OUP user menu

High-purity enrichment of functional cardiovascular cells from human iPS cells

Bo Lin, Jong Kim, Yanxin Li, Haiying Pan, Xonia Carvajal-Vergara, Guy Salama, Tao Cheng, Yong Li, Cecilia W. Lo, Lei Yang
DOI: http://dx.doi.org/10.1093/cvr/cvs185 327-335 First published online: 6 June 2012


Aims A variety of human inherited heart diseases affect the normal functions of cardiomyocytes (CMs), endothelial cells (ECs), or smooth muscle cells (SMCs). To study human heart disease and generate cardiac cells for basic and translational research, an efficient strategy is needed for production of cardiac lineages from human stem cells. In the present study, a highly reproducible method was developed that can simultaneously enrich a large number of CMs and cardiac SMCs and ECs from human induced pluripotent stem (iPS) cells with high purity.

Methods and results Human multipotent cardiovascular progenitor cells were generated from human iPS cells, followed by selective differentiation of the multipotent cardiovascular progenitor cells into CMs, ECs, and SMCs. With further fluorescence-activated cell sorting, each of the three cardiovascular cell types could be enriched with high purity (>90%). These enriched cardiovascular cells exhibited specific gene expression signatures and normal functions when assayed both in vitro and in vivo. Moreover, CMs purified from iPS cells derived from a patient with LEOPARD syndrome, a disease characterized by cardiac hypertrophy, showed the expected up-regulated expression of genes associated with cardiac hypertrophy.

Conclusions Overall, our technical advance provides the means for generating a renewable resource of pure human cardiovascular cells that can be used to dissect the mechanisms of human inherited heart disease and for the future development of drug and cell therapeutics for heart disease.

  • Induced pluripotent stem cells
  • Embryonic stem cells
  • Differentiation
  • Cardiovascular cells

1. Introduction

Heart disease is the leading cause of death in the world. To study heart disease and repair the diseased heart, efficient production of unlimited cardiovascular cells is required. Recent advances in stem cell biology have established the feasibility of reprogramming human fibroblasts into induced pluripotent stem (iPS) cells.1,2 The iPS cells (iPSCs), like embryonic stem (ES) cells, can differentiate into cells of all three germ layers, including cardiovascular cells,1,37 such as cardiomyocytes (CMs), endothelial cells (ECs), and smooth muscle cells (SMCs).3,5,812 CMs derived from iPSCs of patients with LEOPARD syndrome4 or long QT syndrome6,7 displayed some aspects of disease phenotypes. Thus, iPSCs offer an unprecedented opportunity to generate patient-specific cardiovascular cells for modelling heart disease and for the development of novel therapeutic strategies. However, human ES/iPS cell differentiated cell cultures are highly heterogeneous and have the potential of teratoma formation following transplantation, which has prevented further study of mechanisms of heart disease and direct clinical applications using patient iPSC-derived cell mixtures. There is therefore a major emphasis on the purification of contractile CMs from human ES/iPS cells, such as using Percoll gradient fractionation,13 drug selection of ES cells engineered with CM-promoter-driven selectable markers,14 fluorescence-activated cell sorting (FACS) selection of CMs by mitochondrial fluorescent dye labelling (>90%),15 by surface expression of signal-regulatory protein α (SIRPA; >90%),16 or activated leukocyte cell adhesion molecule (ALCAM; otherwise referred to as CD166; ∼60%).17 In addition to the diseased CMs, the mechanisms involving non-CM cardiac cells, including ECs and SMCs, play an essential role in the pathogenesis of human inherited vascular heart disease; for example, the inherited form of pulmonary arterial hypertension is characterized by the abnormal proliferation of SMCs and ECs.18 Therefore, to understand the complex mechanisms of inherited cardiovascular diseases that affect both muscle and vascular cells, it is of significance simultaneously to enrich CMs, SMCs, and ECs from the same patient-derived iPSC line. However, such a method has not previously been described.

In the present study, we delineate a strategy to enrich ECs, SMCs, and CMs efficiently from human iPS or ES cells with high purity (>90%). The enriched cardiovascular cells exhibited normal phenotypes and functions both in vitro and in vivo. In addition, we observed up-regulation of the cardiac hypertrophy genes in CMs enriched from iPSCs of LEOPARD syndrome,4 a heritable disease associated with cardiac hypertrophy. Together, these results show that our technical advance will significantly enhance future efforts in the modelling and dissecting of the mechanisms of human cardiovascular disease and facilitate development of novel cell-based therapeutics.

2. Methods

2.1 Animals

All of the animal studies conformed to the principles of the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH publication no. 85-23, revised 1996), and all protocols were approved by our Institutional Animal Care and Use Committee. Nonobese diabetic/severe combined immune deficiency (NOD/SCID) mice were anaesthetized in a chamber with the introduction of 100% CO2 for 7–10 min. Euthanasia was accomplished by cervical dislocation.

2.2 Cell culture and differentiation

H1 (NIH Registration Number: 0043) and RUES2 (NIH Registration Number: 0013) human ES cells were obtained from WiCell and Rockefeller University, respectively. L1 iPS and L2 iPS6 human iPSC lines were derived from LEOPARD syndrome patients.4 Y1 iPSCs were reprogrammed from human dermal fibroblasts (HDF-α; Cellapplications, San Diego, CA, USA) as previously described,1 and exhibited characteristics of pluripotent stem cells (see Supplementary material online, Figure S1). All ES and iPSC lines were maintained on mouse embryonic fibroblasts and differentiated as previously described.3 To generate CMs and SMCs, day 6 embryoid bodies (EBs), pre-induced with vascular endothelial growth factor A (VEGF; 10 ng/mL) and dickkopf homologue 1 (DKK1; 150 ng/mL) from day 4 (Figure 1A), were dissociated and the low-KDR, c-kit-negative multipotent cardiovascular progenitors (MCPs) were isolated by FACSAria II cell sorter (BD Biosciences: San Jose, CA, USA) and cultured as a monolayer for an additional 14 days with VEGF (10ng/ml) and DKK1 (150ng/ml) before FACS selection.The CD166+ population was cultured with the basal differentiation medium.3 The CD166 population was cultured with SmGM2 medium (Lonza, Allendale, NJ, USA). To generate ECs and SMCs, day 6 EBs, pre-induced with VEGF (10 ng/mL) and basic fibroblast growth factor (bFGF; 10 ng/mL) from day 4, were dissociated, and the MCPs were isolated and cultured as a monolayer for an additional 14 days with VEGF (20 ng/mL) and bFGF (20 ng/mL; Figure 1A) before FACS selection. CD31+ and CD31 populations were cultured with EGM2 (Lonza) and SmGM2 medium (Lonza), respectively. All cytokines were from R&D Systems, Minneapolis, MN, USA.

Figure 1

Cardiovascular differentiation from human induced pluripotent stem cells (iPSCs). (A) A scheme of the differentiation and enrichment of cardiovascular cells from human iPSCs. (B) The highest ratio of CTNT+ cardiomyocytes (CMs) obtained from human iPSC and embryonic stem (ES) cell lines after optimization of bone morphogenetic protein 4 (BMP4) and activin A. (C) Q-PCR analysis of Y1 iPSC-derived whole embryoid bodies (EBs), not the cell cultures from the isolated multipotent cardiovascular progenitor cells (MCPs), at different stages of cardiovascular differentiation. The Y1 EBs were treated with vascular endothelial growth factor (VEGF; 10 ng/mL) and dickkopf homologue 1 (DKK1; 150 ng/mL) from day 4. The ratio of ΔΔCT was analysed using cyclophinin as the control. The average expression was normalized to T0 undifferentiated iPSCs. Bars represent the SEM of three independent experiments. CM, enriched Y1 iPS-CMs.

2.3 Immunofluorescence

The following antibodies were used: anti-human PECAM1 (CD31) and anti-human VE-cadherin (R&D Systems); anti-cardiac troponin T (CTNT) and anti-smooth muscle actin (SMA; Lab Vision, Kalamazoo, MI, USA); anti-human-connexin 43 and anti-human ALCAM (Chemicon, Billerica, MA, USA); anti-human α-actinin (Sigma) and anti-von Willebrand factor (vWF; DakoCytomation, Carpinteria, CA, USA); anti-human MLC2A and anti-human MLC2V (SYSY, Goettingen, Germany); and anti-SIRPA and mouse IgG1 control (Biolegend, San Diego, CA, USA). All secondary antibodies were from Invitrogen, Grand Island, NY, USA.

2.4 Tube formation on Matrigel

A small volume of cell suspension (0.2 mL, containing 0.5 × 105 to 1.0 × 105 ECs) was placed on top of the Matrigel (BD Biosciences, San Jose, CA, USA) coated coverslip. The cultures were incubated at 37°C with 5% CO2, and observed at 24 h for rearrangement of cells into capillary-like structures. To detect the EC uptake of Dil-Ac-LDL (1,19-dioctadecyl-1,3,3,39,39-tetramethylindocarbocyanine perchlorate acetylated low-density lipoprotein), Dil-Ac-LDL (5 μg/mL; Biomedical Technologies Inc., Stoughton, MA, USA) was added into culture media for 4–6 h, fixed and observed under a microscope.

2.5 Matrigel plug assay

NOD/SCID mice were injected subcutaneously with 0.25 mL ice-cold Matrigel (250 μL) containing VEGF (75 ng/mL), bFGF (300 ng/mL), and 0.5–1.0 × 106 Y1 iPS-ECs. The Matrigel plugs were removed after 4 weeks. Vessel-like structures were observed by intravenous injection of high-molecular-weight fluorescein isothiocyanate–dextran (Sigma) 30 min before plug removal. Matrigel plugs were fixed, sectioned and immunostained.

2.6 SMC contraction assay

Human aortic smooth muscle cells (HASMCs; Invitrogen) and Y1 iPS-SMCs were cultured with SmGM2 medium (Lonza), washed with phosphate-buffered saline, and incubated with a bath solution consisting of 155 mM NaCl, 4.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM glucose, and 10 mM Hepes. Contraction was elicited by incubation with 5 µM carbachol (Calbiochem, Billerica, MA, USA). Bright-field images and videos were obtained by using a Leica inverted microscope.

2.7 Optical mapping

Intracellular Ca2+ transients were optically recorded with a high spatiotemporal resolution CMOS camera (Scimedia, Ultima, Costa Mesa, CA, USA; up to 1.5 µm × 5 µm pixel resolution and up to 200 frames per second temporal resolution). Briefly, cells were treated with Tyrode' solution containing a Ca2+ indicator (Rhod-2 AM; 10 µg/mL; Molecular Probes) at 37°C for 15 min. A CMOS camera was mounted on an upright microscope (Olympus, BX61WI) to acquire fluorescent images of intracellular Ca2+ transients. The cells were illuminated with a 520 ± 30 nm excitation beam, and the fluorescence from the preparation was collected via a microscope objective lens (×10 or ×40), passed through a dichroic mirror (565 nm long-pass) and an emission filter (580 ± 30 nm band-pass) to focus on the CMOS camera to measure intracellular Ca2+ transients through the Rhod-2 fluorescence. Data were analysed using custom-made software (IDL, http://www.exelisvis.com/ProductsServices/IDL.aspx).

2.8 Quantitative PCR analysis

Quantitative PCR (Q-PCR) was performed on a 7900HT Fast Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) with Fast SYBR Green Master Mix (Applied Biosystems). The results were analysed using Excel, normalized to cyclophinin gene expression, and compared with the undifferentiated iPSCs. Primer sequences are described in Supplementary material online, Table S1.

2.9 Illumina mRNA deep sequencing

Total RNAs from the human RUES2, MCPs, and the enriched CMS, SMCs, and ECs were extracted using the RNeasy Kit (Qiagen, Valencia, CA, USA) and subjected to Illumina mRNA deep sequencing using a service from LC Sciences (Houston, TX, USA). The sequence results were obtained as the FPKM (fragment per kilobase of exons per million reads) for each transcript.

2.10 Heart tissue

Foetal heart tissue was obtained from a newborn patient at Children's Hospital, University of Pittsburgh Medical Center. The protocol was approved by the ethics committee, the Institutional Review Board on the use of human tissues (Belmont Report, 1979) at University of Pittsburgh, and conforms to the principles outlined in the Declaration of Helsinki.

2.11 Data analysis

Data are shown as means ± SEM of three independent experiments. Statistical analysis was performed with Student's paired t-test.

3. Results

3.1 Cardiovascular differentiation from human pluripotent stem cells

Here we described an experimental strategy to induce and enrich CMs, SMCs, and ECs from human iPSCs in serum-free conditions (Figure 1A). This cardiac induction protocol utilizes the known role of different growth factors in cardiovascular development to manipulate the differentiation and specification of the cardiovascular cells from human iPSCs and was an adaptation of our previous cardiac differentiation protocol developed for human ES cells.3 We found that bone morphogenetic protein 4 (BMP4) and activin A played a key role in determining the final ratio of iPSC-derived CMs (iPS-CMs). To ensure a high ratio of CM conversion, the concentrations of those factors were optimized for each iPS and ES cell line (see Supplementary material online, Figure S2 and Movie S1). Figure 1B indicates the ratio of CMs achieved following optimization. We found that a higher dose of BMP4 was needed on day 0 to initiate cardiac differentiation with human iPSCs in comparison to human ES cells (see Supplementary material online, Figure S2b). In addition, to support CM growth, a low dose of BMP4 (1 ng/mL) was added from day 12, consistent with the finding that the BMP signal is required for CM proliferation.19 Most importantly, we modified the previous protocol (see Supplementary material online, Figure S2a)3 by adding bFGF, instead of DKK1, into the day 4 medium, which efficiently induced the early specification of MCPs to the EC/SMC fate, instead of the CM/SMC fate (Figure 1A).

Q-PCR analysis of the developing whole embryoid bodies, which were induced with the protocol shown in Supplementary material online, Figure S2a, revealed dynamic changes in gene expression indicative of developmental progression in cardiovascular differentiation (Figure 1C). Kinase insert domain receptor (KDR) expression was increased around day 8 and decreased to levels similar to that of iPSCs after 40 days. NK2 homeobox 5 (NKX2.5),20 an early cardiac transcription factor, was up-regulated from day 6, indicating the initiation of cardiovascular development. CTNT was detected from day 8 and persisted until day 50. Myosin light chain 2A (MLC2A)21 and iroquois homeobox 4 (IRX4),22 markers of atrial and ventricular CMs respectively, reached the highest expression level from day 24 and remained highly expressed throughout, indicating the emergence of atrial/ventricular CM subtypes. Two SMC markers, SMA and calponin, were detected from day 8 onwards. The increasing level of calponin from day 15 is likely to reflect the segregation of CMs and SMCs after 2 weeks of differentiation. CD31, an EC marker, was detected from day 8. Expression of CD166, a CM surface marker,17 started from day 8 and was detectable up to day 50.

3.2 Isolation and specification of MCPs from human iPSCs

An MCP population was identified from a day 6 human ES-cell-derived mesoderm population characterized by surface expression of KDR and c-kit.3 FACS analysis with day 6 human iPSC-derived EBs exhibited an identical KDR profile and a similar MCP population (Figure 2A, Supplementary material online, Figure S3). Substantial CMs could only be generated from the iPSC-derived low-KDR, c-kit-negative MCPs, not from the high-KDR, high-c-kit or KDR-negative, high-c-kit population (see Supplementary material online, Movie S2), the same as described before.3 Given that the day 6 low-KDR, c-kit-negative cells are a heterogeneous population containing MCPs at dynamic stages of heart development, the early addition of DKK1 from day 4 predominantly pre-induced MCP specification towards the CM/SMC fate (Figure 2A-I), while the presence of bFGF from day 4 biased the specification of MCPs into ECs/SMCs (Figure 2A-II). Interestingly, the presence of DKK1 or bFGF from day 4 did not alter the KDR/c-kit expression patterning of day 6 iPSC cultures. Therefore, using these different treatments on day 4, MCP differentiation could be manipulated towards either CMs (>40%) and SMCs (>50%), or ECs (>18%) and SMCs (>80%). Figure 2B shows the immunostaining of human iPS-CMs (CTNT, α-actinin, and gap junction protein CX43), SMCs (SMA), and ECs (VE-CAD and vWF). Similar to the finding from human ES cells,3 iPSC-derived CMs and SMCs both exhibited SMA expression, consistent with their common developmental origin. Q-PCR analysis confirmed the expression of cardiovascular cell-specific genes in the Y1-MCP-derived adherent cell cultures and no expression of endoderm or ectoderm marker genes (see Supplementary material online, Figure S4).

Figure 2

Isolation and specification of the iPSC-derived MCPs. (A) Y1 EBs were treated with VEGF (10 ng/mL) and DKK1 (150 ng/mL; panel I) or VEGF (10 ng/mL) and basic fibroblast growth factor (bFGF; 10 ng/mL; panel II) from day 4, followed by isolation of the day 6 MCP population and further specification of MCPs with VEGF (10 ng/mL) and DKK1 (150 ng/mL; panel I) or VEGF (20 ng/mL) and bFGF (20 ng/mL; panel II) for 14 days, respectively. The proportions of CTNT+, SMA+, and CD31+ cells generated from the above two sets of conditions were calculated by FACS analysis. (B) Immunostaining of the MCP-derived populations. The purple arrows indicate the smooth muscle cells (SMCs) and the white arrows indicate the merged CTNT+, SMA+ CMs. Magnification, ×400 (α-actinin, CX43, VE-CAD), or ×200 (SMA). Bars represent the SEM of three independent experiments. Scale bars represent 20 µm.

3.3 Identification of CD166 as a CM surface marker

CD166 (ALCAM) has been found to be transiently expressed in the murine embryonic heart tube from embryonic day 8.25 to 10.5 and on human ES-cell-derived CMs.17,23 Expression of CD166 in human heart has not been described. We assessed CD166 expression in human heart tissue and found that CD166 is expressed on the CMs, but not on the SMCs or ECs (Figure 3A). Consistent with the findings for the human heart tissue, human Y1 iPSC-derived CMs also were positive for CD166 immunostaining, but CD166 immunostaining was not seen in the Y1 SMCs and Y1 ECs (Figure 3B). The distribution of CD166 on CMs was observed along the cell surface at regions of cell–cell contact or sporadically on the surface. This same pattern was observed both in CMs of the heart tissue and in the iPS-CMs (Figure 3, red and green arrows). Co-expression of CD166 and SIRPA was observed on the human iPS-CMs (Figure 3C). In addition, we randomly picked out five views of CD166/SIRPA double-immunostained iPS-CMs, and found that nearly 96% CD166+ cells were also SIRPA+, and nearly 98% SIRPA+ cells were CD166+.

Figure 3

Identification of CD166 as a human CM surface marker. (A) Immunostaining analysis of the human foetal heart tissue. Magnification, ×100 (panels I, III, and IV), or ×450 (panel II). (B) Immunostaining analysis of the Y1 iPSC-derived cardiovascular cells. Red arrows indicate CD166 expresses along the cell surface at regions of cell-cell contact. Green arrows indicate CD166 expresses sporadically on the surface. Yellow arrows indicate the CD166+, CTNT+, SMA+ CMs. SMCs are SMA+, CTNT, CD166. Magnification, ×600 (panel I), or ×200 (panels II, III, and IV). (C) Co-immunostaining analysis of the expression of CD166 and SIRPA on human iPSC-CMs. White arrow indicates a CD166, SIRPA cell. Magnification ×400. Scale bars represent 20 µm.

3.4 Cardiovascular cell enrichment

We used the conditions shown in Figure 2A-I to induce CMs and SMCs from Y1 MCPs. The resulting day 20 cell cultures, containing 40% CMs, 50% SMCs, and very few ECs, were dissociated into single cells, and the CD166+ and CD166 populations were isolated (Figure 4A). The CD166+ population contained over 90% CTNT+ CMs. Over 93% CD166+ cells were SMA+ cells, because SMA is expressed on early CMs and SMCs. In the CD166 population, we found 10% CMs and over 90% SMA+ cells, indicating that at least 80% of the CD166 cells were SMCs (Figure 4A, Supplementary material online, Movie S3 and S4). The conditions shown in Figure 2A-II were used to induce EC/SMC specification, which generated over 20% ECs and over 70% SMCs, with a very low ratio of CMs after additional 14 days in culture. Both the CD31+ and CD31 populations were isolated (Figure 4B). The CD31+ population contained over 96% CD31+ ECs, while the CD31 population contained over 95% SMCs. A similar method was used to enrich cardiovascular cells from human H1 ES and L2 iPS6 cells (see Supplementary material online, Figure S5). Figure 4C illustrates immunostaining of day 20 cells before and after enrichment, indicating the high purity of each enriched cardiovascular cell type. The Illumina mRNA deep-sequencing profile confirmed the signature gene expression in each enriched cell type (Figure 4D). Immunostaining with antibodies to MLC2A and MLC2V and optical mapping of Y1 EBs further showed that the enriched CMs contained both atrial and ventricular subtypes (Figure 4E, Supplementary material online, Figure S6).

Figure 4

Enrichment of cardiovascular cells from human iPSCs. (A) Enrichment of CMs and SMCs from day 20 monolayer cells with the selection of CD166 expression. Frequencies of CMs and SMCs within the CD166+ and CD166 populations were analysed by FACS. Mouse IgG1 isotype was used as the control. (B) Enrichment of ECs and SMCs from day 20 monolayer cells with the selection of CD31 expression. Frequencies of ECs and SMCs within the CD31+ and CD31 populations were analysed by FACS. (C) Immunostaining analysis of the cardiovascular cells before and after FACS enrichment. (D) A heat map showing the relative expressions of marker genes from human RUES2 cells and the enriched CMs, SMCs, and ECs by using the Illumina mRNA deep sequencing. The values of the gene expressions are shown as the normalized and log2 transformed FPKM (fragment per kilobase of exons per million reads). (E) Immunostaining analysis of the CM subtypes of the enriched Y1 CMs. Magnification ×400. Scale bars represent 20 µm.

3.5 Functional analysis of human iPSC-derived cardiovascular cells

The enriched Y1 ECs formed a tube-like capillary structure on Matrigel within 24 h (Figure 5A). The tube-forming cells expressed CD31 and displayed LDL endocytosis (Figure 5A-I, II), confirming that iPS-derived ECs exhibited a similar phenotype and function to that of other established EC lines.24 Matrigel assay further demonstrated the in vivo function of enriched ECs. Histological examination of the Matrigel plug revealed vessel-like structures (Figure 5A-IV) that were observed in other established EC lines.24

Figure 5

Functional analysis of human iPSC-derived cardiovascular cells. (A) CD31 immunostaining and Dil-AC-LDL uptake of enriched Y1 iPS-ECs cultured on Matrigel-coated coverslips (panels I–III) and immunostaining of a section of the Matrigel plug formed by Y1 iPS-ECs (panel IV). Fluorescein isothiocyanate–dextran was injected intravenously before the remove of the Matrigel plug. (B) Contractions of human aortic smooth muscle cells (HASMCs) and iPS-MSCs in response to carbachol. Images were taken before and 30 min after carbachol loading. Red arrows indicate the shrinkage of SMCs. Bars represent the SEM of three independent experiments; *P =0.27. (C) Optical measurement of Ca2+ handling in HASMCs and human iPS-SMCs during contraction–relaxation induced by membrane depolarization with a rapid transition of extracellular [K+] from 4 to ∼60 mM. The cytosolic free Ca2+ level rose immediately after the elevation in extracellular [K+]. A black arrow indicates the time of extracellular [K+] elevation. (D) Simultaneous measurement of membrane potential (Vm) and intracellular Ca2+ transients in a beating monolayer of human iPS-CMs. The rise of Vm transients preceded the rise of intracellular Ca2+ transients, and action potential repolarization was terminated before complete intracellular Ca2+ reuptake. Yellow arrows indicate the beating iPS-CMs. (E) Action potential prolongation and initiation of early after-depolarization as a result of IKr inhibition with 500 nM E4031. The action potential duration was gradually prolonged after addition of E4031 and, eventually, an early after-depolarization was initiated. A black arrow indicates the early after-depolarization. Scale bars = 200 µm.

Next, we investigated whether human iPSC-derived SMCs (iPS-SMCs) had functional properties similar to those of control HASMCs. Carbachol (5 µM) induced significant contraction of both human iPS-SMCs and HASMCs. We also found that a similar proportion of iPS-SMCs (39%) responded to carbachol when compared with HASMCs (35%; P = 0.27; Figure 5B, Supplementary material online, Movie S5). In addition, intracellular Ca2+ transients were recorded from both human iPS-SMCs and HASMCs. Cells were depolarized by increasing the K+ concentration in the culture medium from 4 to 60 mM to initiate contraction, while intracellular Ca2+ transients were recorded during cycles of contraction and relaxation (Figure 5C, Supplementary material online, Figure S7). Intracellular Ca2+ transients detected in the human iPS-SMCs were similar to those observed in HASMCs.12

Finally, human iPS-CMs were stained with a voltage-sensitive dye (PGH1; 2 µL of 5 mg/mL in dimethyl sulfoxide) and a Ca2+ indicator (Rhod 2-AM; 2 µL of 5 mg/mL in dimethyl sulfoxide) to allow simultaneous measurement of membrane potential (Vm) and intracellular Ca2+ transients (Figure 5D). Excitation–contraction coupling (e.g. time delay between Vm and intracellular Ca2+ upstrokes) in spontaneously beating human iPS-CMs was consistent with Vm–intracellular Ca2+ coupling of typical CMs.25,26 E403127 is an inhibitor of the rapid rectifier outward K+ channel, IKr. With the addition of E4031 (500 nM), the action potential duration of the human iPS-CMs was markedly prolonged, and an early after-depolarization was induced (Figure 5E). These observations are similar to those obtained with inhibition of IKr of CMs isolated from the heart.2527 Both action potential dynamics and intracellular Ca2+ handling in the human iPS-CMs are functionally similar to that of primary CMs and suggest the potential of iPSC-derived CMs for drug toxicity testing and patient-specific drug screening.

3.6 Studying mechanisms of human inherited heart disease with patient-derived iPS-CMs

A previous study demonstrated the feasibility in generating hypertrophic CMs from LEOPARD syndrome iPSCs.4 CMs derived from L2-iPS6, a Leopard syndrome iPSC line, exhibited larger cell size compared with CMs from the normal H1 ES cells.4 However, the underlying molecular mechanism of LEOPARD syndrome-associated cardiac hypertrophy was not examined in that study because EBs of L2 and H1 cells contained different CMs ratios. By utilizing our cardiac enrichment protocol, a pool of highly purified CMs was obtained from the L2 iPS6 and the H1 cells. Q-PCR analysis detected the increased expression of cardiac hypertrophy-associated genes, natriuretic peptide type A (ANF), B-type natriuretic peptide (BNP), myosin heavy chain (MHC), phospholamban (PLN), and transforming growth factor β (TGF-β), as well as the down-regulation of sarcoplasmic reticulum Ca2+-ATPase (SERCA2A) in the L2 CMs in comparison with the expression of those genes in the H1 CMs (Figure 6).

Figure 6

Studying the mechanism of cardiac hypertrophy associated with LEOPARD syndrome. Quantitative RT-PCR analysis of cardiac hypertrophy-associated gene expression within CMs from L2 iPS6 cells and H1 ES cells. The ratio of ΔΔCT was analysed using cyclophinin as the control. Bars represent the SEM of three independent experiments; *P <0.05, **P <0.01.

4. Discussion

CMs, SMCs, and ECs have been derived together with other cell types from human ES/iPS cells. Although the heterogeneous population has been utilized for human heart disease modelling,4,6,7 drug toxicity testing,28 and even cell-based therapy in animal models,29 lack of the purified cardiovascular cells is a major obstacle for studying the mechanisms of heart disease, as well as for the potential translational applications. Recent progress in CM enrichment from human ES/iPSCs achieved CMs with high purity (>90%) by FACS selection of mitochondrial fluorescent dye-labelled CMs15 or SIRPA-expressing CMs.16 Considering the requirement of clinical-grade CMs for the future heart disease therapy, the CMs enriched by surface marker selection would provide a safer cell resource than the dye-stained CMs. In the present study, a similar CM purity (>90%) was achieved with selection of CD166 expression compared with that from the SIRPA selection.16 Both SIRPA and CD166 expressions are highly overlapping on human iPSC-derived CMs. Interestingly, 8–17% CMs existed in the SIRPA-negative population,16 which was close to our finding that the CD166 population contained approximately 12% CMs. These results indicate the similar dynamic expressions of both CD166 and SIRPA on human iPSC-derived CMs. SIRPA is human specific, and its function in cardiogenesis is not clear.16 CD166 is broadly expressed in chickens,30 Xenopus,31 mice,23 and humans, and plays a reserved role in cardiogenesis.30 Murakami et al.32 described the CD166 expression on embryonic day 8.5 murine yolk sac cell-derived CMs and ECs. Ohneda et al.33 observed CD166 expression on both mouse embryonic ECs and adult myocardium, but not on adult endocardium, which is consistent with our observation of CD166 expression on human CMs, but not on human ECs, suggesting the dynamic expression of CD166 on different cell types during embryonic development.

Although a previous report described CM enrichment from human ES cells with selection of CD166 expression,18 the purity of enriched CMs in that study was lower (∼60%) than ours (>90%). A possible reason is that the lack of a well-defined CM differentiation protocol in that study and the broad expression of CD166 would lead to the enrichment of CD166+ non-CMs from ES cell-derived cell mixtures. The technical advantage of our work is that we have an established system to induce and isolate MCPs, to specify MCPs towards only three cardiovascular cell types, which minimizes the possible contamination from other CD166+ non-CM cells and ensures the high purity following enrichment. Both SMC and EC formations are coupled with organogenesis, such as in haematopoiesis,34 cardiogenesis,3 and hepatogenesis,35 suggesting the existence of organ-specific SMCs and ECs. Here we described a method to enrich cardiac-specific SMCs and ECs from human ES and iPS cell-derived MCPs, which will facilitate studies of human vascular heart diseases, as well as the development of vascular cell-targeted drugs.

In this study, for the first time, detailed molecular profiling was conducted to elucidate underlying disease mechanisms using patient derived iPS-CMs. CMs derived from patient-specific iPSCs have been utilized to model cardiac hypertrophy in LEOPARD syndrome.4 However, the disease mechanism was not described in that study. With our established CM enrichment protocol, we found that L2 iPS6-CMs showed altered expression of various hypertrophy-related genes when compared with the H1 CMs. Significantly, these gene expression changes were similar to those seen in a mouse LEOPARD syndrome model containing a LEOPARD syndrome-causing PTPN11 knock-in allele.36 In addition, we identified higher TGF-β37 expression in the L2 iPS6-CMs. The consistent gene expression pattern from the LEOPARD syndrome patient and the mouse LEOPARD syndrome model suggested that a similar regulatory network controls cardiac hypertrophy in both human and mouse LEOPARD syndrome CMs.

Overall, we established a system to generate and enrich major cardiovascular cell types from human iPS/ES cells with high efficiency and high purity. Most importantly, this was achieved without any genetic modification, drug selection, or virus integration. Our method will accelerate future studies seeking to understand human heart development and human heart disease, and will provide a renewable cell resource for cell replacement therapy, as well as for drug safety testing and drug screening.


This work was supported by American Heart Association (AHA) 2010 SDG Grant (11SDG5580002) and University of Pittsburgh Start-up support to L.Y. This work was partially supported by grants from the Ministry of Science and Technology of China 2011ZX09102-010-04, 2010CB945204 and 2010DFB30270 to T.C.


We thank Jenny Jiao from Stem Cell Core, University of Pittsburgh for Y1 iPSC characterization and for providing the human neuron cell cDNA; Ira Fox for the gift of the human liver cDNA; Ihor Lemischka for providing the LS-iPSCs; and Michael Tsang and Cecilia W. Lo for critical reading of the manuscript.

Conflict of interest: none declared.


View Abstract