iPSC-derived cardiac organoids for drug cardiotoxicity evaluation and efficacy prediction in myocardial infarction and cardiac hypertrophy models

Abstract

Background

Cardiac organoids (COs) derived from induced pluripotent stem cells (iPSCs) constitute a three-dimensional microphysiological system that more faithfully recapitulates the structural and functional complexity of human heart tissues compared to traditional two-dimensional cultures or animal models. Therefore, it highlights the significance of iPSC-derived COs in providing more precise platform for assessing drug cytotoxicity, cardiac functional toxicity, developmental toxicity, and therapeutic efficacy in disease models.

Methods

Human iPSCs, after generation from somatic cells using reprogramming factors, were differentiated into COs using dynamic culture system. COs were characterized by cell composition, structure, spontaneous beating, electrophysiological properties. Furthermore, drug cytotoxicity was evaluated by assessing cell viability and lactate dehydrogenase release after exposure of COs to doxorubicin. Cardiac functional toxicity was assessed by monitoring changes in beating after exposure to different compounds. In addition, developmental toxicity was tested using thalidomide. Finally, cardioprotective drug efficacy was evaluated based on the construction of myocardial infarction (MI) and cardiac hypertrophy (CH) models.

Results

The reprogrammed iPSCs have successfully generated COs with complex structure and function. Myocardial toxicity induced by anthracycline drugs was faithfully reproduced in organoids. This work also examined the effects of isoproterenol, MYK461, E-4031, and nifedipine on cardiac beating. Long-term exposure to thalidomide triggered abnormal atrial-ventricular differentiation in COs. In MI model, cyclosporine A exerted cardioprotective effects by downregulating apoptosis and inflammation. In CH model, on the other hand, apocynin outperformed metoprolol in alleviating Ang II-induced cardiac hypertrophy.

Conclusions

This study establishes an efficient and stable method for generating uniform and functionally mature COs by utilising dynamic culture systems. Also, we proposes a reliable approach for drug-induced cardiotoxicity assessment and efficacy evaluation on the organoid platform, offering a multipurpose and cutting-edge tool for cardiac drug development.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-026-05079-1.

Introduction

Cardiovascular diseases (CVDs) encompass a wide range of occurring disorders involving the heart and blood vessels. Their common clinical phenotypes include arrhythmias, heart failure, valvular heart disease, congenital heart defects, and vascular diseases. CVDs represents a significant global health challenge [1, 2], as proposed by the World Health Organization. According to current epidemiological data, approximately 17.9 million people worldwide die from CVDs annually [3]. Nowadays, risks of certain types of CVDs have been mitigated owing to advancements in early detection technologies, the development of novel therapeutic strategies, and improvements in lifestyle. However, their persistently high prevalence and mortality rates remain a major public health concern.

Among multiple subtypes of CVDs, myocardial infarction (MI) and cardiac hypertrophy (CH) stand out for their prevalence. MI is primarily caused by coronary artery occlusion, which leads to myocardial ischemia, cell necrosis, and impaired cardiac function, and thus makes it one of the leading causes of heart failure [4]. Its occurrence may be attributed pathologically to inflammatory cytokine release, increased oxidative stress, and cardiomyocyte (CM) apoptosis. In contrast, CH is a compensatory structural adaptation of the heart in response to prolonged pressure overload, such as hypertension or valvular heart disease [5]. The typical clinical manifestations of CH patients generally include CM hypertrophy, increased fibrosis, and impaired contractile function. Pathological CH can ultimately progresses into heart failure and is often associated with severe complications such as arrhythmias [6]. For their in-depth investigations, the current major obstacle lies in the challenge of complex physiological environment of human cardiac tissues by employing conventional two-dimensional (2D) cell models and animal models [7]. It significantly hinders an accurate prediction of drug efficacy and toxicity, contributing to the failure of many candidate drugs in clinical trials. Therefore, it is pivotal and necessary to develop in vitro models with higher accuracy to recapitulate human cardiac tissue, thereby improving the precision of research on CVDs and enhancing the success rate of novel drug development.

The development of induced pluripotent stem cells (iPSCs) has shed light on the research for CVDs. Recent advances have leveraged iPSCs to create highly sophisticated self-organizing cardiac organoids(COs), known as cardioids, which recapitulate the intrinsic morphogenetic principles of human heart development [8].In 2006 Shinya Yamanaka, for the first time, reported a method for inducing iPSCs from adult mouse cells [9], and next year the same lab achieved successful reprogramming of human somatic cells into iPSCs [10]. As a new era in stem cell research, this breakthrough enables the differentiation of iPSCs into specific cell types for applications in disease modeling, drug screening, and regenerative medicine [11, 12]. In cardiovascular research, iPSCs can be directed to differentiate into CMs (iPSC-derived CMs, iPSC-CMs) by regulating the Wnt signaling pathway [13, 14]. These cells can further self-assemble into three-dimensional (3D) COs, which facilitate a more efficient mimicking of the structural and functional properties of human cardiac tissues [15–17]. Hyung et al. successfully generated human iPSC-derived COs with chamber-like structures, epicardial/myocardial layers, atrial/ventricular-like areas, as well as mechanical and electrophysiological functions and in vivo vascularization capacity by optimizing the culture protocol, providing a powerful tool for constructing more physiologically relevant cardiac models [18]. More importantly, multi-cell type interaction COs enable the modeling of cardiotoxicity mechanisms [19, 20]. For instance, by employing COs to assess the effects of drugs with varying pregnancy risk classifications on embryonic cardiac development, Hoang et al. [21] found significant structural and functional impairments caused by high-risk teratogens, such as thalidomide, in organoids. Mills et al. [22] established a high-throughput, CO-based screening platform, with the identification of several compounds with cardiac regenerative potential, and the superiority of COs over traditional 2D models in functional drug evaluation. Beck et al. [23] observed CM atrophy and a reduction in fibroblast content after exposing COs to trametinib, with partial recovery following drug withdrawal. All these findings collectively underscore the potential of COs as a powerful tool for evaluating drug cardiotoxicity. These advancements provide a physiologically more relevant in vitro model for studying CVDs. At the same time, iPSC-CMs are difficult to accurately recapitulate the complex pathological features of MI and CH, given their lack of a 3D microenvironment. In contrast, COs can self-organize into 3D structures to reconstruct the myocardial microenvironment, thereby facilitating a more faithful mimicking of the pathophysiological processes of MI and CH. Critically, MI-like injury has been documented to be induced successfully using oxygen content modulation or mechanical cryoinjury [24–26], while CH models can be established via mechanical stress induction and neurohormonal stimulation with angiotensin II (Ang II) or endothelin-1 [27].

With respect to the above, the present study was conducted to evaluate drug efficacy on the basis of the construction of MI models and CH models using iPSC-derived COs in vitro. Supported by the optimization of the culture system, this study proposes an improved suspension culture strategy to enhance the efficiency and reproducibility of CO generation, enabling high-throughput and consistent organoid production. This study also presents a comprehensive characterization of the morphological and electrophysiological properties of COs, combined with the assessment of their potential applications in acute cardiotoxicity assessment (D10–D20) and efficacy prediction, using metabolically more mature COs after day 30 for developmental cardiotoxicity assessment and maturation analysis. This system will provide a relative stable and physiologically relevant in vitro model, help advancing precision medicine research and novel drug development for CVDs, while facilitating the progress of personalized therapy.

Methods

Reprogramming and expansion of hiPSCs

The routine culture of human dermal fibroblasts-adult (iXCells, #10HU-014) employed fibroblasts expansion medium [Dulbecco’s modified eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS)]. On day 0, cells at the density of around 0.5 × 10⁵ to 1 × 10⁵ were plated on a 2.5% Matrigel (Corning, #354230)-coated 6-well plate in fibroblasts expansion medium and cultured for overnight. On day 1, the spent medium was replaced with mTeSR1 medium (Stemcell Technologies, #100–0276) and cells were incubated for 9 h. Subsequent processing was cell transfection using NM-RNA-reprogramming cocktail (Stemgent, #00–0076) and Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific, #13778030) in Opti-MEM reduced serum medium overnight, which was repeated daily for 3 days. From day 5 onwards, the spent medium was refreshed daily, with the first observation of iPSC-like colonies on day 10. Individual iPSC-like colonies were manually separated into multiple parts, and each part was transferred respectively to a 1% Matrigel-coated 12-well plate to expand cultivation for identification. Positive iPSCs were passaged by ReleSR (Stemcell Technologies, #100–0276) digestion or cryopreserved by CS10 (Stemcell Technologies, #7930) at a ratio of 1:10 every 4 to 6 days. In addition, 10 µM Y-27,632 were supplemented in the medium within 24 h after cell passage or recovery.

Alkaline phosphatase staining of hiPSCs

As instructed, the staining of iPSCs was performed using the Pluripotent Stem Cell Alkaline Phosphatase Color Development Kit (Beyotime, #C3250S) after the plating of iPSC at glass-bottom dishes.

Triploblastic differentiation of hiPSCs

Expandable liver organoids were constructed by following a previously reported differentiation protocol [28], with advanced cerebral organoids established according to a prior research [29]. Noticeably, optimization based on existing differential method were intended to enhanced generation large quantities of COs [20]. iPSCs were dissociated into single cells with Accutase to seed into 6-well plate at 6 × 10⁵ cells/well in 3 mL mTeSR1 medium supplemented with 10 µM Y-27,632. The plate was placed on the shaker with rotation speed 65 rpm for overnight incubation at 37 ℃ in 5% CO₂ incubator. To allow the cell aggregates to settle naturally, the plate was tilted and left undisturbed for 3 min during each medium change throughout the entire procedure, after which 2.7 mL of the spent medium was carefully removed and 3 mL of fresh medium. mTeSR1 medium was replenished to support embryoid body (EB) formation, with the shaking speed increased to 75 rpm on Day − 2 and 85 rpm on Day − 1, respectively. Cardiac differentiation was initiated when EBs were observed in dimension of 300–350 μm in diameter on Day 0, and the medium was refreshed daily in the early culture period from Day 0 to Day 19. Regarding the differentiation medium, it was RPMI 1640 supplemented with 2% B27 minus insulin, 4 µM CHIR99021, 1.25 ng/mL BMP4, 1 ng/mL Activin A, and 0.1 mg/mL Primocin for Day 0; RPMI 1640 supplemented with 2% B27 minus insulin and 0.1 mg/mL Primocin for Day 1; and it remained the same on Day 2 (with an additional 2 µM Wnt-C59 added) for Day 4 as that used on Day 1. Meanwhile, Day 6, the differentiation medium was RPMI 1640 supplemented with 2% B27 and 0.1 mg/mL Primocin; while it on Day 7 was Day 6 medium supplemented with 2 µM CHIR99021 for 1 h, and the post-treatment medium was replaced with fresh Day 6 medium. In the subsequent differentiation stage from Day 20 to Day 30, the medium was refreshed every 2 days. On Day 20, the differentiation medium was glucose-free RPMI 1640 supplemented with 2% B27, 1% palmitate-BSA, 1% oleate-linoleate-BSA, 120 µM L-carnitine, 30 nM T3 hormone, 0.4 mM 2-phospho-L-ascorbic acid trisodium salt, 4 mM glucose, and 50 ng/mL LONG R3 IGF-1. In addition, on Day 26, the differentiation medium was the same as Day 20 medium without the supplementation of LONG R3 IGF-1. The differentiation of COs was terminated by Day 30, and the obtained COs were maintained in Day 26 medium subsequently.

Teratoma formation

iPSCs were dissociated into single cells with Accutase (Stemcell Technologies, #7922) and resuspend in Matrigel diluent at a ratio of equate volume of Matrigel to DMEM/F12. Then, two 2‑month‑old female NOD/SCID mice (Hangzhou Ziyuan Laboratory Animal Technology Co., Ltd.) were subjected to subcutaneous injection of 1 × 10⁶ cells each flank. These experimental animals were observed for 8 weeks and, when the teratomas reached approximately 1–2 cm in diameter, were euthanized by trained personnel via cervical dislocation. All animal experiments in this study were in compliance with the ARRIVE 2.0 guidelines. Tumor tissues were then excised and fixed in formalin and prepared for tissue sections for hematoxylin-eosin (H&E) staining to assess tissue differentiation across all three germ layers.

Karyotype and short tandem repeat (STR) analysis

iPSCs at passage 10 were used for karyotype and STR analysis by Suzhou Feifan Testing Technology Services Co., Ltd. following corresponding standard protocols.

Immunofluorescence staining

Cells were cultured on sterile coverslips, and organoids were collected in 1.5 mL centrifuge tubes. Samples were fixed in 4% paraformaldehyde (20 min), permeabilized with 0.25% Triton X-100 (30 min), and blocked with 5% BSA (1 h). Primary antibodies (diluted in 1% BSA) were incubated overnight at 4 °C, followed by fluorescent-dye conjugated secondary antibodies (1% BSA dilution) overnight at 4 °C in the dark. Nuclei were stained with DAPI for 5 min. Cells were mounted on slides with anti-fade medium, and organoids were embedded in concave slides. Imaging was performed using a Zeiss LSM 900 confocal microscope, and data processed and analyzed with Fiji software.

Flow cytometry

Cells were dissociated into single cells with Accutase, filtered through 70 μm cell strainers, fixed with 4% paraformaldehyde (20 min), and permeabilized with 0.25% Triton X-100 (10 min). After blocking with 5% BSA (30 min), cells and organoids were incubated with primary antibodies for 1 h at 4 °C, then with secondary antibodies for 45 min at 4 °C in the dark. Stained cells were analyzed using a Thermo Fisher Scientific Attune™ CytPix Flow Cytometer, with data processed via FlowJo v10 (BD Biosciences).

Real-time quantitative PCR (qPCR)

Following the extraction of total RNA using the RNA Extraction Kit (Tiangen, #DP430), reverse transcription, using 1 µg of total RNA, was performed with the PrimeScript RT reagent kit (Takara, #RR037A). In a 10 µL reaction volume, qPCR was conducted on a QuantStudio 5 system (Applied Biosystems) with the SuperReal PreMix Plus (Tiangen, #AQ-601). The thermal cycling included an initial denaturation at 94 °C for 30 s, followed by 40 cycles of 94 °C for 5 s and 60 °C for 15 s. The relative gene expression, normalized to that of 18 S rRNA, was calculated using the 2^−ΔΔCt method. All reactions were performed in triplicate, with the amplification specificity confirmed by melt curve analysis.

Morphological analysis of COs

The growth and morphology of COs were monitored in real time under inverted microscope (Olympus, CKX53), with representative images captured at defined key time points throughout the culture. Eventually, growth kinetics were assessed by quantifying the number and diameter of COs in each batch.

Beating analysis of COs

The beating of COs was monitored in real time using high-content analysis (Perkin Elmer, Operetta CLS). Prior to imaging, temperature-induced interference with beating activity was minimized by adopting suitable measures. Quantitative analysis of beating amplitude and mean beating frequency were completed in Fiji and Microsoft Excel. Meanwhile, individual cardioids were randomly selected from the acquired images using the freehand selection tool, combined with the quantification of their mean pixel intensity (Grey Value). The baseline pixel intensity (G0) was defined as the average of the lowest 50 values within the dataset. To calculate the percentage of beating COs, we observed the presence of at least one spontaneous beating during the entire period of recording. Finally, the beating frequency was determined by measuring the time interval between consecutive beating peaks.

Drug testing

For the drug testing experiment, COs exhibiting robust and stable beating after 10–20 days of culture were selected for acute cardiotoxicity assessment. Stock solutions of the tested compounds were prepared as instructed by the manufacturer’s protocol. Doxorubicin (Dox) of different concentrations were added for an incubation of 48 h to evaluate the cardiac toxicity. Furthermore, the effect on beating dynamics were assessed by analyzing changes in cardiac beating after 20 min exposures to 10 nM isoprenaline, 1 µM MYK-461, 50 nM E-4031, and 1 µM nifedipine. Following the metabolic switch, COs were cultured until day 30 to achieve a more mature phenotype. Next, the developmental toxicity was assessed through 10 days of thalidomide administration, with medium changes every 48 h. Post-treatment analysis included assessments of cell viability and cardiac injury biomarkers, as detailed in subsequent sections.

Detection of cardiac injury biomarkers

According to the manufacturer’s protocol, this experiment further quantified the release of lactate dehydrogenase (Beyotime, #C0017) in the culture supernatant. Briefly, 120 µL of supernatant from each sample was mixed with 60 µL of lactate dehydrogenase (LDH) working solution for 30 min of incubation at room temperature in the dark. Absorbance at 490 nm was measured.

For cell viability assessment, COs were incubated with CCK-8 (Beyotime, #C0037) working solution at 37 °C for 2 h. Absorbance at 450 nm was subsequently measured using the microplate reader (TECAN, SPARK).

CH model

The CH model comprised three sequential phases: (1) Model induction using Ang II, (2) molecular and functional phenotyping, and (3) therapeutic intervention assessment. Organoids (10–20 days post-differentiation) were allocated to four experimental cohorts. Corresponding protocols were described as follows: vehicle: Standard culture conditions; hypertrophy model: 800 nM Ang II stimulation for 72 h; drug intervention 1: Ang II + 100 µM apocynin (NOX inhibitor); and drug intervention 2: Ang II + 10 µM metoprolol (β1-adrenergic antagonist). Beating dynamics were continuously monitored throughout the 3-day protocol. Terminal analysis included the quantification of hypertrophy markers (NPPA, NPPB, and MYH7) via qPCR, with 18 S as endogenous control.

MI model

The MI model consisted of three key components: (1) model establishment; (2) molecular and functional characterization; and (3) evaluation of the protective effect of drug intervention on MI. COs cultured for 10–20 days were selected for MI modeling, with the establishment of four cohorts, including (1) the vehicle: Normoxic culture; (2) MI model: hypoxia/reoxygenation (H/R) only, exposure of the COs to 0.1% O₂ + 1 µM norepinephrine for 72 h to simulate ischemic stress and restoration to normoxia (21% O₂, and 5% CO₂) in maintenance medium for 24 h to mimic reperfusion damage; (3) drug intervention 1: H/R + 30 nM selatogrel; and (4) drug intervention 2: H/R + 2 µM cyclosporin A. The beating rhythm of the COs was monitored throughout the experiment. LDH and TUNEL staining were used to determine the degree of damage and apoptosis, respectively. Besides, qPCR was employed to assess the expression levels of inflammation and apoptosis markers in each group.

Statistical analysis

The sample size “n” in this study is rigorously defined as the number of independent biological replicates, specifically, individually generated cardiac organoids. This is explicitly stated in each figure legend. Quantitative data derived from three independent technical replicates in all experiments were presented as mean ± standard deviation (SD). Statistical significance was evaluated using one-way ANOVA with Dunnett post-test corrections, or using unpaired t-test. All analyses were performed in GraphPad Prism 9.5 and p < 0.05 was considered statistically significant.

Results

iPSCs reprogramming and characterization of HDF-A cell-derived iPSCs

Initially, this experiment summarizes the reprogramming process in a schematic diagram, depicting the transition from HDF-A to iPSCs under defined conditions. Five transcription factors (i.e., OCT4, SOX2, KLF4, cMyc and NANOG) were transfected intracellularly four times at 24-h intervals (Fig. 1A). Specifically, small and loosely packed colonies emerged by Day 10. Between Days 15 and 20, the colonies exhibited compact, round morphologies with sharp borders, resembling human embryonic stem cells, characterized by large nuclei and high nucleus-to-cyt oplasm ratio (Fig. 1B). Our subsequent processing focused on the expansion and continuous purification of these iPSC clones that were picked and mechanically isolated. According to alkaline phosphatase staining for verifying the acquisition of pluripotency, there were strong positive signals in iPSC colonies, indicative of an undifferentiated state (Fig. 1C). Further immunofluorescence staining confirmed the expression of OCT4, SOX2, and NANOG, serving as core pluripotency markers, with appropriate and homogeneous nuclear localization (Fig. 1D). Flow cytometry demonstrated high expression levels of pluripotency markers-NANOG (99.6%), OCT4 (99.2%), and SOX2 (99.9%) in the iPSCs population, thereby supporting the homogeneity and pluripotent identity of iPSCs (Fig. 1E). Subsequently, the differentiation capacity of iPSCs was evaluated both in vitro and in vivo. iPSCs were directly differentiated into brain, heart and liver organoids. mRNA expression analysis verified the upregulation of TUBB (ectoderm), PECAM (mesoderm) and AFP (endoderm), all of which were representative lineage-specific markers, demonstrating that iPSCs enabled the generation of derivatives of all three germ layers (Fig. 1F). In addition, teratoma formation assays further validated the pluripotency of the generated iPSCs, given the observation of diverse tissue types, such as neural epithelium (ectoderm), cartilage (mesoderm), and gut-like epithelium (endoderm) (Fig. 1G). Meanwhile, karyotype analysis to assess genetic stability indicated normal chromosomal structure maintained by the iPSCs, without detectable abnormalities (Fig. 1H). In addition, to ensure the authenticity of the reprogrammed cells, STR profiling across 20 loci revealed consistent allelic patterns (single or double peaks), supporting high-quality human cell typing without cross-contamination. Subsequent comparison with major cell line databases (ATCC, DSMZ, and ExPASy) showed no matches, establishing the uniqueness and clonal integrity of this iPSC line (Supplementary Fig. 1, and Supplementary Table 1). Taken together, our experiment successfully reprogrammed iPSCs from somatic cells, with stable pluripotency, normal karyotype, and the potential to differentiate into all three germ layers, supporting their suitability for downstream applications.

Fig. 1.

Establishment of HDF-A-derived iPSCs and corresponding pluripotency validation. A Schematic of the reprogramming protocol using OCT4, SOX2, KLF4, cMyc, and NANOG. B Representative bright-field images showing morphological changes between iPSC colonies and HDF-A. Scale bar: 50 μm. C Alkaline phosphatase staining of iPSC colonies. Scale bar: 100 μm. D Immunofluorescence images of NANOG, OCT4, and SOX2 expression in iPSC colonies. Scale bar: 50 μm. E Flow cytometry analysis of NANOG, OCT4, and SOX2 in iPSC colonies. F qPCR analysis of lineage-specific markers TUBB, PECAM, and AFP after differentiation into brain, heart, and liver organoids (n ≥ 10 independent organoids per condition across three independent experiments, mean ± SD, Student’s t test, ***p < 0.001). G H&E-stained teratoma sections showing ectodermal (neural epithelium), mesodermal (cartilage), and endodermal (gut-like epithelium) derivatives. Scale bar: 100 μm. H Karyotype analysis showing normal chromosomal structure

Generation and characterization of iPSC-derived COs with a dynamic culture strategy

Critically, our experiment employed continuous agitation using a shaker throughout cultivation and differentiation processes to reduce the heterogeneity in the development of COs caused by uneven nutrient supply, pH buffering capacity, and oxygen distribution, as well as cell loss during medium exchange. This dynamic culture system provided gentle mechanical stimulation, which facilitated organoid maturation, with improvement in the uniformity and yield as well. After the first 24 h of dynamic culture, iPSCs spontaneously aggregated into EBs, with uniform size and smooth edges. These EBs rapidly expanded in size and were efficiently directed towards early CO formation during Day 0–7 of differentiation by modulating the Wnt signaling pathway. In details, the proposed process involved three sequential phases, which involved the activation of the Wnt signaling for cardiac mesoderm induction, the inactivation of the Wnt signaling for definitive cardiac lineage specification, and a final activation step promoting proepicardial tissue development [30]. On Day 20, the adjustment of the glucose-to-fatty acid ratio in the culture medium facilitated a metabolic shift from glycolysis to fatty acid oxidation, a hallmark of postnatal CM maturation [20, 30, 31], thereby enhancing CO structural and functional development (Fig. 2A). Across six independent batches, each 6 × 10⁵ iPSCs generated approximately 225 COs (Fig. 2B), of 253 μm in average diameter (Fig. 2C), representing a 80% decrease over the original protocol [20]. By Day 6, COs began to exhibit spontaneous beating, exhibiting progressive intensify and stabilization over time. By Day 20, 100% of the organoids displayed strong and rhythmic beating, indicating early functional maturation (Supplementary Movies 1, and 2). After 60 days, the beating frequency stabilized and gradually approximated that of the human heart, with an average rate of between 1.0 and 1.5 Hz (Fig. 2D). Notably, in long-term culture, COs exhibited continuous beating for over 150 days, demonstrating sustained structural and functional stability until the end of the experiment. The beating frequencies of the cardiac organoids were recorded on days 10, 30, and 90. The results showed a progressive increase in the beating rate over time (Fig. 2E). Immunostaining revealed vascular structures (CD31⁺ and α-SMA⁺) and chamber-specific CM differentiation (MYL7⁺ atrial CMs and MYL2⁺ ventricular CMs) (Supplementary Fig. 2A). To further investigate the development of ventricular- and atrial-like structures in differentiated human cardiac organoids, we performed staining of the atrial marker MYL2 and the ventricular marker MYL7 in the cardiac organoids (Fig. 2F). A progressive reduction in the co-localized areas of atrial and ventricular markers was observed by immunofluorescence as differentiation proceeded. Overall, COs showed early differentiation into atrial and ventricular CMs, along with vascular cell differentiation. Furthermore, there existed significant upregulation in ATP2A2 (encoding the ATPase Sarcoplasmic / Endoplasmic Reticulum Ca²⁺ Transporting 2), KCNH2 (encoding the hERG potassium channel), and SCN5A (encoding the cardiac sodium channel Nav1.5) (Fig. 2G), all of which were key cardiac ion channel genes. Eventually, the above investigations confirmed the electrophysiological functional properties of organoids, which was critical for sustaining their functional integrity. Overall, our optimized strategy could preserve the physiological relevance of COs, while enabling scalable production, providing a multipurpose platform for disease modeling, drug screening, and mechanistic exploration of cardiac development and pathology.

Fig. 2.

Efficient differentiation and functional maturation of COs supported by dynamic culture. A Schematic of the cardiac organoid differentiation protocol as well as continuous stirring and representative bright-field image of COs. Scale bar: 500 μm. B Quantification of CO yields across six independent batches. C Diameter distribution of COs on Day 30 (n ≥ 30 independent organoids per condition across six independent experiments). D Dynamic changes in beating frequency of COs over time (n = 2 independent organoids and 2 random regions in each organoid per condition). E Representative calcium transients of COs on Day 10 visualized by Fluo-4 AM. F Immunofluorescence staining of MYL7 and MYL2 in COs on Day 30. Scale bar: 100 μm. G qPCR analysis of ATP2A2, KCNH2, and SCN5A in COs on Day 30(n = 15 independent organoids per condition across three independent experiments, mean ± SD, Student’s t test, ***p < 0.001)

Compared to static culture in low-adhesion 96-well plates [15, 20], the dynamic culture system demonstrates significant advantages in organoid yield. Under the same cell seeding density, the number of organoids generated in a single well of a 6-well plate under dynamic culture is more than six times that of static culture; in terms of organoid output per unit of cells used, dynamic culture yields three times more than static culture.

In terms of function and viability, the beating ratio of organoids under both culture systems approaches 100%, with no significant difference in survival rate, both maintained at nearly 100% (Supplementary Fig. 2C).

Regarding morphological characteristics, the cardiac organoids produced by dynamic culture have diameters ranging primarily between 150 and 400 μm, with 89% measuring 200–300 μm, 6% measuring 300–400 μm, and only 4% smaller than 150 μm. Core necrosis is relatively rare. Although solid organoids with diameters exceeding 1000 μm may develop central necrosis, the majority of organoids generated in this system possess a cavity structure, resulting in only a very small proportion exhibiting such issues.

Assessment of drug safety with iPSC-derived COs

Our further investigation emphasized on the assessment of the drug-induced cardiotoxicity utilizing COs, with the use of the anthracycline DOX, serving as a representative compound, which is a common chemotherapeutic agent severely limited clinical application owing to its dose-dependent cardiotoxicity characterized by mitochondrial dysfunction, oxidative stress, and CM apoptosis, ultimately leading to dilated cardiomyopathy and heart failure [32]. In the examination of the morphological effects of DOX on COs firstly, COs maintained an intact three-dimensional structure with well-defined boundaries and uniform morphology prior to drug treatment. However, DOX exposure resulted in the progressive shrinkage and structural disintegration of organoids. Severe cardiotoxic effects were observed considering the presence of complete disintegration in the edges of the organoid following 10 µM DOX exposure for 48 h (Fig. 3A). Next, CCK-8 assay for quantifying DOX-induced cytotoxicity demonstrated a dose-dependent reduction in cell viability, suggesting significant impairment on CM survival by DOX (Fig. 3B). Additionally, LDH release assay revealed obvious increase in LDH levels, indicating compromised cell membrane integrity and increased necrosis, further confirming DOX-induced cytotoxicity (Fig. 3C). Furthermore, four representative drugs affecting cardiac beating properties, i.e., isoproterenol (a β-adrenergic receptor agonist), MYK461 (a myosin inhibitor), E-4031 (an hERG potassium channel blocker), and nifedipine (an L-type calcium channel blocker), were tested for functional assessments. Consequently, COs exhibited stable, spontaneous, and synchronized beating rhythms under baseline conditions (Fig. 3D); while distinct functional responses were observed following drug administration. To be specific, isoproterenol led to remarkable increase in the beating amplitude and accelerated beating frequency, mimicking its positive inotropic and chronotropic effects [33, 34]. MYK461 markedly reduced beating amplitude, consistent with its role in inhibiting myosin ATPase activity [35]. E-4031 induced early afterdepolarizations and arrhythmic beating patterns, unveiling its known proarrhythmic properties [36]. Besides, nifedipine completely suppressed beating activity, abolishing both beating frequency and amplitude [37]. Taken together, COs could accurately replicate DOX-induced cardiotoxicity and distinct beating responses to pharmacological agents, providing a comprehensive evaluation of cardiac safety at all three levels of morphology, biochemistry and function. COs may serve as a physiologically relevant in vitro model for cardiotoxicity, offering reliable platform for preclinical drug assessment, which may contribute to reducing the reliance on animal models and enhancing the efficiency of drug development.

Fig. 3.

Evaluation of cardiotoxic and electrophysiological responses in organoids. A Representative bright-field images showing morphological changes in COs before and after DOX treatment. Scale bar: 200 μm. B CCK-8 assay showing dose-dependent decrease in cell viability following DOX treatment (n = 3 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test, **p < 0.01, and ***p < 0.001). Significance was determined by comparing with the 0-concentration group. C LDH release assay indicating increased cytotoxicity and membrane damage in response to DOX exposure (n = 3 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test, ns p > 0.05,**p < 0.01, and ***p < 0.001). Significance was determined by comparing with the 0-concentration group. D Representative motion analysis plots of COs after treatment with four pharmacological agents

Assessment of drug-induced developmental toxicity with iPSC-derived COs

As evidenced by our experiments, iPSC-derived COs exhibited atrial and ventricular chamber structures by Day 30 of differentiation. Therefore, through the use of thalidomide as a representative drug, this study continued to explore the potential of COs as a platform for assessing drug-induced developmental cardiotoxicity. Thalidomide, originally used for pregnancy-related morning sickness, has canceled its listing on the market due to its teratogenic effects, triggering the risk of severe limb and cardiac malformations, particularly during the first trimester [38, 39]. Subsequent research has revealed that thalidomide can inhibit angiogenesis and disrupt CM development, leading to structural and functional cardiac abnormalities, and heightened risk of heart failure consequently. In our study, organoids were subjected to exposure to 20 µg/mL thalidomide during Days 20–30 of differentiation, with the culture medium changed every 48 h to ensure continuous drug exposure over the 10-day treatment period. Then, our experiment focused on the assessment of the impact of thalidomide-treated organoids on their structural and functional characteristics. Upon continuous observation, compared to the untreated controls, thalidomide-treated COs were observed with significantly narrowed size, along with notable morphological alterations. The previously irregular structures became more spherical in shape, accompanied by signs of structural disintegration in several organoids, indicating thalidomide exposure-induced compromise in the structural integrity (Fig. 4A). Morphometric analysis showed an around 10% reduction in the average diameter of these organoids (Fig. 4B), reinforcing the structural impairments observed under thalidomide exposure. Furthermore, thalidomide-treated organoids had decreased average beating frequency from 0.31 Hz to 0.21 Hz (Fig. 4C, E), exhibiting marked beating abnormalities, yet without significant change in the beating amplitude (Fig. 4D, E). Simultaneously, there was irregular beating activity, indicating that thalidomide exposure would disrupt COs’ functional integrity. Immunofluorescence staining further demonstrated significantly reduced proportion of MYL2-positive cells in the thalidomide-treated group, suggesting impaired differentiation of both atrial and ventricular CMs by thalidomide (Fig. 4F, Supplementary Fig. 2B). Thalidomide treatment also induced highly downregulated expression levels of SCN5A (sodium channel), KCNJ2 (potassium channel), and ATP2A2 (calcium pump) in COs, resulting in functional impairments, including impaired depolarization, disrupted repolarization, and dysregulated calcium cycling, attributable respectively to SCN5A reduction, KCNJ2 suppression, and ATP2A2 attenuation. Collectively, all these molecular alterations would diminish the beating frequency and arrhythmic beating patterns. Thalidomide exposure also impaired the expression of cardiac structural and transcriptional markers. There was a broad suppression of both ventricular (MYH7, and MYL2) and atrial (MYH6, and MYL7) sarcomeric genes, suggesting impaired subtype-specific CM maturation. Besides, our detection also noticed reduced levels of HEY1 and HEY2, two critical mediators of Notch signaling, revealing disrupted transcriptional programs required for chamber morphogenesis and lineage specification. In a word, the iPSC-derived COs developed in our experiment would replicate the teratogenic effects of thalidomide, which might trigger cardiac structure and function (Fig. 4G). It is important to note that these conclusions are based on comparative assessments of morphology, contractile function, and marker expression, as direct quantification of cell viability was not performed in this assay. This model thus provides a robust platform for assessing the multifaceted impact of drugs on simulated embryonic cardiac development, enabling the evaluation of their morphological, functional, and molecular effects.

Fig. 4.

Structural and functional assessment of COs following exposure to thalidomide. A Representative bright-field images showing COs from vehicle and thalidomide-treated groups at Day 30 of differentiation. Scale bar: 50 μm. B Diameter measurements of COs in vehicle and thalidomide-treated groups (n ≥ 16 independent organoids per condition across one independent experiment, mean ± SD, Student’s t test, ns p > 0.05). C Quantification of average beating frequency of COs in vehicle and thalidomide-treated groups (n = 12 independent organoids per condition across one independent experiment, mean ± SD, Student’s t test, *p < 0.05). D Quantification of beating amplitude of COs in vehicle and thalidomide-treated groups (n = 12 independent organoids per condition across one independent experiment, mean ± SD, Student’s t test, ns p > 0.05). E Representative motion analysis plots showing beating patterns of COs from vehicle and thalidomide-treated groups. F Immunofluorescence staining of MYL2 and MYL7 in COs from vehicle and thalidomide-treated groups. Scale bar: 50 μm. G qPCR analysis of gene expression in COs from vehicle and thalidomide-treated groups, including markers for cardiac structure, transcriptional regulation, and ion channel function (n = 15 independent organoids per condition across three independent experiments, mean ± SD, Student’s t test, *p < 0.05, ***p < 0.001). Significance was determined by comparing with the Vehicle group

Evaluation of the cardioprotective effect in MI model

H/R and Norepinephrine stimulation were employed to the organoids for establishing an MI model in vitro (Fig. 5A) after successful construction of COs. Generally, the 3D model outperforms traditional 2D cultures, which may facilitate the interaction between CMs and surrounding cells, thereby enhancing electrical signal conduction and improving the physiological relevance of intercellular communication. Eventually, it may enable a more comprehensive assessment of functional parameters, particularly in terms of electrophysiological coupling and myocardial beating coordination that are impossible to be fully captured in 2D culture systems.

Fig. 5.

Pathological characterization and cardioprotective drug assessment of hypoxia-reoxygenation-induced MI in COs. A Vehicle and organoids infarction protocol timeline using altered oxygen (O₂) and norepinephrine in culture. B Representative bright-field images of Vehicle and MI organoids. Scale bar: 200 μm. C Representative beating rhythm of Vehicle and MI organoids. D Long diameters of the Vehicle and MI group on Day 15 and Day 19 (n = 23 independent organoids per condition). E Relative expression levels of LDH (n = 3 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test, **p < 0.01, ***p < 0.001). F mRNA expression of inflammation andapoptosis genes (n = 20 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test, ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001). G Representative TUNEL staining images for each condition. Scale bar: 50 μm. H Quantitative beating analysis from the SLT intervention experiment (n = 3 independent organoids and 3 random regions in each organoid per condition). I Quantitative beating analysis from the CsA intervention experiment (n = 3 independent organoids and 3 random regions in each organoid per condition). J mRNA expression of CM marker and functional genes (n = 15 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test*p < 0.05, ***p < 0.001). Significance was determined by comparing with the MI group

We assessed the structural and functional changes of the COs in MI model. After treatment, the vehicle group maintained self-organized morphology and regular beating, while the COs in the MI group exhibited significant morphological collapse accompanied by volume reduction (Fig. 5B, D). Simultaneously, COs in the MI group showed increased beating frequency with reduced amplitude (Fig. 5C), and some of them eventually exhibited progressively weakened beating until complete cessation. This beating pattern is consistent with clinical observation wherein an increased heart rate accompanied by reduced contraction amplitude thereby triggers compensatory tachycardia to maintain systemic circulation, indicating impaired cardiac function. To quantitatively evaluate cardiomyocyte injury, LDH was measured. The results showed a significant increase in LDH levels in the MI group, indicating substantial cellular damage and loss of membrane integrity (Fig. 5E). To elucidate the molecular mechanisms driving this damage, we assessed the expression of key pathological markers by qPCR. This comprehensive pathological characterization revealed upregulated levels in inflammatory markers IL-6 and TNF-α, increased Bax and decreased Bcl-2 expression (two apoptotic markers) (Fig. 5F). The occurrence of apoptosis was further confirmed and spatially resolved by TUNEL staining (Fig. 5G), which showed that COs within the MI group exhibited a distinct outside-in apoptotic pattern. This gradient of cell death ultimately contributed to the disruption of the overall architecture in COs. Beyond these structural alterations, we also observed changes in the beating patterns of the MI group, characterized by a reduced beating amplitude and an increased beating frequency (Fig. 5H, I). Due to the biological variability between experimental batches, the absolute beating amplitudes and frequencies differed across independent experimental replicates. Therefore, all statistical comparisons were performed within-experiment. These suggest that the MI group may impair CM contractile proteins and electrophysiological stability, thereby compromising the pump function of COs. To explore this possibility at the molecular level, we further examined the expression of key genes involved in contraction and electrophysiology. qPCR results confirmed downregulated mRNA levels of SCN5A (Na⁺ channel) and TNNT2 (troponin T) (Fig. 5J), which validates that our MI model exhibits pathophysiological relevance to clinical myocardial infarction.

Our subsequent evaluation focused on the cardioprotective effect of Cyclosporin A (CsA) and selatogrel (SLT) against myocardial damage in the MI model. As a mitochondrial permeability transition pore inhibitor, CsA can block pathological permeability transition pore opening to attenuate mitochondrial damage and CM apoptosis [40, 41]. While SLT is a novel P2Y₁₂ receptor antagonist [42] that can exert direct cardio protection via suppressing the activation of NLRP3 inflammasome and reducing the release of TNF-α/IL-1β [43]. LDH analysis confirmed both drugs significantly reduced myocardial injury (Fig. 5E). qPCR quantified sustained TNNT2 expression in CsA-treated organoids, comparable to control levels (Fig. 5J); whileSLT groups exhibited partial but incomplete TNNT2 recovery, which were further confirmed by immunofluorescence staining (Fig. 5G). As a result, Both CsA and SLT are effective against MI resulting from H/R-treated, but they may differ in their protective efficacy or mechanism of action. CsA may preferentially preserve myocardial proteins integrity, while the cardio-protection of SLT primarily targeted inflammatory/apoptotic pathways with delayed structural restoration.

Overall, COs could respond to H/R and Norepinephrine stimulation, as evidenced by increased cell death, as well as structural and contraction abnormalities in COs, faithfully mimicking the pathophysiological activities of MI CsA could effectively reduce the release of inflammatory factors and inhibit apoptosis. Beyond the construction of a more physiologically and pathologically relevant in vitro research platform, this COs-MI modelcan also serve as a valuable tool for pharmacological evaluation and mechanistic study of CsA. It may offer a more precise experimental foundation for future development of novel cardioprotective drugs.

Evaluation of the cardioprotective effect in CH model

Finally, we established a CH model using Ang II, with a purpose of demonstrating that our COs can recapitulate the pathophysiological mechanisms of human heart disease and serve as a pharmacological evaluation platform which closely mimics human physiological responses. Ang II is a well-characterized hypertrophic stimulus that induces cardiac hypertrophy and fibrosis primarily through activation of the angiotensin type 1 receptor, initiating a cascade of key pro-inflammatory and pro-fibrotic signaling pathways(including NF-κB, MAPK, and TGF-β) which play a central role in cardiac hypertrophy [43]. It is of great significance to construct physiologically relevant models for the prevention and potential reversal of hypertrophic progression given an intimate association of CH with heart failure and cerebrovascular diseases.

In this part, COs were treated with 800 nM Ang II to induce CH, with simultaneous administration of two pharmacological agents, metoprolol (a β-adrenergic receptor antagonist) and apocynin (a NADPH oxidase inhibitor) [44, 45], to assess their cardioprotective effects. This study monitored the beating behaviors of COs, including beating rhythm and frequency (Fig. 6A). In real-time analysis, the Ang II-induced CH model exhibited prolonged beating intervals and increased amplitude, indicating functional impairments associated with hypertrophic remodeling. In contrast, both metoprolol- and apocynin-treated groups showed a restoration of beating frequency and amplitude to levels comparable to the vehicle group (Fig. 6B).

Fig. 6.

Pathological characterization and cardioprotective drug assessment of Ang II-induced hypertrophy in COs. A Schematic of CH modeling. B Representative beating rhythm of individual organoids under: (i) Vehicle, (ii) Ang II-induced hypertrophy, (iii) Ang II + metoprolol, (iv) Ang II + apocynin. C Representative immunofluorescence images of F-actin for each condition. F-actin (red), DAPI (blue). Scale bar: 50 μm. D aspect ratio of each condition (n = 15 independent organoids per condition). E mRNA expression of hypertrophy-associated marker genes (n = 15 independent organoids per condition across three independent experiments, mean ± SD, One-way ANOVA with Dunnett’s multiple comparisons test, ***p < 0.001). Significance was determined by comparing with the CH group

Previous studies have demonstrated that CH is associated with sarcomere disorganization [46], which directly impacts CM beating function. Sarcomeres, composed primarily of actin and myosin filaments, are essential for coordinated heart beating and relaxation Phalloidin staining revealed disrupted cytoskeletal organization and reduced cardiomyocyte aspect ratio (cellular widening) in the CH model, compared to uniformly distributed actin filaments in controls (Fig. 6C, D).This morphological pattern is consistent with pressure-overload hypertrophy [46]. Furthermore, disrupted cytoskeletal structure and reduced cellular continuity were observed in the CH model, further supporting the impairment of structural integrity induced by Ang II.

According to further mRNA expression analysis, the expression of hypertrophy-associated marker genes NPPA, NPPB, and MYH7 was significantly upregulated in the Ang II model group, confirming the establishment of CH (Fig. 6E). Metoprolol treatment partially reduced the expression of these genes compared to the Ang II group, yet without restoration to the normal level. In contrast, Apocynin treatment significantly downregulated the expression levels of NPPA, NPPB, and MYH7, suggesting superior protective effect against Ang II-induced hypertrophy. This differential effect can be attributed to the distinct pharmaceutical mechanisms of these two drugs. To be specific, metoprolol, as a β1-adrenergic receptor blocker, primarily reduces sympathetic activation to lower heart rate and myocardial beating, thereby alleviating cardiac load indirectly; while Apocynin can directly inhibit the activity of NADPH oxidase (NOX), reduce reactive oxygen species (ROS) production, as well as mitigate oxidative stress and inflammation. In Ang II-induced hypertrophy, NOX activation induce excessive ROS generation, activating a cascade of key pro-inflammatory and pro-fibrotic signaling. By directly suppressing NOX activity, Apocynin attenuates ROS production. This effectively interrupts the oxidative stress-driven pathological remodeling process, thereby preventing cardiomyocyte hypertrophy. Hence, Apocynin may be more effective than metoprolol in mitigating Ang II-induced CH.

Discussion

It has been established that iPSCs-derived COs serve as a type of highly relevant in vitro model for the human heart, enabling the replication of its complex structure, in vivo cellular distribution, responsiveness to external stimuli, and electrophysiological functions [47, 48]. These organoids have emerged as indispensable tools for assessing drug-induced cardiotoxicity, elucidating the mechanisms underlying CVDs, and conducting preclinical drug testing [24, 47]. On April 10, 2025, the U.S. Food and Drug Administration issued guidance to encourage the integration of human-relevant test systems (e.g., organoids and organ-on-a-chip) into drug safety evaluation and regulatory submissions, offering streamlined review for qualifying non-animal testing data under its New Approach Methodologies advancement framework [49]. Such a policy shift highlights a well-established recognition of the significance of organoid platforms as cutting-edge tools for reducing animal use, improving predictive accuracy, and accelerating drug approval processes. Consequently, the efficient generation of robust and homogeneous COs capable of faithfully modeling cardiac diseases has become an inevitable trend in pursuing novel therapies and drug development strategies for cardiovascular disorders.

In the present study, we established a comprehensive platform detailing the full workflow from iPSC generation to drug testing. A key innovation is the optimization of CO construction based on a dynamic culture system. In contrast to reported methods using 96-well ultra-low attachment plates in static culture (one organoid per well), this modified method we proposed employed an optimal seeding density for 6-well plates and incorporated continuous shaking throughout the process. Through decreased consumption of multi-well plates and culture medium, it can contribute to significant reduction of the experimental and labor costs, as well as handling time associated with medium changes. It can also accelerate organoid maturation by enhancing the distribution of nutrients, oxygen [50], pH and exposure of the organoids to continuous mechanical stimulation [51–53]. The proposed dynamic culture system demonstrates clear advantages during late differentiation. The organoids consistently maintain a diameter of 200–300 μm, resulting in a remarkable reduction in the likelihood of central necrosis and supporting coordinated beating [53]. In contrast, static culture in multi-well plates yields a broader diameter range (1–1.6 mm) [20] and irregular beating patterns. Therefore, this approach can also facilitate more efficient large-scale cultivation, in addition to strengthening the reproducibility and functionality of COs.

Furthermore, this study validated the cardiotoxicity of anthracyclines by determining an IC₅₀ of 1.384 µM for doxorubicin in COs following 48 h of exposure. It was lower than the IC₅₀ reported in AC16 human ventricular cardiomyocyte cell line (2.5-5 µM) [54], and closely approximated the value observed in neonatal rat primary CMs (≤ 1 µM) and iPSC-CMs (1–3 µM) [55]. The increased sensitivity in COs observed in our study may be attributed to the presence of multiple cardiac cell types, preserved micro-environmental signaling, and enhanced tissue-level responses that are absent in immortalized cell lines. The strong concordance across these systems highlights the robustness and translational relevance of our CO-based platform in modeling chemotherapy-induced cardiac injury. Moreover, this platform could also accurately recapitulate arrhythmias induced by various compounds, including isoproterenol, MYK461, E-4031, and nifedipine, as determined by video-based motion analysis of COs through the quantification of beating frequency, amplitude, and beat interval variability. Thalidomide was demonstrated here to inhibit the differentiation of both ventricular CMs and vascular endothelial cells in COs, accompanied by a marked reduction in spontaneous beating activity. Furthermore, embryotoxicity screening, beside functional and toxicological evaluation, was also scheduled for this platform by capturing critical aspects of early cardiac development, such as chamber formation, cell lineage specification, and vascular morphogenesis. Through comprehensive assessments, it boosted our in-depth identification of structural and differentiation abnormalities that are often undetectable in conventional 2D culture systems. This system enables a faithful modeling of the complicated process of early heart development, thus enhancing the predictive power for developmental cardiotoxicity and formulating a more reliable tool for detecting the teratogenic potential of drug candidates, especially those that could impair long-term cardiac maturation and morphogenesis. These findings further suggest the utility of COs in teratogenicity screening frameworks aligned with ICH S5(R3) and OECD developmental toxicity guidelines.

Another pivotal advancement of this study is the establishment of COs as a pathomimetic platform for modeling MI and CH. Through preserved cellular interactions, electrical coupling, and 3D tissue architecture, these disease models recapitulate enhanced physiological relevance. In our study, COs responded to H/R injury and Norepinephrine stimulation, as demonstrated by elevated LDH release, structural damage, and abnormal beating, which recapitulated the core pathological features of MI, covering cell apoptosis, tissue injury, and dysfunction in beating activity. The introduction of CsA and SLT in the experimental setup contributed to significant alleviation in cell apoptosis and mitigation in tissue damage, thus supporting their potential therapeutic value in ameliorating H/R-induced injury. As mentioned previously, Ang II-induced COs were observed with arrhythmic beating and interstitial fibrosis deposition, both of which were pathological hallmarks of clinical CH [56]. Critically, mRNA analysis demonstrated marked upregulation of NPPA, NPPB, and MYH7 after CH modeling in relative to the normal, validating the establishment of CH model within COs. Metoprolol and Apocynin were individually administered to assess their therapeutic effects on Ang II-induced COs. Notably, Apocynin exhibited significantly greater efficacy in attenuating hypertrophy-associated gene expression, similar to previously reported roles of oxidative stress and metabolic dysregulation in driving CH [5]. The imbalance between sarcomere synthesis and degradation is a hallmark of hypertrophy, and our findings support the potentially vital function of Apocynin with antioxidant properties in mitigating this imbalance and protecting against hypertrophic remodeling. Those findings provide valuable insights for exploring the cardioprotective mechanisms of different drugs, though the specific mechanisms and clinical relevance require further investigation.

Still, it is noticed that our CO platform also exhibits several challenges, despite the demonstration of significant potential in drug screening, disease modeling, and toxicity assessment. For example, its limitations may include the inability to fully replicate the QRS complex observed in clinical electrocardiograms, insufficient vascularization, and the lack of an immune system, eventually restricting the modeling of cardiac inflammatory, immune, and repair processes. Furthermore, quantitative morphometric analysis of cardiomyocytes—a valuable metric for assessing hypertrophy, disarray, or other structural pathologies—remains technically challenging within complex 3D organoids and represents an important direction for future methodological development. In the future, the key lies in improving electrophysiological monitoring by incorporating more quantitative electrophysiological measurements-such as field potential duration, arrhythmia index, and conduction velocity-to accurately assess electrophysiological changes during cardiac disease modeling. This will enable a more precise evaluation of the physiological relevance and applicability of organoid disease models. Moving forward, we will investigate deeper disease mechanisms and employ more comprehensive detection indicators to evaluate the efficacy of small molecule candidates or assess drug-induced cardiotoxicity.

Conclusions

In summary, this study establishes a robust and scalable platform for generating and applying iPSC-derived COs, offering a physiologically relevant in vitro model for cardiovascular research. This system may facilitate the achievement of enhanced organoid maturation, structural integrity, and functional consistency by optimizing culture conditions, particularly by implementing dynamic shaking systems. The advantages of this platform may involve a successful recapitulation of drug-induced cardiotoxicity, including anthracycline sensitivity and compound-specific arrhythmias, and capture of the critical aspects of developmental and embryotoxic cardiac responses, such as chamber formation and vascular differentiation. Findings in our study also document the translational utility of COs in modeling MI and CH, providing disease-relevant readouts and therapeutic response profiles. Altogether, the proposed CO-based platform may offer a powerful and versatile tool for preclinical drug screening, developmental toxicity evaluation, and mechanistic studies in CVD research, even with existing limitations in vascularization, immune competence, electrophysiological complexity, etc.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CVD

Cardiovascular disease

MI

Myocardial Infarction

CH

Cardiac Hypertrophy

hiPSCs

Human induced pluripotent stem cells

hiPSC-CMs

Human induced pluripotent stem cells-derived cardiomyocytes

COs

Cardiac organoids

STR

Short tandem repeat

Ang II

Angiotensin II

HDF-A

Human Dermal Fibroblasts-adult

EBs

Embryoid bodies

DOX

Doxorubicin

LDH

Lactate dehydrogenase

H/R

Hypoxia/reoxygenation

SLT

Selatogrel

CsA

Cyclosporin A

Met

Metoprolol

Apo

Apocynin

NOX

NADPH oxidase

ROS

Reactive oxygen species

Author contributions

HCJ and ZQ conceived and designed the study. HCJ and ZQ performed all experiments, data analysis and graphing, and drafted the manuscript. ZHF, ZZY, LXY, PHP, and WF provided experimental assistance. ZHF, WJH, and SBF contributed to the critical revision of the manuscript. LJF and ML conducted a comprehensive review and language refinement of the entire manuscript. All authors have read and approved the final version of the manuscript.

Funding

The research was funded by Bio-Sincerity Pharma-Tech Co., Ltd., and all authors were supported by the company.

Data availability

All data generated or analyzed in this study have been included in this manuscript and supplementary file.

Declarations

Ethics approval and consent to participate

iXCells Biotechnologies has confirmed that there was initial ethical approval for the collection of HDF-A, and that signed informed consent had been provided by the donors. All animal-related procedures involved in the organoid platform construction project underwent rigorous ethical review and were formally approved by the Institutional Animal Care and Use Committee (IACUC) of Hangzhou Bio-Sincerity Pharma-Tech Co., Ltd. (Approval No.: ACU2025-002), with the authorization issued on January 2, 2025. The approval confirms full compliance with international animal welfare standards and ethical research protocols. Throughout project execution, we strictly implement the 3Rs principles (Replacement, Reduction, and Refinement) to ensure humane treatment of research animals.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.