Molecular and electrophysiological evaluation of human cardiomyocyte subtypes to facilitate generation of composite cardiac models

Abstract

Paucity of physiologically relevant cardiac models has limited the widespread application of human induced pluripotent stem cell (hiPSC)-derived cardiomyocytes in drug development. Here, we performed comprehensive characterization of hiPSC-derived cardiomyocyte subtypes from 2D and 3D cultures and established a novel 3D model to study impulse initiation and propagation. Directed differentiation approaches were used to generate sinoatrial nodal (SANCM), atrial (ACM) and ventricular cardiomyocytes (VCM). Single cell RNA sequencing established that the protocols yield distinct cell populations in line with expected identities, which was also confirmed by electrophysiological characterization. In 3D EHT cultures of all subtypes, we observed prominent expression of stretch-responsive genes such as NPPA. Response to rate modulating drugs noradrenaline, carbachol and ivabradine were comparable in single cells and EHTs. Differences in the speed of impulse propagation between the subtypes were more pronounced in EHTs compared with 2D monolayers owing to a progressive increase in conduction velocities in atrial and ventricular cardiomyocytes, in line with a more mature phenotype. In a novel binary EHT model of pacemaker-atrial interface, the SANCM end of the tissue consistently paced the EHTs under baseline conditions, which was inhibited by ivabradine. Taken together, our data provide comprehensive insights into molecular and electrophysiological properties of hiPSC-derived cardiomyocyte subtypes, facilitating the creation of next generation composite cardiac models for drug discovery, disease modeling and cell-based regenerative therapies.

Keywords: cardiac differentiation; electrophysiology; engineered heart tissues; iPSC; stem cell.

Figure 1.

Differentiation and characterization of cardiomyocyte subtypes. (a) Schematic representation of protocols utilized for the differentiation of cardiomyocyte subtypes from hiPSCs. (b) Representative histograms and summarized data showing % of TNNT2⁺ cells in the three groups at day 20 of differentiation. A corresponding IgG isotype antibody was used as negative control for flow cytometry (gray). n = 4 independent differentiations. Error bars, s.e.m. Mann-Whitney U test. (c–f) qRT-PCR analysis demonstrating expression of cardiomyocyte genes (c), SAN-associated genes (d), atrial-associated genes (e), and ventricular-associated genes (f). N = 5 independent differentiations. Expression corrected to GEOMEAN of reference genes RPLP0 and GUSB. Error bars, s.e.m. Kruskal Wallis Test followed by Mann Whitney U-test for post hoc comparison. *p < 0.05, **p < 0.01. (g) Immunofluorescence stainings demonstrating the expression of SHOX2 and TNNT2, NR2F2 and TNNT2 and, MYL2 and ACTN2, in SANCM, ACM, and VCM. Scale bars, 20 μm.

Figure 2.

Single cell RNA-sequencing of SANCM and ACM. (a) UMAP representation of single cell transcriptomes of SANCM and ACM at day 20 of differentiation. (b) UMAP showing the original identifier of each cardiomyocyte subtype and plate (1–3). (c–e) Violin plots showing cardiac sarcomeric genes used to identify cardiomyocyte clusters (c), genes used to identify the SANCM populations (d) and genes used to identify ACM populations (e).

Figure 3.

Characterization of cardiomyocyte subtypes by single cell patch-clamp. (a) Illustration of analyzed action potential parameters. (b) Representative spontaneous action potential traces of day 21 SANCM, ACM, and VCM (c–g) cycle length (c) membrane diastolic potential, MDP (d), upstroke velocity, dV/dt_Max (e), maximal action potential amplitude, APA_Max (f) and action potential duration at 20%, APD₂₀, 50%, APD₅₀ and 90% repolarization, APD₉₀. N = 9 cells from four independent differentiations. Error bars, s.e.m. Kruskal Wallis Test followed by Mann Whitney U-test for post hoc comparison. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4.

Single cell response to Noradrenaline (NA), Carbachol (CCh) and Ivabradine (IVA). (a) Representative action potential traces of SANCM, ACM and VCM in response to 100 nM NA (top panel). Baseline traces are in gray and individual groups are color coded; Bar graphs demonstrating cycle length in baseline (BL) and after addition of NA (bottom left panel), summarized as % change (bottom right panel). (b) Representative action potential traces of SANCM, ACM, and VCM following addition of 1 µM CCh (top panel); Bar graphs demonstrating cycle length in baseline (BL) and after addition of CCh (bottom left panel), summarized as % change (bottom right panel). (c) Representative action potential traces of SANCM, ACM, and VCM upon treatment with 3 µM Ivabradine (top panel); Bar graphs demonstrating cycle length in baseline (BL) (bottom left panel), summarized as % change (bottom right panel). N = 6 from four independent differentiations. Error bars, s.e.m. Wilcoxon signed-rank test, *p < 0.05.

Figure 5.

Generation and characterization of subtype-specific engineered heart tissues (EHTs). (a) Schematic representation of EHT fabrication. (b) Timeline of samples taken for qRT-PCR analysis (c) qRT-PCR analysis of gene expression in EHTs. Expression corrected to GEOMEAN of reference genes RPLP0 and GUSB. N = 4 independent differentiations. Error bars, s.e.m. Kruskal Wallis Test followed by Mann Whitney U-test for post hoc comparison. *p < 0.05, **p < 0.01.

Figure 6.

EHT response to Noradrenaline, Carbachol and Ivabradine. (a) Representative beating profiles and summary of beating rates recorded at 7, 14, 21, and 28-days post fabrication in SANCM-, ACM-, and VCM EHTs. (b–d) Beating rate in response to Noradrenaline (b), Carbachol (c) and Ivabradine (d). N = 6 EHTs/subtype from four independent differentiations (e) Beating rate in baseline (BL), in response to 100 µM carbachol (CCh) and subsequent treatment with 1 µM Noradrenaline (NA). (f) Beating rate in baseline (BL), in response to 100 µM carbachol (CCh) and subsequent treatment with 10 µM isoproterenol (Iso). N = 6 from four independent differentiations. Error bars, s.e.m. Pairwise error calculated with Wilcoxon’s Test. Multiple group comparison with Kruskal Wallis Test followed by Mann Whitney U-test for post hoc comparison, *p < 0.05.

Figure 7.

Impulse propagation in monolayers and EHTs. (a and b) Representative activation maps (left. panels) and summarized bar graphs (right panels) of spontaneous (a) and paced at 1 Hz (b) monolayers at 7 and 21 days post seeding. (c and d) Representative activation maps (left panels) and summarized data (right panels) of spontaneous (c) and paced (d) engineered heart tissues on 7-, 14-, 21-, and 28-days post fabrication. Error bars s.e.m. N = 4–6 from four independent differentiations. Multiple group comparison with Kruskal Wallis Test followed by Mann Whitney U-test for post hoc comparison, *p < 0.05, **p < 0.01 Mann Whitney Test. *p < 0.05, **p < 0.01 versus indicated sample.

Figure 8.

Fabrication of binary engineered heart tissues (BIN-EHTs). (a) Schematic illustrating the method to combine hiPSC-SANCM (cyan) and hiPSC-atrial cells (magenta) in BIN-EHTs. (b) Bright field and fluorescent images of a heteropolar EHT, where SANCM were labeled with DiI prior to fabrication. (c) Representative activation map (left) and summarized data (right) showing conduction velocities in SANCM, and ACM parts of BIN-EHTs. (d) Bar graph depicting site of impulse initiation in 8 BIN-EHTs. (e) Isochronal maps (left) demonstrating impulse initiation under baseline (BL) and 10 min after treatment with 3 µM ivabradine in a representative EHT. Summarized data from 8 BIN-EHTs depicts site of impulse initiation. Error bars s.e.m. N = 8 from four independent differentiations. Wilcoxon’s Test, *p < 0.05.