An in vitro model for cardiac organoid production: The combined role of geometrical confinement and substrate stiffness

Figures

• 

  Fig. 1. Platform and (a) Organoid culture platform is composed by a glass coverslip (G) covered by a silicone stencil, with through holes of 500 μm diameter. To modulate substrate stiffness, an acrylamide gel (PAAg, -H or -L) of known stiffness (H: E = 30 kPa; L: E = 4 kPa) is firmly attached on the glass coverslip. iPSC seeding and culture into this platform leads, following the summarized protocol (b), to the generation of cardiac organoids. (c) Constructs cultured under geometrical confinement are spheroids characterized by reproducible dimension (area) and shape (circularity factor), independently from substrate stiffness, and were compared to spheroids cultured in absence of geometrical confinement (EB). Differences in both area and circularity factor between ORG-G (n = 17), ORG-H (n = 46), ORG-L (n = 32) and EB (n = 42) were assessed by Kruskal-Wallis test followed by Dunn's multiple comparison test. The stars indicate the significance that withstood adjustment for the comparison of each ORG vs. EB. ∗∗P < 0.01; ∗∗∗∗P < 0.0001.

• 

  Fig. 2. Bulk RNA-seq analysis. (a) Scatterplot of the first two principal components (PC1 and PC2) obtained from the principal component analysis (PCA) performed on all expressed genes in the bulk RNA-seq dataset. PC1 and PC2 together explained 38 % of the variance and allowed the four spheroid types to be clearly distinguished. (b–d) Cultures grown under geometric confinement show a large number of significantly differentially expressed (DE) genes compared to samples grown under non-confined conditions. The scatterplots show the log2 fold-change (FC) and the significance (x and y axis, respectively) for the confined (ORG-G, -H and -L, respectively) vs. non-confined (EB) condition comparisons. Red dots represent significant DE genes at adjusted P-Value <0.05 and |log2FC|>1; blue dots = adjusted P-Value <0.05; green dots = |log2FC|>1; grey = Not significant. The first seven over- and under-expressed genes ranked by a combined score, calculated as the product of the log2-FC x -log10(P-value) obtained by the DEA, were highlighted in boxed labels. (e) Functional enrichment analysis shows a positive association of confinement conditions with the regulation of several functions related to…

• 

  Fig. 3. Box plots of selected cardiac marker gene expression. Box plots display the normalized expression levels (y-axis) of selected cardiac marker genes across three different ORGs and EBs. These plots highlight significant expression differences, particularly the elevated levels in ORG-L compared to EBs. The markers include atrial cardiomyocyte genes (MYL7, MYL4), ventricular cardiomyocyte genes (MYH7, MYL3), sarcomeric genes (TNNT2, ACTN2), cardiomyocyte-enriched mitochondrial genes (MT-ND3, MT-CO3), fibroblast markers (FN1, COL4A1), epicardial cell markers (TBX18, WT1), a pacemaker cell marker (HCN4), and a long non-coding RNA associated with the cardiac apex (AC020909.2). Statistical analyses were performed using the negative binomial generalized linear model (GLM), as detailed in the Materials and Methods section, with overall results provided in Supplementary Data 1a. Significant differences for pairwise comparisons between ORGs and EBs are denoted as follows: ∗ nominal P-value <0.05; ∗# adjusted P-value <0.05.

• 

  Fig. 4. scRNA-seq analysis. (a) UMAP plot does not suggest the presence of major differences between the analyzed sample batches. (b) Data integration of ORG-H and ORG-L conditions led to the identification of 9 clusters. (c) Cell type annotation of scRNA-seq clusters. d) Functional annotation by cluster top markers, based on Gene Ontology (GO) for biological processes confirms the presence of clusters with a transcriptome signature linked either to contractile or other functions. e) The heatmap illustrates the average relative expression of the top 10 non-redundant marker genes for each cluster identified in the integrated scRNA-seq datasets of ORG-L and ORG-H, ranked by summary log fold change (logFC) (see Supplementary Data 3). Both clusters (columns) and genes (rows) are hierarchically grouped using Euclidean distance and the complete-linkage clustering method. The color gradient indicates relative gene expression levels, ranging from higher (dark red) to lower (dark blue). Dendrograms were cut to define three primary subgroups for both clusters and genes, demarcated by thicker white lines on the heatmap. (For interpretation of the references to color in this figure legend, the…

• 

  Fig. 5. Violin plots of scRNA-seq gene expression. Violin plots depict the relative expression levels (y-axis) of a subset of cardiac and non-cardiac marker genes across all nine cell clusters identified in the integrated scRNA-seq datasets of ORG-L and ORG-H. Distinct expression patterns are observed for genes that are highly enriched in specific clusters, such as epicardial cell markers, which are predominantly expressed in cluster 7 while absent in others. The black line in each violin plot marks the median of the expression value of each gene calculated on all cells in the cluster.

• 

  Fig. 6. Enrichment networks of bulk RNA-seq data ORG-L vs. ORG-H. For the comparison of ORG-H vs. ORG-L samples, enrichment networks were drawn to show the Gene Ontology (GO) biological processes (BP) gene sets (nodes) that are significantly associated (false discovery rate <0.01) with ORG-L and ORG-H conditions. Node color refers to the association with the phenotype (ORG-L, red; ORG-H, green); node gradient color is proportional to the gene-set normalized enrichment score (NES), from lower (light) to higher (dark); node size is proportional to the gene-set size. Edges connect related GO-BPs. Edge thickness is proportional to the similarity between two GO-BPs, for a cutoff = 0.25 of the combined Jaccard plus Overlap coefficient. ORG-H vs. ORG-L comparison suggests that processes related to cardiac maturity and cardiomyocyte ultrastructure are significantly regulated by fine tuning of substrate stiffness. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

• 

  Fig. 7. Enrichment networks of bulk RNA-seq data ORG-L vs. EB. For the comparison of ORG-L vs. EB samples, enrichment networks were drawn to show the Gene Ontology (GO) biological processes (BP) gene sets (nodes) that are significantly associated (false discovery rate <0.01) with ORG-L and EB conditions. Node color refers to the association with the phenotype (ORG-L, red and EB, blue); node gradient color is proportional to the gene-set normalized enrichment score (NES), from lower (light) to higher (dark); node size is proportional to the gene-set size. Edges connect related GO-BPs. Edge thickness is proportional to the similarity between two GO-BPs, for a cutoff = 0.25 of the combined Jaccard plus Overlap coefficient. Culture under confinement on soft gels as per ORG-L vs. EB comparison positively regulates processes related to cardiac maturity, function and 3D organization (e.g. chamber development). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

• 

  Fig. 8. Characterization of ORG-L samples. Immunofluorescence staining supports results of transcriptome. Optical sections, of both top and equator level, of organoid stained for cardiac marker (NKX2.5) show the presence of internal cavities, non located in the spheroid center and divided by a septum; ORG-L are characterized by a single cell layer of WT1 (green) and TBX18 (red) positive cells; staining for cell-cell junctions indicates presence of ZO1 (green) and Cx45 (green) (present in cardiomyocytes since the initial stages of differentiation) in both EB and ORG-L, while Cx43 (red) (specific of cardiomyocytes undergoing ventricular specification) is expressed only in ORG-L samples; staining for cTnT (red) confirms the presence of CMs in both EB and ORG-L, although staining for myosin light chain isoforms (MYL) suggests enhanced commitment in direction of ventricular CMs in cell subsets populating ORG-L. Upregulation in ORG-L of functions related to matrix remodeling and attributable to the functionality of non-contractile cardiac cell subsets is shown by staining for collagen (Coll1A1) and fibronectin (FN1). (For interpretation of the references to color in this figure legend, the…

• 

  Fig. 9. Electrophysiology study. a) Hierarchical clustering and heatmap based on a selected GOBP gene set, indicative of cardiac muscle cell potential suggesting differential regulation in ORG-L vs. EB samples. b) Representative voltage traces of spontaneous action potential recorded in single cardiomyocytes at 21 day of differentiation 24 h post dissociation from Organoids (ORG-L red trace) and Embryoids (EB, blue trace). c) Scatter plot of the firing rate (Hz) in ORG-L (red dots; N = 8)) and EB-derived cardiomyocytes (blue dots; N = 13). d) Scatter plots of maximum diastolic potential (MDP), action potential amplitude (APA) color as in C. e) scatter plots of action potential duration at 30 %, 50 % and 90 % of repolarization (APD30, APD50 and APD90, respectively) in ORG-L and EBs (color as in c) Mean ± SEM values are reported as black squares and whiskers. f) Representative traces of TTX-sensitive currents elicited by depolarizing voltages from ORG-L (red traces, top panel) and EB-derived cardiomyocytes (blue traces, bottom panel). g) Mean sodium current density-voltage relations in ORG-L (red dots; N = 11) vs. EB-derived cardiomyocytes (blue dots; N = 10). h) Mean activation and…

• 

  Fig. 10. Dose response to drug administration. Organoids cultured onto low substrates show functional response to cardioactive drugs in a dose dependent manner. (a) Representative contraction amplitude plots and (b) average beating rates of untreated ORG-L (n = 47) versus drug-treated ORG-L with increasing doses of Isoproterenol, i.e. 0.001 (n = 5), 0.01 (n = 9), 0.1 (n = 10) and 1 μM (n = 21); and Verapamil, i.e. 0.001 (n = 5), 0.01 (n = 5), 0.1 (n = 5) and 1 μM (n = 5). Kruskal-Wallis test (P < 0.0001) followed by Dunn's multiple comparison test was used to assess differences between non-treated vs. treated ORGs. The stars indicate the significance that withstood adjustment for each condition vs. non-treated ORG. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001.

• 

  figs1.

• 

  figs2.