In vitro generation of human embryonic stem cell-derived heart organoids possessing physiological ion currents Authors: Jiyoung Lee, Hiroshi Matsukawa, Kohei Sawada, Rin Kaneko, Fumitoshi Ishino Published: 2026 ABSTRACT IntroductionRecently, methods for in vitro organogenesis have been widely developed due to their strong potential for medical applications. Cardiac organoids with functional ion channel activity are necessary for non-clinical applications.MethodsIn the present study, we optimized the culturing method of human heart organoids (hHOs) from human embryonic stem cells (hESCs) in the presence of the highly concentrated laminin-entactin and fibroblast growth factor 4 by the supplementation of BMP4, BMP10 and ROCK inhibitor (Y27632).ResultsThe resulting hHOs showed distinctive cardiac morphology with atrium- and ventricle-like chambers composed of cardiac cells, smooth muscle and endothelial cells as well as expressed the integral proteins of gap junctions like Connexin 40 (Cx40), Cx43 and Cx45, and ion channel proteins including Kir2.1(KCNJ2), hERG (KCNH2), Nav1.5 (SCN5A), and Cav1.2 (CACNA1C). In addition to the contractile ability in hHOs, isolated cardiomyocytes from these hHOs exhibited Na and Ca currents by patch clamp analysis.ConclusionsThese results indicate that hHOs generated in this study retain physiological cardiac functions, as they recapitulate in vivo cardiac ion currents and cardiac muscle contraction, although their abilities remain limited. The present method for hHOs will contribute to understanding the fundamental mechanism of cardiogenesis and provide an alternative tool for cardiac safety assessment of newly developed drugs as an in vitro human ESC-derived test system. Abbreviations human heart organoid human embryonic stem cell human ether-a-go-go-related gene laminin-entactin 6-bromoindirubin-3′-oxime Introduction Heart organoid-generating systems have potential applications in drug screening, regenerative medicine and structural and functional analyses of heart defects. Various differentiation protocols of cardiac cells have been developed: 1) in vitro cardiomyocyte differentiation from pluripotent stem cells [induced pluripotent stem cells (iPSCs) and embryonic stem cells (ESCs)] [[1], [2], [3], [4], [5]] and 2) in vivo and in vitro direct reprogramming of cardiomyocyte-like cells from fibroblasts [[5], [6], [7]]. Given the increasing demand for the generation of structurally organized and functional heart organoids, we recently developed a method for three-dimensional (3D) heart organoids [8] with in vivo heart-like structures as well as cardiac functions implying electrophysiological properties by culturing mouse embryoid bodies (EBs) from ESCs with the laminin-entactin (LN/ET) complex [9] and fibroblast growth factor 4 (FGF4) [10]. Translational studies of the mouse heart organoid-generating system are necessary for human applications, and we established a method for human PSC-derived heart organoids based on the previous laminin and FGF4 method. Using human ESCs was a reasonable choice for easier maintenance of the undifferentiated state than human iPSCs; nevertheless, hESCs have been under ethical issues. In this study, we improved heart organoid culture conditions and generated hESC-derived heart organoids with cardiac structural features and electrophysiological properties. The generated hESC-derived heart organoids express gap junction proteins required for the cardiac intercellular conduction system, as well as critical ion channel proteins for generating action potentials, including Kir2.1, human ether-a-go-go-related gene (hERG), Nav1.5, and Cav1.2. Furthermore, the observation of functional Na+ and Ca2+ ion currents using patch-clamp methods indicates that the human heart organoids generated by the present method are feasible for nonclinical cardiotoxicity screening. Methods Cell culture The hESCs (gifted from Dr. Akutsu, National Center for Child Health and Development, Japan) were cultured and maintained in Primate ES cell medium (Reprocell) supplemented with bFGF (10 ng/mL) on SL10 feeder cells (Reprocell). The hESC lines SEES1 and SEES2 were used to generate heart organoids. Prior to EB formation, the hESCs were cultured twice on Matrigel (0.3 mg/mL per 1-well of a 6-well plate, BD) to remove feeder cells. For hESC-derived EB formation, the cultured hESCs on a Matrigel coat, washed with PBS (++) and incubated with 0.5 mL of CTK solution (Reprocell) for 3 min at 37°C. The dissociated hESCs were collected by a cell scraper and seeded in a U-bottomed 96-well plate (Prime surface 96-well U, Sumiron) in KSR- and NEAA-supplemented DMEM Glutamax medium (KSR-EB medium) with Y27632 (10 μM) and incubated at 37°C for 1 day. On culturing day 1, the medium was changed to fresh KSR-EB medium without Y27632 and incubated until culturing day 4. For the generation of hESC-derived heart organoids, we performed the in vitro culture of hESC-derived EBs by using a protocol previously described [8]. Briefly, hESC-derived EBs were placed onto 85.7 μg/cm2 LN/ET gel (BD 354259)-coated chamber slides (Falcon 354118) in 200 μL of heart organoid basic medium [50 μg/mL penicillin/streptomycin, 20% KSR, 1 mM sodium pyruvate, 100 μM β-mercaptoethanol, 2 mM l-glutamine, 60 ng/mL progesterone, 30 ng/mL β-oestradiol, 5 μg/mL insulin, 20 μg/mL transferrin, and 30 nM selenite in DMEM/F12 (Gibco 11320033)] containing FGF4 (60 ng/mL). Subsequently, the hESC-derived EBs were incubated at 37°C and 5% CO2 for 15 days or further with medium changes on days 3, 5, 7, 9, 11, 13, and 15. Starting on day 9, BMP4 (50 ng/mL), BIO (2.5 μM), LIF (1000 units/mL) [7] (BBL), BIO, BMP4, BMP10, and LIF (BBBL) and additional supplementation with the ROCK inhibitor Y27632 (Y) were added to FGF4 heart organoid basic medium. Additionally, the hESC-derived heart organoids were transferred onto 10.7 μg/cm2 iMatrix-411 (Laminin 411, Nippi)-coated chamber slides at culturing day 11. After culturing, the hESC-derived heart organoids were collected for further analyses, including immunofluorescence staining, ultrastructural analysis and patch clamp analysis. Quantification of the sizes of heart organoids and statistical analysis To measure and quantify the sizes (areas) of EBs and heart organoids, the microscopic images at day 0 (for EBs) and day 15 (for heart organoids) were processed with Fiji software (an open source distribution of ImageJ). The images were first converted to 8-bit grayscale and then segmented into binary images using thresholding. For area quantification, the region of interest (ROI) was primarily delineated using the wand tool; for samples where the wand tool could not accurately capture the boundaries, the freehand tool was employed as an alternative. The ROI area was then calculated using the measurement tool. For the statistical analysis, GraphPad Prism 11.0 was used. A nested t-test was applied to the size change from EBs to day 15 heart organoids for each culture condition. Area data were presented as a superplot with individual data points (each EB and each heart organoid; small transparent dots) and the batch mean (large solid dot). A nested one-way ANOVA was applied to compare the size among culture conditions and violin plot was presented results. To show chamber-formation success rate and batch-to-batch variability in organoid generation, we plotted a bar plot with individual data points (indicating the mean for each batch). An ordinary one-way ANOVA was applied for statistical analysis. Also, the batch-to-batch variation in the success rate of chamber formation was assessed using the Brown-Forsythe test. Immunofluorescence staining Immunofluorescence staining was performed as reported previously [8]. hESC-derived heart organoids were collected after 15 days of culture and frozen in optimal cutting temperature (OCT) compound (Tissue-Tek) in a plastic tissue mould (Cryo dish, Shoei). The sections were sliced at 5 μm using a cryostat set at −16°C and transferred to MAS-coated glass slides (Matsunami). Additionally, mouse embryonic, postnatal and adult heart sections were prepared from C57BL/6 mouse embryos from 9.5 to 13.5 days postcoitum and postnatal day 1 (P1) and adults (over 9 weeks). Prior to freezing in OCT, the mouse hearts were immersed in sucrose-PBS solutions (4%, 10%, 15%, and 20%). The cryosections were fixed in 4% paraformaldehyde-PBS at room temperature (RT) for 15 min. The fixed sections were washed three times in PBS for 5 min and incubated with a blocking buffer (PBS containing 5% normal goat serum and 0.3% Triton X-100 or PBS containing 5% bovine serum albumin (BSA) and 0.3% Triton X-100) at RT for 1 h. The sections were subsequently incubated overnight with primary antibodies in dilution buffer (PBS containing 1% BSA and 0.3% Triton X-100) at 4°C. After the sections were washed with PBS, they were incubated with secondary antibodies at RT for 1 to 2 h. The slides were counterstained with DAPI (diluted 1:1,000, Dojindo Laboratories) and mounted using a VECTASHIELD HardSet Antifade Kit (VECTOR Laboratories). Primary antibodies against the following proteins were used: MLC2A (MYL7, Synaptic System #311 011), MLC2V (MYL2, Synaptic System #310 003), Cx43 (GJA1, Sigma, C6219), Cx40 (GJA5, Invitrogen, #37-8900), Cx45 (GJC1, ab78408), Kir2.1 (KCNN4, IK1, KCNJ2, GTX54786), and potassium channel hKv11.1 (hERG, KCNH2, Sigma Aldrich, #P9497). The appropriate secondary antibodies (all from Invitrogen) for each primary antibody were also used. Whole-mount immunostaining by the tissue clearing method For whole-mount immunostaining by tissue clearing, we used the tissue clearing reagent CUBIC-L (TCI) as described in the manufacturer's protocol. The collected hESC-derived heart organoids were cultured for over 20 days, and mouse embryonic hearts were fixed in 4% PFA/PBS overnight at 4°C and washed with PBS three times for 2 h each at RT. The washed samples were delipidated by incubating in 750 mL of CUBIC-L solution in a shaker for 2 days at 37°C and then washed with PBS three times for 2 h each. For immunostaining, the samples were incubated in mixed antibodies (Alexa Fluor 647 Mouse Anti-cTnT; BD Phar, BV480 Rat Anti-Mouse CD31; BD Horizon, anti-aSMA antibody Alexa Fluor 568; Abcam) diluted with staining buffer (0.01% sodium azide and 0.5% Triton X-100 in PBS) for 3 days at RT and washed with PBS three times for 2 h each. For RI matching, the samples were incubated in 750 mL of CUBIC-R+ solution overnight at RT. The immunostained hESC-derived heart organoids were analyzed with a laser confocal microscope (Lightsheet, Zeiss). Comparison of cardiac marker gene expression with the public scRNA-sequencing dataset of human developing hearts To determine the developmental stage of heart organoids based on cardiac marker expression by immunostaining results, we analyzed the expression levels of these genes in the scRNA-sequencing dataset of human developing hearts. Data was obtained from NCBI GEO (Accession: GSE106118) [11]. Raw UMI counts were normalized and scaled to 10,000 (Counts Per 10,000, CP10K). The normalized values were then log-transformed using log(x + 1), and the values of HE5W, HE6W, HE7W, HE9W, and HE13W for left atrium (LA), left ventricle (LV), right atrium (RA), and right ventricle (RV) in target genes were presented as a dot plot (mean ± SEM). HERG channel trafficking inhibition assay For the inhibition assay of hERG channel trafficking, 17 days cultured hESC-derived heart organoids were cultured for over 16 h (overnight) under conditions in the absence or in the presence of 30 mM of pentamidine prepared in ddw. After incubation with/without pentamidine, hESC-derived heart organoids were collected, embedded in OCT compound, and cryosectioned at 5 mm using a cryostat. The cryosections were used for immunostaining with anti-mouse KDEL (10C3) antibody (Novusbio, #NBP1–97469SS) and hERG antibody (Sigma Aldrich, #P9497) and appropriate secondary antibodies. The cryosection slides were counterstained with DAPI, mounted and observed the subcellular localization of hERG channels under laser confocal microscope (Zeiss, LSM710). Electrophysiology Cell preparation Cells were prepared from hESC-derived heart organoids 18–30 days post-EB formation with an enzymatic digestion method as described previously [12]. Briefly, hESC-derived heart organoids were washed with cell isolation buffer (CIB) twice, 200 μL of enzyme mix [collagenase II (1 mg/mL, Worthington), collagenase IV (1 mg/mL, Sigma), 0.3 mM CsCl2 in CIB] was added, and the cells were incubated at 37°C for 15 min. The samples were mixed by pipetting (10 times), 500 μL of CIB–Ca2+-BSA was added, and small pieces of samples were broken up into single cells by pipetting 7 times. The dissociated cells were centrifuged at 2000 rpm for 2 min, washed with CIB–Ca2+-BSA, and finally centrifuged at 2000 rpm for another 2 min. The cell pellets were suspended in 100 μL of normal Tyrode solution (Sigma T2397). For attachment, isolated cells were transferred on round cover slips precoated with collagen type I (IWAKI, Tokyo, Japan) and maintained in Tyrode solution in a humidified culture dish for over 30 min until use for electrophysiological experiments. Patch-clamp A conventional patch clamp method was used in the whole-cell voltage-clamp mode [13]. Cover slips with the isolated cells from hESC-derived heart organoids were transferred to a temperature-controllable recording chamber and superfused with the extracellular solutions at a rate of 2–3 mL/min. Pipette electrodes were fabricated from borosilicate glass capillaries by a filament using a P-97 micropipette puller (Sutter Instrument, Novato, CA, USA). The pipette electrodes with electrical resistance between 2 and 5 MOhm were used. Membrane currents mediated by ion channels were filtered at 2 or 5 kHz using an Axopatch 200B amplifier and digitized at 20 kHz using pCLAMP 10 Software Suite (Molecular Devices, San Jose, CA, USA). Whole-cell voltage-clamp configuration was obtained by visual access under an upright microscope (Olympus, Tokyo, Japan). All the recordings were performed at 30–33°C. The intracellular solution contained the following (in mM): 50 CsCl, 60 CsF, 10 NaCl, 2 MgCl2, 20 ethylene glycol-bis (2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), and 10 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid (HEPES); the pH was adjusted to 7.25 with 1 M CsOH. The extracellular solution contained the following (in mM): 140 NaCl, 4 CsCl, 1 or 2 CaCl2, 1.2 MgCl2, 10 glucose, 10 HEPES; the pH was adjusted to 7.40 with 1 M NaOH. When Ca currents were recorded, another intracellular solution was used and contained the following (in mM): 112 CsCl, 27 CsF, 2 NaCl, 8.2 EGTA, 4 MgATP, 10 HEPES; the pH was adjusted to 7.25 with 1 M CsOH, and tetrodotoxin (TTX) was additionally included at 20 μM externally or a different extracellular composition was used as follows (in mM): 130 N-methyl-d-glucamine (NMDG), 4 CsCl, 4 BaCl2, 1.2 MgCl2, 10 glucose, 10 HEPES; the pH was adjusted to 7.40 with 1 M HCl. To ensure that the remaining currents after removing Na± currents were mediated by Ca channels, we added cobalt (Co) ions, which are known as an inorganic Ca channel blocker [14], to the extracellular solution at 4 mM. Organic and inorganic ion channel blockers, TTX and CoCl2, were purchased from Fujifilm and Wako Pure Chemical. For elicitation of inward currents through voltage-dependent ion channels, depolarizing step pulses were delivered at 0.1 Hz. Leak and capacitive currents were subtracted using a P/N protocol (8 times). P/N corrected waveforms were used to measure the peak amplitudes of the respective ion channels. Peak currents attained as incremental voltage pulses increased stepwise were normalized to the maximum value and plotted as a function of the voltages to generate current-voltage curves relationship (I–V curves). Peak analysis was performed using pCLAMP 10 Software Suite. Data are shown as the mean ± standard deviation. Ethics approval All animal experiments were approved by the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University (TMDU) (A2019-069A). All procedures were performed according to the guidelines of the institution and the national guidelines on the care and use of laboratory animals. All experiments using hESCs were approved by the institutional hESC Use Committee of TMDU (Title: Reconstitution of organogenesis by using hESC, Date: 20200508) and were performed according to the guidelines of this institution and the national guidelines. Results Heart organoid generation from human embryonic stem cells (ESCs) In a previous study, we generated functional 3D heart organoids from mouse ESCs in the presence of the laminin-entactin (LN/ET) complex via the exogenous growth factor, fibroblast growth factor 4 (FGF4) in vitro. These mouse heart organoids exhibited functional activities, including beating movement, cardiac muscle contraction and electrophysiological characteristics such as action potential propagation. Given that in vitro mouse heart organoids showed appropriate drug responses in electrophysiological studies, these organoids will be useful for preclinical cardiotoxicity screening. However, because of the different composition of ion channels and the dynamics of intracellular Ca as well as the different heart rate between rodents and human cardiac muscle [15], it is necessary to use biomimetic materials that reproduce human cardiac tissue instead of rodents, in order to overcome the problem of species difference (especially rodents) in the field of drug discovery. Particularly, the use of human ES derived heart organoids as surrogates for human cardiomyocytes (CMs) as well as human hearts would be more appropriate methodology for the evaluation of cardiac toxicity in newly developed drug screening. Therefore, we generated heart organoids by using human ESCs (hESCs). The hESCs were cultured and maintained with bFGF (10 ng/mL) in primate ESC culture medium, and then, embryoid bodies (EBs) were formed without bFGF in KSR-EB medium. To determine whether the LN/ET complex and FGF4 signalling are also key factors for the self-organization of human heart development, we cultured intact hESC-derived EBs on gelated LN/ET complexes with FGF4. Given that laminin 111 of the LN/ET complex is abundantly expressed during early heart development and laminin 411 is highly expressed in mouse embryonic day 12.5 (E12.5) hearts [8,16] (Supplemental Fig. 1A), we exchanged extracellular matrix from the LN/ET complex with laminin 411 (iMatrix-411) at culturing day 11. As expected, the hEBs showed cardiac morphogenic changes, as shown in the mEB-derived heart organoids. Consequently, hESC-derived heart organoids with multiple chambers were efficiently generated in the human heart organoid-generating system via a two-step extracellular environment (Fig. 1A and Supplemental Fig. 1B).Fig. 1In vitro recapitulation of developing hearts. (A) Experimental scheme to generate hESC-derived heart organoids (HOs) in vitro. The hEBs derived from hESCs were cultured on the gelated LN/ET complex with HO medium containing FGF4. (B) Morphogenetic shift during HO formation. The hEBs were cultured with FGF4 only (batch n = 6, hEBs n = 36, day 15 HOs n = 36), BBFL(batch n = 2, hEBs n = 12, day 15 HOs n = 11), BBBFL (batch n = 8, hEBs n = 80, day 15 HOs n = 74), and Y (batch n = 2, hEBs n = 12, day 15 HOs n = 10) conditions on the LN/ET complex. The batch indicates independent differentiation. Top: The sizes (areas) of EBs and HOs were measured from microscopic images at day 0 (for EBs) and day 15 (for HOs) using Fiji (ImageJ). For the statistical analysis, GraphPad Prism 11.0 was used. A nested t-test was applied to the size change from EBs to day 15 heart organoids for each culture condition. Area data were presented as a superplot with individual data points (each EB and each heart organoid; small transparent dots) and the batch mean (large solid dot). ∗∗P = 0.0019, ∗P = 0.0323, ∗∗∗∗P < 0.0001. Bottom: A nested one-way ANOVA was used to compare the sizes across culture conditions, and a violin plot with individual data points was used to display the results (left). Chamber formation was observed under the above conditions. The chamber-formation success rate in organoid generation, presented as a bar plot (mean with SD) with individual data points (indicating the mean for each batch), was analyzed using an ordinary one-way ANOVA (right). (C) Representative hESC-derived HOs with multiple heart chambers generated using the above culture system after culturing for 15 days. The left panel indicates hESC-derived HOs in the presence of FGF4 only, and the right panel indicates hESC-derived HOs in the presence of BBBFL. Both hESC-derived HOs showed remarkable morphogenic changes. See also Supplemental Fig. 1B. Scale bar: 400 μm. In vitro recapitulation of developing hearts. (A) Experimental scheme to generate hESC-derived heart organoids (HOs) in vitro. The hEBs derived from hESCs were cultured on the gelated LN/ET complex with HO medium containing FGF4. (B) Morphogenetic shift during HO formation. The hEBs were cultured with FGF4 only (batch n = 6, hEBs n = 36, day 15 HOs n = 36), BBFL(batch n = 2, hEBs n = 12, day 15 HOs n = 11), BBBFL (batch n = 8, hEBs n = 80, day 15 HOs n = 74), and Y (batch n = 2, hEBs n = 12, day 15 HOs n = 10) conditions on the LN/ET complex. The batch indicates independent differentiation. Top: The sizes (areas) of EBs and HOs were measured from microscopic images at day 0 (for EBs) and day 15 (for HOs) using Fiji (ImageJ). For the statistical analysis, GraphPad Prism 11.0 was used. A nested t-test was applied to the size change from EBs to day 15 heart organoids for each culture condition. Area data were presented as a superplot with individual data points (each EB and each heart organoid; small transparent dots) and the batch mean (large solid dot). ∗∗P = 0.0019, ∗P = 0.0323, ∗∗∗∗P < 0.0001. Bottom: A nested one-way ANOVA was used to compare the sizes across culture conditions, and a violin plot with individual data points was used to display the results (left). Chamber formation was observed under the above conditions. The chamber-formation success rate in organoid generation, presented as a bar plot (mean with SD) with individual data points (indicating the mean for each batch), was analyzed using an ordinary one-way ANOVA (right). (C) Representative hESC-derived HOs with multiple heart chambers generated using the above culture system after culturing for 15 days. The left panel indicates hESC-derived HOs in the presence of FGF4 only, and the right panel indicates hESC-derived HOs in the presence of BBBFL. Both hESC-derived HOs showed remarkable morphogenic changes. See also Supplemental Fig. 1B. Scale bar: 400 μm. To produce heart organoids with perfect structural similarity to the in vivo heart, we investigated several culture conditions, for example, FGF4 only; additional supplementation with 6′ bromoindirubin 3′ oxime (BIO, a Wnt activator), bone morphogenetic protein 4 (BMP4), and LIF (BBFL) from day 9 to day 15; BIO, BMP4, BMP10, and LIF (BBBFL) from day 9 to day 15; and additional supplementation with the ROCK inhibitor Y27632 (Y) (Fig. 1B). Because BMP10 is expressed in E9.5 and E11.5 mouse embryonic hearts despite loss of expression in previous mouse heart organoids, BMP10 is believed to be a critical factor for adequate heart development. The hEBs cultured under all conditions showed growth and morphological changes. Statistically, the size (Area) between hEBs and human heart organoids at day 15 under the conditions FGF4, BBFL, and BBBFL was significantly different (Fig. 1B, top), although the size of human heart organoids at day 15 among all conditions was not significantly different (Fig. 1B, bottom left). The success rates of chamber formation under the conditions of FGF4, BBFL, and BBBFL were over 75% (Fig. 1B, bottom right). In particular, the morphology of heart organoids at day 11 of culture from hEBs matched that of the mouse embryonic heart composed of atria and ventricles, or the human fetal heart at 29 days of gestation. These conditions appear to be tailored to promote self-organization during the culture of heart organoids. In addition, the cultured heart organoids in the presence of BBFL and BBBFL showed vigorous proliferation compared to those with FGF4 only, suggesting that BIO, BMP4, BMP10, and LIF promote growth in FGF4-mediated heart development in this culture system. The comparison of the morphology of heart organoids between the FGF4 and FGF4-BBBFL conditions showed that BBBFL is efficient for the generation of more mature and differentiated heart organoids (Fig. 1C). Regarding the acquisition of cardiac functionality, these hESC-derived heart organoids by FGF4 and FGF4-BBBFL conditions showed beating movements (movies S1 and S2). Evaluation of hESC-derived heart organoids by histological analysis and ultrastructural analysis Next, we performed histological evaluation of the generated hESC-derived heart organoids. Major sarcomere proteins in CMs, atrial myosin light chain 2 (MLC2A) and ventricular myosin light chain 2 (MLC2V) [17], were expressed in 15-day cultured hESC-derived heart organoids (Fig. 2A), although the signal of MLC2V was weak in these heart organoids, similar to that in E10.5 mouse hearts approximately corresponding to 6-7W fetal human heart [11], much earlier than 19W fetal human heart [18]. In addition to CMs, the heart consists of other cell types, smooth muscle cells (SMCs) and endothelial cells (ECs), which are important for angiogenesis (blood vessel formation) during cardiogenesis. Therefore, we investigated the expression of a CM marker, cardiac troponin T (cTnT) [19]; an EC marker, platelet/endothelial adhesion molecule 1 (PECAM; CD31) [20]; and an SMC marker, ɑ-smooth muscle actin (αSMA), to determine whether the organization of cardiac structural cells occurs properly in hESC-derived heart organoids. As a result, whole-mount immunostaining using an optimized clear, unobstructed brain/body imaging cocktail and computational analysis (CUBIC) method [21] showed the embryonic heart-like organization of cardiac structural cells in the inner structure of the hESC-derived heart organoids (Supplemental Fig. 2), suggesting that this culture system enables the in vitro generation of heart organoids from hESCs in vitro.Fig. 2Evaluation of hESC-derived HOs. (A) Immunofluorescence staining for MLC2V (red) and MLC2A (green) in the hESC-derived HOs. Top: a representative hESC-HO (n = 6). Scale bar: 25 μm. Bottom: mouse embryonic heart from E10.5. Scale bar: 50 μm. (B) Immunostaining to detect Cx43 (red)/Cx40 (green), (left panels), and Cx43 (red)/Cx45 (green) (right panels). Top: hESC-HOs (n = 6 for Cx43/Cx40, n = 4 for Cx43/Cx45). Bottom: mouse embryonic hearts from E10.5. Scale bar: 50 μm. (C) Immunostaining to detect one of potassium channel proteins, Kir2.1 in a representative HO (top, n = 10), E10.5 mouse embryonic heart (bottom left) and adult mouse heart (right, bottom). Scale bar: 150 μm. (D) hERG immunostaining in a representative HO (top and right, bottom, n = 10), E10.5 mouse embryonic heart (left, bottom). Scale bar: 75 μm. (E) Gene expression levels of MLC2V, MLC2A, Cx40, Cx43, Cx45, Kir2.1 and hERG in human fetal hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). The scRNAseq dataset was obtained from NCBI GEO (Accession: GSE106118) [11]. The log-transformed normalized values of the target genes were presented as a dot plot (mean ± SEM). log1p represents log (x + 1), and CP10K represents counts per 10,000. Evaluation of hESC-derived HOs. (A) Immunofluorescence staining for MLC2V (red) and MLC2A (green) in the hESC-derived HOs. Top: a representative hESC-HO (n = 6). Scale bar: 25 μm. Bottom: mouse embryonic heart from E10.5. Scale bar: 50 μm. (B) Immunostaining to detect Cx43 (red)/Cx40 (green), (left panels), and Cx43 (red)/Cx45 (green) (right panels). Top: hESC-HOs (n = 6 for Cx43/Cx40, n = 4 for Cx43/Cx45). Bottom: mouse embryonic hearts from E10.5. Scale bar: 50 μm. (C) Immunostaining to detect one of potassium channel proteins, Kir2.1 in a representative HO (top, n = 10), E10.5 mouse embryonic heart (bottom left) and adult mouse heart (right, bottom). Scale bar: 150 μm. (D) hERG immunostaining in a representative HO (top and right, bottom, n = 10), E10.5 mouse embryonic heart (left, bottom). Scale bar: 75 μm. (E) Gene expression levels of MLC2V, MLC2A, Cx40, Cx43, Cx45, Kir2.1 and hERG in human fetal hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). The scRNAseq dataset was obtained from NCBI GEO (Accession: GSE106118) [11]. The log-transformed normalized values of the target genes were presented as a dot plot (mean ± SEM). log1p represents log (x + 1), and CP10K represents counts per 10,000. To identify the ultrastructural features of the hESC-derived heart organoids, we performed transmission electron microscopy (TEM) analysis. The intercalated disk (ID) is a cardiac muscle-specific ultrastructure required for coordinated muscle contraction and conduction of electrical excitability in the function of hearts [8]. Additionally, the ID is composed of gap junctions, desmosomes, and facia adherens (intermediate junction) and arranged by a simple network or complex network [22]. The hESC-derived heart organoids showed cardiac ultrastructural properties, including large mitochondria and abundant desmosomes around cardiac muscle filaments, glycogens, and secretory vesicles in atrial CMs. Notably, transverse sections of the hESC-derived heart organoids exhibited a complicated arrangement of IDs, suggesting that these human heart organoids may possess properties in terms of cardiac muscle contraction and conduction (Supplemental Fig. 3A). To quantify cardiac contraction in hESC-derived heart organoids, we used MUSCLEMOTION software [23] with the images obtained by high-speed camera. The regular contractility was observed in hESC-derived heart organoids (Supplemental Fig. 3B), although the frequency of contraction was lower (6.8 times/20 s, SEM = 0.77, n = 5) than that in mES-derived heart organoids (22.6 times/20 s, SEM = 1.1, n = 4), suggesting the physiological discrepancy between human and mouse hearts. For proper conduction, gap junction proteins such as Connexin 43 (Cx43, GJA1), Cx40 (GJA5) and Cx45 (GJC1) play essential roles in impulse propagation in the heart [24]. In the human heart, Cx43 is expressed in atrial and ventricular working myocardial cells, Cx40 is mainly expressed in the atria, and Cx45 exists in His-bundle and bundle branches. To confirm the expression of gap junction proteins at the histological level, we performed immunofluorescence staining with the generated hESC-derived heart organoids. As expected, differential localization of Cx43, Cx40 and Cx45 was detected in the hESC-derived heart organoids, and the expression pattern of these gap junction proteins was similar to that of the mouse embryonic heart (E10.5) (Fig. 2B), suggesting adequate recapitulation of cardiac gap junctions in the hESC-derived heart organoids. Because four types of ion channels containing Kir2.1, hERG, Nav1.5 and Cav1.2 are especially important for the myocardial function as well as the development and the maturation of myocardial tissue in the heart, it is notable to determine whether these major ion channels are expressed in hESC-derived heart organoids. Inwardly rectifying potassium (K) channel Kir2.1, is predominantly expressed in the heart, and its expression during heart development gradually increases in the late embryonic and postnatal hearts. Importantly, Kir2.1 confers membrane stability and action potential repolarization in the cardiac cells, and produces the inward rectifier potassium current, IK1 current. Also, mutations of the gene encoding human Kir2.1 (KCNJ2) are responsible for cardiac arrhythmia in Andersen's syndrome [25,26]. Therefore, we examined the expression of Kir2.1 in the hESC-derived heart organoids to determine whether these heart organoids indicate appropriate electrophysiological features like developing hearts in vivo. Immunofluorescence staining showed expression of Kir2.1 in hESC-derived heart organoids (Fig. 2C), suggesting that the resting membrane potential and cardiac excitability of individual cardiomyocytes of these heart organoids might be controlled by IK1. Additionally, mutations of the human ether-a-go-go-related gene (HERG), which encodes the major subunit for the hERG channel, also cause cardiac arrhythmia as well as mutations of KCNJ2 [27]. Given that hERG as one of voltage gated K channels produces IKr which is the rapidly activating component of IK, the major outward current involved in the repolarization of action potential in human heart muscle cells [28], we investigated the expression of hERG channel in the hESC-derived heart organoids. Immunofluorescence staining exhibited clear expression of hERG channels in hESC-derived heart organoids (Fig. 2D), suggesting that the hESC-derived heart organoids generated in this study might produce IKr via hERG channel. To determine the developmental stage of human heart organoids, we compared cardiac marker expression levels between immunostaining results and the scRNA-sequencing dataset of human developing hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). Data was obtained from NCBI GEO (Accession: GSE106118) [11]. Based on the gene expression levels of the examined genes, the immunostaining results for MLC2V and MLC2A in heart organoids corresponded to 6W human fetal hearts, Cx40 to LA and RA of 9W human fetal hearts, Cx43 to LA of 9W or 13W and to RV of 13W human fetal hearts, and Cx45 to LA of 9W or 13W and to RV of 6-7W human fetal hearts. Although the expression of Kir2.1 and hERG in human fetal hearts was low until 13 weeks of gestation, and the expression of Kir2.1 and hERG in heart organoids was detectable by immunostaining, it is difficult to specify the developmental stage of heart organoids, due to the modality gap between immunofluorescence and scRNA-seq analyses and conflicting findings, such as weak MLC2V expression and undetectable K+ currents. (Fig. 2E). Regarding the development of new therapeutic drugs, the investigation of hERG channels’ block which causes a harmful side effect in human hearts is important for the preclinical test of drugs for their safety. For example, pentamidine, an antiprotozoal compound treated in Pneumocystis pneumonia causes acquired long QT syndrome by the direct block of IKr current. This block of IKr current occurred by the inhibition of hERG channels export from endoplasmic reticulum (ER) by pentamidine [29]. Therefore, we examined whether hESC-derived heart organoids exhibit the inhibition of hERG channel trafficking in cell surface by the treatment of pentamidine. Thirty mM pentamidine were added in 17 days cultured hESC-derived heart organoids for over 16 h (overnight). To detect the localization of hERG in the cells of heart organoids, we immunostained hESC-derived heart organoids with a hERG antibody and an ER marker, KDEL antibody. Expectedly, the export of immature hERG from ER was inhibited in hESC-derived heart organoids incubated with pentamidine, while the strong hERG signals were detected in the outside of ER in hESC-derived heart organoids without pentamidine (Supplemental Fig. 4), suggesting the reflection of drug response on hERG channel trafficking block in hESC-derived heart organoids. In normal cardiac functions, not only the outward currents but also the inward currents like the Na current and the Ca current are important. Among them, the inward Na current, which initiates action potentials in the heart, plays a crucial role in fast impulse propagation in cardiac tissue. The main cardiac sodium channel is Nav1.5, which is encoded by SCN5A [30]. Mutations in SCN5A are associated with multiple heart diseases, such as Brugada syndrome, long QT syndrome, conduction disease and cardiomyopathy. Additionally, genetic variation in SCN5A is associated with differences in cardiac conduction and the risk of arrhythmia. To determine the expression of Nav1.5 in the hESC-derived heart organoids, we performed immunofluorescence staining with an anti-SCN5A antibody. We detected clear expression of SCN5A in the hESC-derived heart organoids, suggesting the fast depolarization phase of cardiac action potential (Fig. 3A).Fig. 3Expression of sodium and calcium channel proteins (Nav1.5 and Cav1.2) in hESC-derived HOs. (A) Nav1.5 immunostaining to detect sodium channel in hESC-derived HOs (left and center, n = 9), E10.5 mouse embryonic heart (top right), and mouse adult heart (bottom right). Scale bar: 90 μm. (B) Cav1.2 immunostaining to detect calcium channel in hESC-derived HOs (left and center, n = 11), E12.5, E13.5, E15.5, and E18.5 mouse embryonic heart (right). Scale bar: 75 μm for left, 50 μm for center and 200 μm for right panels. (C) Gene expression levels of Nav1.5 and Cav1.2 in human fetal hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). The scRNAseq dataset was obtained from NCBI GEO (Accession: GSE106118) [11]. The log-transformed normalized values of the target genes were presented as a dot plot (mean ± SEM). log1p represents log (x + 1), and CP10K represents counts per 10,000. Expression of sodium and calcium channel proteins (Nav1.5 and Cav1.2) in hESC-derived HOs. (A) Nav1.5 immunostaining to detect sodium channel in hESC-derived HOs (left and center, n = 9), E10.5 mouse embryonic heart (top right), and mouse adult heart (bottom right). Scale bar: 90 μm. (B) Cav1.2 immunostaining to detect calcium channel in hESC-derived HOs (left and center, n = 11), E12.5, E13.5, E15.5, and E18.5 mouse embryonic heart (right). Scale bar: 75 μm for left, 50 μm for center and 200 μm for right panels. (C) Gene expression levels of Nav1.5 and Cav1.2 in human fetal hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). The scRNAseq dataset was obtained from NCBI GEO (Accession: GSE106118) [11]. The log-transformed normalized values of the target genes were presented as a dot plot (mean ± SEM). log1p represents log (x + 1), and CP10K represents counts per 10,000. Another major inward current is Ca current. In cardiovascular physiology, voltage gated calcium channels including Cav1.2 and Cav 1.3 (L-type Cav channel) play important roles for calcium influx by Cav1.2 in adult cardiomyocytes and contribute to inward current during cardiac action potential and excitation-contraction coupling [31]. Cav1.2 which is encoded by CACNA1C expresses in ventricular and atrial cardiomyocytes. Mutations of CACNA1C are associated with long-QT syndrome (gain of function) and Brugada syndrome (loss of function). Given the significant contribution of Ca channel to normal cardiac function, we next examined the expression of Cav1.2 in hESC-derived heart organoids. Immunofluorescence staining with an anti-CACNA1C antibody showed strong expression of CACNA1C in hESC-derived heart organoids (Fig. 3B), suggesting the cardiac contractility triggered by Ca influx via Ca channels. On Nav1.5 and Cav1.2, the developmental stage of human heart organoids was determined by comparing these gene expression levels between immunostaining results and the scRNA-sequencing dataset from NCBI GEO: GSE106118 [11] of human developing hearts (HE5W, HE6W, HE7W, HE9W, and HE13W for LA, LV, RA, and RV). As a result, although Nav1.5 and Cav1.2 in heart organoids were highly expressed by immunostaining, in contrast to the low expression in 13W human heart in scRNA-seq data, it is difficult to determine whether the developmental stage of heart organoids is later than 13W in human hearts, for the same reasons as the results of Kir2.1 and hERG mentioned above, such as cross-modal constraints and limited contractibility. (Fig. 3C). Physiological ion currents in the hESC-derived heart organoids Recent studies showed the importance of normal Nav1.5 channel activity for not only cardiac excitability but also histological development [32], therefore, the identification of functional Nav1.5 ion current is thought to be essential to ensure the normal cardiac development of hESC-derived heart organoids. To characterize the electrophysiological properties of ion current, patch clamp methods are golden standard and most reliable, especially, for the isolation of specific ion current from multiple ion currents in stem cell derived cardiomyoctes. Therefore, to analyze the electrophysiological features of the CMs prepared from the hESC-derived heart organoids cultured 19 to 36 days after EB formation, we enzymatically isolated the CMs and performed whole-cell voltage-clamp experiments with the isolated CMs. In many cells from heart organoids, typical inward currents, which had the maximal peak current at −20 mV (n = 4), were induced in response to depolarizing step pulses (<100 pA: n = 6; >100 pA, <500 pA: n = 13; >500 pA: n = 8; n = 26 in total), suggesting that the induced inward currents were mainly mediated by voltage-dependent Na channels (Fig. 4A). These results suggested proper recapitulation of impulse propagation in the heart organoids by action potential initiation via Na currents. The inward currents were substantially reduced by changing to a Na-free extracellular solution (Fig. 4B top) or by inclusion of tetrodotoxin (TTX), a specific inhibitor of Na channels (Fig. 4B bottom). Ca currents were barely but clearly observed in a few cells from the hESC-derived heart organoids, each of which had an amplitude of approximately 40 pA following the blockade of Na currents (Fig. 4B top and bottom). The remaining inward currents showed even slower activation kinetics, which are characteristic of voltage-dependent Ca channels. Further reduction in the remaining inward currents was achieved by cobalt, an inorganic Ca channel blocker. Thus, the isolated cells from the hESC-derived heart organoids express ion channels that play integral roles in shaping cardiac action potentials; that is, voltage-dependent Na and Ca channels contributing to the initial rapid depolarization (phase 0) and plateau phase (phase 2) in cardiac action potentials, respectively.Fig. 4Electrophysiological analyses of CMs isolated from hESC-derived HOs in the whole-cell patch-clamp. (A, left) Representative Na current traces were elicited by the depolarizing voltage protocol, as shown in the inset. (A, right) Current-voltage relationships of peak INa with maximal inward currents at −20 mV (n = 4). Normalized peak amplitudes of INa to the maximal value were plotted as a function of the membrane potential. (B) Inward currents (blue) were substantially reduced by replacing Na+ ions with NMDG (top, red) or inclusion of TTX at 20 μM (bottom, red), and the addition of CoCl at 4 mM abolished the residual currents (top and bottom, black). Electrophysiological analyses of CMs isolated from hESC-derived HOs in the whole-cell patch-clamp. (A, left) Representative Na current traces were elicited by the depolarizing voltage protocol, as shown in the inset. (A, right) Current-voltage relationships of peak INa with maximal inward currents at −20 mV (n = 4). Normalized peak amplitudes of INa to the maximal value were plotted as a function of the membrane potential. (B) Inward currents (blue) were substantially reduced by replacing Na+ ions with NMDG (top, red) or inclusion of TTX at 20 μM (bottom, red), and the addition of CoCl at 4 mM abolished the residual currents (top and bottom, black). Discussion Heart development starts at a very early stage of embryogenesis (cardiac crescent at the E7.5 heart in mice corresponding to 20 days of gestation in humans) with a limited cell number and simple extracellular environment and proceeds to distinctive morphogenesis with functionality. Thus, elucidation of the spatiotemporal regulation of cardiogenesis is important for understanding not only heart development but also the structure-related specific functions of each part of the developing heart as well as the complete regulatory system of organogenesis in the body. In our previous study, we successfully generated mouse ESC-derived 3D heart organoids comprised of CMs, ECs and SMCs by inducing self-organization of cells in the presence of exogenous growth signals and extracellular matrix [8]. The mouse heart organoids exhibited morphological changes, such as chamber formation similar to the E10.5 heart in mice, involving multiple structural heart cell types in a developmental stage-specific manner, and showed cardiac ultrastructural properties and autonomous beating with myocardial contraction, suggesting functional acquisition. As biomimetic models, organized heart organoids can be promising tools for understanding cardiogenesis across many species. The demand for human heart organoids in drug discovery, as well as cardiac etiology and pathology, has increased. Therefore, we applied the self-organized heart organoid-generating method to human ESCs with slight modifications, obtaining hESC-derived heart organoids with multiple chambers, expression of heart-specific proteins, including gap junction proteins, and electrophysiological properties. Our results indicated that FGF4 and LN/ET extracellular matrix trigger early cardiac development in humans and mice, suggesting that these two factors play conserved roles in mammalian cardiac morphogenesis. Importantly, Na+ and Ca2+ ion currents were detected in CMs of hESC-derived heart organoids using patch-clamp analysis, suggesting in vitro reconstitution of the human heart, along with action potential propagation and cardiac contraction. In particular, the Na+ ion channel protein, Nav1.5, was detected in hESC-derived heart organoids due to the normal Na+ ion current. These findings might be useful for developing a system to block Na+ ion current using hESC-derived heart organoids for drug toxicity screening; however, for reliable pharmacological screening, hESC-derived heart organoid cells must be assessed for functional maturity with K+ currents, including differentiation into cardiac cell types, developmental stages, and excitation-contraction (EC) coupling. Although MUSCLEMOTION analysis demonstrated that they exhibit regular, repeated contractile and relaxant responses, the heartbeat of the hESC-derived heart organoids was weaker than that of the mouse heart organoids, suggesting that further maturation will be necessary for application in an in vitro drug-safety testing system. The hESC-derived heart organoids exhibit cardiac morphology similar to atrial- and ventricular-like structures reflecting adequate function, suggesting dramatic morphogenic changes in cardiogenesis by beating CMs or contraction that induce morphogenic shifts by mechanical force stress load. This issue might be solved by a mechanical stress culturing system that accelerates (stimulates) mechanical force and elongation stretch. Lipid metabolism is also important for heart function and CM maturation. During mouse embryonic development, glycogen is abundant in embryonic CMs and serves as an energy source. However, cAMPs are generated by lipid metabolism in mature CMs, showing normal cardiac performance during the postnatal period. Additionally, palmitates were shown to be involved in the maturation of CMs induced from human pluripotent stem cells [33]. Thus, further studies will be necessary to improve the maturity of CMs in hESC-derived heart organoids by switching the energy source from glucose to lipids, depending on the developmental stage, and applying mechanical loads in the culture system. The heart is an essential organ for life maintenance, together with the blood and circulatory system, to carry nutrients and waste around the body to maintain its cells. Many patients suffer from heart diseases and defects, and the adult human heart cannot completely recover from damage. From another medical point of view, cardiovascular toxicity is one of the major barriers to drug development: unanticipated heart toxicity has been a major reason for the withdrawal of many anticancer drugs and anti-inflammatory drugs from the market. Therefore, a new, safe medical treatment for many diseases and an effective nonclinical heart-toxicity evaluation system closer to the in vivo heart are eagerly anticipated. Although pluripotent stem cell-derived heart organoids are feasible for nonclinical heart toxicity evaluation systems, including both patch clamp analysis of isolated cells and drug responses of the heart organoids as a whole, there remain limitations, such as the immature characteristics of CMs in heart organoids. Finally, understanding the heart regulatory system using hESC-derived heart organoids may provide important insights into not only the etiology of cardiovascular diseases, thereby enabling the development of a new therapeutic method, but also the developmental relationships among organs in our body. Funding This work was supported by the Grant-in-Aid for Scientific Research (C) (20K06652) to J. Lee. Declaration of competing interest Authors J. Lee and F. Ishino are co-inventors and co-owners of three granted patents (Patent Nos. JP7315923, US 12, 163, 158 B2 and EP3699275) that are related to the technology described in this manuscript. These patents may be subject to commercial licensing. All other authors declare no competing interests.