5000+ ORGANOIDS
5000+ ORGANOIDS
5000+ ORGANOIDS
Comprehensive pre-clinical assessment of dose-dependent cardiotoxicity in the SmartHeart platform
-
HIGHLIGHTS
→ The SmartHeart platform enables multiple read-outs, including contractility, electrophysiology and calcium transients, allowing a comprehensive assessment of drug cardiac effects.
→ SmartHeart platform can detect single and multiple ion channels effects
→ SmartHeart platform discriminates different classes of risks of arrhythmia (TdP risk) in the CiPA panel
→ SmartHeart platform can recapitulate the effect of ranolazine and terfenadine, which has not been reproduced in 2D
→ SmartHeart platform recapitulates the variations in contractile force (inotropic effects) induced by cardioactive drugs
-
SUMMARY
Cardiotoxicity remains a leading cause of drug attrition, and, although significant progress has been made in the assessment of drug-induced cardiac risks, we still lack easy-to-use physiologically-relevant models, scalable to carry out a fast and systematic drug screening on the 3 major and complementary cardiac read-outs: contractility, electrophysiology and calcium transients. In this white paper we evaluate the effect of nine cardioactive drugs in the SmartHeart platform, which enables the self-assembly of cardiac cells in ring-shaped organoids. The drug responses on electrophysiology, contractility and calcium transients were assessed, and the expected effects were recapitulated for all the drugs. Therefore, we show here that this platform captures and discriminates complex drug mechanisms, distinguishing different classes of risk of arrhythmia (TdP risk), and recapitulates phenotypes not reproduced in 2D, offering a comprehensive and physiologically accurate in vitro drug screening assay.
-
INTRODUCTION
Drug development remains one of the most resource-intensive processes in modern medicine, requiring over a decade and $1-2 billion in investment per approved compound [1]. Cardiotoxicity stands as a primary contributor to drug failure, 24% of drugs encountering safety issues during clinical trials and 32% of drugs having identified cardiotoxicity after reaching the market. These unforeseen cardiac adverse effects present serious risks to patients’ wellbeing and impose substantial financial burdens on pharmaceutical organizations.
The assessment of drug-induced cardiac risks has undergone considerable transformation over the past decades. Early approaches concentrated on QT interval prolongation and hERG channel inhibition, prompted by the withdrawal of compounds like terfenadine following reports of dangerous arrhythmias (Torsades de Pointes). Subsequent regulatory guidelines (ICH S7B, E14) established standardized testing protocols, yet excessive reliance on these singular biomarkers resulted in the elimination of compounds that may have possessed acceptable safety profiles [2].
The Comprehensive in vitro Proarrhythmia Assay (CiPA) initiative, launched in 2013 [3], addressed these limitations through a more comprehensive mechanistic framework. CiPA incorporates multi-channel ion current assessment, computational modeling of human cardiac electrophysiology, validation using human induced pluripotent stem cell (hiPSC)-derived cardiomyocytes, and integration with early clinical data. This paradigm represents a transition toward human-relevant methodology designed to improve the accuracy of drug safety predictions through standardized advanced testing protocols.
Conventional preclinical models have demonstrated significant limitations in their ability to predict clinical outcomes. Standard in vitro toxicity screening approaches, including two-dimensional cell cultures and animal studies, inadequately represent the complexity of human cardiac tissue and the heterogeneity of patient populations encountered in clinical practice. The limitations of animal models in predicting human responses became particularly evident through clinical failures such as the TGN1412 incident [4]. This drug, used to cure leukemia and rheumatoid arthritis, induced life-threatening conditions involving multiorgan failure in all the healthy human volunteers of the phase 1 of the clinical trials. This highlighted fundamental species-specific differences in drug response.
New Approach Methodologies (NAMs) have gained prominence in drug development, driven by scientific progress, regulatory changes, and ethical considerations. The FDA, EPA, and NIH have reduced requirements for mandatory animal testing, recognizing accumulating evidence regarding the limited predictive value of traditional animal models for human outcomes. This regulatory shift, mirrored in European initiatives, indicates a permanent transition toward human-relevant testing methodologies and represents a renewed emphasis on mechanistic understanding of human-specific biological processes.
Human induced pluripotent stem cells (iPSCs) provide expanded opportunities for in vitro testing through the generation of cell lines from healthy donors or patients. The reprogramming of accessible adult cells into iPSCs, followed by directed differentiation into cardiac lineages, enables the creation of human cardiac cell resources previously unavailable for research applications. These cells can be organized into three-dimensional tissue architectures that more accurately reflect the cellular composition and functional characteristics of native human cardiac tissue.
Organoid technologies bridge the gap between simplified cellular models and complex human physiology. Cardiac organoids and heart-on-chip platforms integrate advances in stem cell biology, tissue engineering, and microsystem technologies to provide controlled access to human cardiac physiology in vitro. Three-dimensional tissue organization enables functional assessment through clinically relevant, non-destructive measurements including contractile force generation, beating patterns, and electrophysiological responses.
The SmartHeart platform advances organoid technology through its ring-shaped architectural design, which more accurately reproduces the mechanical forces experienced by cardiac tissue in physiological conditions. This technology addresses the need for scalable platforms capable of supporting high-throughput screening workflows while maintaining the biological fidelity required for accurate human cardiac response assessment.
The SmartHeart platform addresses fundamental challenges in cardiotoxicity evaluation through several technical innovations. The ring-shaped geometry facilitates physiologically appropriate force distribution within engineered cardiac tissues, while the scalable design enables production of up to 2016 individual tissue constructs per 96-well plate, compatible with automated high-throughput screening systems. Controlled substrate stiffness parameters ensure physiological relevance for contractile force measurements, and the glass-bottom configuration supports high-resolution imaging and live cell monitoring for detailed functional analysis.
The platform integrates three essential cardiac assessment modalities: contractility measurement, action potential recording, and calcium handling analysis. This multi-parameter approach provides comprehensive cardiac function evaluation that exceeds the capabilities of traditional single-endpoint assays. Through the use of hiPSC-derived cardiomyocytes organized within self-assembled 3D tissue structures, SmartHeart reproduces key aspects of native cardiac physiology, including appropriate responses to inotropic stimulation and ion channel modulation.
In this whitepaper we display the dose-responses of nine cardioactive drugs with different mechanisms on the SmartHeart platform in electrophysiology, contractility and calcium transients. This whitepaper illustrates how the expected responses to these nine drugs were recapitulated, showing that the SmartHeart platform can detect single- and multiple-channel effects, discriminate different classes of torsade de pointes (TdP) risks, and recapitulate the effects of inotropic drugs. Therefore, we establish here the potential of the SmartHeart platform for next-generation cardiac safety assessment, offering enhanced physiological relevance and improved translational potential compared to existing methodologies.
-
METHODOLOGY
Ring-shaped tissues were generated with SmartHeart 96 well-plates, in which a mix of cardiomyocytes derived from induced pluripotent stem cells and cardiac fibroblasts was seeded.The cells then self-assembled into contractile cardiac rings, and their responses to several drugs were assessed in electrophysiology, contractility and calcium transients, offering a thorough overview of the drugs effects on cardiac physiology.
In electrophysiology and calcium transients analyses, fluorescent dyes were used, and the traces of their fluorescence in time were analyzed to retrieve numerous metrics, described in Figure 1. Tissue contractility was assessed by tracking the ring shape in time, and several contractile metrics could be derived from the traces. The study of the evolution of the different metrics described in Figure 1 was carried out for the tested drugs and is detailed in the Results section. A more detailed Materials & Methods section can be found at the end of this paper.
Figure 1: Description of the multiple electrophysiology, contractility and calcium transients descriptors available the SmartHeart platform
-
RESULTS
Nine cardioactive compounds, with diverse actions, were tested on the SmartHeart platform, and their dose-response was assessed on electrophysiology (8 drugs), contractility (9 drugs) and calcium transients (2 drugs), according to the relevance of the read-outs. Table 1 recapitulates the effects on the different readouts of the tested drugs.
Table 1: Summary of the effects of tested compounds
Seven drugs from the CiPA panel were tested. These drugs are issued from different classes of TdP risks and involve distinct mechanisms.
Dofetilide is a selective potassium current (IKr) blocker. Potassium ion channels are involved in the repolarization of the heart. Dofetilide is also known to induce a high risk of arrhythmia. In agreement with this, we observed a prolongation of action potential duration, with a strong increase (>2 fold) in late action potential durations APD50-90 (durations from peak to 50% to 90% repolarization), as well as a strong decrease in frequency (negative chronotropic effect) and contraction stress (negative inotropic effect). At the highest dose, up to 38% of the rings became quiescent, and the beating irregularity increased up to 9 fold.
Figure 2: Dofetilide dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=29 tissues; and in contractility (lower panel): N=2 experiments, n=27 tissues
Disopyramide, a molecule targeting sodium and potassium currents (INa and IKr), engenders a dual effect. First, the sodium channel blockade affects the depolarization of the heart: here we observe a decrease in the depolarization speed and action potential (AP) amplitude. Secondly, the potassium channel blockade hampers the repolarization, and we see a strong increase in late action potential durations, as shown for dofetilide.
Concurrently, the tissues displayed a strong decrease in contraction stress and beating irregularity with increasing doses.
Figure 3: Disopyramide dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=25 tissues; and in contractility (lower panel): N=2 experiments, n=30 tissues
Domperidone is a dopamine antagonist which has been proven to inhibit the activity of potassium and sodium channels (hERG and Nav1.5) [5], [6]. The SmartHeart platform recapitulates this dual effect as we observe the same effects on depolarization and repolarization as with disopyramide: a significant decrease in the depolarization speed (maximum rise slope of the action potential), the action potential amplitude and the beating frequency, due to Nav1.5 blockade, as well as a strong increase of the late repolarization durations (APD50-90), due to hERG inhibition. This observed delayed depolarization and repolarization resulted in a decrease in contraction stress, and maximum contraction and relaxation speeds, as well as an increase in interpeak irregularity. Also, for the highest concentration, the ratio of quiescent rings increased up to 33%.
Figure 4: Domperidone dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=29 tissues; and in contractility (lower panel): N=2 experiments, n=16 tissues
Metoprolol is a beta-blocker, selective to 𝛽1 receptor and has been shown to block the hERG potassium channel at high concentrations [7]. In SmartHeart, we observed the expected decrease in frequency and contraction stress, as well as an increase in contraction irregularity, depolarization speed, and late APDs (APD50-90). Metoprolol also induces a side effect on calcium transients by decreasing the activity of L-type calcium channels and ryanodine receptors RyR2. This could be recapitulated in a preliminary experiment as we observed a decrease of the calcium transients amplitude and an increase in the time of calcium release, explained by L-type calcium channels’ decreased activity.
Figure 5: Metoprolol dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=35 tissues; in contractility (middle panel): N=2 experiments, n=25 tissues; and in calcium, paced at 1.5Hz (lower panel): N=1 experiment, n=9 tissues.
Nifedipine is an L-type calcium channel blocker. L-type calcium channels play an important role in the plateau phase of the action potential, as the entrance of calcium compensates the potassium release, leading to a stable cell potential. A blockade of these channels induces a decrease in the plateau phase duration and an overall decrease in action potential duration. This is recapitulated in the SmartHeart as we observe shortened APDs and an impaired depolarization, with a decrease in depolarization speed and action potential amplitude, in which calcium current is also involved in iPSC-CMs. The shortening of the action potentials goes along with an increase in the beating frequency. Moreover, a preliminary calcium transient experiment shows a decrease in the calcium transient duration, amplitude and the calcium release rate (max rise slope), in agreement with a calcium channel blockade.
Figure 6: Nifedipine dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=25 tissues; in contractility (middle panel): N=2 experiments, n=25 tissues; and in calcium (lower panel): N=1 experiment, n=3 tissues.
Terfenadine is an antihistaminic drug, which blocks potassium and sodium channels (hERG and Nav1.5), impeding both the depolarization and repolarization phases, as seen with disopyramide. Accordingly, we could observe a significant decrease in the depolarization speed and the action potential amplitude and we also observed an increase of the late repolarization (APD50-90). The contraction frequency was also decreased and the interpeak irregularity increased with increasing doses.
Figure 7: Terfenadine dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=31 tissues; and in contractility (lower panel): N=2 experiments, n=25 tissues.
Ranolazine is a molecule inhibiting the late sodium current INa,L and the potassium current IKr. It also has a negative effect on calcium influx. The electrophysiological readouts recapitulate this multiple ion channels effect, as they exhibit a decrease in action potential amplitude and depolarization speed, reflecting INa,L blockade, as well as a moderate late repolarization prolongation from APD50, due to the compensating effects of IKr blockade, prolonging the repolarization and ICa,L blockade, shortening the plateau phase.
Figure 8: Ranolazine dose-response in SmartHeart. In electrophysiology (upper panel): N=3 experiments, n=34 tissues.
In addition to these drugs from the CiPA panel, 2 inotropic (affecting contraction force) drugs were tested.
Isoprenaline is a 𝛽 receptor agonist. The activation of these receptors induce an increase in contraction rate and force. In agreement with this, increasing doses of isoprenaline induced a positive chronotropic effect (increased frequency) and a positive inotropic effect (increased contraction stress). Increased contraction and relaxation speeds were also observed.
Figure 9: Isoprenaline dose-response in SmartHeart. In contractility: N=2 experiments, n=15 tissues.
Verapamil is a calcium inhibitor. Calcium is a crucial ion which enables the cardiomyocyte contraction. Here, we could recapitulate the decreased contraction stress (negative inotropic effect), as well as the decrease in contraction speed.
Figure 10: Verapamil dose-response in SmartHeart. In contractility: N=3 experiments, n= 32 tissues.
-
DISCUSSION
First, we could first validate that the SmartHeart platform enables distinguishing different mechanisms. Indeed, we could show the expected response with single channel or dual- to triple-channel effects, proving that SmartHeart can capture complex pharmacological mechanisms.
Also, we see that the strength of the APD increase in the K+ blockers is stronger in the High TdP and Intermediate TdP drugs than in the low TdP ones. The CiPA initiative was created to discriminate hERG blockers which induce a high risk of TdP from promising drugs which happen to be hERG blockers but do not induce any TdP.[3] In this prospect, the SmartHeart platform completes this goal as it can discriminate against different classes of TdP drugs.
Additionally, the SmartHeart platform could recapitulate the effects of terfenadine and ranolazine, which are typically not well recapitulated in 2D.
Indeed, the limitations of traditional 2D hiPSC-cardiomyocyte models in accurately predicting clinical arrhythmogenic risk are exemplified by their misclassification of well-characterized compounds. Ranolazine, widely accepted as having no TdP risk in clinical practice, is frequently misclassified as medium to high-risk due to its hERG blocking activity and QT prolongation effects [8], despite its protective anti-arrhythmic properties through late sodium current inhibition [9]. Conversely, terfenadine—a known TdP-inducing agent—often fails to demonstrate expected proarrhythmic effects in 2D systems, with studies reporting no early afterdepolarization (EAD) occurrence and only beat arrest responses [10]. These discrepancies highlight the inadequacy of simplified cellular models to capture the complex multi-channel interactions and integrated electrophysiological responses that determine true arrhythmogenic potential. The SmartHeart platform's 3D tissue architecture and physiological cell-to-cell coupling may provide the integrated cellular environment necessary to accurately recapitulate both the protective effects of ranolazine and the proarrhythmic mechanisms of terfenadine, addressing fundamental limitations in current cardiac safety assessment paradigms
Finally, the SmartHeart platform could recapitulate both the positive chronotropic and inotropic responses to increasing concentrations of isoprenaline. The contractile response to 𝛽-adrenergic receptor stimulation and the positive force-frequency relationship are hallmark features of the tissue maturation [11],[12] , and are not recapitulated in conventional 2D culture of hiPSC-CMs, and can be shown only in mature 3D cardiac models [13]. The SmartHeart platform could also recapitulate verapamil’s expected negative inotropic effect. Demonstrating that the SmartHeart can detect inotropic effects in addition to electrophysiological alterations confirms the clinical relevance of this platform, offering the opportunity of a comprehensive and physiologically accurate drug screening.
-
MATERIALS & METHODS
Tissues generation and culture
For the following experiments the tissues were generated with the SmartHeart® 96 well-plate, in which a total of 65k cells were seeded in each well. The cell mix was a ratio of 3:1 hiPSC-CMs:fibroblasts. The required numbers of Human Ventricular Cardiac Fibroblasts (Innoprot ; P10453) and iCell® Cardiomyocytes² (FUJIFILM Cellular Dynamics ; Catalog #: R1059) were mixed and resuspended in 65 μL per well of iCMM + 1:1000 Y-27632 . The plate was then kept at 37°C, 100µL of iCMM per well was added the next day and the medium was changed every 2 days. The dose-responses were evaluated after 14 days in culture.
Contractility Recording of 4Dcell SmartHeart cardiac organoids
SmartHeart cardiac rings were maintained at 37°C throughout the experiment using a heated microscope stage with temperature feedback control (±0.5°C).. Test compounds and vehicle control (0.1% DMSO) were pre-warmed to 37°C before administration.
Contractile activity was recorded using an inverted microscope equipped with a 10× objective (no binning, 1 ms exposure) and high-speed camera operating at minimum 35 frames/sec. Baseline contractility was recorded for 11 seconds per cardiac ring before drug treatment. Compounds were administered in sequential doses with 10 minute incubation periods between recordings. At each dose level, contractile activity was recorded for 11 seconds focusing on the outer edge of cardiac rings where motion was most visible.
Videos were analyzed with the SmartX2.0 Software, based on a deep-learning algorithm, tracking the rings through time. The software then provides multiple contractility metrics such as beating frequency, contraction strain, contraction stress, contraction and relaxation speeds and frequency regularity.
Optical mapping using voltage-sensitive fluorescent dye in 4Dcell SmartHeart® cardiac organoids
To assess action potential kinetics in 4Dcell SmartHeart® cardiac organoids, samples from day 14 to day 17 of culture were prepared as follows: organoids were gently washed once with Tyrode’s solution (Sigma-Aldrich, J67607.AP) to remove residual medium. Organoids were then incubated for 30 minutes at 37°C and 5% CO₂ in Tyrode’s solution containing Fluovolt™ (1:1000 ; Invitrogen™ F10488) supplemented with PowerLoad™ concentrate (1:100 ; Invitrogen™ F10488). Following incubation, excess dye was removed by a single wash with fresh Tyrode’s solution. Membrane potential imaging was performed using a Leica THUNDER imaging system equipped with a 20× objective. Fluorescence was captured at 475 nm excitation (15% laser intensity) with 2×2 binning and a 9 ms exposure time, acquiring 525 frames. All imaging was conducted at 37°C to maintain physiological conditions, and samples were protected from prolonged light exposure to minimize photobleaching. Data were analyzed with an in-house code to quantify AP amplitude, frequency, durations and kinetics, ensuring robust assessment of cardiac function and drug responses in the engineered tissue platform.
Calcium transient imaging in 4Dcell SmartHeart® cardiac organoids
To assess calcium transients in 4Dcell SmartHeart® cardiac organoids, samples at day 25 of culture were prepared as follows: organoids were gently washed once with Tyrode’s solution (Sigma-Aldrich, J67607.AP) to remove residual medium. Organoids were then incubated for 20 minutes at 37°C and 5% CO₂ in Tyrode’s solution containing 5 µM Cal-520 AM (AAT Bioquest, Cat#21130) supplemented with 0.02% acetoxymethyl (Ac) Puronic F-127 (AAT Bioquest, Cat#20053) to enhance dye dispersion and cellular uptake. Following incubation, excess dye was removed by a single wash with fresh Tyrode’s solution. Calcium transient imaging was performed using a Leica THUNDER imaging system equipped with a 20× objective. Fluorescence was captured at 475 nm excitation (15% laser intensity) with 2×2 binning and a 5 ms exposure time, acquiring 525 frames. All imaging was conducted at 37°C to maintain physiological conditions, and samples were protected from prolonged light exposure to minimize photobleaching. Data were analyzed with an in-house code to quantify calcium transient amplitude, frequency, and kinetics, ensuring robust assessment of cardiac function and drug responses in the engineered tissue platform.
Drug test
The 9 following compounds were tested in this study, at different doses. The diluted compound solutions were warmed up at 37°C before administration.
Statistical analysis
Differences between experimental and DMSO control groups were analyzed with Wilcoxon paired tests. p-Values <0.05 were considered significant for all statistical tests.
Report of statistical significance: *:p<0.05; **:p<0.01; ***:p<0.001; ****:p<0.0001.
-
REFERENCES
First, we could first validate that the SmartHeart platform enables distinguishing different mechanisms. Indeed, we could show the expected response with single channel or dual- to triple-channel effects, proving that SmartHeart can capture complex pharmacological mechanisms.
Also, we see that the strength of the APD increase in the K+ blockers is stronger in the High TdP and Intermediate TdP drugs than in the low TdP ones. The CiPA initiative was created to discriminate hERG blockers which induce a high risk of TdP from promising drugs which happen to be hERG blockers but do not induce any TdP.[3] In this prospect, the SmartHeart platform completes this goal as it can discriminate against different classes of TdP drugs.