Modeling human skeletal muscle with microphysiological systems: a narrative review
Article information
Abstract
Purpose
Conventional 2-dimensional cultures and animal models have limited ability to reproduce the structural complexity, dynamic mechanical cues, and sustained functionality of native skeletal muscle tissue. To overcome these limitations, skeletal muscle-on-a-chip platforms have been developed as advanced in vitro systems for studying muscle physiology, pathology, and regeneration.
Current Concepts
These microengineered systems incorporate essential features of skeletal muscle, including 3-dimensional architecture, cellular alignment, contractile function, and responsiveness to biochemical and mechanical stimuli. Recent advances, such as vascularization, multi-organ integration, and spaceflight-compatible designs, have expanded their applications in disease modeling and drug screening.
Discussion and Conclusion
This review examines key engineering strategies, biological performance metrics, and representative applications of skeletal muscle-on-a-chip systems. It also addresses technical challenges, including long-term functionality, measurement standardization, and clinical translation, and considers future prospects for their integration into preclinical testing and regenerative medicine.
Introduction
Microphysiological systems (MPS) are engineered in vitro models designed to replicate the structure and function of living tissues under physiologically relevant conditions [1]. Typically, they consist of cells or tissues of human or animal origin maintained within a microenvironment engineered to simulate the biochemical, mechanical, and architectural features of native organs. These systems often incorporate microfluidics, biomaterials, and mechanical or electrical stimuli to reproduce the dynamic cues encountered in vivo [2]. Organs-on-a-chip (OoCs), which are miniaturized, chip-based platforms that use microfabricated chambers and controlled fluid flow to model organ-level architecture and function, comprise a key subset of MPS [3].
The development of MPS has been driven by the limitations of traditional 2-dimensional (2D) cell cultures and the translational gap associated with animal models [4]. Although animal studies have historically formed the basis of preclinical research and drug development, interspecies differences in drug metabolism, immune signaling, and cellular organization often limit their predictive value for human outcomes [5]. In contrast, MPS offer human-relevant platforms that support long-term culture, enable real-time monitoring, and allow quantitative and multiplexed functional readouts [6–8]. These capabilities have led to their growing adoption in drug screening, toxicity testing, disease modeling, and regenerative medicine [9,10]. Their emergence has also been supported by regulatory frameworks that increasingly recognize MPS as scientifically valid non-animal alternatives in certain preclinical contexts.
Skeletal muscle-on-a-chip is a specialized application of MPS technology aimed at modeling the structure and function of human skeletal muscle tissue. Accounting for approximately 40% of total body mass, skeletal muscle plays vital roles in locomotion, posture, metabolism, and systemic homeostasis [11]. However, conventional models are limited in their ability to accurately reproduce muscle tissue structure and function. Skeletal muscle-on-a-chip systems address these shortcomings by recreating the aligned architecture, contractile properties, and responsiveness of native muscle within a miniaturized, controllable device [12]. These platforms typically use muscle precursor cells embedded in biomimetic hydrogels, anchored between flexible posts or micropatterned structures, to guide the formation of 3-dimensional (3D) aligned myotubes [13]. Functional parameters such as force generation, calcium transients, and metabolic activity can then be monitored in real time [14].
Recent advances have expanded the capabilities of muscle-on-a-chip systems by incorporating vascular networks, integrating with other organ models, and adapting them for spaceflight experiments to study disuse-induced atrophy (Figure 1) [15]. These platforms are now used to evaluate drug efficacy and toxicity, model muscle-related diseases such as Duchenne muscular dystrophy (DMD) and sarcopenia, and investigate strategies for tissue regeneration. This review outlines the engineering principles and biological characteristics of skeletal muscle-on-a-chip platforms, presents representative applications, and explores future directions toward clinical translation and personalized medicine.
Overview of skeletal muscle on a microphysiological system (MPS), created with support from the Medical Illustration & Design (MID), Yonsei University College of Medicine. The platform supports vascularization, multi-organ integration, real-time monitoring, external stimulation, and disease modeling. It recapitulates native muscle structure, providing a versatile tool for skeletal muscle tissue engineering and therapeutic screening.
Key Features of Organ-on-a-Chip Systems
OoC platforms combine microfluidic technology with tissue engineering to recreate key aspects of organ-level function in vitro. Unlike traditional static cultures, these devices integrate dynamic flow, spatial compartmentalization, and tissue-specific mechanical and biochemical cues [16]. These features allow researchers to model tissue physiology and disease in a more realistic, controllable, and scalable manner.
A defining feature of OoC platforms is the use of microfluidic channels to simulate vascular perfusion or interstitial fluid flow [17]. These channels, typically fabricated through methods such as soft lithography or 3D microprinting, enable precise control over geometry and flow dynamics. Continuous perfusion allows nutrient and gas exchange, waste removal, and the generation of spatial or temporal gradients of soluble factors—processes poorly replicated in static 2D or bulk 3D cultures. For instance, endothelialized channels can mimic vascularization, facilitating studies on barrier function, immune cell trafficking, and drug transport across tissue interfaces.
Material selection is another critical aspect of OoC design [18,19]. While polydimethylsiloxane (PDMS) is widely used due to its optical transparency, gas permeability, and compatibility with soft lithography, alternative materials such as thermoplastics and hybrid hydrogels are increasingly explored to address limitations in drug absorption and mechanical durability. Within the tissue chambers, cells are often embedded in biomimetic hydrogels such as collagen, fibrin, Matrigel, or gelatin methacrylate (GelMA), which replicate the extracellular matrix (ECM) and provide structural support. The stiffness, porosity, and biochemical composition of these hydrogels can be tuned to promote the differentiation, alignment, and functional maturation of various cell types.
For excitable or mechanically active tissues such as the muscle, heart, or lung, integrating mechanical and electrical stimulation is particularly important [12]. Many devices use stretchable membranes or pneumatic actuators to apply cyclic strain, simulating physiological motions such as breathing, peristalsis, or locomotion. In skeletal muscle-on-a-chip models, electrical stimulation, delivered through embedded electrodes or conductive coatings, is routinely applied to promote myotube maturation and induce contractions [13]. Real-time monitoring of electrophysiological activity can be achieved using microelectrode arrays and other biosensors, which provide quantitative, non-invasive readouts of muscle excitability, action potential propagation, and electromechanical coupling [14]. Some platforms incorporate optogenetic stimulation as an alternative to electrodes, enabling precise spatial and temporal control of contractile activation via light-sensitive ion channels [20].
Advances in engineering have also enabled the creation of modular, interconnected systems in which multiple tissue types, such as liver, gut, pancreas, and muscle, are cultured on separate chips but linked through a shared circulatory circuit [21]. These multi-organ platforms make it possible to investigate inter-tissue communication, systemic toxicity, and metabolic crosstalk under physiologically relevant flow conditions. For example, muscle-liver or muscle-pancreas combinations can be used to study endocrine signaling and metabolic regulation [22].
The functional versatility of OoC systems has made them increasingly valuable in drug development pipelines, both for assessing efficacy and identifying toxicity. Compared with conventional models, they offer superior replication of tissue-specific drug responses, reduced compound requirements, and compatibility with high-content imaging and automated assays. Their adoption is further supported by growing regulatory interest in non-animal testing methods and recognition of OoCs as tools that can improve the predictive accuracy of human drug response data.
Engineering Strategies for Skeletal Muscles-on-a-Chip
1. Implementation of engineered skeletal muscle tissues in organ-on-a-chip systems
Replicating the spatial organization of skeletal muscle is a central consideration in designing physiologically relevant muscle-on-a-chip systems. Native skeletal muscle consists of densely packed, highly aligned myofibers arranged into fascicles, and this anisotropic structure is critical for efficient contraction and directional force generation [23]. To reproduce this architecture in vitro, various microengineering strategies have been developed to impose biophysical constraints or provide anisotropic guidance cues to muscle precursor cells (Figure 2) [24–28].
Strategies and platforms for 3-dimensional (3D) skeletal muscle tissue modeling. (A) Schematic overview of the fabrication and transplantation of 3D skeletal muscle constructs using induced myogenic progenitor cells and decellularized skeletal muscle extracellular matrix (ECM). Adapted from Jin Y et al. Adv Funct Mater 2021;31:2006227, with permission [24]. (B) Schematic illustration of skeletal muscle organoid-on-a-chip fabrication. Adapted from Li J et al. Adv Funct Mater 2024;34:2401564, with permission [25]. (C) Schematic representation of a microfluidic multi-organ-on-a-chip with two chambers designed to co-culture skeletal muscle and pancreatic tissues, enabling inter-organ communication. Adapted from Fernández‐Costa JM et al. Adv Mater Technol 2023;8:2200873, according to the Creative Commons license [27]. (D) Left: Schematic representation of a microfluidic skeletal muscle chip showing media and gel channels. Right: Hydrogel compaction driven by cell contraction force under varying physical conditions, including differences in distance, size, or stress. Adapted from Kim J et al. Adv Funct Mater 2024;34:2410872, according to the Creative Commons license [28]. (E) Schematic illustration of mold design for vascularization in 3D muscle tissue engineering. Adapted from Wan L et al. Micromachines (Basel) 2020;11:907, according to the Creative Commons license [26].
One early approach employed 3D photopatterning to fabricate GelMA-based muscle bundles within a microfluidic device [29]. This method promoted uniaxial alignment of C2C12 myoblasts and allowed passive tension to be measured via pillar deflection. Upon differentiation, the constructs formed cylindrical, multinucleated bundles that exhibited progressive strain over time. The same system proved useful for modeling muscle injury, as cardiotoxin exposure produced dose-dependent architectural disruption and reduced tension output, underscoring its pharmacological testing potential. Another study utilized collagen type I tubes of varying diameters to guide myotube alignment within channel-like structures [30]. The study demonstrated that tube diameter strongly influenced myotube organization, with smaller channels (~75 μm) producing more fascicle-like arrangements than larger ones. Both the reproducibility of myotube formation and the degree of alignment were markedly improved in smaller-diameter constructs, illustrating the role of physical boundary conditions in shaping cytoskeletal organization.
In another strategy, anisotropic mechanical patterning was applied to decellularized muscle-derived ECM hydrogels polymerized under uniaxial strain on a stretchable PDMS substrate [24]. This process aligned fibrillar proteins and other ECM components, creating a biomimetic environment that directed primary muscle progenitor cell orientation and fusion. The resulting fascicle-like tissues exhibited improved alignment, sarcomeric organization, and contractility in vitro. When implanted in mouse models of volumetric muscle loss and DMD, these constructs promoted tissue regeneration and functional recovery, demonstrating translational potential for regenerative medicine.
An anisotropic organoid-on-a-chip model was also developed using PDMS boundary constraints combined with fibrin–Matrigel matrices to generate a uniaxial microenvironment conducive to aligned myofiber formation [25]. C2C12 myoblasts cultured in this system developed organized sarcomeric structures, higher fusion indices, and enhanced contractile kinetics under electrical stimulation. Myosin heavy chain isoforms and sarcomeric α-actinin were upregulated, and activation of YAP/TAZ and Rho/ROCK signaling pathways indicated mechanically driven muscle maturation. The platform was further adapted to model intermittent hypoxia by cycling tissues through 10%, 5%, or 1% O2. Mild to moderate hypoxia promoted adaptive hypertrophy and increased force generation, whereas severe hypoxia impaired contractility, altered calcium dynamics, and disrupted mitochondrial function. Antioxidant treatments such as N-acetylcysteine and resveratrol partially restored muscle performance and reduced oxidative stress, highlighting the model’s suitability for studying oxygen-related pathology and therapeutic interventions.
Functional validation of skeletal muscle-on-a-chip systems typically involves evaluating contractility, calcium transients, and electrical responsiveness. For precise control of muscle activation, optogenetically modified C2C12 myoblasts expressing channelrhodopsin-2 were embedded in collagen–Matrigel and organized into aligned 3D microtissues suspended between elastic microposts [31]. This configuration enabled real-time contractile force measurement via post deflection. Under pulsed blue light stimulation, tissues produced consistent twitch responses with ~20 ms latency and an average active force of 1.41 µN. The system also supported spatially targeted stimulation and complex actuation patterns, enabling applications in soft robotics and scalable functional assays.
Vascularization, an essential feature for replicating native muscle environments, remains underrepresented in current muscle-on-a-chip designs. One approach employed thermo-responsive sacrificial templates made from a polyester–paraffin wax mixture to create circular microfluidic channels within a collagen-based ECM scaffold [26]. These channels retained structural integrity during ECM remodeling induced by C2C12 differentiation and were later endothelialized with human umbilical vein endothelial cells to mimic capillary networks. The construct supported myotube formation between PDMS pillars and maintained structural stability after template removal, providing a viable strategy for integrating perfusable vasculature into 3D muscle tissues.
Expanding on this physiological complexity, other platforms have been developed to study inter-organ communication. In one example, electrically stimulated skeletal muscle tissues were co-cultured with pancreatic beta-cell spheroids (MIN6) within a microfluidic chip equipped with continuous perfusion and integrated plasmonic biosensors [27]. Upon stimulation, C2C12 muscle bundles secreted interleukin (IL)-6 in a time-dependent manner, triggering insulin release from the pancreatic compartment. Temporal correlation and control experiments confirmed IL-6 as the primary mediator. This configuration enabled time-resolved analysis of myokine–endocrine signaling in a controlled setting, providing a versatile platform for studying muscle–pancreas interactions relevant to metabolic diseases such as type 2 diabetes.
To improve standardization and reproducibility, another study systematically assessed how design and hydrogel parameters influence skeletal muscle-on-a-chip performance [28]. This work introduced a validated framework evaluating pillar geometry, gelation time, collagen–Matrigel ratios, and cell seeding density individually and in combination. Key performance metrics, such as muscle band compaction, weak-point thickness, contraction force, and response time under electrical stimulation, were quantitatively measured. Optimal conditions (2 mg/mL collagen with 0.82 mg/mL Matrigel, 30-minute gelation, and 6×106 cells/mL) produced compact, well-aligned tissue with the highest contractile strength (∼55 μN) and enhanced mechanical stiffness, as confirmed by tensile testing. Comparative biochemical and transcriptomic profiling revealed that optimized tissues upregulated genes associated with muscle development, mechanotransduction, and repair, whereas suboptimal tissues showed increased expression of atrophy- and fibrosis-related cytokines, including IL-6 and ALK1.
2. Disease modeling and therapeutic evaluation using muscles-on-a-chip
Skeletal muscle-on-a-chip platforms have shown significant utility in modeling genetic muscle disorders such as DMD (Figure 3) [32,33]). In one study, a micropatterned chip was used to co-culture dystrophin-deficient myoblasts from DMD patients with donor-derived mesoangioblasts, enabling controlled evaluation of cell-based therapeutic potential [34]. Mesoangioblasts produced substantially higher levels of dystrophin protein compared with donor myoblasts and achieved greater membrane localization, forming dystrophin-positive segments over 200 μm in length. This chip-based approach allowed quantitative assessment of functional protein restoration under defined mechanical and spatial conditions, providing a human-relevant in vitro model for evaluating stem or progenitor cell therapies for DMD.
Strategies for disease modeling using a skeletal muscle-on-a-chip. (A) Left: Schematic of the microfluidic platform used to generate engineered muscle tissues. Right: Representative image of a formed muscle tissue and immunofluorescence staining of muscle tissue. Red: myosin heavy chain; blue: nuclei. Top: Representative image of an old sedentary-derived myobundle; middle and bottom: images of young athletic-derived and old sedentary-derived myobundles, respectively. Adapted from Giza S et al. Aging Cell 2022;21:e13650, according to the Creative Commons license [32]. (B) Timeline and workflow for a skeletal muscle-on-a-chip experiment conducted aboard the International Space Station (ISS). Adapted from Parafati M et al. NPJ Microgravity 2023;9:77, according to the Creative Commons license [33].
Another study investigated age-associated muscle decline using donor-specific human satellite cells [32]. CD56⁺ cells isolated from young active (YA) and old sedentary (OS) donors were embedded in a collagen–Matrigel hydrogel and seeded into a PDMS microfluidic chip containing two cylindrical posts spaced 5 mm apart. The cells self-organized into aligned myobundles anchored between the posts and were maintained under perfusion. Chronic low-frequency electrical stimulation (3 V, 2 Hz, 2 ms for 5 days) was applied through integrated platinum electrodes. YA-derived tissues exhibited synchronous contractions and higher displacement amplitudes (~1 μm), whereas OS-derived bundles demonstrated reduced or asynchronous responses. Transcriptomic analysis revealed stronger upregulation of myogenic and hypertrophic markers, including ACTN3, TNNI2, and IGF-1, in YA constructs, while OS tissues displayed diminished responses and disrupted sarcomeric organization.
Skeletal muscle plays a crucial role in maintaining body shape and supporting essential physiological processes. Musculoskeletal function is influenced by multiple factors, including nutrition, infection, injury or trauma, and the accumulation of advanced glycation end-products (AGEs), which contribute to tissue degeneration, particularly in aging and diabetic populations [35]. A recent study employed a skeletal muscle-on-a-chip system composed of 3D muscle bundles generated from C2C12 myoblasts embedded in a collagen–Matrigel matrix and anchored between PDMS micropillars. Following AGE treatment, the tissues exhibited decreased force generation and increased stiffness, accompanied by transcriptional signatures indicative of ECM disruption and cellular stress. These findings underscore the pathological role of AGE accumulation in impairing muscle function, especially in the context of aging and diabetes-related muscle wasting.
Physiological deterioration caused by extreme environments such as spaceflight induces changes similar to those seen in age-related sarcopenia [36]. In one study, 3D muscle tissue chips derived from human muscle biopsies of young and aged male donors were cultured autonomously aboard the International Space Station (ISS) for 10 days [33]. Transcriptomic analysis revealed downregulation of genes involved in myogenesis, metabolism, and calcium signaling in flight-exposed tissues compared with ground controls, with more pronounced effects in aged donor-derived constructs. Specifically, expression of myosin heavy chain isoforms (MYH1, MYH2, MYH6), ACTN3, and MYL10 was significantly decreased in older samples under microgravity, indicating fiber-type shifts and reduced contractile potential. Genes associated with glycolysis, such as phosphofructokinase (PFKP) and phosphoenolpyruvate carboxykinase 2 (PCK2), were upregulated, suggesting a metabolic shift toward glycolytic pathways. These results highlight the value of muscle tissue chip platforms for modeling sarcopenia-like molecular alterations under microgravity and identifying age-specific therapeutic targets.
Environmental stressors such as spaceflight, which elicit physiological changes analogous to age-related sarcopenia, have also been examined using muscle-on-a-chip technology [37]. Engineered human skeletal muscle constructs were fabricated with aligned collagen nanofibrils seeded with primary muscle cells and launched to the ISS. After seven days in microgravity, transcriptomic and proteomic profiling revealed reduced myotube formation, downregulation of Wnt and Notch signaling pathways, and elevated markers of apoptosis and metabolic reprogramming, including increased GDF-15 and lipid metabolism–related genes, mirroring clinical features of sarcopenia. Gene set enrichment analysis confirmed overlap between the transcriptional signatures of microgravity-exposed constructs and sarcopenic muscle samples. Notably, treatment with IGF-1 or a 15-PGDH inhibitor partially reversed these changes, restoring expression of key genes involved in cell adhesion and myogenesis. These findings further support the use of space-adapted muscle-on-a-chip platforms as preclinical models for aging-related muscle degeneration and for screening regenerative therapeutics.
Conclusion
Skeletal muscle-on-a-chip platforms have evolved from simple contractile constructs to physiologically relevant systems capable of modeling vascularization, inter-organ communication, and muscle degeneration induced by environmental or disease-related factors. To facilitate broader adoption in preclinical research and eventual clinical translation, several key challenges must be addressed.
Despite increasing interest in musculoskeletal therapeutics, clinical translation remains limited by the absence of predictive preclinical models. Conventional animal studies often fail to replicate human-specific muscle responses, particularly in complex conditions such as sarcopenia, cachexia, and age-related degeneration. Furthermore, few models reliably predict muscle toxicity associated with investigational drugs. Muscle-on-a-chip platforms offer a potential solution by more accurately recapitulating human tissue architecture and functional outputs, thereby narrowing the gap between preclinical findings and clinical outcomes.
Standardization is a critical priority. Current variability in device design, hydrogel composition, electrical stimulation protocols, and functional output measurements undermines reproducibility and hampers cross-study comparisons [38,39]. Although recent studies have identified optimal parameters for tissue compaction and force generation, universally accepted protocols are still lacking. Establishing shared reference models, standardized measurement criteria, and benchmarking strategies will be essential to improving data consistency and fostering regulatory acceptance.
Regulatory integration is another essential step toward translation. Muscle-on-a-chip platforms capable of replicating disease-relevant phenotypes and therapeutic responses—such as dystrophin restoration in DMD or tissue regeneration in sarcopenic models—represent promising alternatives to animal testing. To meet unmet regulatory needs, including the demand for more human-relevant models of drug-induced myotoxicity and sarcopenia progression, these systems must undergo rigorous validation for predictive accuracy. Analytical performance, biological relevance, and scalability must be clearly demonstrated before such platforms can be incorporated into formal regulatory workflows.
Advancing the functional capabilities of muscle-on-a-chip technology will also require integration with diverse analytical modalities. Electrophysiological recording, optical sensors for metabolic activity, and chemical detection of inflammatory or myogenic markers can enable real-time, label-free monitoring of tissue state [14]. Incorporating transcriptomic, proteomic, and metabolomic profiling will further enhance the ability to capture molecular-level changes and identify pathway-specific responses under pathological or pharmacological conditions. Continuous, real-time signal acquisition over extended culture periods can provide dynamic insights into tissue behavior, reduce reliance on endpoint analyses, and improve both sensitivity and throughput.
Ultimately, improving clinical relevance and translational potential will depend on close alignment between engineering innovation and the pressing needs of muscle disease research and drug development. As these technologies advance, their applications in drug screening, mechanistic studies, and regenerative medicine are expected to expand. The convergence of engineering precision, biological fidelity, and integrated analytics positions skeletal muscle-on-a-chip systems as a powerful platform for advancing muscle biology and supporting translational research.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Funding
This work was supported by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (NRF-RS-2021-NR061930). MID (Medical Illustration & Design), as a member of the Medical Research Support Services of Yonsei University College of Medicine, for providing excellent support with medical illustration.
Data Availability
Not applicable.