A patterned human primitive heart organoid model generated by pluripotent stem cell self-organization

Abstract

Pluripotent stem cell-derived organoids can recapitulate significant features of organ development in vitro. We hypothesized that creating human heart organoids by mimicking aspects of in utero gestation (e.g., addition of metabolic and hormonal factors) would lead to higher physiological and anatomical relevance. We find that heart organoids produced using this self-organization-driven developmental induction strategy are remarkably similar transcriptionally and morphologically to age-matched human embryonic hearts. We also show that they recapitulate several aspects of cardiac development, including large atrial and ventricular chambers, proepicardial organ formation, and retinoic acid-mediated anterior-posterior patterning, mimicking the developmental processes found in the post-heart tube stage primitive heart. Moreover, we provide proof-of-concept demonstration of the value of this system for disease modeling by exploring the effects of ondansetron, a drug administered to pregnant women and associated with congenital heart defects. These findings constitute a significant technical advance for synthetic heart development and provide a powerful tool for cardiac disease modeling.

Fig. 1. Developmental induction methods for improving human heart organoid developmental modeling.

a A schematic diagram depicting the differentiation protocol for creating human heart organoids and the media conditions for the four maturation strategies (control, MM, EMM1, and EMM2/1). Created with BioRender.com. b Brightfield images of organoids throughout the 30-day culture period. Two representative organoids are shown for each condition (data representative of 23–24 independent organoids per condition across five independent experiments). Scale bar = 400 μm. c Quantification of organoid long diameter and short diameter at day 30 of culture for each maturation strategy (n = 7–9 independent organoids per condition). d Quantification of organoid area at day 30 of organoid culture for each maturation strategy (n = 17–20 independent organoids per condition across two independent experiments). Data presented as a violin plot with all points, one-way ANOVA with Dunnett’s multiple comparisons test. e Quantification of percentage of organoids visibly beating under brightfield microscopy in each condition from 5 different organoid batches (n = 22–24 independent organoids in each condition across five independent experiments). f TEM images displaying sarcomeres, myofibrils (M) and I-bands (arrows) in day 15 organoids and in organoids from each maturation condition at day 30 (n = 4 independent organoids per condition). Scale bars = 1 μm. g Quantification of sarcomere length within TEM images (n = 4 independent organoids per condition, n = 18, 15, 69, 58, and 35 measurements for D15, control, MM, EMM1, and EMM2/1, respectively). Data presented as mean ±  s.e.m, one-way ANOVA with Dunnett’s multiple comparisons tests. h mRNA expression of key sarcomeric genes involved in cardiomyocyte maturation between days 20 and 30 of culture for each condition (n = 14 independent organoids per day per condition per gene across three independent experiments). Data presented as log₂ fold change normalized to Day 20. Values = mean ±  s.e.m. Source data are provided as a Source Data file.

Fig. 2. Single-cell RNA sequencing of human heart organoids reveals distinct cardiac cell populations.

a UMAP dimensional reduction plots of integrated single-cell RNA sequencing data for each condition in day 34 organoids. Cluster identities are in the legend below. b Quantification of total cell count percentages per cluster. Colors of regions correspond to those found in the legend in (a). c Differential expression heatmap displaying the top 10 differentially expressed genes for all clusters. d Feature plots displaying key marker genes for each cluster. Color intensity represents the relative value of gene expression per gene.

Fig. 3. Cluster identity and cell–cell communication networks highlight the importance of self-organization in heart organoid development.

a Dot plots of differentially expressed marker genes in each cluster for each condition. Color is indicative of the average expression level across all cells, and the size of the circle is indicative of the percentage of cells within a particular cluster that express the respective gene. b Visualization of cell-cell ligand-receptor communication networks for each condition. Colors of clusters (exterior) match that of UMAP projections. Ligands are indicated as blue bands and receptors are indicated by red bands. Arrows within depict pairing from ligands to receptors.

Fig. 4. Human heart organoids develop increasingly mature metabolic profiles following developmental induction conditions.

a Mitochondrial labeling within day 30 human heart organoids in each condition (n = 6 independent organoids per condition). White = Mitotracker, blue = NucBlue. Scale bars = 10 μm. Detailed images of mitochondria are shown below each main image. b Quantification of mitochondrial area surrounding each individual nucleus (n = 6 independent organoids per condition, n = 50, 59, 74, and 70 measurements for control, MM, EMM1, and EMM2/1, respectively). Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons tests. c TEM images displaying mitochondria in day 15 organoids and in organoids from each maturation condition at day 30 (n = 4 independent organoids per condition). Yellow arrows indicate mitochondria, LD = lipid droplets, Gg = glycogen granules. Scale bars = 1 μm. d Quantification of mitochondrial area from TEM images. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons tests (n = 4 independent organoids per condition, n = 40, 91, 154, 117, and 81 mitochondria measured for D15, control, MM, EMM1, and EMM2/1, respectively). e mRNA expression of metabolic genes PPARGC1A and CPT1B between days 20 and 30 of culture for each condition (n = 8 independent organoids per condition across three independent experiments). Data presented as log₂ fold change normalized to Day 20. Values = mean ± s.e.m. f Oxygen consumption rate measurements from Agilent Seahorse XFe96 metabolic stress test assay in all conditions (n = 8 independent organoids per condition across two independent experiments). Values = mean ±  s.e.m. Quantifications from oxygen consumption rate assay (n = 8 independent organoids per condition. Values = mean ±  s.e.m., one-way ANOVA) including g Basal respiration, h Maximal respiration, and i Spare respiratory capacity. j Feature plots displaying key metabolic genes upregulated in the VCM and ACM clusters. k Expression heatmaps of key metabolic genes in the VCM and ACM clusters in each condition. Data displayed as log₂ and is normalized to each column (for each gene and cluster). Source data are provided as a Source Data file.

Fig. 5. Developmental induction conditions promote progressive electrophysiological maturation in human heart organoids.

a Representative calcium transient traces within d30 human heart organoids from each condition (n = 12 independent organoids per condition across three independent experiments). Traces represent data from an individual cardiomyocyte within human heart organoids. b Quantification of peak amplitude of calcium transient traces from each condition (n = 12 across three independent experiments). Values = mean ±  s.e.m. c Quantification of calcium transient peak frequency for each condition (n = 12 independent organoids per condition across three independent experiments). Values = mean ±  s.e.m. d Feature plots displaying key electrophysiological genes differentially expressed in the VCM and ACM clusters in each condition. e mRNA expression of key electrophysiological genes between d20 and d30 for each condition (n = 8 independent organoids per day per condition per gene across three independent experiments). Data presented as log₂ fold change normalized to d20. Values = mean ±  s.e.m. f Representative voltage tracings of organoids in the EMM2/1 and control conditions depicting atrial-, nodal-, and ventricular-like action potentials (n = 8 (EMM2/1, Ventricular), 9 (EMM2/1, Atrial and Nodal; Control, Atrial), and 10 (Control, Nodal) individual cells from three independent organoids across three independent experiments). Representative immunofluorescence images (g) and quantification (h) of caveolin-3 puncta for each condition (n = 15 independent organoids per condition across three independent experiments). Green = caveolin-3, red = TNNT2, blue = DAPI. Scale bar = 20 μm. Data presented as fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons tests. Representative immunofluorescence images (i) and quantification (j) of KCNJ2+ puncta for each condition (n = 14 independent organoids per condition across three independent experiments). KCNJ2 = green, TNNT2 = red, DAPI = blue. Scale bar = 20 μm. Data presented as fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons tests. Source data are provided as a Source Data file.

Fig. 6. Developmental induction promotes the emergence of a proepicardial organ and formation of distinct atrial and ventricular chambers by self-organization.

a Representative surface and interior immunofluorescence images of individual day 30 organoids in all conditions displaying WT1 (green), TNNT2 (red), and DAPI (blue). Three organoids are displayed for each condition (n = 12–15 independent organoids per condition across two independent experiments). Yellow arrows represent WT1+ cells on outer surface. White arrows represent TNNT2+ cells on lower chamber wall. Scale bars = 200 m. Quantification of TNNT2+ chamber area (b) and WT1+ chamber area (c) in each condition from images presented in (a) (n = 12, 13, 13, and 15 (b) and n = 12 (c) independent organoids for control, MM, EMM1, and EMM2/1, respectively, across two independent experiments). Values are presented as fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. d Representative surface and interior immunofluorescence images of individual day 30 organoids in all conditions displaying MYL2 (green), MYL7 (red), and DAPI (blue). Three organoids are displayed for each condition (n = 13 independent organoids per condition across three independent experiments). Scale bars = 200 μm. e Quantification of MYL2+ area in each organoid in each condition from images presented in (d) (n = 13 independent organoids per condition across three independent experiments). Values are presented as fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. f Representative immunofluorescence images of individual day 30 organoids in all conditions displaying atrial marker NR2F2 (green), ventricular marker MYL3 (red), and DAPI (blue). Three organoids are displayed for each condition (n = 12 organoids per condition across two independent experiments). Scale bars = 200 μm. g Quantification of colocalization (Pearson’s coefficient) between NR2F2 (green) and MYL3 (red) from images presented in (f). Values = mean ± s.e.m., unpaired t test. h Feature plot highlighting scRNA-seq VCM and ACM clusters. Feature plots displaying hallmark atrial chamber identity genes (i) and ventricular chamber identity genes (j) that are differentially expressed in the ACM (i) or VCM (j) cluster. Source data are provided as a Source Data file.

Fig. 7. An endogenous retinoic acid gradient is responsible for spontaneous anterior-posterior heart tube patterning.

a Schematic of in vivo heart tube formation, highlighting the localization and intensity of retinoic acid from anterior (arterial pole) to posterior (venous pole) of the primitive heart tube. Created using BioRender.com. b Raman spectroscopy intensity plots for all conditions at d30. Data representative of n = 3 organoids per condition. c mRNA expression of ALDH1A2 in all conditions at d30 (n = 7 independent organoids/condition, two independent experiments). Data presented as log₂ fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. d Feature plot displaying expression of ALDH1A2. e Immunofluorescence images of individual day 30 organoids in the control and EMM2/1 conditions displaying ALDH1A2 (green), TBX18 (red), and DAPI (blue). Three organoids displayed per condition (n = 22-24 independent organoids per condition, three independent experiments). Scale bar = 200 μm. f High magnification images of organoids shown in (c), displaying ALDH1A2 (green), TBX18 (red), and DAPI (blue). Yellow square (Scale bar = 200 μm) represents the area of high magnification. Scale bar = 50 μm. g Quantification of ALDH1A2⁺ TBX18⁺ area within organoids in each condition from (e) and in Supplementary Fig. 30 (n = 22, 22, 22, and 24 independent organoids for control, MM, EMM1, and EMM2/1, respectively, across three independent experiments). Data presented as fold change normalized to control. Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. h Immunofluorescence images of d30 organoids displaying WT1 (green), MYL3 (red), and DAPI (blue). Three organoids displayed per condition (n = 12 organoids/condition, two independent experiments). Scale bars = 200 μm. i Immunofluorescence images of d30 EMM2/1 organoids following exposure to either deoxyaminobenzaldehyde (DEAB), retinoic acid (RA), or Untreated. Staining was performed for ventricular marker MYL3 (pink), atrial marker NR2F2 (green), and DAPI (blue). Two organoids displayed for each condition (n = 9, 10, and 9 independent organoids for Untreated, +DEAB, and +RA, respectively, across two independent experiments). Scale bar = 200 μm Quantification of MYL3+ area (j) and NR2F2+ area (k) from organoids presented in (i). Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. Source data are provided as a Source Data file.

Fig. 8. Primitive heart tube organoids demonstrate a direct link between ondansetron administration and ventricular cardiac defects.

a Representative immunofluorescence images of individual day 30 EMM2/1 organoids following exposure to varying concentrations of ondansetron (1 μM, 10 μM, or 100 μM) or no treatment (untreated) from day 9 to day 30 of culture. Staining was performed for ventricular marker MYL2 (green), atrial marker MYL7 (red), and DAPI (blue). Three organoids are displayed for each condition (n = 12 independent organoids per condition across two independent experiments). Scale bar = 200 μm. Quantification of MYL2+ area (b) and MYL7+ area (c) for each condition (n = 12 independent organoids per condition across two independent experiments). Data presented as fold change normalized to untreated. Values = mean ± s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. d mRNA expression of ventricular marker MYL2 in all conditions at day 30 (n = 6 independent organoids per condition across two independent experiments). Data presented as log₂ fold change normalized to control. Values = mean ± s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. Representative voltage tracings of organoids (e) showing three voltage traces from independent organoids in each condition (f) (representative of n = seven independent organoids per condition across two independent experiments). g–j Quantification of voltage tracings from individual organoids in each condition from traces presented in (e) and (f) (n = 7 independent organoids per condition across two independent experiments), displaying frequency (g), amplitude (h), APD30 (i), and APD90 (j). Values = mean ±  s.e.m., one-way ANOVA with Dunnett’s multiple comparisons test. Source data are provided as a Source Data file.