In general, uniaxial cell alignment is a representative morphological feature of the skeletal muscles that can be induced by applying shear (Mozetic et al., 2017; Kim and Kim et al., 2019; Kim et al., 2019) and tensile force (Choi et al., 2016; Agrawal et al., 2017; Cvetkovic et al., 2017; Das et al., 2019; Kong et al., 2021). The shear flow inside a cylindrical nozzle induces shear-thinning of the polymer solution. The wall shear forces exerted on the shear-thinning polymer induce the alignment of inherent polymer molecules (Compton and Lewis, 2014; Schwab et al., 2020). In particular, elongational forces are produced in the stretching region of the extruded polymer that can further accelerate the alignment and Kim et al. (2019) studied the shear-induced cell-laden collagen alignment and demonstrated that a fibrillating buffer solution (KCl and l-glycine) incorporated in the bioink induces the formation of collagen fibril bundle, closely mimicking its native structure (Kim and Kim et al., 2019). Based on their previous research, Kim et al. (2019) included gold nanowire as a nanofiller that was stably aligned during the extrusion process and strongly aligned upon electric field stimulation (Figure 1A). In addition, the orientation of the distribution of nanowires were manipulated under varying volume flow rates and nozzle speeds (Kim et al., 2019).

In order to apply tensile force, the muscle cell-laden hydrogels are confined around the two micropillars or posts. Upon tissue maturation and cell-mediated gel compaction, the hydrogel remain anchored to the posts and the cells align along the longest axis of the system. Choi et al. (2016) induced uniaxial muscle cell alignment by depositing polycaprolactone (PCL) anchors as a geometric constraint at both ends of the muscle construct (Figure 1B). The PCL and cell-containing mdECM bioink were 3D printed using an integrated composite tissue/organ building system (Choi et al., 2016). Alternatively, Agrawal et al. (2017) fabricated a hydrogel pillar using a photopatterning method. After injecting the hydrogel precursor solution into an inlet of the microfluidic device, a prefabricated photomask was positioned on the bottom of the device, followed by exposure to ultraviolet (UV) radiation, which resulted in the formation of two circular pillars. Thereafter, the muscle cell-laden gelatin methacrylate solution was injected into the device for photopolymerization, which created patterns between the cells and around the pillars (Agrawal et al., 2017).

In native skeletal muscle, a single fascicle comprises multiple muscle fibers arranged in a bundled form. Several biofabrication techniques have attempted to mimic the muscle bundle-like structures for recapitulating the hierarchical arrangement ranging from a single muscle fiber to a fascicle. This physiological aspect can be achieved as follows: fabrication of bundle-shaped 3D scaffold (Kim et al., 2018a; Kim et al., 2020), 3D bioprinting cell-laden construct with built-in channels (Miller et al., 2012; Kang et al., 2016; Kolesky et al., 2016; Kim et al., 2018b), and 3D embedded printing (Bertassoni et al., 2014; Choi et al., 2019; Skylar-Scott et al., 2019; Gao et al., 2021).

A 3D scaffold comprising multiple polymeric struts or fibers adjacent to each other has been fabricated to mimic the bundle structure. Kim et al. (2018a) demonstrated shear-induced poly (vinyl alcohol) (PVA) alignment during the printing of PVA/PCL mixture (Figure 2A). Subsequent leaching of PVA yielded aligned micropatterns on the surface of the PCL scaffold, which was seeded with muscle cell-laden collagen solution (Kim et al., 2018a). Alternatively, Jin et al. (2021) fabricated the thermally drawn PCL fiber-based microchannel scaffolds (Figure 2B). After the scaffolds were soaked in the skeletal muscle dECM/muscle cell mixture and gelated at 37°C, the nuclear elongation was observed along the orientation of fibers, indicating the cell-guiding effect of the aligned fiber structures. In addition, the highly porous scaffold surface obtained from the salt-leaching method allowed the absorption of the dECM pre-gel solution into the scaffolds, which was evident from the presence of collagen mesh-like structures on the inner and outer sides of the fiber (Jin et al., 2021).

Microchannel-guided formation of aligned units has been utilized in a broad range of target tissues, including bone, nervous tissues, and vasculatures (Daly et al., 2018; Huang et al., 2018; Lee et al., 2020; Hwangbo et al., 2021). Sacrificial hydrogels—deposited as a temporary structure for structural support or generation of certain structures—are often removed by variations in temperature once the crosslinking of the engineered construct is complete; therefore, they have been widely adopted in 3D bioprinting for establishing built-in channels. In addition, thermoresponsive sacrificial polymers (i.e., gelatin and Pluronic F-127 (PF-127)) are printed as a temporary structure that dissolve in response to temperature variations after the bulk hydrogel has been crosslinked. Kang et al. (2016) printed supporting PCL pillars and sacrificial hydrogel patterns between cell-laden hydrogel patterns to yield a muscle bundle-like structure. Moreover, stretching and aligning muscle cells were observed inside these bundle structures 3 days after printing structure (Kang et al., 2016). Subsequently, Kim et al. (2018b) developed a larger construct (mm³–cm³) with multilayered muscle bundles to utilize cells from human biopsies and assess if the construct was applicable for severe muscle defect replacement (Figure 2C; Kim et al., 2018b).

Furthermore, 3D embedded printing utilizes a recently developed printing method named freeform reversible embedding of suspended hydrogel (FRESH). The major limitation of soft biomaterials such as ECM hydrogels is attributed to their poor shape-retention during the printing process that renders the desired resolution and 3D geometries unachievable (Hinton et al., 2015). The thermoreversible support bath prevents the collapse of the printed hydrogels and is subsequently melted away after the printing process. Moreover, complex structures such as branched tubular networks and trileaflet heart valves have been fabricated using the FRESH method (Bhattacharjee et al., 2015; Lee et al., 2019). Based on this method, Choi et al. (2019) directly printed a bundle-like structure in a granule-based support bath. Initially, they prepared the support bath by blending an 8% (w/v) gelatin solution to generate gelatin granules. Thereafter, mdECM struts were directly printed into the support bath with horizontal and vertical distances of 200 µm from each other. Subsequently, PVA was added into the gelatin granule to induce rapid polymerization of mdECM bioink during its extrusion (Choi et al., 2019).

To date, several vascularization strategies have been developed to recapitulate the native vascular function (i.e., nutrient and oxygen supply to the highly oxygen-consuming organ) in the 3D-engineered muscle. The coaxial printing method and direct ink-writing method have been studied for the vascularization of engineered muscle as well as engineered tissues of other types.

Among the conventional vascularization methods, the hydrogel mold-casting method can generate hollow channel structures that are subsequently seeded with endothelial cells (ECs), forming endothelialized lumen structures (Bersini et al., 2018; Osaki et al., 2018; Redd et al., 2019). Osaki et al. (2018) observed emergent angiogenic sprouting from the surface of the endothelialized channels towards the muscle bundle between two channels, which mimicked myokine-regulated muscle angiogenesis. However, this method requires sequential stereolithography-based steps for obtaining the micropatterned mold. The multiple channel structures should be individually obtained to allow active diffusion of nutrients, thereby adding complexity to the mold-preparation step.

In general, 3D printing enables the iterative one-step fabrication of composite structures. For instance, the coaxial printing method enables direct fabrication of two- (Zhang et al., 2016; Gao et al., 2017; Choi et al., 2019) or three- (Gao et al., 2019; Gao et al., 2021) layered 3D conduits of varying geometries and dimensions (Figure 3A). Zhang et al. (2016) bioprinted a 3D microfibrous scaffold structure using a composite bioink comprising alginate (GelMA) and photoinitiator. The alginate in the core part of the nozzle was exposed to a chemical crosslinking agent (Ca²⁺) provided from the shell part of the nozzle during the printing process. The printed scaffold manifested mechanical stability after the UV crosslinking of GelMA. Ultimately, the cardiomyocytes were seeded onto the stiffened microfibrous scaffold after the bioprinted ECs self-assembled into a layer of confluent endothelium, thereby resembling the structure of native myocardium (Zhang et al., 2016). In another research, Choi et al. (2019) fabricated a compartmentalized core–shell structure in which a layer of ECs surrounded the muscle fiber-like struts, thereby closely mimicking the hierarchical architecture of vascularized muscle. Within a coaxial nozzle, the syringe containing the muscle cell-laden mdECM bioink and the syringe containing the EC-laden vascular dECM bioink were attached to the core and shell parts, respectively. In comparison to the muscle-only construct, the higher twitch and tetanic peak forces of the vascularized muscle confirmed its enhanced maturity (Choi et al., 2019).

The direct ink writing method proposed by Skylar-Scott et al. (2019) enables the fabrication of complex vasculature in a cell-filled support bath via 3D embedded printing (Figure 3B). This method is novel owing to the high cellular density of the support bath (up to 10⁸ cells/ml) (i.e., living tissue matrix) which is the prerequisite for physiologically relevant tissue models. The living tissue matrix exhibited a self-healing, viscoplastic behavior that allowed the printing of sacrificial ink in the desired network structure without crevasses formation on the printing path. Subsequently, the temperature-mediated melting of sacrificial gelatin ink and the simultaneous stiffening of the matrix yielded hollow vasculature networks, followed by the perfusion of EC suspension (Skylar-Scott et al., 2019). Based on a similar approach, Neufeld et al. (2021) fabricated a 3D glioblastoma model with a perfusable vasculature. Initially, the PF-127-based vascular bioink was deposited on the surface of cancer cell-laden fibrin bioink, followed by casting another layer of fibrin bioink onto the vasculature. Once the fibrin bioink achieved the desired physiological stiffness, the PF-127 was washed out with cold PBS, and a mixture of ECs and pericytes were injected into the vascular channels (Neufeld et al., 2021). Overall, the direct ink-writing method allows freeform fabrication of tissue-specific vascular structures. However, the mechanical properties of the support bath inks and vascular inks should be carefully manipulated in combination to achieve the desired geometry of vasculatures.

Motor neuron innervates the skeletal muscle via neuron–myofiber contacts or neuromuscular junctions (NMJs), which mediate the conversion of neuronal signal into muscle contraction. Functional NMJs have been engineered by multiple biofabrication methods including 3D co-culture (Morimoto et al., 2013; Bakooshli et al., 2019) of neuron and myofibers, and 3D bioprinting of structures with a spatial organization that induces spontaneous neurite formation (Cvetkovic et al., 2017; Kim et al., 2020; Kong et al., 2021).

In particular, the neurons can self-organize to form NMJs in the 3D co-culture of motor neurons (MNs) and muscle cells. Multiple biofabrication skills, including the usage of patterned stamps and molds, facilitate the integration of MN spheroids with differentiated myofibers. Morimoto et al. (2013) used polydimethylsiloxane (PDMS) stamps for patterning Matrigel to form aligned muscle bundle, and subsequently, added mouse neural stem cell-derived spheroids (mNSCs) onto the muscle bundle. The mNSCs were successfully differentiated into motor neurons, and acetylcholine receptors were formed at the muscle bundle-spheroid contacts. Moreover, the muscle bundle contracted unidirectionally in response to glutamic acid activation of neurons, indicating the formation of functional NMJs (Morimoto et al., 2013). Based on a similar approach, Bakooshli et al. (2019) prepared a dumbbell-shaped PDMS mold that was seeded with a mixture of human pluripotent cell-derived MN spheroids and human muscle progenitor cells suspended in fibrin-based hydrogel (Figure 4A; Bakooshli et al., 2019). The expression of NMJ-localized protein, such as synaptic MuSK protein, confirmed the recapitulation of early NMJ synaptogenesis in this model.

In addition, 3D bioprinting technology facilitates the precise positioning of multiple cell types in a desired organization and array. Kim et al. (2020a) generated a functional NMJ inside a 3D-printed multilayered bundle-shaped construct containing a hollow microchannel structure (Figure 4B). Fibrinogen/gelatin/hyaluronic acid/glycerol bioink containing human muscle progenitor cells and human neural stem cells (hNSCs), acellular sacrificial bioink, and a supporting PCL pillar were deposited using multidispensing modules. Consequently, βIIIT⁺ neurites contacting the acetylcholine receptor clusters were formed on the myofibers, which confirm the formation of NMJs (Kim et al., 2020). In an alternative approach, Cvetkovic et al. (2017) used a stereolithography-based 3D-printing technique to fabricate a modular co-culture system (Figure 4C). A 3D tissue ring, previously formed by integrating two layers comprising differentiated myofibers and motor neuron embryoid bodies (EBs) in a 3D-printed ring mold, was transferred to the hydrogel skeleton to mimic the physiological muscle–tendon–bone arrangement. This co-culture arrangement exhibited the extension of the neurite toward the differentiated muscle and the formation of the acetylcholine receptor cluster (Cvetkovic et al., 2017).

Contractility is one of the most widely accepted muscle functional readouts for distinguishing between normal and diseased muscle phenotypes (Sharples et al., 2012; Juhas et al., 2018). In context, the early detachment of spontaneously contracting myofibers allowed only partial contractility analysis in 2D cultures, which were confined to the variations in related gene expressions (Nikolić et al., 2012; Tarum et al., 2017). Various external stimuli have been applied to improve the contractility of the 3D-engineered muscle. These stimuli aim to mimic the native stimuli the developing skeletal muscle receives (Westerga and Gramsbergen, 1994; Eken et al., 2008), which induce muscle hypertrophy. In general, electrical stimulation and mechanical loading are frequently utilized in combination.

First, electrical stimulation is applied as a surrogate of neural input that are delivered to individual myofibers to induce muscular contraction. In addition, the regenerating muscles in the self-repair process require electrical signals from motor neurons, implying that the electrical stimulation can enhance the function of the engineered muscle (Kim et al., 2019). Upon applying the electrical stimulation, the pulse energy and width should be optimized to minimize the tissue damage (Khodabukus and Baar, 2012). Grosberg et al. (2011) applied an electrical field stimulation for cantilever-type thin muscular films and measured the transmembrane potential of cardiomyocytes. This revealed an action potential morphology, indicating the contractility achieved in the tissue (Grosberg et al., 2011). Additionally, Khodabukus et al. (2019) demonstrated that a 3D-engineered muscle electrically stimulated with 1 and 10 Hz of frequency for 1 week resulted in an increased contractile force (nonstimulated: 9.1 ± 0.38 mN mm⁻², 1 Hz: 19.3 ± 0.63 mN mm⁻², 10 Hz: 18.9 ± 0.69 mN mm⁻²). More importantly, the stimulation of 1 Hz enabled the highest specific force generation among the outputs of engineered human muscle reported to date (Khodabukus et al., 2019).

On the other hand, mechanical loading mimics the consistent passive stretch exerted upon the embryonic muscle. Kim et al. (2019) reported that combined electrical and mechanical stimulation induced the remodelling of ECM network structures in engineered muscle tissue, such that the force transmission was facilitated without impeding muscle contraction. In particular, 20 min of alternating electrical and mechanical stimulations resulted in 31% higher contractile force (Kim et al., 2019). Moreover, Aguilar-Agon et al. (2019) studied the influences of mechanical loading on anabolic shift based on the results of in vivo resistance exercise regimens. Insulin-like growth factor-1, which is associated in skeletal muscle growth, was upregulated 21 h after the cessation of mechanical load (15% elongation of engineered construct); after 45 h from cessation, the atrophic gene MAFbx underwent significant downregulation. In addition, maximal isometric force exhibited an increase of 140% (21 h) and 265% (45 h) as compared to nonloaded constructs (Aguilar-Agon et al., 2019). Collectively, the electrical stimulation and mechanical loading could result in significant muscle hypertrophy and increase in force generation, which can potentially enable the maturation of engineered muscle to a similar extent as that in human adult muscle.

Muscular dystrophies arise from mutations in genes encoding various proteins such as muscle ECM, cytoplasmic and plasma membrane proteins that result in progressive muscle weakness or wasting (Tedesco and Cossu, 2012). The most common and severe type of dystrophy, DMD, is an X-linked recessive disorder that occurs almost exclusively in males. The underlying mechanism of DMD is attributed to the mutation in the gene encoding dystrophin, a membrane-bound cytoskeletal protein that links the intracellular actin cytoskeleton to extracellular matrix. Therefore, the absence of dystrophin destabilizes the membranes and causes consistent muscle fiber damage during contraction, which results in consistent fiber degeneration (Nowak and Davies, 2004; Allen and Whitehead, 2010).

Thus, patient-derived cells are cultured under uniaxial tension or micropatterned substrates to form a 3D myofiber for modeling DMD. During and after the culture, the cells are screened for degenerative morphologies such as less sarcomeric α-actinin organization and decreased contractile force. Nesmith et al. (2016) fabricated a thin, contractile muscular film composed of DMD patient-derived cells in a cantilevered design by printing an array of thin bands using a microcontact printing method (Figure 5A). Under electrical stimulation with increasing frequencies ranging from 1 to 4 Hz, the dynamic stress on the healthy tissues on day 6 was at least one order of magnitude higher in comparison to the diseased tissues, as evaluated based on the variations in the radius of film curvature. The results implied that the impaired force transmission caused by the absence of dystrophin was recapitulated in this system (Nesmith et al., 2016). Subsequently, Khodabukus et al. (2019) used fibrin-based 3D “myobundles” confined to rectangular Teflon frames to demonstrate the correct membrane-bound localization of dystrophin. Moreover, the electrical stimulation increased the dystrophin, myosin heavy chain, and sarcomeric α-actinin content (Khodabukus et al., 2019). Overall, the 3D-engineered muscle models can represent more adult-like phenotypes as compared to the developmentally less matured cytoplasmic localization of dystrophin in conventional 2D culture (Serena et al., 2016).

Maffioletti et al. (2018) applied a tensile force to fibrin-based 3D engineered tissue and generated a multilineage patient-specific model using a patient-derived iPSC of laminopathy (LMNA), which is one of the MD types (Figure 5B). LMNA results from the mutation of the LAMIN A/C gene encoding a protein that composes the nuclear lamina. Under uniaxial tension, hiPSCs were successfully differentiated into myofiber-resembling structures. Notably, the genotype–phenotype correlation was achieved as hiPSCs displayed significant nuclear abnormalities such as nuclear elongation, which was absent in normal muscles and other dystrophic hiPSC lines (Maffioletti et al., 2018). Moreover, Bersini et al. (2018) demonstrated a multicellular DMD model composed of cells from human biopsy to mimic the aberrant fibrosis. They fabricated a spatially organized mesoscale system where differentiated myofibers were surrounded by 3D microvascular network and a fibroblast layer. In particular, collagen and fibronectin secretion levels were significantly higher in patient-derived fibroblasts as compared to normal and TGF-β1 treated fibroblasts. This trend was not displayed in 2D triculture system, indicating that the 3D mesoscale system supported more physiologically relevant fibrosis progression such as muscle-specific fibroblast recruitment. Furthermore, the α-smooth muscle actin, which is expressed in the fibroblasts residing in the dystrophic muscle, was expressed in high levels in the DMD fibroblasts, thereby indicating an accurate representation of the DMD-affected tissue microenvironment (Bersini et al., 2018).

Skeletal muscle self-repair in response to mild injuries is modulated by the interaction between the muscle-resident satellite cells and the immune system via the process of satellite cell activation, proliferation, and differentiation. However, excessive and unregulated inflammatory responses are associated with muscle loss and weakness, which has been reported in several types of diseases including sarcopenia, cachexia, chronic obstructive pulmonary disease and rheumatoid arthritis. In particular, several clinical studies have reported that elevated IFN-γ levels in autoimmune diseases and chronic inflammation contribute toward muscle wasting.

Based on a 3D fibrin-based engineered cardiac muscle tissue composed of human primary cells, Chen et al. (2021) assessed the influence of exercise on IFN-γ-induced muscle wasting (Figure 5C). In this model, muscle atrophy was characterized by weaker and slower muscle contraction and reduced calcium-handling protein expression. Notably, the exercise-mimetic electrical stimulation with IFN-γ treatment rescued the sarcomeric protein organization and elongated the myotube, thereby implying an attenuated muscle atrophy condition. In particular, inhibited expression of IFN-γ-induced JAK (Janus kinase)/STAT (signal transducer and activator of transcription) pathway downstream proteins suggested the previously unknown, exercise-mediated cell autonomous anti-inflammatory response that occurs in the muscle tissue, besides the paracrine signalling which has been traditionally suggested as the fundamental mechanism. Furthermore, prevention of muscle atrophy after treatment of FDA-approved small-molecule inhibitors of JAK/STAT pathway confirmed the utility of this platform for preclinical drug screening (Chen et al., 2021).

Juhas et al. (2018) established a 3D co-culture system based on previously referred 3D engineered muscle to assess the role of macrophages in skeletal muscle regeneration process. An inhibited endogenous self-repair process was successfully mimicked as progressive myofiber degeneration was induced after cardiotoxic injury, which resembled a severe muscle injury. Then a robust regenerative response was induced upon incorporating bone-marrow derived macrophages with lower levels of activated caspase 3, which indicated that the anti-apoptotic paracrine signalling between the muscle and macrophages were mimicked in this model. Furthermore, the implantation of macrophage-incorporated engineered muscles in a mouse dorsal window-chamber model resulted in robust blood vessel ingrowth and functional muscle regeneration. As such, the in vitro incorporation of macrophages can be considered as an effective approach for preconditioning an implantable construct in a pro-regenerative milieu before implantation (Juhas et al., 2018).