OUP user menu

Pharmacological response of human cardiomyocytes derived from virus-free induced pluripotent stem cells

Ashish Mehta, Ying Ying Chung, Alvin Ng, Fahamy Iskandar, Shirhan Atan, Heming Wei, Greg Dusting, William Sun, Philip Wong, Winston Shim
DOI: http://dx.doi.org/10.1093/cvr/cvr132 577-586 First published online: 12 May 2011


Aims Generation of human induced pluripotent stem cell (hiPSC) lines by reprogramming of fibroblast cells with virus-free methods offers unique opportunities for translational cardiovascular medicine. The aim of the study was to reprogramme fibroblast cells to hiPSCs and to study cardiomyogenic properties and ion channel characteristics of the virus-free hiPSC-derived cardiomyocytes.

Methods and results The hiPSCs generated by episomal vectors generated teratomas in severe combined immunodeficient mice, readily formed embryoid bodies, and differentiated into cardiomyocytes with comparable efficiency to human embryonic stem cells. Temporal gene expression of these hiPSCs indicated that differentiation of cardiomyocytes was initiated by increasing expression of cardio/mesodermal markers followed by cardiac-specific transcription factors, structural, and ion channel genes. Furthermore, the cardiomyocytes showed characteristic cross-striations of sarcomeric proteins and expressed calcium-handling and ion channel proteins, confirming their cardiac ontogeny. Microelectrode array recordings established the electrotonic development of a functional syncytium that responded predictably to pharmacologically active drugs. The cardiomyocytes showed a chronotropic dose–response (0.1–10 µM) to isoprenaline and Bay K 8644. Furthermore, carbamycholine (5 µM) suppressed the response to isoprenaline, while verapamil (2.5 µM) blocked Bay K 8644-induced inotropic activity. Moreover, verapamil (1 µM) reduced the corrected field potential duration by 45%, tetrodotoxin (10 µM) shortened the minimal field potential by 40%, and E-4031 (50 nM) prolonged field repolarization.

Conclusion Virus-free hiPSCs differentiate efficiently into cardiomyocytes with cardiac-specific molecular, structural, and functional properties that recapitulate the developmental ontogeny of cardiogenesis. These results, coupled with the potential to generate patient-specific hiPSC lines, hold great promise for the development of an in vitro platform for drug pharmacogenomics, disease modelling, and regenerative medicine.

  • Induced pluripotent stem cell
  • Differentiation
  • Cardiomyocyte
  • Drug screening

1. Introduction

Human embryonic stem cells (hESCs),1 with their ability to differentiate into various cell lineages, including cardiomyocytes,2,3 have provided researchers with a unique tool for the study of developmental biology, toxicology, and regenerative medicine.46 Cell therapy for cardiovascular repair is being studied in clinical trials.7 However, a major obstacle in the clinical application of hESC-derived cardiomyocytes for myocardial cell transplant is immune rejection of these allogenic cell grafts.

Takahashi and colleagues,8 for the first time, demonstrated that fibroblasts could be reprogrammed by ectopic retroviral expression of transcription factors (Oct-4, Sox2, c-Myc, and Klf4) to generate human induced pluripotent stem cells (hiPSCs) that shared similar pluripotency to hESCs. Likewise, Yu and co-workers9 also demonstrated reprogramming of fibroblasts with Oct-4, Sox2, Nanog, and Lin28 using lentiviral vectors. This virus-based reprogramming technique is able efficiently to generate hiPSCs that could be differentiated into various cell types, including cardiomyocytes.1012 However, viral vectors are associated with major drawbacks, such as genomic integration,13 insertional mutations, residual transgene expression, and tumourigenesis,14,15 which may limit the clinical application of hiPSC-derived cell types. In the clinical setting, therefore, vector integration-free methods are preferred. Recently, virus-free and vector-free hiPSCs have been derived by employing episomal vectors,13 piggyBac,16 mini-circle,17 and synthetic modified mRNA,18 which may overcome the numerous drawbacks of virus-derived hiPSCs, as well as the ethical and immunological issues surrounding use of hESCs.

There is, however, a paucity of information on the cardiomyogenic potential of virus-free hiPSCs. Our objective in the present study was to generate virus- and vector-free hiPSCs, to characterize their cardiomyogenic differentiation potential, and to define developmental steps involved in the differentiation process. Furthermore, we attempted to study the molecular, structural, and functional properties of hiPSC-derived cardiomyocytes (hiPSC-CMs).

2. Methods

2.1 Cell culture and reprogramming of fibroblasts

For detailed methodology of the maintenance of cell lines, please see Supplementary material online. Reprogramming of human foreskin fibroblasts (HFF; Lonza, Basel, Switzerland) were co-transfected with two oriP/EBNA1-based episomal plasmids (pEP4 EO2S CK2M EN2L, pEP4 EO2S ET2K; Addgene Inc., Cambridge, MA, USA) via nucleofection (NHDF – VPD-1001 with U-20 program; Amaxa, Walkersville, MD, USA). Transfected fibroblasts (∼1.0 × 106 cells per nucleofection) were directly plated to 10 cm mitomycin-C inactivated mouse embryonic fibroblast-seeded dishes with foreskin fibroblast culture medium.13 Culture medium was changed every other day. On day 4 post-transfection, the foreskin fibroblast culture medium was replaced with human embryonic stem cell culture medium19,20 containing 100 ng/mL of basic fibroblast growth factor (R&D Systems, Minneapolis, MN, USA). Colonies with morphology similar to hiPS colonies were readily visible by day 21–25 post-transfection. Individual colonies were picked manually and plated on 1% Matrigel-coated dishes and maintained in mTeSR1 medium (Stemcell Technologies, Vancouver, BC, Canada). Enzymatic passaging with dispase was employed for expansion, characterization, and differentiation experiments as described in the Supplementary material online.

2.2 Embryoid body formation and cardiomyocyte differentiation

Pluripotent stem cell colonies were dispersed into small clumps with dispase (1 mg/mL) and placed in low-adhesion culture dishes in embryoid body (EB) medium,20 with or without 5 µM SB203580 (Calbiochem, La Jolla, CA, USA) for 8 days.21 Subsequently, EBs were plated on 0.1% gelatin-coated dishes in EB medium without SB203580. Beating areas were typically observed around day 11–14 from EB formation. Beating areas were manually cut after day 21 of differentiation and utilized for various experiments.

2.3 Immunostaining

Colonies of iPSCs and single cells generated from beating clusters were seeded on Matrigel- or gelatin-coated glass slides, respectively. For detailed methodology, please refer to the Supplementary materials online.

2.4 Teratoma formation

All animal experiments were conducted following experimental protocols approved by the SingHealth Institutional Animal Care and Use Committee, in full compliance with Singapore laws and regulations, and followed the guidelines of the US National Institutes of Health Guide for the Care and Use of Laboratory Animals. Severe combined immunodeficient (SCID) mice, 6 weeks old, weighing 20–23 g, were obtained from SingHealth Experimental Medicine Centre, and were anaesthetized with 2% isoflurane initially followed by 1% isoflurane with oxygen with a flow rate of 1 ml/min during surgery. Approximately 1 × 106 hiPSCs were injected into the kidney capsule as previously described.22 Mice were killed with carbon dioxide asphyxiation at 8 weeks after cell injections, and tumours were collected, fixed, and processed for haematoxylin and eosin staining following conventional protocols.

2.5 Real-time PCR

For quantitative real-time reverse-transcription PCR (qRT-PCR) analysis, adult human heart RNA (Clontech, Mountain View, CA, USA), undifferentiated hiPSCs/hESCs (day 0) and differentiating EBs at different time points were used. The RNA was isolated with the RNeasy kit (Qiagen GmbH, Hilden, Germany). One microgram of total RNA was converted to cDNA by Superscript II first-strand synthesis system (Invitrogen, Carlsbad, CA, USA). The cDNA template (5 ng) was used from each sample, and SYBR green real-time PCR studies were performed using Quantifast kit (Qiagen GmbH) and primer (see Supplementary material online, Table S1) according to the kit instructions. Samples were cycled with Roto Gene Q (Qiagen GmbH) as follows: 5 min at 95°C, followed by 40 cycles of 10 s at 95°C and 30 s extension at 60°C. All experiments were performed in triplicate. Relative quantification was calculated according to the ΔΔCt method for quantitative real-time PCR (using an endogenous control gene, GAPDH). For each gene, the expression on a specific day was then normalized by its baseline values.

For semi-quantitative RT-PCR and genomic PCR please refer to the Supplementary material online.

2.6 Microelectrode array recordings

To characterize the electrophysiological properties of the hiPSC-CMs, a microelectrode array (MEA) recording system (Multichannel Systems, Reutlingen, Germany) was used. Contracting areas were microdissected and plated on gelatin-coated MEA plates. The clusters were allowed to adjust for 72 h before performing any recording. All clusters were monitored for their beating abilities (beats/min) under the microscope during the 72 h period. Clusters that maintained relatively uniform beating rates were then subjected to drugs. The MEA system allows simultaneous recordings from 60 titanium nitride-coated gold electrodes (30 µm) at high spatial (200 µm) and temporal resolutions (15 kHz). To assess the effects of different drugs on the electrophysiological properties, the stock drugs were diluted in medium (2 mL). The MEA clip along with the beating clusters was maintained at 37°C throughout the duration of the experiments. Care was also taken that all buffers, including the medium used during all experiments, were pre-warmed to 37°C. The tested drugs included isoprenaline hydrochloride (Iso), carbamylcholine (CCh), verapamil (Ver), Bay K 8644 (Bay), tetrodotoxin (TTX), and E-4031 (all from Sigma-Aldrich, St Louis, MO, USA). All extracellular recordings were performed for 180 s at baseline and at 5 min after drug application at 37°C. Data were recorded using MC Rack software (Multichannel Systems) for all drugs; however, for measurement of conduction velocity and generation of local activation maps, Cardio2D software (Multichannel Systems) was used. The recorded electrograms were also used to determine the local field potential (FP) duration (FPD). The FPD measurements were normalized (corrected FPD; cFPD) to the beating rate of the contracting areas with the following Bazzet correction formula: cFPD = FPD/√(RR interval), as described previously.11

2.7 Statistical analysis

Comparisons at each time point were conducted using analysis of variance (ANOVA) followed by Tukey's post hoc test, and all data are presented as mean values ± SEM. Differences were considered statistically significant at P ≤ 0.05.

3. Results

3.1 Generation and characterization of human iPSCs

We generated five virus-free hiPSC clones using two episomal vectors (containing Oct-4, Sox2, Klf4, c-myc, Nanog, Lin28, and SV40L) according to the protocol (see Supplementary material online, Figure S1A). The visual morphology of these hiPSCs was similar to hESC colonies (e.g. compact colonies, high nucleus-to-cytoplasm ratios, and prominent nucleoli), on feeders as well as feeder-free culture, with normal karyotype (see Supplementary material online, Figure S1B and C). Strong nuclear staining for Oct-4 and surface antigens, SSEA4, Tra-1-60, and Tra-1-81 (Figure 1A) confirmed the pluripotency of these clones. Furthermore, flow cytometric analysis of these undifferentiated colonies revealed that more than 95% of cells expressed hallmark markers (Oct-4, SSEA4, Tra-1-60, and Tra-1-81) of pluripotent stem cells (see Supplementary material online, Figure S1D).

Figure 1

Characterization of virus-free iPSCs. (A) Immunostaining of the undifferentiated hiPSC colonies with Oct-4, SSEA4, Tra-1-60, and Tra-1-81 antibodies followed by counterstaining with 4',6-diamidino-2-phenylindole. Scale bar represents 100 µm. (B) Relative gene expression levels of Oct-4, Sox-2, and Nanog in five iPSC clones (C1–C5) in comparison with the hESC line, H9 (positive control) and human foreskin fibroblasts (HFF; negative control). The mean Ct values of duplicate measurements were calculated and subsequently normalized against housekeeping gene (GAPDH) for the same sample. After normalization, the means of triplicate samples from three independent experiments were plotted relative to H9 for all the markers. Data represented are mean ± SEM (n = 3). *P < 0.05 vs. H9. (C) Haematoxylin and eosin staining of teratoma sections of clone MSnviPSNF3 showing the presence of ectoderm (neural rosettes), mesoderm (cartilage), and endoderm (secretory tubule). Scale bars represent 200 μm.

Real-time gene expression study of pluripotency markers, Oct-4, Sox2, and Nanog, clearly demonstrated that all the five hiPSC clones maintained higher/comparable levels of pluripotency markers in comparison with the standard hESC line, H9 (Figure 1B), while these markers were not detected or very low for foreskin fibroblast cells (Figure 1B). Despite reprogramming of the naive fibroblasts using non-integrating episomal plasmids, it was imperative to rule out genetic incorporation in these hiPSC clones. Genomic DNA PCR indicated no extraneous genetic incorporation of any vector components or exogenous genes in our virus-free iPSC clones (see Supplementary material online, Figure S2A).

One of the hallmarks for pluripotent stem cells is their ability to differentiate into three developmental germ layers in vitro as well as in vivo. In vitro differentiation of hiPSCs was performed by generating EBs (see Supplementary material online, Figure S2B), which demonstrated marked up regulation of genes indicative of ectodermal (Nestin and Pax6), mesodermal (Isl1 and GATA4), and endodermal markers (AFP and HNF4α), confirming ESC-like differentiation capabilities (see Supplementary material online, Figure S2C). Furthermore, transplantation of the undifferentiated hiPSCs into the kidney capsule of SCID mice generated teratomas at 8 weeks, which on haematoxylin and eosin staining demonstrated the presence of neuroepithelial rosettes (ectoderm), cartilage (mesoderm), and secretory tubules (endoderm), representing the three germinal layers (Figure 1C).

3.2 In vitro differentiation and characterization of hiPSCs as cardiomyocytes

To induce cardiomyocyte differentiation, the hiPSCs were cultivated in suspension, where they formed three-dimensional differentiating cell aggregates (EBs) in the presence of SB203580 for 8 days. Embryoid bodies were plated, and rhythmically contracting areas appeared as early as 11 days post-differentiation (Figure 2 and Movie 1 on Supplementary file 2). These beating clusters increased with time, and the differentiation efficiency of our clones based on beating frequency ranged from 5 to 25% by day 21 of differentiation. The contracting areas had a diameter ranging from 0.2 to 1.2 mm and continued to beat vigorously for several weeks (45–60 days as observed) in culture. Based on the number of beating clusters, our clone MSnviPSNF3 demonstrated 20–25% contracting efficiency, comparable to the standard hESC line, H9 (20–30%). Interestingly, the appearance of beating clusters was less frequent in other clones (MSnviPSNF2, 10–15%; MSnviPSNF5, 5–10%; and MSnviPSNF1and MSnviPSNF6, 0–5%). Henceforth, all experimental evaluations were performed using the MSnviPSNF3 clone and the hESC line, H9.

Figure 2

Temporal gene expression pattern during cardiomyogenesis. A spontaneous contracting cluster developed following differentiation. Scale bar represents 50 µm. Graphs show real-time RT-PCR data showing various hallmark markers for cardiomyocyte differentiation (days0, 7, 14, and 21). The data shown here are a comparison of MSnviPSNF3 (C3) with respect to the hESC line (H9). The mean Ct values of duplicate measurements were calculated and subsequently normalized against housekeeping gene (GAPDH) for the same sample. After normalization, the means of triplicate samples from three independent experiments were plotted relative to the baseline values of each individual marker for both cell lines. Data represented are means ± SEM (n = 3). Details of gene abbreviations are provided in the Supplementary material online, Table S1.

Real-time PCR analysis (Figure 2) demonstrated that initiation of differentiation was accompanied by a significant decrease in pluripotency marker (Oct-4) levels by day 7, followed by up-regulation of markers of mesodermal commitment towards cardiogenesis. Cardiac progenitor marker (Isl1, a marker of the secondary heart field) and cardiac-associated transcriptional factors (Nkx.2.5, Hand1, and GATA4) increased significantly by day 7 in comparison to day 0 of differentiation (Figure 2). Concomitant with the increased expression of cardiac transcriptional factors, significant up-regulation was also observed for cardiac-specific structural and sarcomeric proteins (actinin, troponins, MYH7, and MLC2v) by days14 and 21 (Figure 2). Furthermore, ion channel proteins [Cav1.3, encoding the α1D subunit of the L-type calcium channel; KCNH2 (hERG), mediating the rapid delayed rectifier potassium current (IKr); and the hyperpolarization-activated cyclic nucleotide-gated potassium channel (HCN2), responsible for the If pacemaker current]11 were significantly up-regulated by day 21 of differentiation (Figure 2). A similar trend was also observed in the control hESC line, H9.

3.3 Molecular and cellular characterization of hiPSC-CMs

Spatial organization for cardiac-specific proteins in hiPSC-CMs was performed by immunostaining. We observed that 61.5 ± 9.9% (range 50–70%, n = 5) of the cells within each beating cluster stained positive with cardiac transcriptional factor, Nkx2.5 (Figure 3A). Furthermore, hiPSC-CMs also stained positive for cardiac α-actinin, myosin light chain 2, β-myosin heavy chain, cardiac troponin C, and titin (Figure 3A). Positively stained cardiomyocytes demonstrated an immature striated pattern, indicative of the early stages of myocyte development. However, Z-bands and A-bands were clearly visible in the cardiomyocytes (Figure 3A; indicated by arrows). Our hiPSC-CMs also stained positively for connexin 43, Na+–Ca2+ exchanger and sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA2; Figure 3A), suggesting their ability to form cell–cell interactions23 and regulate calcium.24

Figure 3

Structural and molecular characterization of iPSC-CMs. (A) Immunostaining of transcriptional factor, Nkx2.5, structural proteins, sarcomeric α-actinin, myosin light chain 2 (MLC2), β-myosin heavy chain (β-MHC), cardiac troponin C (cTn C), titin, gap junction connexin 43 (Cx43), and ion channels, namely, Na+–Ca2+ exchanger (Na-Ca Ex) and sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA). Nuclei were counterstained with 4',6-diamidino-2-phenylindole in all images. All images were obtained at ×63 magnification except Nkx2.5 (×20). Bar = 50 µm for all images. (B) Semi-quantitative RT-PCR analysis of undifferentiated and day 21 differentiated cardiomyocytes of clone MSnviPSNF3 and H9 with adult human heart. Abbreviations: iPSC-UD, undifferentiated MSnviPSNF3 cells; iPSC-CM, MSnviPSNF3-derived cardiomyocytes; hESC-UD, undifferentiated H9 cells; hESC-CM, H9-derived cardiomyocytes; and AHH, adult human heart.

However, semi-quantitative (Figure 3B) and real-time PCR analysis revealed that sarcomeric structural proteins and ion channels were expressed at lower levels in hESCs and hiPSC-CMs in comparison to the adult heart, while the reverse was true for early developmental markers (Nkx2.5 and GATA4; Figure 3B; Supplementary material online, Table S2).

3.4 Functional characterization of cardiomyocytes

Despite the presence of hallmark molecular and cellular characteristics, did these hiPSC-CMs display functional electrophysiological properties and form a functional syncytium? Extracellular ECGs were recorded from electrodes underlying the EBs on a MEA chip (Figure 4A and B). Our hiPSC-CMs demonstrated a conduction velocity ranging from 1.5 to 2.5 cm/s, depending on the EB (n = 9). Electrical activation maps (Figure 4C) demonstrated development of a functional syncytium with stable pacemaker activity and synchronized field potential propagation. The conduction patterns were relatively stable and reproducible within each beating cluster.

Figure 4

Microelectrode array mapping and chronotropic effects of pharmacological interventions. (AC) Contracting clusters on the MEA chips with extracellular electrograms from all electrodes used to generate colour-coded activation maps. Scale bar represents 200 µm. (DI) Dose-dependent chronotropic changes induced by Bay K 8644 (D and E), isoprenaline (F), verapamil (G and H), and carbamycholine (I). *P < 0.05 vs. control (no treatment) and #P < 0.01 vs. control (no treatment).

3.5 Chronotropic responses of hiPSC-CMs to pharmacologically active compounds

One of the most important properties of cardiac tissue physiology is the ability to respond appropriately to pharmacologically active compounds. Human iPSC-CMs demonstrated a dose-dependent increase in spontaneous beating frequency when treated with the L-type calcium activator, Bay K 8644, and the β-adrenergic agonist, isoprenaline (Figure 4DF). While treatment with 10 µM Bay K 8644 increased the beating rate by 2.5 times [0 µM Bay K 8644 (control) vs. 10 µM Bay K 8644: 0.921 ± 0.002 vs. 2.610 ± 0.040 Hz; P < 0.05; Figure 4E], 1 µM isoprenaline (Iso) doubled the beating rate [0 µM Iso (control) vs. 1 µM Iso: 0.799 ± 0.007 vs. 1.910 ± 0.010 Hz; P < 0.05; Figure 4F]. In contrast, the chronotropic effects of the L-type calcium channel blocker, verapamil (1 µM), reduced the beating frequency by 30–40%, and at 5 µM arrested beating (Figure 4G and H). Similar results were observed with carbamycholine, but at comparatively lower doses (Figure 4I).

Activation–inhibition studies with hiPSC-CMs demonstrated a significant increase in the beating frequency with 10 µM Bay K 8644, which could be significantly reduced by 2.5 µM verapamil (Figure 5A and B). Interestingly, exposure to verapamil also resulted in irregular beating rates, and prolonged exposure arrested spontaneous beating (Figure 5A). Likewise, muscarinic inhibition following adrenergic stimulation was performed. Isoprenaline (1 µM) treatment significantly increased the beating frequency, which could be inhibited with carbamycholine (5 µM; Figure 5C and D).

Figure 5

Inhibition studies of iPSC-CMs. (A and B) Changes in beating frequency induced by verapamil in the presence of Bay K 8644. *P < 0.05 vs. control, and #P < 0.05 vs. Bay group. (C and D) Muscarinic inhibition of beating frequency during adrenergic stimulation. *P < 0.05 vs. control, #P < 0.05 vs. Iso group, and +P < 0.05 vs. control. Data are represented as the means ± SEM of three independent experiments. Abbreviations: Bay, Bay K 8644; Ver, Verapamil; ISO, isoprenaline; and CCh, carbamycholine.

Furthermore, verapamil demonstrated a dose-dependent effect on the cFPD that at a concentration of 1 µM caused a 45% shortening of cFPD (Figure 6A and B). Likwise, TTX, a reversible, selective blocker of Na+ channels, was observed to shorten the FPmin (40–45%) in a dose-dependent manner and reduce conduction velocity by 33–42% (Figure 6C and D). Finally, E-4031, a methanesulfonanilide class III anti-arrhythmic drug that blocks the hERG-type potassium channel, caused a statistically significant QT prolongation at a concentration of 50 nM, which did not increase significantly at a concentration of 100 nM (Figure 6E and F). These results validated the presence of functional ion channels in our hiPSC-CMs.

Figure 6

Human iPSC-CMs respond to pharmacologically active compounds. (A and B) Extracellular field potential (FP) recordings following verapamil treatment demonstrate shortening of the field potential duration (FPD). (C and D) Tetradotoxin (TTX), a sodium channel blocker, reduces minimal field potential (FPmin) and conduction velocity. (E and F) Field potential recording demonstrates an increase in field repolarization following treatment with E-4031. Note that the small black arrows show the normal baseline recordings, whereas large grey arrows show changes after treatment. (G) Inset, schematic diagram of a multi-electrode array trace, showing how results were analysed to calculate FPD and field repolarization (QT interval). All graphs represent values as means ± SEM of three independent experiments. *P < 0.05 vs. control, and #P < 0.01 vs. control. Abbreviation: cFPD, corrected FPD.

4. Discussion

The reprogramming of differentiated somatic cells to pluripotency holds great promise for drug discovery and developmental biology. This opens an opportunity to offer hope that patient-specific hiPSCs could generate clinically useful cell types for autologous therapy aimed at repairing defects arising from injury, illness, and ageing. Virus-based methods of generating iPSCs, however, have demonstrated high reprogramming efficiency, viral backbone integration, and continuous expression of reprogramming genes,18 which present formidable obstacles for their eventual therapeutic use.13 Thus, application of reprogramming without incurring genetic changes and subsequent characterization of their derivatives has become a focus of intense research effort. While there are limited studies showing cardiogenic potential for virus-based iPSCs,11,25 no report is currently available for cardiogenic differentiation of virus-free iPSCs. In the present study, we reprogrammed human fibroblast cells using virus-free methodology, and differentiated the hiPSCs to a defined cardiac lineage.

Our initial attempts to use only one episomal plasmid failed to generate any iPSC colonies (data not shown), which probably could be attributed to the lack of sufficient levels of pluripotency genes (Oct-4 and Sox2). Subsequent transfection with two plasmids in combination resulted in generation of iPSCs, suggesting that high levels of these transgenes improve reprogramming. These iPSC colonies exhibited typical hESC morphology (see Supplementary material online, Figure S1B), and exhibited pluripotency markers at gene and protein levels similar to hESC lines (Figure 1). Similar to hESCs, when injected into SCID mice, these iPSCs formed teratomas consisting of differentiated derivatives of all three primary germ layers (Figure 1). Moreover, PCR analysis failed to detect episomal vector integration in the genome (see Supplementary material online, Figure S2A). Exogenous DNA is not integrated into the human iPSC genome with oriP/EBNA1-based episomal vectors, which is gradually lost when cells are cultured in the absence of drug selection pressure.13 However, the reprogramming efficiency with these vectors is reported to be low (∼3–6 colonies per 106 input cells).13 These low frequencies are, however, sufficient to recover iPSCs from a reasonable number of starting cells.

We next investigated the ability of these hiPSCs to differentiate into the cardiomyocyte lineage. Our results demonstrate that reprogrammed somatic cells could differentiate into cardiomyocytes with apt molecular, structural, and functional properties, like hESCs2,26 and viral hiPSC-CMs.11 In order to increase the number of cardiomyocytes in culture, we used a specific p38 MAP kinase inhibitor, SB203580, that could enhance cardiomyocyte generation, probably by favouring early mesoderm formation.21 We observed that the presence of SB203580 enhanced cardiomyocyte differentiation (with SB203580 vs. without SB203580 for MSnviPSNF3: 20–25 vs. 5–7%). Importantly, cardiomyogenesis from hiPSCs mimics the human developmental pathway, and the resulting cardiomyocytes form a functional cardiac syncytium that responds appropriately to pharmacologically active compounds. We also noted that all our five clones demonstrated differential cardiogenic bias, which is in concordance with hESC biology and may be linked to epigenetic and intrinsic properties of each cell line.20,2730

Temporal gene expression (Figure 2) of our hiPSCs associated with cardiomyogenesis demonstrated a similar trend to that observed in hESC differentiating systems31 and followed the cardiac developmental programme. Down-regulation of Oct-4, the master regulator of pluripotency,20,32 with simultaneous up-regulation of brachyury (marker for primitive streak; data not shown), indicated a trend towards differentiation. Expression of transcriptional factors, GATA4 and Nkx2.5, by day 7–14 suggested commitment towards the cardiac fate. Finally, up-regulated gene expression levels of cardiac-specific structural genes, including sarcomeric-related proteins and ion channels, confirmed terminal differentiation towards cardiomyocytes. These findings may have important implications for future understanding of cardiac developmental pathways and suggest that current differentiation systems and protocols may be subjected to manipulations to trigger known cardio-mesoderm signalling pathways in an attempt to augment cardiomyocyte yield.3,33

Our hiPSC-CMs also demonstrated positive staining for most sarcomeric proteins (Figure 3). However, it was noteworthy that not all cardiomyocytes demonstrated mature cross-striations. While some cardiomyocytes showed Z- and A-bands, suggesting maturity, others did not show clear striations. We believe this may be due to lack of mature myofibril structural organization in de novo cardiomyocytes.34 This may also be deduced from cardiac gene expression levels in hiPSC-CMs and hESC-CMs in comparison with adult human heart (see Supplementary material online, Table S2). Instead, the hiPSC-CMs are phenotypically reminiscent of previously reported 16-week-old fetal hearts.2 Interestingly, despite having irregular sarcomeric structures, these cardiomyocytes expressed functional ion channels coupled with downstream signalling pathways that could be modified by specific cardiac drugs.

Human pluripotent stem-cell-derived cardiomyocytes in various stages of maturity offer the opportunity to create appropriate human test models with high biological relevance in drug safety testing, preventing pharmaceutical attrition.5,35,36 Lack of clinically relevant cell lines and costly animal models further accentuate this problem. In proof-of-concept studies using the MEA mapping technique, we demonstrated that hiPSC-CMs responded appropriately to pharmacological compounds known to affect chronotropy, conduction, repolarization, and QT prolongation (Figures 46). Although the conduction velocity of iPSC-CMs was variable (1.0–2.5 cm/s), it was comparable to hESC-CMs.37 We assume that this variability may be due to the unpredictable geometry (size and shape) of the contracting tissue in a beating cluster, resulting in more heterogeneous measurements in different EBs.23 However, within each EB relatively reproducible measurements were observed. This drawback is counterbalanced by the ability to study the same tissue for prolonged periods in different experimental settings, using each EB as its own control.23 We also observed a significant increase in amplitude following Bay K 8644 treatment (Figure 4E). Bay K 8644, a known inotrophic drug and specific L-type Ca2+ channel agonist, increased Ca2+ influx, resulting in an increase in the amplitude of contractions as well as an increase in ICa.38 Furthermore, the response of our iPS-CMs to verapamil (a blocker of ICaL), tetrodotoxin (a blocker of INa), and E-4031 (a blocker of IKr) was consistent with previous findings of virus-mediated iPSC-derived cardiomyocytes,11,39 confirming pharmacological maturity of the derived cardiomyocytes.

Microelectrode arrays permit stable and long-term FP recordings that enable evaluation of relationships between dose dependency and adverse effects in the drug discovery process.40 Although FP does not resolve individual ion currents that contribute to the action potential, we demonstrated that addition of specific channel blockers altered the FP of our iPSC-CMs, indicating expression of functional Na+, Ca2+ and K+ ion channels, which lends credence to its use in drug testing. Indeed, generation of iPSCs from QT patients, resulting in QT-defective CMs,10 could serve as an immensely valuable in vitro model for drug development and related pharmacogenomics.41,42

The most exciting prospects of virus-free iPSC-derived cardiomyocytes are their potential for personalized cell therapy for myocardial repair. Although, in this present study, we have not addressed the regenerative property of these cardiomyocytes, it is conceivable that these cells could circumvent the problem of immune rejections43,44 when derived autologously. Studies from other groups have demonstrated that pluripotent stem-cell-derived cell types do home and engraft in the host tissue,6 without rejections.45 Nevertheless, up-scaling and purification of cardiomyocytes, in vivo maturation, electrical coupling, teratoma formation, and arrhythmogenicity still need to be addressed before this technology becomes a clinical reality.

In conclusion, we generated virus-free hiPSCs that are capable of differentiating into functional cardiomyocytes with similar efficiency to hESCs, and which recapitulate the cardiac developmental process with well-developed electrical properties. This renewable source of cardiomyocytes may serve as an important in vitro tool for drug screening and possibly regenerative medicine.


This study was supported by National Research Foundation, Singapore grant, NRF2008-CRP001-68.


The authors acknowledge Dr Ralph Bunte, Office of Research, Duke-NUS Graduate Medical School, Singapore for his opinion on tissue pathology in the mouse teratoma assay. G.D. (Principal Research Fellow of NHMRC of Australia) was supported by a grant from the JO and JR Wicking Trust.

Conflict of interest: none declared.


View Abstract