We used the directed differentiation protocol to generate ventricular-like CMs from iPSCs derived from 3 healthy individuals27. On day 15, iPSC-CMs were digested and distributed to 4 experimental groups (Fig. 1a). B27 medium, routinely used for iPSC-CM culture, served as a control. To unravel the synergistic effects of MM, NP, and ES on the maturation of iPSC-CMs, we systematically applied NP and ES to MM in a stepwise parallel manner (Fig. 1a). MM was designed based on a published FA-supplemented medium that enhances the metabolic maturation of iPSC-CMs13, with some modifications (Suppl. Table 1). NP was used to induce cell alignment and ES was applied to induce a beating frequency of 2 Hz (Suppl. Video 1).
Combined approach enhances structural maturation of iPSC-CMs
We observed that NP application induced changes in cell shape and a significant increase in the alignment of iPSC-CMs in the MM + NP and MM + NP + ES groups compared to the B27 and MM groups (Fig. 1b). In all three groups (MM, MM + NP, and MM + NP + ES) there was a noticeable increase in cell volume and granularity of iPSC-CMs compared to the B27 control. Compared to the MM group, NP did not induce an additional increase in cell volume and granularity, but the combination of MM + NP + ES induced further significant increases (Fig. 1c,d), suggesting that ES plays an important role in the hypertrophic growth of iPSC-CMs. A comparable proportion of cardiac troponin T (cTNT)-positive cells was found in all conditions, with the highest cTNT mean fluorescence intensity in iPSC-CMs from the MM + NP + ES group (Fig. 1e,f). Co-immunostaining for the sarcomeric protein α-actinin and cardiac ryanodine receptor (RYR2) revealed well-organised sarcomeric structures in iPSC-CMs under all conditions, but marked differences for the RYR2 localisation and the α-actinin/RYR2 colocalisation (Fig. 1g,h). In B27-cultured CMs, robust RYR2 staining was observed in the nucleus and punctate staining in the cytosol, with a low degree of α-actinin/RYR2 colocalisation. In contrast, CMs from the other three groups revealed reduced nuclear RYR2 staining and an augmented presence of striated patterns. The α-actinin/RYR2 colocalisation was enhanced in both the MM and MM + NP groups, and this effect was further augmented by ES. In the B27 and MM groups, the gap junction protein connexin 43 (Cx43) was partially localised to perinuclear regions rather than the plasma membrane, whereas Cx43 membrane localisation was increased in CMs of the MM + NP group and further significantly improved by ES (Fig. 1i). These results provide evidence for the additive effects of NP and ES to MM on the structural maturation of iPSC-CMs.
Combined approach improves electrophysiological maturation of iPSC-CMs
To evaluate the effect of MM, NP and ES on electrophysiological properties, we performed patch-clamp and multi-electrode array (MEA) studies to investigate the action potential (AP) and field potential (FP) parameters of single and monolayer iPSC-CMs, respectively. We observed the ‘notch-and-dome’ AP morphology only in iPSC-CMs (43%) of the MM + NP + ES group (Fig. 2a). The resting membrane potential (RMP) was found to be progressively more negative in CMs from the MM (-49.7 ± 1.7 mV), MM + NP (-58.2 ± 1.6 mV) and MM + NP + ES (-65.6 ± 2.1 mV) groups compared to the B27 group (-44.1 ± 2.1 mV), while the maximum AP upstroke velocity (Vmax) was gradually increased in iPSC-CMs from the MM (5.0 ± 0.2 V/s), MM + NP (6.6 ± 0.6 V/s) and MM + NP + ES (11.0 ± 2.0 V/s) groups compared to the B27 control (4.2 ± 0.3 V/s). Similarly, a gradual increase in AP amplitude (APA) was observed in the four groups (Fig. 2b). The AP duration at 90% repolarisation (APD90) was significantly shorter in iPSC-CMs paced at 0.5 Hz in the MM + NP + ES group than in the B27 control (Fig. 2c). As the transient outward K+ current (Ito) underlies the prominent phase 1 repolarisation of cardiac APs and the ‘notch-and-dome’ AP morphology, we measured Ito and found a significantly higher Ito density in iPSC-CMs from the MM + NP + ES group, but only a slight increase in MM and MM + NP conditions compared to the B27 group; and NP itself has less effect on Ito when comparing the MM + NP group with the MM group (Fig. 2d,e).
Intercellular electrotonic coupling and conduction velocity (CV) across CMs are largely dependent on Cx43 expression at the gap junction28. Consistent with the Cx43 localisation data (Fig. 1i), heatmaps of electrical signal propagation analysed by MEA illustrate the stepwise increase in CV in iPSC-CM monolayers from the MM (22.3 ± 0.8 cm/s), MM + NP (25.6 ± 0.9 cm/s) and MM + NP + ES (27.8 ± 1.5 cm/s) groups, compared to the B27 condition (12.5 ± 1.2 cm/s). Similar stepwise changes in spike amplitude and slope were observed in the four groups (Fig. 2f,g).
To analyse which changes in specific ion currents underlie the improved electrophysiological functionality, we recorded INa, IK1, and IKr using the patch-clamp technique. We found that INa density was significantly higher in MM-cultured iPSC-CMs with a mean peak current density of -87.9 ± 8.9 pA/pF at -20 mV compared to the B27 group (-40.2 ± 7.3 pA/pF). Notably, NP induced only a small increase in INa density with a mean peak of -100.1 ± 9.6 pA/pF at -25 mV when compared to the MM group, but INa density was further significantly induced by ES in the MM + NP + ES group with a mean peak of -136.9 ± 13.5 pA/pF (Fig. 3a,b). These data are consistent with the AP- (Vmax, APA) and FP-metrics (CV, spike amplitude and slope) shown in Fig. 2.
The electrophysiological immaturity of iPSC-CMs compared to adult CMs is partly attributed to the low density of the hyperpolarising K+ current IK1, which is important for stabilising the RMP29. We found a very low IK1 density in B27-cultured iPSC-CMs (-2.8 ± 0.4 pA/pF at -130 mV), and only a slight increase in IK1 in the MM group (-6.1 ± 1.2 pA/pF). Interestingly, NP induced a significant increase in IK1 in the MM + NP group (-13.1 ± 2.6 pA/pF), which was further induced by ES (-15.5 ± 2.6 pA/pF in the MM + NP + ES group) (Fig. 3c,d). HERG channels conducting the rapid delayed rectifier K+ current IKr are involved in phase 3 repolarisation of cardiac APs. Similar to IK1, both E-4031-sensitive IKr step and tail current densities were only slightly induced by MM, but significantly induced by NP and further enhanced by ES (Fig. 3e-g). These data are consistent with the most negative RMP and the shortest APD90 in iPSC-CMs from the MM + NP + ES group (Fig. 2b,c).
Taken together, these findings highlight the distinct effects of the three stimuli on specific ion currents. This is evidenced by the strong influence of NP on IK1 and IKr, with less or no effect on INa and Ito, and the robust effect of MM on INa, but less on IKr. Importantly, the data underline that the combined approach significantly enhanced electrophysiological maturation of iPSC-CMs.
Combined approach improves calcium handling and contractility of iPSC-CMs
Since excitation-contraction coupling in CMs involves calcium cycling to convert electrical signals into mechanical output (contraction), we next examined L-type calcium channel (LTCC) current ICa−L and calcium transients in iPSC-CMs. ICa−L densities exhibited similar reductions in both the MM and MM + NP groups when compared to the B27 group, which were further reduced by ES (Fig. 4a,b). To quantify intracellular Ca2+ dynamics, we performed Fura-2-based calcium imaging in iPSC-CMs paced at 0.5 Hz (Fig. 4c-e). Significantly reduced diastolic and systolic Ca2+ levels were observed in the MM, MM + NP and MM + NP + ES groups compared to the B27 group, but Ca2+ transient amplitudes were comparable between all groups despite the reduced ICa−L density in the MM, MM + NP and MM + NP + ES groups. The Ca2+ transient decay time constant (tau) is also significantly shortened in the MM group compared to the control, whereas no further shortening was observed in the MM + NP and MM + NP + ES groups. Application of 10 mM caffeine resulted in significantly increased Ca2+ release from the sarcoplasmic reticulum (SR) in the MM, MM + NP and MM + NP + ES groups compared to the B27 group (Fig. 4e). These data suggest a more efficient coupling between ICa−L and SR Ca2+ release, enhanced Ca2+ decay kinetics and a higher SR calcium content in these three groups.
Video-based analysis of iPSC-CM beating properties5 showed that the changes in calcium handling were associated with improved contractile function. Stopping ES in the MM + NP + ES group resulted in cessation of beating, followed by regaining of spontaneous beating activity within 15–30 minutes at a rate comparable to the other three groups (Fig. 4f). Significantly shorter beating duration, contraction and relaxation times were observed in the MM + NP + ES group compared to the other groups (Fig. 4g; Extended Data Fig. 2a,b). We found a similar trend in the MM and MM + NP groups compared to the B27 control, but no significant difference between the two groups. Consistent with this observation, all cultures in the MM + NP + ES group successfully captured high-frequency (2 Hz) field stimulation, whereas none of the B27 cultures demonstrated this capability (Fig. 4h; Extended Data Fig. 2c). This improved contractile function was accompanied by an increased gene expression ratio of TNNI3/TNNI1 and MYL2/MYL7, whereas the expression of MYH6, encoding the fast-twitch MHC isoform, was upregulated in response to sustained ES, leading to a reduced MYH7/MYH6 ratio (Fig. 4i). These results demonstrate that MM, NP and ES individually and synergistically induce electrophysiological and functional maturation of iPSC-CMs.
Combined approach improves drug response of iPSC-CMs
To investigate whether the maturation state of iPSC-CMs influences their drug response, we chose verapamil (calcium-channel blocker), E-4031 (hERG-channel blocker) and isoprenaline (β-adrenergic stimulus) as model substances to detect pro-arrhythmic activity based on changes in FP parameters (Fig. 5; Extended Data Fig. 3a,b). We observed beating arrest in cultures from the B27 (all cultures), MM (9/17) and MM + NP (7/18) groups at 1 µM verapamil. Concentration-dependent reductions in spike amplitude were observed in the B27 group and, to a lesser extent, in the MM and MM + NP groups. In contrast, no effect of verapamil on beating activity and spike amplitude was observed in the MM + NP + ES group (Fig. 5b; Extended Data Fig. 3c). Verapamil-induced shortening of FP duration (FPDc), which corresponds to QT-shortening in the clinic, was comparable in iPSC-CMs from the MM, MM + NP and MM + NP + ES groups. However, this effect was more pronounced compared to iPSC-CMs cultured in B27 (Fig. 5c,d). Previous studies showed that immature iPSC-CMs failed to produce APD prolongation after E-4031 treatment, even at high concentrations30. Similarly, we found that E-4031 induced only minor changes in FPDc in B27-cultured iPSC-CMs, whereas significant concentration-dependent FPDc prolongation was detected in the MM, MM + NP and MM + NP + ES groups (Fig. 5e,f). Furthermore, we observed a more pronounced positive-chronotropic response of iPSC-CMs to isoprenaline in these three groups than in the B27 group, which correlates with FPDc-shortening (Fig. 5g,h). The EC50 of isoprenaline for the chronotropic effect was also much lower in these three groups than in the B27 group (Fig. 5h). These experiments demonstrate the substantial impact of the maturation state of iPSC-CMs on their response to various cardioactive drugs. They emphasise the importance of utilising iPSC-CMs with more adult-like electrophysiological properties for accurate drug risk assessment.
Combined approach downregulates MAPK/PI3K-AKT pathways
Previous studies have shown that FA-enriched media induce iPSC-CM maturation by regulating key genes involved in FA metabolism, mitochondrial function, calcium cycling, ion channels and sarcomere13,20. To gain insight into the molecular mechanisms driving iPSC-CM maturation by NP and ES, we performed RNA sequencing (RNA-seq) analysis. Surprisingly, NP had little synergistic effect when combined with MM, whereas the addition of ES strongly influenced gene expression (Fig. 6a-d; Extended Data Fig. 4a,b). Comparing the MM and MM + NP groups, only 163 differentially expressed genes were identified, of which 56 were upregulated and 107 downregulated in the MM + NP group. In contrast, 1,370 significantly upregulated and 1,657 downregulated genes were identified in the MM + NP + ES group compared to the MM group, of which 747 significantly upregulated and 990 downregulated genes were also identified when compared to the MM + NP group, indicating the synergistic effects of NP and ES.
Pathway enrichment analysis of the downregulated genes in the MM + NP + ES group mainly mapped to MAPK/PI3K-AKT, TNFR2-NFκB, G-protein-coupled receptor (GPCR), and cytokine/chemokine signalling (Fig. 6e; Extended Data Fig. 4d; Suppl. Table 2). Using the transcription factor target database, we identified the SRF cluster, which includes many genes involved in cell cycle regulation and cell proliferation (Fig. 6f,g; Extended Data Fig. 4c). We also found a decrease in SRF protein levels in CMs from the MM + NP + ES group compared to the other groups (Fig. 6h). These findings encouraged us to evaluate the expression of genes that regulate cell cycle progression. Notably, activators of G2/M checkpoints including cyclins (CCNB1-3), cyclin-dependent kinase 1 (CDK1) were downregulated, whereas CDK inhibitors (CDKN1A, CDC20) were upregulated in the MM + NP + ES group compared to the other two groups (Fig. 7a,b). Interestingly, genes (ANLN, SEPTIN7/2) encoding activators of cytokinesis were also downregulated. We did not observe any significant changes in the gene sets (CCND1-3, CCNE1/2, CCNA1/2, CDK2/4/6, CDKN2A-D, CDKN1B/C, CDH1) controlling the G1 and S progression and the G1/S checkpoint (Fig. 7a,b). These data suggest a cell cycle arrest after S phase and before exit from M phase in the MM + NP + ES group, which may lead to bi-nucleation or nuclear polyploidy. To confirm this, we examined DNA content and found that the number of diploid iPSC-CMs was significantly reduced and polyploid cells significantly increased in the MM + NP + ES group compared to the other three groups (Fig. 7c,f). Interestingly, we observed no difference in the proportion of 5-ethynyl-2’-deoxyuridine (EdU)-incorporated iPSC-CMs between all four groups, which can detect DNA synthesis during S phase (Fig. 7d,g). However, the proportion of Ki67+ iPSC-CMs and Ki67− polyploid iPSC-CMs was higher in the MM + NP + ES group than in the other groups (Fig. 7e,h,i). Ki67 is widely expressed throughout the entire cell cycle, except in G0, and reaches a maximum in S/G231. These results indicate that the downregulation of MAPK/PI3K-AKT and SRF-related genes is involved in G2/M arrest and polyploidy development of iPSC-CMs.
Changes in gene expression profile associated with metabolism and electrophysiology
The pathway enrichment analysis of genes upregulated in the MM + NP + ES group revealed that the combined approach induced the upregulation of genes involved in electron transport chain (ETC), TCA cycle, mitochondrial biogenesis, NRF2 signalling, glucose metabolism, N-glycan biosynthesis, FA oxidation, tRNA aminoacylation, etc. (Fig. 8a; Extended Data Fig. 4e; Suppl. Table 3). These gene clusters were only slightly upregulated in the MM + NP group compared to the MM group (Extended Data Fig. 4e).
Using the transcription factor target database we identified two clusters, TFAM (Fig. 8b) and HMCES (Fig. 8c), which were enriched in the MM + NP + ES group (Extended Data Fig. 4c). In these two clusters, mitochondrial DNA (mtDNA)-encoded NADH dehydrogenase subunits (MTND2-6) and pseudogenes (MTND4P12, MTND5P11, MTND6P4), cytochrome c oxidase subunits (MT-CO1/3) and pseudogenes (MT-CO1P2, MT-CO1P12, MT-CO3P12), cytochrome b (MTCYB), 12S rRNA (MT-RNR1) and tRNAs (MT-TP/-TM/-TE/-TI/-TW/-TC/-TY/-TR/-TN/-TQ) were upregulated in the MM + NP + ES group compared to the MM group.
Subsequently, we found increased mitochondrial mass in CMs from the MM + NP + ES group compared to the other groups (Fig. 8d). Furthermore, CMs from the MM + NP + ES group showed a higher expression of OPA1, PPARGC1α and PPARα (Fig. 8g), confirming an enhanced mitochondrial development in response to ES. Measurements of oxygen consumption rate as a surrogate for mitochondrial function showed an increased basal and maximal respiration, ATP production and spare capacity of CMs from the MM group compared to the B27 control. There is only a marginal additional increase in these parameters in the MM + NP + ES and MM + NP groups compared to the MM group (Fig. 8e,f). These findings indicate that the combination of MM + NP + ES leads to upregulation of oxidative phosphorylation, activation of mtDNA-encoded components and an overall enhancement of mitochondrial development and function.
Further examination of individual changes in key ion channel components revealed a clear correlation between gene expression and channel function (Fig. 8h). We found upregulation of genes contributing to Ito (KCNA7, KCNC3, KCNC4, KCND2, KCND3), IK1 (KCNJ2, KCNJ4, KCNJ6, KCNJ11, KCNJ12), IKr (KCNH2, KCNE2) and IKs (KCNQ1) currents, as well as genes encoding calcium-activated (KCNN1) and voltage-gated (KCNS3, KCNAB2, KCNIP3) potassium channels in the MM + NP + ES group compared to the MM group. In contrast, genes encoding the LTCCs (CACNA1C, CACNA1D, CACNA1A, CACNA2D3, CACNA2D4) and regulating the LTCC activity (CACNG7) were downregulated, whereas the genes encoding the T-type calcium channels (CACNA1H, CACNA1I) were upregulated. No significant differences in SCN5A expression were found between the three groups, but SCN1B was upregulated and SCN3B was downregulated in the MM + NP + ES group. In addition, the expression of the genes encoding calcium handling proteins were not significantly altered. Collectively, these results are consistent with the significantly increased INa, Ito, IK1, and IKr but decreased ICa−L currents in iPSC-CMs from the MM + NP + ES group.