Magnetic torque-mediated early asymmetric mechanical stimulation directs structural and functional maturation in human cardiac organoids

Abstract

The human heart establishes functional left–right asymmetry through region-specific mechanical cues; however, current methodologies provide limited capacity for spatially localized force stimulation within 3D tissues. To address this limitation, we developed a magnetic torque stimulation (MTS) platform that enables remote, non-contact mechanical loading of human induced pluripotent stem cell (hiPSC)-derived cardiac organoids functionalized with magnetic microbeads covalently conjugated to the membrane-binding lectin, allowing stable anchorage to cell-surface glycoconjugates. We then applied this system to asymmetric fusion organoids composed of bead-conjugated (Bead⁺) and bead-free (Bead⁻) regions to recapitulate spatially localized mechanical cues that mimic endogenous region-specific signaling. A 96-h rotating magnetic field (organoid day 3–7) generated controlled torque forces in individual organoids, resulting in robust activation of mechanotransduction pathways, including transcriptional upregulation of ITGB1, FN1, ZYX, VCL, ACTN2, and TEAD1, along with increased fibronectin and vinculin protein accumulation. In fusion organoids, the same mechanical stimulation elicited domain-specific activation, with mechanotransduction markers selectively elevated in the Bead⁺ region immediately after stimulation (Day 7), confirming spatially restricted force transmission within a 3D tissue environment. Following an additional culture period without further stimulation (until Day 15), MTS-treated fusion organoids exhibited enhanced maturation, characterized by increased expression of cardiomyocyte structural and ion-channel genes, enhanced fibronectin deposition, reinforced focal adhesion assembly, upregulated Piezo1 expression, and improved sarcomere organization. Collectively, these results demonstrate that an early asymmetric mechanostimulation can contribute to the emergence of maturation-related phenotypes in cardiac organoids. This work suggests that magnetic torque–based actuation can serve as a useful platform for studying biomechanical contributions in cardiac organoid models.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-51332-1.

Introduction

Heart development is governed by the coordinated integration of biochemical signals and mechanical cues within the local microenvironment, which together direct tissue morphogenesis and functional maturation^1–5. Accumulating evidence has established that mechanical forces are not merely passive consequences of growth but actively regulate key developmental processes, including cardiomyocyte alignment, cytoskeletal organization, and tissue-scale remodeling^6–8. Central to these processes is integrin-mediated focal adhesion signaling, which links extracellular mechanical inputs to intracellular mechanotransductive pathways and downstream biochemical responses^9–11. Importantly, integrins function as mechanosensing receptors not only in the heart but also during embryonic organogenesis. In developing mouse liver, for example, vascular perfusion activates endothelial β1 integrin signaling, which is required for organ growth and survival, demonstrating that integrin-mediated mechanotransduction can directly regulate embryonic organogenesis¹². More broadly, integrin-based cell–matrix adhesions are proposed as key mechanosensing structures that translate physical forces into biochemical signaling during diverse morphogenetic events in the embryo¹³. In addition, mechanical cues can also originate at the cell membrane through glycocalyx-associated components and subsequently engage integrin-mediated signaling. Such membrane-level perturbations can be transmitted to integrin-linked cytoskeletal networks, thereby influencing focal adhesion–mediated signaling¹⁴. Consistent with this framework, disruption of integrins or focal adhesion components impairs cardiac morphogenesis and compromises tissue integrity^15–17, underscoring the essential role of integrin-dependent mechanotransduction during heart development.

Human induced pluripotent stem cell (hiPSC)-derived cardiac organoids have emerged as robust in vitro platforms for modeling early human heart development by recapitulating key aspects of three-dimensional (3D) tissue architecture and multicellular organization. These systems enable interrogation of developmental processes that are difficult to resolve in conventional two-dimensional (2D) cultures and have therefore been widely applied to studies of cardiac lineage specification, early morphogenesis, and disease modeling^18–20. More recently, increasing emphasis has been placed on understanding how physical and mechanical cues, in concert with biochemical signaling, contribute to cardiac development^21,22. Despite these advances, most existing cardiac organoid platforms provide limited control over how mechanical stimuli are delivered and distributed within a 3D tissue context. Consequently, membrane-mediated mechanotransductive signaling pathways, as well as their associated structural and transcriptional responses, remain insufficiently characterized in cardiac organoid systems, particularly in the context of spatially confined mechanical inputs that are critical during early cardiac development.

Importantly, mechanical and signaling responses during heart development are inherently spatially heterogeneous. For example, spatiotemporally controlled mechanical cues during early cardiac morphogenesis regulate localized progenitor cell behaviors, such as mesenchymal-to-epithelial transition, demonstrating that mechanical cues are differentially distributed and interpreted across developing cardiac tissues²³. Early developmental events, such as heart tube formation and looping, are characterized by pronounced tissue asymmetry and the emergence of localized microenvironments that enable region-specific signal activation^24,25. However, current in vitro systems typically apply mechanical stimulation in a global and homogeneous manner—such as cyclic stretch^26,27, substrate stiffness modulation^28,29, or microfluidic flow-based shear stress applied uniformly across the tissue³⁰—precluding the ability to model localized or asymmetric mechanotransductive responses within a single 3D tissue. This limitation constrains the ability to model aspects of spatial patterning and regional specialization driven by mechanical cues.

To address these challenges, we previously developed a magnetic torque stimulation (MTS) platform that applies controlled mechanical torque to membrane-coupled magnetic microbeads through a uniformly rotating magnetic field, thereby generating localized mechanical inputs at the bead–cell interface without inducing bulk tissue deformation^31,32. Unlike conventional mechanical loading systems that rely on global stretch or modulation, this approach allows differential mechanical stimulation between bead-conjugated and bead-free regions within a 3D tissue. In the present study, we integrate this MTS platform with hiPSC-derived cardiac organoids to investigate how spatially localized mechanical stimulation regulates mechanotransductive signaling and downstream maturation. To enable effective actuation, we establish a stable magnetic bead conjugation strategy compatible with early cardiac differentiation and further engineer asymmetric fusion organoids composed of bead-conjugated and bead-free regions, allowing spatially distinct mechanical loading within a single organoid. Using this approach, we examine the immediate mechanotransductive responses to early asymmetric mechanical stimulation—defined here as spatially localized torque transmission confined to bead-conjugated regions within an otherwise uniform rotating magnetic field—and assess how such localized cues prime long-term structural and mechanosensitive maturation of cardiac organoids. Together, this work provides a framework for investigating spatially biased mechanobiological responses in cardiac organoids.

Results

Embryoid body formation under magnetic bead conjugation conditions and pluripotency assessment

Stable incorporation of magnetic microbeads into embryoid bodies (EBs) is a prerequisite for controlled MTS. Wheat germ agglutinin (WGA) is a lectin that binds N-acetylglucosamine- and sialic acid–containing glycoproteins abundantly present on the cell membrane³³, enabling stable interactions with the glycocalyx and facilitating membrane-level transmission of mechanical cues. Based on this property, carboxylated magnetic microbeads were functionalized under three surface conditions to establish an appropriate bead conjugation strategy: Non-treated, WGA-only, and EDC-NHS-mediated WGA-conjugated (Fig. 1a). During EB formation, bead distribution and stability varied markedly depending on surface modification. Two days prior to EB formation (Day -2), Non-treated and WGA-only beads exhibited heterogeneous peripheral clustering, consistent with weak and transient cell-bead interactions. In contrast, WGA-conjugated beads showed uniform attachment across individual cells within the aggregates. After medium exchange at Day-1, the majority of Non-treated and WGA-only beads detached from the EBs, whereas WGA-conjugated beads remained stably associated. As a result, uniform bead retention during EB formation was observed exclusively in EBs formed under the WGA-conjugated condition (Fig. 1b). To evaluate whether bead incorporation affected the initial differentiation state, pluripotency was assessed at the onset of differentiation. RT-qPCR analysis revealed comparable expression levels of NANOG, POU5F1, and SOX2 between bead-conjugated and bead-free EBs (Fig. 1c). Consistently, protein-level analysis by Western blotting and immunofluorescence staining showed no detectable differences in Nanog or Oct4 expression or spatial distribution between the two conditions (Fig. 1d–g and Supplementary Fig. 1). Together, these data demonstrate that EDC-NHS-mediated WGA conjugation enables stable and uniform bead attachment throughout EB assembly, while bead incorporation does not compromise pluripotency at the onset of differentiation, supporting its use as a robust platform for subsequent MTS-based mechanotransduction studies.

Fig. 1.

Bead conjugation–dependent stability during embryoid body formation and maintenance of pluripotency. (a) Schematic overview of magnetic bead conjugation conditions and experimental groups during embryoid body formation. (b) Time-dependent morphological changes of embryoid bodies under different bead conjugation conditions before and after medium exchange, including Non-treated (NT), WGA-only (WGA), and WGA-conjugated (WGA–EDC/NHS). Scale bars: 500 μm (4x), 200 μm (10x). (c) mRNA expression levels of pluripotency markers (NANOG, POU5F1, and SOX2) in EB Bead⁻ and EB Bead⁺ groups at Day 0. (d) Western blot analysis of pluripotency markers (Nanog and Oct4) in EB Bead⁻ and EB Bead⁺ groups at Day 0. Nanog, Oct4, and GAPDH were detected on separate membranes. Full-length uncropped western blots corresponding to the cropped bands shown in this figure are provided in Supplementary Fig. 1. (e) Quantification of Nanog and Oct4 protein levels normalized to GAPDH in EB Bead⁻ and EB Bead⁺ groups at Day 0. (f) Immunofluorescence staining of pluripotency markers (Nanog and Oct4) in EB Bead⁻ and EB Bead⁺ groups at Day 0. (g) Quantification of Nanog and Oct4 mean fluorescence intensity (MFI) in EB Bead⁻ and EB Bead⁺ groups at Day 0. Data are presented as mean ± SD. Statistical significance was determined by unpaired two-tailed Student’s t-test (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Bead stability and beating activity following 96 h MTS

To evaluate the effects of MTS under different bead conjugation conditions, EBs containing Non-treated, WGA-only, or EDC-NHS-mediated WGA-conjugated beads were analyzed using a single MTS platform. MTS was applied continuously for 96 h during early cardiac differentiation, spanning the transition from cardiac mesoderm to cardiac progenitor stages (Fig. 2a). Stimulation was initiated after mesoderm formation and prior to the onset of spontaneous contraction, thereby minimizing confounding effects from beat-driven mechanical forces. All experiments were conducted under identical MTS conditions, and organoids were embedded in growth factor-reduced Matrigel to facilitate effective force transmission during stimulation. Bead stability under continuous MTS was assessed using Alexa Fluor 488-labeled beads. Sustained fluorescence was observed exclusively in the WGA-conjugated group throughout the 96 h stimulation period, indicating stable bead retention under dynamic torque loading. In contrast, Non-treated and WGA-only groups showed no persistent fluorescence signal, consistent with bead detachment during stimulation (Fig. 2b). Following confirmation of bead stability, spontaneous beating activity was evaluated at Day 7. Spontaneous contractions were observed in all groups (Supplementary Videos 1–3). Quantitative analysis of beating parameters—including beats per minute (BPM), peak-to-peak interval, and contraction-relaxation duration—revealed no significant differences among Non-treated, WGA-only, and WGA-conjugated conditions (Fig. 2c). Together, these results indicate that EDC-NHS-mediated WGA conjugation is required to maintain stable bead attachment during prolonged MTS, while 96 h of stimulation does not induce immediate alterations in beating behavior, supporting the use of this stimulation window to probe early mechanotransductive responses rather than overt functional maturation.

Fig. 2.

Bead stability and functional beating responses after 96 h of MTS. (a) Schematic overview of magnetic torque stimulation (MTS) applied to NT, WGA, and WGA–EDC/NHS groups. (b) Morphological changes at the onset (Day 3) and completion (Day 7) of MTS in NT, WGA, and WGA–EDC/NHS groups, with fluorescent signals indicating bead retention. Scale bars: 200 μm. (c) Quantification of beating rate (i), peak-to-peak time (ii), and contraction–relaxation duration (iii) in NT, WGA, and WGA–EDC/NHS groups at Day 7. Data are presented as mean ± SD. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test, except for (c) (ii) Peak-to-peak time, which was analyzed using the Kruskal–Wallis test with Dunn’s multiple comparisons test (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Bead-conjugation-dependent focal adhesion and mechanotransduction responses to 96 h MTS

To examine bead-conjugation-dependent responses to MTS, focal adhesion- and mechanotransduction-related markers were analyzed immediately after completion of the 96 h stimulation period (Day 7). RT-qPCR analysis showed that transcript levels of ITGB1, FN1, VCL, ACTN2, and TEAD1 were increased by approximately 1.5-fold, 1.7-fold, 7.2-fold, 23.4-fold, and 1.3-fold, respectively, compared with the NT condition, with the most pronounced increases observed in the EDC–NHS–mediated WGA-conjugated bead condition following 96 h MTS. ZYX expression was increased by approximately 1.3-fold compared with the WGA-only condition, whereas no significant difference was observed relative to NT. In contrast, ITGA5 and LMNA mRNA levels remained unchanged across bead conjugation conditions (Fig. 3a). Immunofluorescence staining performed immediately after stimulation revealed that fibronectin signal intensity was increased by approximately 2.0-fold relative to NT in the WGA-conjugated group following 96 h MTS (Fig. 3b,d). Similarly, vinculin staining intensity in the same condition was increased by approximately 1.7-fold compared with NT, as observed at both the whole-organoid level and in higher-magnification images (Fig. 3c,e). Taken together, under the 96 h MTS condition, increased expression of selected mechanotransduction-related markers was accompanied by corresponding increases in downstream focal adhesion–associated proteins, including fibronectin and vinculin, in organoids prepared with EDC–NHS–mediated WGA-conjugated beads, indicating that molecular responses to mechanical stimulation under MTS conditions were observed in parallel with changes in focal adhesion protein expression. Based on these results, EDC–NHS–mediated WGA conjugation was used as the bead preparation strategy for subsequent MTS experiments.

Fig. 3.

Bead conjugation–dependent enhancement of mechanotransduction-associated markers after 96 h of MTS. (a) mRNA expression levels of mechanotransduction-related markers (ITGB1, ITGA5, FN1, ZYX, VCL, ACTN2, TEAD1, and LMNA) in NT, WGA, and WGA–EDC/NHS groups at Day 7. (b) Immunofluorescence staining of fibronectin in NT, WGA, and WGA–EDC/NHS groups at Day 7. Scale bars: 200 μm. (c) Immunofluorescence staining of vinculin in NT, WGA, and WGA–EDC/NHS groups at Day 7. Scale bars: 200 μm (10x, scan area 0.45x), 10 μm (63x), 2 μm (63x, scan area 3x). (d) Quantification of fibronectin MFI in NT, WGA, and WGA–EDC/NHS groups at Day 7. (e) Quantification of vinculin MFI in NT, WGA, and WGA–EDC/NHS groups at Day 7. Data are presented as mean ± SD. Statistical significance was determined by one-way ANOVA with Tukey’s multiple comparisons test (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Fusion of bead-conjugated and bead-free regions and beating activity following 96 h MTS

Fusion organoids composed of bead-conjugated (Bead⁺) and bead-free (Bead⁻) regions were generated to assess whether asymmetric and region-specific mechanical cues could be delivered within an organoid. Fusion was initiated immediately following completion of mesoderm differentiation (Day 2), and successful integration of the two regions was observed within one day. The Bead⁺ region was visualized using Alexa Fluor 488-labeled beads, enabling clear identification of the fusion interface over time. At Day 3, all fusion organoids were embedded in growth factor–reduced Matrigel. Continuous magnetic torque stimulation (MTS) was applied only to the MTS group from Day 3 to Day 7, whereas fusion organoids in the No MTS group were maintained without stimulation (Fig. 4a,b). Beating activity was evaluated at Day 7 in fusion organoids. Consistent with observations in single organoids, quantitative analysis of beating parameters—including beats per minute (BPM), peak-to-peak interval, and contraction–relaxation duration—revealed no detectable differences between fusion organoids subjected to 96 h MTS and those maintained under No MTS conditions (Fig. 4c and Supplementary Videos 4 and 5). Collectively, these results show that Bead⁺ and Bead⁻ regions can be reproducibly fused and that application of 96 h MTS does not alter early beating behavior in fusion organoids, supporting the use of this system as a stable platform for asymmetric mechanical stimulation.

Fig. 4.

Fusion of Bead⁻ and Bead⁺ regions and beating activity following 96 h of MTS. (a) Schematic overview of fusion organoid generation composed of Bead⁻ and Bead⁺ regions and subsequent MTS. (b) Morphological changes at the onset (Day 3) and completion (Day 7) of MTS in fusion organoids with fluorescent bead signals indicating the Bead⁺ region. Scale bars: 500 μm. (c) Quantification of beating rate (i), peak-to-peak time (ii), and contraction–relaxation duration (iii) in No MTS and MTS groups at Day 7. Data are presented as mean ± SD. Statistical significance was determined by unpaired two-tailed Student’s t-test (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Bead-conjugation-dependent mechanotransduction responses in asymmetric fusion organoids

To assess spatially localized mechanotransductive responses under asymmetric MTS, fusion organoids containing Bead⁺ and Bead⁻ regions were analyzed at Day 7. RT-qPCR analysis demonstrated that under MTS conditions, transcript levels of ITGB1, VCL, ACTN2, and TEAD1 in the Bead⁺ region were increased by approximately 1.6-fold, 1.7-fold, 2.2-fold, and 1.4-fold, respectively, compared with the MTS-treated Bead⁻ region. In contrast, expression levels of ITGA5, FN1, ZYX, and LMNA remained comparable between Bead⁺ and Bead⁻ regions (Fig. 5a). Notably, these transcriptional differences between Bead⁺ and Bead⁻ regions were observed only under MTS application, whereas no comparable regional differences were detected under No MTS conditions, and overall lower expression levels compared with the MTS condition, indicating that the observed gene expression changes arise specifically in response to MTS. To evaluate protein-level responses, immunofluorescence staining was performed on whole fusion organoids. Fibronectin signal intensity under MTS condition showed no statistically significant difference compared with No MTS group (Fig. 5b,d). Analysis at Day 15 revealed a mild upward trend in the MTS group (Supplementary Fig. 2). In contrast, vinculin staining intensity was increased by approximately 2.7-fold in MTS-treated fusion organoids relative to No MTS conditions, as observed at both the whole-organoid level and in higher-magnification images (Fig. 5c,e). Taken together, these results suggest that MTS induces region-specific transcriptional responses in the Bead⁺ region, while protein-level changes, particularly in vinculin, are observed at the whole-organoid level. However, as immunofluorescence analysis was not performed in a region-specific manner, these protein-level data should not be interpreted as direct evidence of localized differences between Bead⁺ and Bead⁻ regions.

Fig. 5.

MTS-dependent mechanotransduction responses in fusion organoids. (a) mRNA expression levels of mechanotransduction-associated markers (ITGB1, ITGA5, FN1, ZYX, VCL, ACTN2, TEAD1, and LMNA) in Bead⁻ and Bead⁺ regions of fusion organoids under No MTS and MTS conditions. (b) Immunofluorescence staining of fibronectin in fusion organoids under No MTS and MTS conditions at Day 7. Scale bars: 200 μm. (c) Immunofluorescence staining of vinculin in fusion organoids under No MTS and MTS conditions at Day 7. Scale bars: 200 μm (10x, scan area 0.45x), 10 μm (63x), 2 μm (63x, scan area 3x). (d) Quantification of fibronectin MFI in fusion organoids under No MTS and MTS conditions at Day 7. (e) Quantification of vinculin MFI in fusion organoids under No MTS and MTS conditions at Day 7. Data are presented as mean ± SD. Statistical significance was determined by one-way ANOVA with Dunnett’s multiple comparisons test using the MTS Bead⁺ region as the reference for (a), except for LMNA, which was analyzed using the Kruskal–Wallis test followed by Dunn’s multiple comparisons test. Unpaired two-tailed Student’s t-test was used for (d) and (e) (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Structural, electrophysiological, and mechanosensitive maturation in fusion organoids following MTS

After completion of the 96 h MTS period, the long-term effects of asymmetric mechanical loading were assessed by culturing fusion organoids under identical conditions until Day 15 (Fig. 6a). Both the MTS and No MTS groups maintained stable fused morphology throughout the culture period, and Alexa Fluor 488-labeled beads remained detectable within the Bead⁺ region in both conditions. In the MTS group, the persistent bead signal enabled continuous identification of the mechanically stimulated domain over time (Fig. 6b). However, by Day 15, the previously distinguishable Bead⁺ and Bead⁻ regions had become extensively intermingled following Matrigel removal and continued culture, making manual separation of the two domains technically infeasible. Therefore, subsequent analyses at this timepoint were performed at the whole-organoid level. At Day 15, functional differences between conditions became evident. Fusion organoids subjected to MTS exhibited more stable contraction profiles compared with No MTS condition (Fig. 6c). Quantitative analysis revealed trends toward reduced BPM, increased peak-to-peak interval, and prolonged contraction–relaxation duration in the MTS condition compared with the No MTS group—patterns generally associated with enhanced electrophysiological maturation; however, these parameters remained statistically comparable between groups (Fig. 6d and Supplementary Videos 6, 7). RT-qPCR analysis further supported these maturation-associated changes. Transcript levels of cardiomyocyte structural and maturation-related genes MYL7, MYL2, MYH7, ACTN2, JPH2, and TNNI3 were increased by approximately 1.6-fold, 1.7-fold, 1.2-fold, 1.5-fold, 1.2-fold, and 1.7-fold, respectively, in the MTS group compared with No MTS group (Fig. 6e). In contrast, TNNT2 and cTnT expression showed no significant difference between conditions, indicating comparable overall cardiomyocyte formation between conditions (Fig. 6e and Supplementary Fig. 3). In addition, expression levels of ion-channel genes RYR2 and KCNJ2 were increased by approximately 1.3-fold following MTS, and the mechanosensitive ion channel PIEZO1 was upregulated by approximately 1.5-fold in the MTS group. In contrast, the pacemaker-associated markers TBX3 and HCN4 did not show statistically significant changes, although HCN4 exhibited a modest upward trend (Fig. 6e). Protein-level analyses were consistent with these transcriptional changes. Immunostaining for sarcomeric α-actinin revealed more organized and elongated sarcomeres in MTS-treated fusion organoids, with quantified sarcomere length increased relative to No MTS group, reaching an average length of approximately 1.73 μm (Fig. 6f,g). Sarcomere length measurements were performed without pharmacological relaxation prior to fixation, and future studies incorporating defined relaxation protocols will allow more precise assessment of resting sarcomere length. In addition, Piezo1 immunostaining showed increased numbers of Piezo1-positive puncta and increased total puncta area in the MTS group compared with No MTS fusion organoids (Fig. 6h,i). Collectively, early asymmetric MTS induced consistent changes across genes and proteins associated with structural, electrophysiological, and mechanosensitive maturation in fusion organoids, with these differences becoming clearly evident by Day 15.

Fig. 6.

Long-term effects of MTS on structural, electrophysiological, and mechanosensitive maturation in fusion organoids. (a) Schematic overview of the experimental timeline for long-term culture of fusion organoids following 96 h of MTS. (b) Morphological features of fusion organoids under No MTS and MTS conditions at Day 15, with persistent localization of fluorescently labeled beads indicating the Bead⁺ region. Scale bars: 500 μm. (c) Representative contraction profiles of fusion organoids under No MTS and MTS conditions at Day 15. (d) Quantification of beating parameters, including beating rate (i), peak-to-peak time (ii), and contraction–relaxation duration (iii), in fusion organoids under No MTS and MTS conditions at Day 15. (e) mRNA expression levels of cardiomyocyte-related (MYL7, MYL2, MYH7, ACTN2, JPH2, TNNI3, and TNNT2), ion channel-related (RYR2 and KCNJ2), mechanosensitive channel-related (PIEZO1), and pacemaker-associated (TBX3 and HCN4) markers in fusion organoids under No MTS and MTS conditions at Day 15. (f) Immunofluorescence staining of sarcomeric α-actinin in fusion organoids under No MTS and MTS conditions at Day 15. Scale bars: 200 μm (10x, scan area 0.45x), 2 μm (63x, scan area 3x). (g) Quantification of sarcomere length in fusion organoids under No MTS and MTS conditions at Day 15. (h) Immunofluorescence staining of Piezo1 in fusion organoids under No MTS and MTS conditions at Day 15. Scale bars: 200 μm (10x, scan area 0.45x), 5 μm (63x, scan area 1.5x). (i) Quantification of Piezo1 puncta number and total puncta area in fusion organoids under No MTS and MTS conditions at Day 15. Data are presented as mean ± SD. Statistical significance was determined by unpaired two-tailed Student’s t-test, except for (d) (iii) Contraction–relaxation duration and (e) HCN4, which were analyzed using the Mann–Whitney test (ns: not significant, *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Discussion

In this study, we demonstrate that spatially localized mechanical stimulation delivered during an early stage of cardiac differentiation exerts sustained effects on mechanotransductive signaling and downstream maturation in hiPSC-derived cardiac organoids. By integrating stable WGA-mediated magnetic bead conjugation to cell-surface glycoproteins with MTS, we achieved membrane-proximal mechanical loading without inducing bulk tissue deformation, allowing us to selectively probe early mechanotransductive responses in a 3D tissue context. Applying MTS from the cardiac mesoderm stage through the onset of spontaneous beating (Day 3–7) enabled mechanical inputs to be delivered during a developmental window that precedes the dominance of contraction-generated forces and coincides with mechanically sensitive morphogenetic processes observed in vivo^2,25,34. Consistent with this timing, early MTS did not acutely alter beating behavior at Day 7, indicating that localized mechanical inputs do not directly perturb early contractile function (Figs. 1, 2, and 3). Instead, pronounced differences emerged at later stages, supporting a model in which early mechanical stimulation functions as an instructional developmental cue that primes subsequent maturation programs rather than acting as an acute modulator of electrophysiological output³⁵. This temporal separation between early mechanotransductive activation and later functional outcomes aligns with developmental paradigms in which early microenvironmental cues bias long-term tissue organization and maturation^25,35. However, functional assessments in this study were performed under spontaneous contraction without external rate control, which limits interpretation of rate-dependent parameters. Measures such as contraction–relaxation duration and peak-to-peak interval can be influenced by beating frequency; therefore, future studies should employ electrical field pacing to standardize beating rate and enable more precise linkage of early mechanotransductive activation to subsequent electrophysiological maturation.

The fusion organoid system further enabled direct evaluation of spatially localized mechanotransductive responses under identical culture conditions. By incorporating Bead⁺ and Bead⁻ regions within a single organoid, spatially distinct mechanical inputs could be imposed without altering the biochemical environment during differentiation. Under these conditions, early MTS selectively activated focal adhesion–associated mechanotransduction pathways in the mechanically stimulated Bead⁺ region (Figs. 4 and 5), as evidenced by localized upregulation of integrin- and adhesion-related markers. Notably, despite sharing the same biochemical microenvironment within a single fused construct, only the Bead⁺ region exhibited significant mechanotransductive activation at early time points. This spatial specificity argues against a predominantly secreted factor–mediated mechanism and instead indicates that the observed changes are preferentially driven by localized mechanical stimulation. Importantly, these early regional differences preceded and was associated with structural and functional divergence at latter stages (Fig. 6), demonstrating that localized mechanical inputs can contribute to the induction of spatially distinct activation of integrin-dependent mechanotransduction pathways. These findings are consistent with prior studies showing that anisotropic mechanical loading can guide spatial variations mechanical states in engineered cardiovascular tissues^36–38.

Differential responses among integrin subunits may further suggest the specificity of the mechanotransductive program activated by early MTS. Whereas expression of fibronectin-binding α-subunit ITGA5 remained largely unchanged, ITGB1 was consistently upregulated following mechanical stimulation (Figs. 3 and 5). Given that ITGA5 is primarily associated with early adhesive engagement^39,40, while ITGB1 participates in sustained force transmission through multiple integrin heterodimers^41–43, these results may indicate that early MTS differentially influences adhesion pathways with distinct mechanical roles. In this context, the preferential increase in ITGB1 could reflect a tendency toward pathways associated with more stable or force-bearing focal adhesion complexes, although further analysis will be required to substantiate this interpretation. In parallel, increased TEAD1 expression in the absence of significant changes in LMNA indicates that mechanotransductive signaling in this system does not necessarily require overt nuclear lamina remodeling. Instead, TEAD1 activation may occur through cytoskeleton-mediated regulation of intracellular tension downstream of adhesion reinforcement, as proposed in previous studies of early cardiac differentiation⁴⁴.

ECM-associated responses exhibited a delayed temporal profile relative to focal adhesion signaling. Although fibronectin-related transcriptional changes were readily induced by MTS in single Bead⁺ organoids, similar responses were not immediately apparent in Bead⁺/Bead⁻ fusion organoids at early time points (Figs. 3 and 5). This attenuation may reflect constraints imposed by matrix reorganization associated with tissue fusion, during which dynamic ECM remodeling may transiently limit additional ECM deposition^45–47. Notably, fibronectin expression showed an increasing trend at later stages (Day 15), suggesting that early localized mechanical stimulation may gradually promote ECM remodeling as tissue architecture stabilizes during long-term maturation. Together, these observations indicate that the timing of ECM-related responses to mechanical cues is strongly influenced by the evolving structural and mechanical state of the tissue^48,49.

Despite these remaining questions, the present study establishes that localized mechanical stimulation applied during a defined early developmental window is sufficient to activate focal adhesion–centered mechanotransduction pathways and lead to long-term maturation outcomes in cardiac organoids. By enabling spatially distinct mechanical loading within a single 3D tissue, the MTS-based approach provides a powerful platform for investigating how non-uniform mechanical cues influence mechanotransductive signaling in cardiac organoids. More broadly, this framework may offer opportunities to investigate how early mechanical asymmetries contribute to tissue patterning, functional regionalization, and disease-relevant remodeling in human cardiac tissues.

Conclusions

Our study highlights a spatially localized mechanical stimulation strategy applicable to cardiac organoid systems and demonstrates that localized mechanical cues delivered during an early developmental window can facilitate long-term structural and functional maturation. Together, these findings support the use of magnetic torque–based actuation as a broadly applicable platform for investigating the biomechanical regulation in 3D tissue models.

Materials & methods

Culture of human induced pluripotent stem cells (hiPSCs)

The WTC11 human induced pluripotent stem cell (hiPSC) line (obtained from the Coriell Institute, USA) was cultured on Matrigel-coated plates (Corning, 354277) in mTeSR™ Plus medium (STEMCELL Technologies, #100-0276) at 37℃ in a humidified incubator with 5% CO₂. Matrigel was prepared using the manufacturer-recommended dilution factor, which was adjusted according to the lot-specific protein concentration. The culture medium was replaced every two days, and cells were passaged every 4–6 days using ReLeSR™ (STEMCELL Technologies, #100-0484) according to the manufacturer’s protocol when they reached approximately 85% confluency. Y-27632 (10 µM; STEMCELL Technologies, #72304) was added to the medium for 24 h after each passage.

Wheat germ agglutinin (WGA) conjugated magnetic beads preparation

1.5 × 10⁷ of Dynabeads™ M-270 Carboxylic Acid (magnetic; 2.8 µm diameter; Thermo Fisher Scientific, 14305D) were washed three times with 0.1 mM MES buffer (pH 5.5; Biosolution) and finally resuspended in 0.1 M MES buffer. For conjugation, 0.4 mg/µL of N-hydroxysuccinimide (NHS; Sigma-Aldrich, 130,672) and 0.2 mg/µL of N-(3-Dimethylaminopropyl)-N’-ethyl carbodiimide hydrochloride (EDC; Sigma-Aldrich, E7750) were each dissolved in 0.1 M MES buffer, mixed at a 1:1 (v/v) ratio, and incubated for 5 min at room temperature to form the active crosslinker. The washed bead suspension, the activated EDC/NHS solution, and WGA (Lectin from Triticum vulgaris (wheat); stock 25 µM; Sigma-Aldrich, L9640) solution were combined to form a 100 µL reaction mixture at a 7:2:1 volumetric ratio (beads : WGA : EDC/NHS). The mixture was incubated for 40–50 min at room temperature under gentle shaking to allow covalent conjugation via amide bond formation. After conjugation, the supernatant was removed, and the beads were directly mixed with 7.5 × 10⁵ cells at a 1:20 bead-to-cell ratio for subsequent seeding. For the WGA-only group, beads were washed and incubated with WGA under identical conditions without prior EDC/NHS treatment. The non-treated group consisted of beads washed with 0.1 mM MES buffer only, without EDC/NHS or WGA, and used directly as controls. For optional bead visualization, Alexa Fluor™ 488 Hydrazide dye (Thermo Fisher Scientific, #A10436) was added to the final conjugation mixture at 2% (v/v).

3D Cardiac organoid formation and application of MTS

hiPSCs were detached using Accutase (Sigma-Aldrich, A6964) and conjugated with magnetic beads as described above. Two days before differentiation (Day-2), bead-conjugated cells were then seeded into ultra-low attachment U-bottom 96-well plates (Corning, 7007) at 7500 cells/well in mTeSR™ Plus containing Y-27632 (10 µM). The plates were centrifuged at 180 × g for 4 min to induce embryoid body (EB) formation. After 24 h, the medium was replaced with mTeSR™ Plus without Y-27632.

Differentiation was initiated 48 h after seeding (Day 0) using RPMI 1640 (Thermo Fisher Scientific, 11875-093) supplemented with B-27 minus insulin (Thermo Fisher Scientific, A18956-01) and 6 µM CHIR99021 (TOCRIS, 4423) for 48 h (Day 0–2). On Day 2, the medium was replaced with RPMI/B-27 minus insulin and 2 µM IWP2 (TOCRIS, 3533) for 48 h (Day 2–4). This differentiation scheme was adapted from previously published cardiac differentiation protocols^50–54. Although differentiation was guided toward the cardiac lineage, the resulting three-dimensional constructs contained multiple cardiac-relevant cell populations, including cardiomyocytes and endothelial cells, consistent with previously reported cardiac organoid models.

On Day 3, during IWP2 treatment, each organoid was embedded in a Matrigel mixture composed of growth factor–reduced Matrigel (Corning, 354,230), prepared according to the manufacturer’s instructions, and RPMI/B-27 minus insulin at a 7:3 ratio (v/v). For each condition, approximately 30 organoids were randomly selected and embedded in a 35-mm dish, followed by incubation at 37 °C for 40 min to allow gelation. After gelation, medium was added and MTS was initiated immediately. MTS was applied continuously from Day 3 to Day 7 (96 h) at 60 rpm using a rotating magnetic field system that generates a spatially uniform magnetic field of approximately 200 mT across a 35-mm culture area. In this system, membrane-coupled magnetic beads experience rotational torque in response to the externally applied field, transmitting mechanical forces to the cell membrane. Based on prior finite element analysis (FEA) simulations and theoretical modeling of the magnetic field and bead properties, the torque applied to each bead was estimated to be approximately 30–45 pN^31,32.

On Day 4, IWP2 was removed and the medium was replaced with RPMI/B-27 minus insulin, followed by an additional medium change on Day 5. On Day 7, the medium was switched to RPMI 1640 supplemented with B-27 supplement without vitamin A (Thermo Fisher Scientific, 12587-010), after which MTS was terminated and samples were collected.

Fusion cardiac organoid formation

For fusion organoids, bead-conjugated hiPSCs (Bead⁺) and bead-free hiPSCs (Bead⁻) were seeded separately under identical EB formation conditions. On Day 2, pairs of organoids (one Bead⁺ and one Bead⁻) were combined in the same U-bottom well to allow spontaneous fusion. On Day 3, fused organoids were embedded using the same Matrigel (7:3 with RPMI/B-27 minus insulin) protocol described above.

After completion of MTS on Day 7, Matrigel was removed, and organoids were transferred back to ultra-low attachment 96-well plates with RPMI/B-27 without vitamin A, with medium changes every 2 days until Day 15. Fused organoids were sampled on Day 7 and Day 15.

Morphological analysis

Bright-field images were acquired using an inverted microscope (Nikon Ti2-E, Japan) at 4 × and 10 × magnification. Basic image processing was performed using ImageJ software.

Contraction analysis

Contractile activity of 3D cardiac organoids was recorded using an inverted microscope (Nikon Ti2-E, Japan) at 4 × and 10 × magnification. Videos were acquired at 16.3 frames per second under spontaneous beating conditions. Beating dynamics were quantified using the MUSCLEMOTION plugin for ImageJ, which was installed from the developer repository (available at https://github.com/l-sala/MUSCLEMOTION/tree/master).

Immunofluorescence staining

• i.Cryosection staining

3D cardiac organoids were washed with PBS (Biowest, L0615-500) and fixed in 4% paraformaldehyde (PFA; Biosesang, PC2031-100-00) at 4 °C overnight. Fixed organoids were cryoprotected in 30% sucrose solution until they had completely sunk to the bottom (typically 1–2 days), embedded in OCT compound (SAKURA, 4583), and frozen. Cryosections (12 µm thickness) were obtained using a cryostat and stored at − 80 °C until staining.

For cryosection staining, slides were washed once with PBS and permeabilized with 0.4% Triton X-100 (VWR) in PBS for 15 min at room temperature. Blocking was performed with 5% Bovine serum albumin (BSA; Sigma-Aldrich, A9418) for 1 h at room temperature. Primary antibodies were diluted in 1% BSA and incubated at 4 °C overnight. Sections were washed three times with PBS at 5 min intervals and incubated with secondary antibodies diluted in 1% BSA for 90 min at room temperature in the dark. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific) in PBS (1:1000) for 15 min at room temperature in the dark, followed by five washes with PBS at 5 min intervals. Stained sections were mounted using mounting solution (DAKO, S3023). Fluorescence images were acquired using a Nikon microscope or a confocal fluorescence microscope (Carl Zeiss, LSM900), with confocal z-stack images collected at 1 µm intervals. Details of primary and secondary antibodies used are provided in Supplementary Table 1.

• ii.Whole-mount staining

Organoids were fixed as described for cryosection staining. For whole-mount immunostaining, organoids were washed once with PBST (PBS containing 0.05% Tween-20, Bio-Rad, #1706531) and permeabilized with 0.4% Triton X-100 in PBS for 2 h at room temperature. Blocking was performed with 5% BSA for 1 h at room temperature. Primary antibodies were diluted in 5% BSA and incubated at 4 °C overnight. Organoids were washed five times with PBST at 15 min intervals and incubated with secondary antibodies and DAPI diluted in 5% BSA for 2 h at room temperature in the dark. Following five additional PBST washes at 15 min intervals, organoids were mounted with approximately 20 µL of mounting solution per organoid. Confocal fluorescence images were acquired using a confocal microscope (Carl Zeiss, LSM900) with z-stack acquisition consisting of approximately 40 optical slices at 8 µm intervals. For samples stained with the same primary antibody, images were acquired using identical imaging settings (e.g., laser power and gain settings) across all experimental groups.

Fluorescence intensity was quantified by measuring the mean fluorescence intensity (MFI), and sarcomeric α-actinin organization was assessed by measuring the spacing between α-actinin-positive sarcomeric bands using ImageJ software. Piezo1 puncta were quantified by threshold-based particle counting in ImageJ.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)

For single organoids, whole organoids were collected following Matrigel removal. Fusion organoids were processed by removing Matrigel and manually dissecting the Bead⁺ and Bead⁻ regions using a surgical blade, which were collected separately for RNA isolation. Organoids were lysed in 1 mL TRIzol™ reagent (Thermo Fisher Scientific, 15596018). After addition of 0.2 mL chloroform (Sigma-Aldrich, C2432), the samples were vortexed for 10 s and then incubated for 3 min at room temperature. The samples were centrifuged at 15,000 rpm for 15 min at 4 °C to separate the phases, and the clear aqueous phase was transferred to a new tube. An equal volume of isopropanol (Sigma-Aldrich, 190,764) was added, followed by gentle mixing and incubation on ice for 10 min. Samples were centrifuged at 15,000 rpm for 10 min at 4 °C, and the generated RNA pellet was washed twice with 75% ethanol. After removal of residual ethanol, the pellet was air-dried, resuspended in nuclease-free buffer (Integrated DNA Technologies) and incubated at 58 °C for 10 min to ensure complete dissolution. RNA concentration and purity were measured using a NanoDrop One (Thermo Fisher Scientific). cDNA was synthesized from total RNA adjusted to 5 ng/µL using the PrimeScript™ cDNA Synthesis Kit (Takara, RR037A). RT-qPCR was performed using iQ™ SYBR® Green Supermix (Bio-Rad) with 25 ng of cDNA per well in triplicate reactions. Amplification was conducted on a CFX96 or CFX Opus real-time PCR system (Bio-Rad). Relative gene expression levels were quantified using the ΔΔCt method with β-actin as the housekeeping gene, and expression was normalized to the control group. Details of primer sequences used for RT-qPCR are provided in Supplementary Table 2.

Western blotting

Embryoid bodies (EBs) were washed once with PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, 89900) supplemented with protease inhibitor cocktail (Xpert, P3100-001) and phosphatase inhibitor cocktail (Xpert, P3200-001). Samples were disrupted by at least three cycles of incubation on ice followed by vortexing, and then centrifuged at 15,000 rpm for 15 min at 4 °C to obtain clarified protein lysates. Total protein concentrations were quantified using the bicinchoninic acid (BCA) protein assay kit (Thermo Fisher Scientific, #23227), and equal amounts of protein were prepared for loading. Protein samples (20 µg per lane) were separated on 12% Tris–Glycine polyacrylamide gels and transferred onto 0.2 µm nitrocellulose membranes using the iBlot 3 Dry Blotting System (Thermo Fisher Scientific). Membranes were blocked for 1 h at room temperature with 5% BSA in 1X TBST (prepared from 10X TBS buffer, distilled water, and Tween-20), followed by incubation with primary antibodies (dilutions listed in Supplementary Table 3) overnight at 4 °C. After three washes with 1X TBST at 10 min intervals, membranes were incubated with horseradish peroxidase–conjugated (HRP) secondary antibodies for 1 h at room temperature. Protein bands were visualized using ECL chemiluminescent substrate (Thermo Fisher Scientific, #34577).

Statistical analysis

Statistical analyses and graph generation were performed using GraphPad Prism (version 9.2.0). Data are presented as mean ± standard deviation (SD). Normality was assessed using the Shapiro–Wilk test. Parametric tests (unpaired t-test or one-way ANOVA with appropriate post hoc tests) were applied to normally distributed data, whereas non-parametric tests (Mann–Whitney or Kruskal–Wallis with Dunn’s post hoc test) were used when normality was not satisfied. The specific statistical test used for each dataset is indicated in the corresponding figure legend. A total of seven independent experimental batches (biological replicates) were analyzed. Statistical significance is denoted as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). Detailed sample sizes, statistical tests, and exact p-values are reported within the figure legends for clarity.

Author contributions

Myeongjin Kang: Conceptualization; Data curation; Formal analysis; Investigation; Project administration; Visualization; Writing – original draft. Myeongjin Song: Investigation; Funding acquisition; Supervision; Writing – review & editing. Yongdoo Park: Conceptualization; Funding acquisition; Supervision; Writing – review & editing.

Funding

This study was supported by a Korea University Grant and the National Research Foundation of Korea (NRF) grants, funded by the Korea government (No. RS-2024-00461527 and No. RS-2025-02303718).

Data availability

The data that support the findings of this study are available from the corresponding authors upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Ethics approval not required.