the tissue and the anchor ( Figure 5B, black arrows). Confocal imaging on day 5 revealed a dense cell population on the tissue-anchor interface area and in the matrix projections within the anchors ( Figure 5C; Figure S20, Supporting Information). In the histological sections, cells were found in the anchor areas, suggesting that cell invasion and spreading of tissue formation contributed to rendering the biohybrid interface stable and coherent ( Figure 5D).
To understand the impact of the tissue remodeling process and the applied mechanical tension on the interface stability, we analyzed histological sections of the most internal tissue-anchor interface (green star in Figure 5A), a crucial point to guarantee the structural stability of the overall assembly. Even if sporadic lacunae were present at the border between the SMT and anchors, the biomatrix efficiently adhered to the anchor surface ( Figure 5E). At day 5, the surface contact between the biological tissue and the synthetic components was greater than 90%, independent of the application of tension (duration: 1 day) and the presence of channels in the SMT construct ( Figure 5E,F). On day 15, the interface areas with full contact between the two phases decreased (remaining however >70%), suggesting that the pro-  longed mechanical tension decreases the stability of the assembly to a mild extent. From day 5 to day 15, the contact area between the tissue and the anchor surface decreased differently for each group, corresponding to: −27.6%, −17.35%, and −34.9% for +Ch/+Anc/+Tens, +Ch/−Anc/−Tens, and −Ch/+Anc/+Tens constructs, respectively. In the absence of tension, the tissueanchor interface remained more cohesive, whereas in tensioned constructs, it experienced a certain loss of contact. Interestingly, unchanneled and tensioned constructs displayed a slightly starker cohesion loss than the channeled ones. This fact suggests that the presence of channels does not compromise the interface stability but could even contribute to augmenting it, potentially by softening the tissue structure and allowing for more flexible interaction with the more rigid anchor.
As shown by Live/Dead staining on 3D constructs, viable cells with an elongated morphology adhered and grew on the tissueanchor interface of the constructs with channels, whereas clusters of dead and roundish cells were observed on the interface of constructs without channels ( Figure 6A). The fraction of dead cells in the unchanneled constructs was however reduced as compared to other tissue areas (as shown in Figure 4A), which suggest that at the extremities of the unchanneled constructs, the perfusable design of the anchors allowed for sufficient media exposure. Moreover, by exploring the anchor-tissue interface via confocal imaging along the vertical axis, we observed the alternation of the anchor structures and the tissue layers, which in constructs with channels displayed a high-density cell population with high viability index and SMT-relevant morphology (i.e., myotube-like cell bodies) ( Figure 6B). As shown by scanning electron microscopy (SEM), cells could adhere to the anchors' surface while growing aligned and forming myotubes, however without invading and obstructing the designed cavities ( Figure 6C At the end of the tissue maturation time, the constructs were recovered for structural analysis (Figure 7A,B; Figure S33, Supporting Information). Most of the constructs retained the original biohybrid conformation, with anchors stably integrated into the whole assembly ( Figure 7A). The fraction of constructs that completed the tissue maturation while maintaining their overall integrity was ≈90% and did not vary due to the presence of channels within the tissue or the construct fixation on the maturation template. The overall integrity was higher for the SMT based on the 7% GelMA-8% NaAlg ink than for constructs based on a different bioink formulation (4% GelMA-11% NaAlg). The num-   ber of constructs that got superficially damaged while remaining functional was the same for the two bioinks (<25%) ( Figure 7B).
During culture, the tissue underwent a strong morphological change due to the hydrogel shrinkage, as well as tissue remodeling and maturation. We performed a tensile test on our constructs to assess how the mechanical properties of the tissue changed over time ( Figure 7C). The tensile strength of the construct measured at day 5 (before initiating mechanical tensioning on the maturation template) fell in the 0-20 Pa range (Figure 7D), whereas the tensile strength measured at day 14 attained the 3 kPa range values ( Figure 7E). The fracture strain value increased from 0.025% on day 5 to 0.365% on day 14. Young's modulus calculated from the elastic deformation data domain was ≈3.94 kPa at day 5, which is lower than the value calculated from the rheology of the crosslinked bioink ( Figure S2, Supporting Information). This elasticity decrease could be due to the creation of void spaces for channels, the addition of cells, and the construct's culture for a few days. The value of the matured construct (≈14.7 kPa) was compliant with the ranges expected in actual SMT (5-170 kPa). [27] The structural analysis performed on the construct photographs revealed that the longitudinal size of the muscle tissue reduced from the initial value of 1.5 cm (day 0) to 1.2 ± 0.2 cm on day 15 ( Figure 7F), and the two pillars of the maturation template converged by ≈15°toward the middle point of the construct longitudinal axis ( Figure 7G; Figure S34, Supporting Information). The force exerted by the muscle construct onto the pillars caused an inward tilting of the pillars. We measured the deformation by quantifying with a cantilever system the force acting on a pillar inserted into an agar phantom ( Figure S34, Supporting Information). A shortening of the construct at 3 mm corresponds to a 1.5 mm displacement of each anchor; the total force required for this displacement was found to be 604.1 ± 101.5 μN, which falls in a common value range for passive forces generated by engineered skeletal muscle. [28] These observations suggest that the selected bioink supported an SMT with appropriate mechanical properties for coherent and stable integration with the anchors and the maturation template, as well as for long-term development under mechanical tension.

Muscle Architecture Formation
To assess the outcome of tissue maturation, the constructs were analyzed for the expression of biomolecular and morphological markers of SMT after 15 days of culture. In H&E stainings, the cell nuclei of the mechanically tensioned constructs were oriented along a preferential direction ( Figure 8A) and long myotubes (length > 150 μm) were observed ( Figure 8B). Not subjecting the constructs to mechanical tension resulted in matrices with a heterogeneous distribution of cells growing with multiple directionalities ( Figure 8A,B). In the constructs with channels (+Ch/+Anc/+Tens), the lumen of the channel was seen along with the cells clustered in the proximity of the channels' walls, aligning with their axes ( Figure 8A). During the maturation process, the structure of the channels remained functional ( Figure  S35, Supporting Information), and channels running parallel to the longitudinal axis of the construct were observed in the internal volume of the matured constructs ( Figure 8C).
The expression of the major markers for skeletal muscle differentiation and maturation, namely, MyoD and MyHC (Myosin Heavy Chain), varied among conditions, suggesting different stages of the muscle tissue formation process (Figure 9). Tensioned biohybrid SMT displayed stronger expression of MyHC as compared to other conditions. Myotubes were observed in all conditions, but in constructs lacking mechanical stimuli they were found sporadically and with reduced size.
While myotube width did not vary significantly among conditions, the myotubes of channeled or full constructs under tension (+Ch/+Anc; −Ch/+Anc) were substantially longer (≈200 μm) than those in the other samples (ranging from 50 to 70 μm) (Figure 10A,B). Importantly, the fusion index and the creatinine kinase activity were significantly higher in perfusable muscle under mechanical stress than in all other conditions ( Figure 10C,D).
Myotubes on the mechanically stimulated SMT constructs had significantly lower angle shift (15.2 ± 4.5°and 10.2 ± 6.1°) as compared with unstimulated control groups (up to 45.2 ± 7.2°) (p < 0.05), which pointed to a higher myotube alignment (Figure 10E). Intriguingly, in the unstimulated constructs with channels, the myotubes in the proximity of the channels were highly aligned ( Figure 10F) with a calculated shift angle value of 18.2 ± 6.5°, exhibiting higher overall cell alignment with respect to the other control groups ( Figure 10E,F; Figure S35, Supporting Information). This observation suggests that channels can provide a structural pattern that guides cell organization and the unidirectional growth of myotubes to some extent.
To characterize the myofiber formation process during the SMTs mechanical stimulation, we monitored the +Ch/+Anc/+Tens constructs via confocal microscopy (Figure 11). Two days after starting the mechanical SMT stimulation (day 6), elongated cells and isolated myotubes were visible in confocal microscopy, which coexpressed MyHC and MyoD. This suggests that cells' exposure to the differentiation medium and the initiation of mechanical stimulation induced cell fusion and myotube formation. However, a large fraction of cells were still in an early differentiation stage at this time point. At day 11, the MyoD expression drastically decreased and larger myotubes formed, which aligned with the longitudinal axis of the construct. At day 14, no MyoD expression was detected, while MyHC + fibers were densely packed and uniformly distributed with a straight, net orientation parallel to the longitudinal axis of the construct. On day 15, aligned MyHC + fibers were found at different tissue depths in the tissue ( Figure S36, Supporting Information).

Secretome Analysis
To gain deeper insights into SMT maturation, the cytokine content of the culture media was analyzed via ELISA at different time points (Figure 12). In this analysis, several well-known regulators of myogenic functions have been considered, including (i) four ligands for the cytokine receptors (IL-4, IL-6, IL-12, and IL-13), two ligands for receptor tyrosine kinases (BDNF and VEGF), and two members of the TGF-family (Myostatin and BMP-4).
During early myogenesis, IL-6 promotes satellite cells' activation and proliferation. In the transition from myoblast to nascent myotube and then to mature myofiber, IL-6 stimulates cell cycle withdrawal and the initiation of differentiation, as well as cell migration and fusion. [29][30][31] The highest cumulative concentration of IL-6 was found in the liquid media of constructs with channels and anchors fixed to the agar template to mature under mechanical tension (peak value at day 15: > 1 ng mL −1 ). IL-4 and IL-13 induce second-stage fusion during myogenic differentiation. [29,32,33] Their release was significantly augmented in constructs developed under mechanical tension as compared to the nonfixed control groups. IL-4 production was increased by more than 20% in all conditions developed under tension at all late time points (day 11, 13, and 15), while the production of IL-13 in anchored constructs developed under tension was enhanced by ≈55.6% and 64.3% at day 13 and 15, respectively, as compared to +Ch/+Anc/−Tens constructs. The +Ch/+Anc/+Tens constructs also showed increased production of VEGF, BDNF, and IL-12, which can be arguably correlated to the application of mechanical tension.
During myogenic differentiation, VEGF not only induces cell cycle withdrawal, cell migration, and differentiation, but it also promotes cell survival. [34] The release of VEGF from cells in designs without channels was reduced as compared to the designs with channels (−36.4% and −15.8% in the anchored and nonanchored constructs, respectively, at day 15). BDNF is a recognized promoter of cell cycle withdrawal and myogenic differentiation. [29,35] At the endpoint of the observation range, the cumulative concentration of BDNF in the media of anchored, channeled constructs reached a value of ≈600 pg mL −1 , the highest value observed in the series of samples.
As an additional positive regulator of the differentiation initiation, IL-12 was also analyzed, showing increased production in the second part of the kinetics for all types of constructs, which points out an auto/paracrine role of this interleukin. The SMT developed under tension displayed the highest values of IL-12 cumulative release peaking in the range of 500-700 pg mL −1 at the end of the kinetics and strongly diverging from constructs cultured in suspension.
Myostatin is a negative regulator of proliferation and at the same time, it inhibits the initiation of satellite cells' differentiation. [36] It is expressed in diverse in vivo scenarios, including developing muscle, muscle regeneration following acute injury, satellite cell-dependent compensatory hypertrophy, and muscular dystrophy. In our study, all the constructs expressed myostatin in the initial phase of the culture. Two days after the beginning of the myogenic differentiation phase, the level of released myostatin decreased in all samples, with more evident inhibition effects in the constructs developed under mechanical tension. A similar trend was observed for BMP-4, a cytokine that suppresses cell cycle withdrawal and myogenic differentiation, and it is expressed in the developing muscle. [37,38] Even if the production of BMP-4 was drastically reduced in the second phase of the tissue culture (i.e., differentiation phase) in all constructs, a stronger inhibition was observed in the constructs subjected to mechanical tension (+Anc/+Tens) as compared to those cultured in suspension (−Anc/−Tens; +Anc/−Tens).

Construct Perfusability for In Vitro Drug Testing
We assessed the utility of our perfusable, bioprinted construct for drug testing by determining drug distribution and release from the channeled and unchanneled matrices. We hypothesized that the drug distribution in the construct is governed by both intratissular diffusion and channel-mediated liquid convection (Figure 13A,B). To test this, we added the model drug fluoresceinisothiocyanate (FITC)-dextran to the culture media of the constructs for 1 h and monitored the distribution of the compound over time, during and after incubation.
First, we observed the diffusion of the drug, using confocal microscopy, through the tissue at multiple time points during incubation. We found that the fluorescent drug could diffuse from the culture media to the tissue across the scaffold border ( Figure 13C). Importantly, FITC-dextran spreads throughout the muscle tissue more rapidly in the channeled constructs as compared to the unchanneled controls ( Figure 13D; Figure S37, Supporting Information). The dye highlighted the structure of the channels only a few minutes after starting the incubation (red arrows in 13D), suggesting that the presence of channels, as well as the perfusable design of the anchors and the tissue-anchor interface, enable the rapid convection of liquids within the tissue volume. The size of the channels that were fluorescent in the FITC-dextran exposed construct as evaluated by confocal microscopy (time point = 15 min) was found compatible with the one expected for the channels' diameter (≈200 μm).
After completing the incubation with FITC-dextran (1 h), the scaffolds were washed in PBS and placed back into culture media. We estimated the efficiency of drug loading into the tissue and subsequent release by measuring the fluorescence emission of a 1 mm 3 fragment of the constructs immediately and two days after the incubation ( Figure 13E). As compared to the unchanneled controls, the channeled constructs internalized higher amounts of FITC-dextran (+33%) and released it more efficiently, retaining less than 5% of the original drug content (vs 8% for unchanneled constructs) after two days of incubation in culture media. By sampling the culture media over time and analyzing it for fluorescence emission, we observed an initial, rapid increase in the fluorescence intensity (first 3 h) that was then followed by a minimal signal growth over time ( Figure 13F). Such a trend was found in both tissue types, with the channeled constructs featur- ing more intense fluorescence emission. The lower fluorescence emission in the media of the unchanneled constructs could be due to both the low drug loading efficiency and the high drug retention. Both these effects can be caused by the reduced tissue surface available for exchange with the bulk solution, which limits drug penetration and release. The higher initial, burst drug release of the channeled constructs as compared to the unchanneled constructs could be partially explained by the release of a residual drug solution that remained within the channel space even after washing. The release of the drug contained on the surface layer of the construct would be greater in the channeled constructs due to the larger surface area, which could also play a role in the observed initial burst release. After 8 h, the fluorescence intensity change stabilized and only mildly increased during a long time frame (up to 40 h), suggesting that, after a few hours, the drug release was dominated by molecular diffusion through the tissue.
To further investigate the utility of the bioprinted construct in pharmaceutical research, we studied drug distribution and release after localized delivery within the tissue to understand the effects of the fluid convection and diffusion in a model with higher biomedical relevance for the topical delivery of drugs (e.g., intramuscular injection model) (Figure 14). [39,40] We tested two drug models (i.e., FITC-dextran and Rhodamine B) with different molecular sizes (40 kDa and 479 Da, respectively) by injecting a small solution amount in the central area of the construct (≈3 mm depth) ( Figure 14A,B). The medium of the constructs was then analyzed for fluorescence emission, revealing a faster drug release rate in the channeled constructs as compared to the unchanneled ones ( Figure 14C,D).
In the channeled constructs, the drug release had a biphasic trend with an initial exponential phase followed by a slowing and stabilization of drug liberation over time. By contrast, the kinetics of drug release in the unchanneled constructs were slower. A large fraction of the administered FITC-dextran was retained by the unchanneled tissue. The cumulative rhodamine B release from unchanneled constructs almost matched the one observed in the channeled constructs (release range: 75-80%) but over a longer time period. It is important to note that some of the drug solution injected into the channeled construct may have directly entered into a channel instead of only tissue, contributing to a quicker initial release. Overall, the difference in the pharmacoki- netic profiles suggested that the intratissular drug distribution was strongly affected by the presence of the channels. Moreover, the rhodamine B release was faster as compared to the FITCdextran's (≈18 vs 9% at 30 min in channeled constructs, respectively), potentially because of rhodamine's smaller molecular dimensions. Following the injection procedure, all the constructs preserved their structural integrity and even the channeled constructs did not show any signs of fragmentation, which would have altered the drug release profile and compromised the assessment of the pharmacokinetics. Our data support the hypothesis that the fluid convection mediated by vessel-like channels and the specific anchor design that allows the longitudinal channels to communicate with the external environment and be perfused are key features of a biomimetic model of vascularized muscle tissue. Our perfusable SMT model can also be fabricated with human primary cells, which well tolerate the biofabrication process, and efficiently differentiate into myotubes, as shown by morpho-logical and molecular features ( Figures S38 and S39, Supporting Information). As such, we expect that our construct design will improve data generation in pharmacological research in vitro.

Biohybrid Designs of SMT
SMT biofabrication should recreate perfusable anisotropic structures that mimic native tissue, such as perfusable microchannels aligned in parallel with bundles of mature myofibers. To align fibers, the maturation of engineered, mesoscale (i.e., cm size ranges) SMT requires constant mechanical tensioning, which can be provided by rigid culture templates counteracting the matrix shrinkage. [2,8] Biohybrid designs for SMT maturation should contain anchoring synthetic structures to the ends of thick muscle tissue. These modular designs also facilitate the handling of fragile hydrogel-based 3D SMT constructs. [41] Biofabrication can even model the myotendinous axis by realizing a seamless interface that optimizes force transfer from muscle to tendon. [23] Ladd et al. modeled a myotendinous junction model by culturing myoblasts and fibroblasts on a scaffold featuring electrospun polymers that replicate the gradients in mechanical stiffness at the interface between the tendon and skeletal muscle. [42] Despite its potential for SMT engineering, the multimaterial bioprinting of synthetic 3D structures and their biohybrid interfaces has been scarcely investigated. [22] Merceron et al. printed a myotendinous construct by alternating the deposition of polycaprolactone and polyurethane with muscle cells and fibroblasts. [43] Both polymers were layered in a grid shape to form a backbone with two distinct elasticities that structurally support the growth of both cell types into a construct mimicking the native tissue organization. The synthetic backbone resembled a distributed scaffolding material that facilitated the integration with the cell-laden matrix. Kim et al. used multimaterial bioprinting to form a cubic SMT block (1 cm 3 ), [21,22] in which the bioink was coprinted with a sacrificial gelatin-based ink and synthetic support pillars but this assembly was optimized for simultaneous mechanical tensioning and communication of the channels with the external environment.
By contrast, our print design allows media to passively diffuse through the channels and anchors. The anchors fix the muscle to the pillars in the maturation template, and could support a longer muscle tissue block suitable for the development of long fibers. The printing pattern stably intercalated the tissue and anchor materials during the layers' deposition. By optimizing the printing procedure against various parameters, we achieved a sufficiently high printing resolution to realize constructs with the intended characteristics, such as perfusable channels and interpenetration between SMT and anchors. Despite the shrinkage of the matrix during muscle tissue development, our biohybrid SMT remained functional for two weeks. Our design printed with PEGDA/Pluronic and GelMA/NaAlg was robust enough to withstand forces spanning the μN and mN range, which are typical for bioactuators of similar size. [28,44,45] Most intriguing was that patterning perfusable networks within the constructs not only reduced the occurrence of hypoxic regions without altering the structural stability of the assembly, but also provided contact guidance to the growing myocytes, promoting the formation of unidirectionally aligned muscle fibers.
In our system, synthetic materials and biomaterials were shaped into one construct during the bioprinting process. The synthetic structures were stabilized via crosslinking at the same time as the biomatrix itself. The muscle tissue then matured while being in direct contact with their surfaces, allowing the soft tissue to adapt and tightly adhere to the complex morphology of the anchors. Future studies should address if our one-go fabrication process is advantageous not only to support the SMT development but also their dynamic functions, for example, by increasing the efficiency of force transmission through the tendon-like structures.
The skeletal muscle is a highly vascularized organ, whose activity largely depends on the adaptive nature and fine distribution of its microvasculature. Therefore, our proposed design aimed at incorporating a system that could provide efficient perfusion. The organization of the functional unit for blood flow regulation in the muscle (i.e., muscle microvascular unit) includes arterioles with different diameters (≈20-50 μm) that perfuse capillaries of smaller sizes, which irrigate the muscle tissue and are then collected by venules. [46] The microvascular unit typically comprises five to ten capillaries located in-between three or four adjacent myofibers, and it is highly conserved among species, being similar for small mammals and humans. In our model, the chan-www.advancedsciencenews.com www.advhealthmat.de nels mimicking the vessels were fabricated with an initial diameter of 200 μm, determined by the mechanical properties and the extrudability of the sacrificial ink. The channels' diameter then decreased to 50 μm by the end of the maturation process, thus matching the typical size range of arterioles. The distance between channels reduced from the initial 350-400 μm range to the 200-300 μm range, producing a tight interspersion of the vessels into the tissue.
Although sufficient to support cell survival, growth, and maturation, our perfusable network did not match the small lumen and fine distribution of capillaries found in the native muscle because of the inherent spatial resolution challenges of the 3D extrusion-bioprinting technique. Hence, our model mimicked the distribution of arterioles rather than the native vessel density or its capillary/myofiber ratio. Moreover, we printed channels that run parallel to the myofiber orientation axis so as to not disrupt the functional structure of the muscle and its connection to the tendon-like anchors. However, this design consideration meant our model did not include any branched structures, interconnections, or hierarchical levels of organization of vascular architecture. Interestingly, we observed muscle cells clustering close to the channel borders and aligning during myotube formation with the channels' direction. We argue that such a cellular organization could have contributed to stabilizing the channel's structure and perfusability to the end of the tissue development process. Nevertheless, engineering a vascular endothelium with specialized cells (i.e., endothelial cells, mural cells, fibroblasts, etc.) within the channels would help to increase the biomimetism of the model, and extend its robustness and applicability. For example, endothelial cells would stabilize the channels' walls and also regulate the nutrient and oxygen exchange. [47] In general, our model could serve as a platform to engineer vascular systems or combine diverse muscle cell populations to establish heterocellular communication, originate functional cellular niches, and study their effects on SMT development in vitro.
In this study, we demonstrated a model that supported passive tissue perfusion and its utility in studying drug distribution. However, to extend the applicability horizon of our proposed model, future research should investigate the implementation of systems for active and controlled tissue perfusion under fluidic pressure, which would better mimic the blood circulation dynamics of vascularized tissue. Such a realization will require higher robustness of the whole assembly and more cohesive interfaces that could efficiently endure the applied pressure.

Materials for Bioprinting and Developing SMT
Materials for biofabrication should display biomechanical properties that guarantee muscle cell differentiation and proper SMT formation. Much research has been done to elucidate the role for matrix or tissue stiffness in striated muscle development, [18,20] finding that muscle cells have complex responses to the mechanical properties of scaffolds. For example, myogenic differentiation in culture depends intimately on optimal outside-in signaling mediated by the matrix elasticity. [11] Cell fusion into myotubes might occur independent of substrate flexibility, but myosin/actin striations only emerge in scaffolds displaying stiffness that match the values of actual muscle (passive Young's modulus of ≈10-12 kPa). [48][49][50] Unlike sarcomere formation, the cell adhesion strength augments monotonically versus substrate stiffness with the strongest adhesion on stiff scaffolds. [51] So, the successful development of SMTs requires scaffold materials that display the right balance between elasticity and stiffness. Alginate and gelatine-derived hydrogels provide an excellent environment for myogenic differentiation due to a mechanical profile that mimics one of the native matrices (e.g., elasticity moduli of 10 to 15 kPa). [52,53] GelMA is a UV-based crosslinkable hydrogel that has been widely used in biofabrication and skeletal muscle cell models, due to its excellent thermal stability, biocompatibility, solubility in water, biodegradability, and low antigenicity. [14,54,55] Alginate is a naturally derived polysaccharide hydrogel and another top biomaterial used in printed micro-biofabrication, and widely applied in muscle tissue engineering. [55,56] In bioprinting, alginate and gelatin blends are preferably created to improve the thixotropic and rheological properties of the bioinks. [54] Moreover, alginate-gelatin blends also create multicompartmental hydrogel designs to control cellular orientation (e.g., 3D microfilaments), a crucial technical aspect in muscle tissue biofabrication. [55,57] Even if the component materials were not novel in biofabrication, our specific, 8% GelMA-7% NaAlg ink formulation allowed us to print construct geometries with higher accuracy compared to other formulations. Our constructs showed not only a higher fraction of functional channels with desired open extremities that remained functional for the whole culture duration, but also a more compact and homogeneously distributed bulk matrix, regular interchannel distance, and less fragile scaffold borders. Even if crossed by perfusion networks, the architecture of the whole construct was sufficiently intricate to guarantee a stable and durable interface between the muscle tissue and the synthetic elements. In this environment, we proved that the SMT could develop even if the cells were not embedded in conventional biomaterials (such as collagen, fibrinogen, and Matrigel) [21] in which the differentiation of C2C12 cells and the formation of functional fibers are commonly achieved within one week. [13] However, these materials are not optimal for bioprinting of larger structures, due to their cost, limited versatility, and applicability to high-resolution biofabrication. In our system, we achieved SMT maturation by allowing for more days of differentiation (11 days) after an initial phase required for cell proliferation (4 days) (more discussion on materials for biofabrication is available in Section S9, Supporting Information).
We found that exposing SMT to mechanical stress within our biohybrid system increased autocrine and paracrine loops for myogenesis promotion. Biofactors supporting tissue maturation were exponentially expressed for the duration of the culture. Mechanical tensioning and switching from growth to differentiation media lead to a decrease of BMP-4 and Myostatin in the second half of culture time, which fostered myogenic differentiation. While printing anchors did not affect the secretome, the presence of channels impacted the production of cytokines, suggesting that our biohybrid design preserved cell viability by effectively combining the channels of the soft tissue with the voids of the anchors.
The data herein presented reports on the following novelties: a perfusable design of the anchor's structure; its realization via a combination of several, different materials (biological, synthetic, and sacrificial inks); improved tissue performance in relation to survival, maturation, and mechanical robustness. Such findings are promising for the development of novel techniques to control the maturation process of tissues characterized by anisotropic cell distribution, defined mesoarchitecture, and responsiveness to mechanical stress. Finally, as our proposed model can also be fabricated with human cells and we demonstrated the utility of using our perfusable construct to study drug distribution, our biofabrication approach has a strong potential for biomedical applications.

Conclusions
Our results demonstrated that 3D extrusion-based bioprinting applies to the engineering of multiphase constructs that include living cells, biomaterials, synthetic materials, and void volumes acting as perfusable networks. When realized through an intricate and perfusable design, this complex assembly supported the development of 3D SMT, a process that entails dramatic structural deformation deriving from the shrinkage of the polymeric matrix, cell morphology changes, and cell-mediated matrix remodeling. We expect that layer-by-layer multimaterial bioprinting of biohybrid SMT will impact the field of muscle tissue engineering for use in biomedicine, nutrition, and biorobotics. In fact, we showed how this approach can overcome some of the main challenges in bioengineering of 3D muscle tissue in vitro, including: achieving shape fidelity in biomimetic tissue designs; retaining structural stability during tissue remodeling; providing efficient liquid exchange to core regions of constructs and avoiding hypoxic areas formation; and increasing ease of manipulation of tissue in vitro. [53,58,59] As these challenges are also common for the engineering of other tissue types, [60] we believe that our biohybrid fabrication approach can guide scientists in improving biofabrication also in other tissue domains.
Bioink Preparation: GelMA was synthesized under minimal light exposure by coverture with aluminum foil. Briefly, gelatin was dissolved in PBS at a concentration of 10% w/v at 50°C under constant stirring. 0.6 g methacrylic anhydride per 1 g of gelatin was slowly added to the solution, which was then kept continuously stirring for further 60 min. The solution was then dialyzed using a membrane with a molecular weight cut-off of 12-14 kDa. Dialysis was performed at 40°C against a large volume of deionized water for four days, with daily water renewals. The solution was kept at 37-50°C during the entire synthesis and purification to avoid thermal gelation. The dialyzed GelMA solution was diluted with PBS to a final volume of 40 mL and frozen overnight. Finally, the frozen GelMA was freeze-dried until completely dry and stored at 4°C until used. To prepare the GelMA-NaAlg polymer blends, GelMA was used at final concentrations of 4%, 6%, and 8% w/v, while NaAlg was used at final concentrations that incremented by 3% while remaining in a defined range of total polymer concentration (11-15% w/v). All blends contained 0.5% w/v of LAP photoinitiator. The tested formulations are listed in Table S1 of the Supporting Information. The blends were generated from GelMA precursor solutions at concentrations of 16%, 12%, and 8% w/v obtained by diluting the freeze-dried GelMA into PBS, and adding the LAP photoinitiator. The solution was kept on a thermo-shaker at 1400 rpm and 70°C for 30 min to fully dissolve the solutes, and then filtered under sterile conditions through 0.2 μm supor filters for sterilization. The alginate precursor solutions (concentrations of 6%, 10%, 14%, 18%, and 22% w/v) were obtained by gradually adding appropriate NaAlg amounts to PBS at 65°C, under continuous stirring. The NaAlg solutions were then kept on an incushaker at 250 rpm and 50°C for 8 h, and finally sterilized by autoclaving. GelMA and NaAlg solutions were warmed in a water bath to 37°C before bioprinting. To prepare the pluronic ink, Pluronic F-127 (Sigma-Aldrich) was gradually added to cooled PBS (4°C) to reach a concentration of 40% w/v. The solution was centrifuged at 4°C for 1 min, autoclaved, and stored at 4°C until used.
Rheology of Polymer Formulations: Rheological measurements of the polymer GelMA-NaAlg blends were performed using a rotational MCR series rheometer (Anton-Paar, AT) with a parallel plate geometry. The polymer blends were deposited on the measuring surface of the rheometer plate (radius: 4 mm) and the upper plate was lowered until direct contact with the material was established. Frequency sweep tests were performed with a frequency range of 0.1-100 rad s −1 and a constant shear strain of 0.5%. All measurements were performed at room temperature. All polymeric solutions were tested in both uncrosslinked and crosslinked states. For crosslinking, the material was first irradiated with a handheld UV light (UltraFire 502UV, UltraFire, Piscataway, NJ) for 15 s and subsequently covered with a crosslinking agent (50 mm calcium chloride, CELLINK). Extracted values were imported and analyzed in MATLAB (MathWorks, USA).
Design of the Biohybrid SMT: All bioprinted muscle constructs were fabricated by extrusion-based bioprinting by means of a CELLINK Bio-X6 Bio (CELLINK, Boston, USA). This system, well applicable to multimaterial printing, is composed of multiple dispensing modules (six different printheads), a pneumatic pressure controller, an XYZ stage/controller, and a sterile chamber for printing. The design of the multilayered, biohybrid construct was based on a G-code that facilitates the biofabrication via BIO X6 bioprinter by enabling layer-by-layer crosslinking and the technical adjustments of parameters (such as printing height and speed) during an ongoing print. The constructs were designed with 3D computer-aid design (CAD) modeling using Autodesk Fusion 360°(San Francisco, CA, USA). As converted to a motion program, the CAD modeling provided the design's paths to print the constructs with different morphologies.
Bioprinting of the Biohybrid SMT: All printed materials (i.e., cell-laden and sacrificial bioinks) were aseptically loaded into different cartridges. The inks were extruded through conical nozzles with an inner diameter of 410 μm (22G) (CELLINK, Boston, USA) in a closed aseptic chamber during the printing process. Before printing, the nozzles were aligned to determine their respective X-Y-Z offsets with minimum 0.01 mm accuracy. Printing parameters were optimized against the following variables: pressure, speed, and temperature. The optimal printing conditions for each ink were determined through tests for strand assessment, printing accuracy, and shape fidelity (see the Supporting Information). Table S2 of the Supporting Information lists the used inks with their optimized printing parameters. The parameters were used as a reference point, as their constant adjustment during printing was needed to ensure continuous ink deposition. To maximize the successful outcome of the bioprinting process, three preliminary iterations were run in the absence of cells, before the cell-laden constructs were actually bioprinted. To create the bioink, the cells were counted, and a specific number of cells were centrifuged, collected as pellets, and then mixed with a specific biopolymer volume to prepare the cell-laden bioink at a specific cell seeding density. The printing procedure was optimized to maximize the cell seeding density while retaining good mechanical properties of the resulting bioink that allowed for accurate printing. The optimal cell concentration was found to be 2.5 × 10 7 cells mL −1 . To efficiently mix the cells and the bioink, the two bioink components were loaded into two sterile interconnected syringes, and shuffled for a minimum of 50 times (pace of 1 Hz for a minimal time duration of 50 s). Into detail, the cells were first resuspended into GelMA, and then the NaAlg and the cell-laden GelMA solutions were mixed through the syringe shuffling. The resulting bioink was then transferred to a 3 mL printer cartridge using the Luer lock connection. To keep both the GelMA and the Pluronic F-127 in their thermally gelated state, the printbed temperature was set to 19°C. To generate SMT constructs, the GelMA-NaAlg blends were printed with the temperature-controlled printhead and kept at 21°C throughout the entire print. The cell-laden bioink was printed at a speed of 10 mm s −1 and air pressures of 30 kPa. The sacrificial bioink was printed at a speed of 8 mm s −1 at 75 kPa. To print the anchors, an ink based on PEGDA (M n = 700; Sigma-Aldrich) and Pluronic F-127 was used. The two components were mixed in a hot environment. In a lightprotected beaker, the LAP photoinitiator dissolved in PBS at 70°C under magnetic stirring for a final concentration of 0.5% w/v. After PEGDA addition and full dissolution (40% w/v), Pluronic F-127 (40% w/v) was gradually added to the mixture. The solution was autoclaved and stored at 4°C in a light-protected tube until use. Dispensing speed and pressure of the PEGDA-Pluronic polymer blend was 75 mm min −1 and 780 kPa, respectively. Each layer was photocrosslinked directly after printing using the printer's integrated UV module (365 nm, 30 s). After printing all seven layers, the constructs were submerged in an ice-cold crosslinking agent (50 mm CaCl 2 , 45 s). As Pluronic F-127 (40% w/v in deionized H 2 O) has a low critical gelation temperature (≈14°C), [61] washing the constructs with the ice-cold crosslinking agent simultaneously flushed out the sacrificial ink and crosslinked the alginate portion of the hydrogel.
Culture of 3D Constructs: After printing, myoblast-laden constructs were let into GM for four days to allow for cell proliferation. Half the volume of the culture media was collected and replaced with a new GM. At day 5, anchored constructs were mounted on their maturation templates. [24] These templates were plastic plates with preformed agar beds (3 mL, 1.5% w/v) that were pretreated with culture media to prevent nutrient depletion of the culture media by diffusion of the nutrients into the agar bed. The beds were used to fix the anchors in place during tissue maturation by pinning them down with two pipette tips (length 1.2 mm; upper diameter ≈2 mm; composing material: sterile polypropylene). The constructs were cultured for 11 days in differentiation medium (DM), composed of DMEM/high glucose, supplemented with 2% (v/v) horse serum, 2 mm glutamine, 100 U mL −1 penicillin, 100 μg mL −1 streptomycin, 50 ng mL −1 IGF-1, and 2 mg mL −1 6-aminocaproic acid (all purchased from Sigma-Aldrich). For human cell constructs, the differentiation medium consisted of myotonic differentiation medium supplemented with 2% (v/v) horse serum (PAA Laboratories), 1% (v/v) fetal calf serum, and 0.1% (v/v) gentamicin. Control constructs with no anchors were simply cultured in DM without any mechanical fixation. During culture, half media content was daily renewed and aliquots of the media were collected for protein content analysis at specific time points.
Perfusion Experiments: To assess perfusion, a dynamic staining experiment was performed. The constructs were put on a tilted glass surface (30°C ) and three aliquots of red staining solution were repeatedly added to the upper extremity of the construct (time interval among additions: 1.5 min). The staining solution was composed of a culture medium containing the red dye Erythrosin B (0.1% w/v; Sigma-Aldrich; product number: 200964). The staining solution was left to perfuse or diffuse tissue constructs following gravity. The construct's stained (%) fraction was calculated from optical pictures at different time points to describe the perfusion kinetics. An additional time point for construct observation was added at 10 h to illustrate the staining saturation conditions of all constructs.
Live/Dead Staining in 3D Constructs: To assess cell viability in the constructs, Live/Dead staining was performed immediately after bioprinting and at different time points during the culture time, following the manufacturer's instructions (Thermo Fisher Scientific; R37601). Calcein and propidium iodide were excited at 488 and 561 nm laser wavelengths, respectively, and imaged on a confocal microscope (Zeiss LSM 780 Airyscan, Zeiss AxioObserver.Z1). Imaging analysis was performed on at least three images from each analyzed sample feature (≥3 samples per condition; ≥3 experimental replicates).
Immunofluorescence in 3D Constructs: After in vitro culture, the constructs were fixed overnight in a 4% paraformaldehyde solution at room temperature. After fixation, samples were rinsed in PBS (5 min, 3×), and stained for f-actin, MyoD, and Myosin Heavy Chain (MyHC). Briefly, samples were permeabilized with 0.1% Triton X-100 (20 min) and blocked with a 1% bovine serum albumin solution (1 h). After extensive washing, the constructs were stained with the primary Anti-MyoD Antibody (#ZRB1452, Sigma-Aldrich, dilution 1:1000) and incubated for 4 h with an anti-rabbit secondary antibody coupled with Rhodamine (#SAB3700846, Sigma-Aldrich, dilution 1:200). Cells were incubated in AlexaFluor 488 phalloidin (#R37110, Invitrogen) for 1 h. 4′-6-diamidino-2-phenylindol (DAPI) was finally used to label the nuclei (dilution 1:1000; 15 min). All incubations were performed under gentle shaking and at room temperature. To reveal HIF-1ɑ, the constructs were stained with anti-HIF-1ɑ antibody (ab179483, Abcam, Cambridge, UK), which was used at 1:50 dilution with goat anti-rabbit IgG (Alexa Fluor 488) secondary antibody. The staining of MyHC was performed by incubating the samples for 4 h with the mouse anti-MyHC antibody (Myosin 4, eFluor 660, Clone: MF20, Affymetrix eBioscience), which was used at a 1:10 dilution for a final concentration of 10 μg mL −1 . All images were taken using a confocal microscope (Zeiss LSM 780 Airyscan, Zeiss AxioObserver.Z1, ScopeM), using the same exposure time, and were analyzed in Fiji ImageJ. Differentiated myotubes in a specific microscopic field were observed under ×10 and ×20 magnification. Either the total number of nuclei or the number of nuclei within MyHC-positive myotubes was counted in 5 fields per sample. The fusion index was calculated as a ratio between the number of nuclei within MyHC-stained myotubes and the total number of nuclei. The images were randomly selected from at least three regions from each sample, and the experiments were replicated three times. The alignment of myotubes was evaluated by measuring the angle shift between the long axis of the myotube and the y-axis direction (corresponding to the longitudinal axis of the constructs) by using ImageJ. [62] Nuclear Extraction and Transcription Factor Assays: Constructs were lysed and processed for nuclear and cytoplasmic fractionation according to manufacturer instructions (Abcam, ab113474). Cell-laden hydrogels were mechanically dissolved by pipetting, and resuspended in the pre-extraction buffer. Sample solutions were then centrifuged to remove the cytoplasmic fraction first, and then centrifuged a second time before being resuspended in the nuclear extraction buffer to obtain the nuclear fraction. The HIF-1 transcription assay was performed according to the manufacturer's instructions (Abcam; ab133104).
Histology: After in vitro culture, the constructs were fixed overnight in a 4% Paraformaldehyde solution at room temperature. Then, the samples were embedded in paraffin and sections of 4.5 μm thickness were cut with a microtome (Microm, HM430, Thermo Scientific). The sections were stained with hematoxylin-eosin (#GHS116 and #HT110116, respectively from Sigma-Aldrich). The slides were mounted, and imaged in bright field with a light microscope (Olympus CKX41, Olympus Schweiz AG).
Scanning Electron Microscopy: The samples were fixed with 2.5% (v/v) glutaraldehyde in PBS solution at room temperature for 30 min and then rinsed three times with PBS. Then, the samples were dehydrated in an ascending series of ethanol solutions (30%, 50%, 70%, 90%, and 100% (v/v); 5 min in each solution), followed by three incubations of 10 min each in 100% ethanol, dried over a molecular sieve. After transfer into metal capsules, the samples were inserted into a critical point dryer (Tousimis 931) and the ethanol was substituted against liquid CO 2 . Then the samples were dried over the critical point of CO 2 (31°C/73.8 bar). After the pressure was slowly released, the samples were taken out and mounted on SEM-stubs. For conductivity, the samples were sputter coated with 5 nm Pt/Pd (Safematic CCU-010). The examination was done in a JSM-7100F JEOL SEM at 3 kV by secondary electron detection.
Metabolic Activity in 3D Constructs: To evaluate cell growth within the scaffolds, the AlamarBlue assay was performed to determine the metabolic activity of seeded cells and confirm their proliferation. Briefly, the culture medium was supplemented with 10% (v/v) of AlamarBlue reagent solution at 0.1 mg mL −1 (#R7017, Sigma-Aldrich) and incubation was carried out for 2 h at 37°C before the fluorescence signal (Ex/Em = 530/590) was measured with a Synergy H1 microplate reader (Biotek). Fluorescence intensity values were corrected for the background control (culture medium with resazurin).
Measurement of Creatine Kinase (CK) Activity: The CK activity was used as a marker of myogenic differentiation. [6] To assess CK activity in the constructs, a small amount of tissue (5 mg) was dissected from the constructs at the end of maturation time (day 15) and tested via a colorimetric CK Activity Assay Kit (Ab155901, Abcam, Cambridge, UK) according to the manufacturer's instructions. The tissue was washed with cold PBS, resuspended in 100 μL of ice-cold CK Assay Buffer, and homogenized with ten passes. Insoluble materials were removed by centrifugation (5 min, 2000 rpm, at 4°C). Supernatants were collected and plated at different dilutions into a 96-well plate. The Reaction Mix was added to each standard and sample well, and Background Reaction Mix was added to the Background control sample well. The output in optical density (OD) at 450 nm was measured on a microplate reader in kinetic mode for 10 min at 37°C. CK activity was measured two times at 5 min intervals and each assay was performed in duplicate. The amount of NADH generated by CK during the reaction time (ΔT) was calculated from the OD as follows where B = amount of NADH in the samples well calculated from a standard curve (nmol), T = reaction time (minutes), V = original sample volume added into the reaction well (mL), and D = sample dilution factor. Secretome Analysis: The amount of IL-4, IL-6, IL-12, IL-13, BDNF, VEGF, BMP-4, and Myostatin in the liquid media of constructs was quantified by specific ELISA, according to the manufacturer's protocol. Briefly, media aliquots were collected at different time points and added to an ELISA kit plate coated with an antibody against the protein of interest (capture antibody), then a biotinylated secondary antibody (detection antibody) was added and the plate was incubated at room temperature for 2 h. The reaction catalyzed by horseradish peroxidase was stopped by adding 1 m sulfuric acid and absorbance was measured at 450 nm. Mouse IL-4, IL-6, and IL-13 ELISA kits were purchased from R&D system (Minneapolis, MN, USA; product numbers: M4000B, M6000B, M1300CB). IL-12 and VEGF ELISA kits were purchased from Thermo-Fisher Scientific (Waltham, MA, USA; Product Number: BMS616). BMP-4, Myostatin, and BDNF ELISA kits were purchased from LS Bio (Seattle, WA, USA; product numbers: LS-F23502, LS-F35789, LS-F23504).
Tensile Test: To determine the elastic modulus (E) and strain at the break of the constructs during maturation, the samples were examined with a uniaxial tensile testing machine (Instron 5942, Instron, Norwood, MA, USA) at room temperature. The constructs were washed, and their extremities were then carefully glued to two pieces of paper with a strong tissue adhesive glue. The papers were folded around the construct that was then carefully positioned in a vertical position on the sample holder of the tensile machine. The samples were characterized at a strain rate of 0.01 s −1 up to rupture and break. The stress-strain response data were plotted and the slope of the linear section (elastic deformation region) of the response was used to extract the elastic (Young's) modulus of the specimen.
Drug Distribution Assay: To perform a drug distribution study, the constructs were first incubated with FITC-dextran (Sigma-Aldrich, 46944) and then studied the drug internalization and its progressive release in the media. Briefly, the FITC-dextran (average mol. wt 40 kDa) was reconstituted into PBS, diluted (100×), and then filtered (200 μm filters, Corning). The liquid media of the constructs was replaced with the drug solution (0.1 mg mL −1 ) and the constructs were left in incubation for 60 min at 37°C . Confocal imaging was performed on some of the samples to observe the drug distribution during the incubation time ( ex 488 nm). After the incubation, the samples were extensively washed with PBS and then placed back in the culture media. To quantify the progressive release of the drug from the constructs, the media was then sampled over time and analyzed with a Synergy H1 microplate reader (Biotek) for fluorescent emission ( ex 493 nm; em 517 nm). Fluorescence intensity values were corrected for the background control (culture medium) and graphed. To assess drug loading, a small (≈1 × 1 × 1 mm 3 ) fragment of the construct was dissected with microscissors and disaggregated in the liquid medium until complete dissolution. Fluorescent emission from the sample was measured and converted to molecular quantities with a standard curve realized from GelMA-NaAlg bioink mixed with FITC-dextran. To study drug release after localized injection, a 27 G needle syringe (BD Science, Qume Dr San Jose, CA, USA) was used to slowly inject the central area of the constructs with a small volume (≈30 μL) of two different drug molecules: FITC-dextran (1 mg mL −1 ) and Rhodamine B (Sigma-Aldrich, R6626; 1 mg mL −1 ). Drug release was assessed by measuring the fluorescence emission of culture media over time (for Rhodamine: ex 493 nm; em 517 nm) and converting it to molecule content via a standard curve. To be graphed in the kinetics study, the drug content was expressed as a percentage of the injected amount.
Replication of Construct Deformation: The passive force exerted by the muscle constructs on the pillars during tissue development was estimated via a custom-made cantilever system. The culture template composed of an 8 mm tick layer of agar gel (1.5% w/v agarose, Sigma-Aldrich) and filledin P10 pipette tips were built in a petri dish ( Figure S31, Supporting Information). The dish was then fixed in a vertical position to orient the pillars horizontally. Pressure was applied to the central part of the tip to achieve a later displacement of the pillar of ≈1.5 mm (which was observed on real constructs). The whole set-up was placed on a scale, and weight values were taken before and after the application of pressure. The force value was estimated from the weight difference between the two conditions. Data Analysis and Statistics: Data were analyzed with the Graph Pad Prism 9 Software. All variables are expressed as mean ± standard deviation (SD). Data have been acquired from at least four independent experiments and three technical replicates, unless otherwise stated. Samples (i.e., bioprinted constructs) were at least three per category. To assess statistically relevant differences between the two experimental groups, the t-test was used (p < 0.05 and p < 0.01 are expressed as * and **, respectively). A general linear two-way ANOVA test was used to verify the hypothesis of whether there were changes in various parameters over time among the experimental groups and to identify relevant variations among several experimental groups. For all experiments, at least three samples were used to study each condition, and at least three experimental replicates were performed.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author. biomolecular assays; performed analysis; wrote the manuscript. O.Y. performed rheology and synthesis of GelMA. J.G. and L.S. performed the bioprinting and material characterization. R.G. designed muscle actuators and optimized the bioprinting. A.B. performed COMSOL simulation of the muscle shrinkage. R.K. conceived the study; came up with the idea of building biohybrid muscles. M.F., O.Y., and R.K. revised the manuscript. All authors approved the final draft of the manuscript.

Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.; Keywords bioactuators, bioinks, biointerfaces, bioprinting, skeletal muscle tissue engineering