Scalable Production of Cardiomyocytes Derivedfrom c-Myc Free Induced Pluripotent Stem Cells

Limor Zwi-Dantsis, B.Sc.,1,2 Itzhak Mizrahi, Ph.D.,1 Gil Arbel, M.Sc.,1

Amira Gepstein, Ph.D.,1 and Lior Gepstein, M.D., Ph.D.1

Cardiomyocytes derived from induced pluripotent stem (iPS) cells hold great promise for basic and translationalcardiovascular research. For the successful implementation of this unique technology, however, it is essential toestablish efficient, reproducible, and safe strategies to produce cardiomyocytes in a scalable manner. The aim ofthe current study was to establish scalable bioprocess that allows direct embryoid bodies formation for thedifferentiation of murine iPS cells (generated without the oncogene c-Myc) into cardiomyocytes. The cardio-myocytes’ structural, molecular, and functional properties were then compared to ones derived by the well-established static culture system. Similar gene expression patterns were observed in both differentiation systemswith the sequential expression of mesoderm markers, cardiac transcription factors, and cardiomyocyte structuralgenes. Cells in the contracting embryoid bodies were stained positively for cardiac troponin-I, sarcomeric a-actinin, cardiac troponin-T, and connexin-43. Electrophysiological measurements using multielectrode arrayrecordings demonstrated that the bioreactor-derived cardiomyocytes were functionally similar to static derivedcardiomyocytes and responded appropriately to different drugs, including adrenergic and muscarinic agonists(isoproterenol and carbamylcholine, respectively) and the gap junction uncoupler heptanol. Our study describes,for the first time, a strategy for scalable differentiation of c-Myc-free iPS cells into cardiomyocytes with theappropriate molecular, structural, and functional properties. The result of this study should have importantimplications for several cardiovascular research areas and specifically for the emerging field of regenerativemedicine.

Introduction

The ability of the heart to regenerate is limited andtherefore any significant heart cell loss, for example,

during myocardial infarction, is irreversible and may lead tothe development of progressive heart failure. An experi-mental approach for the treatment of heart failure may be toutilize the emerging technologies of stem cells, cell therapy,and tissue engineering to repopulate the injured heart withnew myocytes.1 This exciting cell-based cardiac repair strat-egy has been hampered, however, by the lack of cell sourcesfor cardiomyocytes.

In 2006, Yamanaka and Takahashi2 were able to repro-gram mouse somatic cells by ectopic retroviral expression of4 transcription factors (OCT4, SOX2, c-Myc, and KLF4),yielding cells with characteristics similar to those of mouseembryonic stem cells (ESCs). These reprogrammed cell linesgenerated were termed induced pluripotent stem (iPS) cells.Application of this approach to human somatic cells fol-lowed shortly, yielding human iPS cell lines with character-istics similar to human embryonic stem cells (hESCs).3,4 The

iPS cells were able to differentiate into cell derivatives of allthree germ layers, including cardiomyocytes.5–8

The iPS technology offers unique opportunities for severalcardiovascular research areas (human disease modeling, in-dividualized drug testing, and in the emerging field of re-generative medicine) by providing a potential source forpatient-derived autologous cardiomyocytes. However, thispromising approach is halted by the limited supply of cells,which greatly depends on the availability of a controlledlarge-scale bioprocess.

In the current study we aimed to establish an efficient,reproducible, and safe strategy to produce cardiomyocytesfrom iPS cells in a scalable manner. Specifically, our objec-tives were threefold: (1) to differentiate mouse iPS cells(generated without the oncogene c-Myc) to functional car-diomyocytes; (2) to establish a strategy for scalable produc-tion of iPS derived cardiomyocytes using a spinner flasks(SFs) system; and (3) to characterize the structural, molecularand functional properties of the differentiating cardiomyocytesin the SF system as compared to the well-established statichanging drops (HD) culture method.

1Sohnis Family Research Laboratory for Cardiac Electrophysiology and Regenerative Medicine (Bruce Rappaport Faculty of Medicine) and2Biotechnology Interdisciplinary Unit, Technion – Israel Institute of Technology, Haifa, Israel.

TISSUE ENGINEERING: Part AVolume 17, Numbers 7 and 8, 2011ª Mary Ann Liebert, Inc.DOI: 10.1089/ten.tea.2010.0235

1027

Materials and Methods

Culture of iPS cells

The mouse iPS cell line utilized (TTF256H18, kindly pro-vided by Shinya Yamanaka) was generated by retroviraltransduction of fibroblasts with OCT4, SOX2, and KLF4 andshown to display all the defining parameters of iPS cells.9

Undifferentiated colonies were cultured on a mitotically in-activated mouse embryonic fibroblasts (MEF) feeder layer, aspreviously described.10 The culture medium consisted of85% Dulbecco’s modified Eagle’s medium (Invitrogen) sup-plemented with 15% fetal calf serum (Hyclone), 0.1% leu-kemia inhibitory factor (Millipore), 1 mM L-glutamine,0.1 mM mercaptoethanol, and 1% nonessential amino acidstock (all from Invitrogen).

Differentiation methods of iPS cells

To induce differentiation, the standard HD method toderive embryoid bodies (EBs) was used. Undifferentiated iPScells were dissociated with 0.25% trypsin–ethylenediamine-tetraacetic acid (EDTA) (Biological Industries) and sus-pended in differentiation medium composed of Iscove’smodified Dulbecco’s medium (Biological Industries), 20%fetal calf serum, 1 mM L-glutamine, 0.1 mM mercaptoetha-nol, 1%-nonessential amino acid stock (Invitrogen), and 1%penicillin–streptomycin (Biological Industries). The cellswere aggregated in HD composed of either 500 or 1000 cellsin 20 or 30mL of the differentiation medium. At 3 days ofdifferentiation, EBs were transferred from the HD into 6-cmbacteriological Petri dishes. At day 7, the EBs were plated on0.1%-gelatin-coated culture dishes and examined for theappearance of spontaneous beating.

For the SFs suspensions, iPS cell colonies were dissociatedwith 0.25% trypsin–EDTA and suspended in the differenti-ation medium (1�105 cells/mL). One hundred milliliters ofcell suspension was then transferred into 250 mL SFsequipped with bulb-shaped glass stirrer (Integra Bios-ciences). The SF was stirred at 60 rpm and a half-mediumexchange was performed every 3 days. After 7 days, the EBswere plated on 0.1%-gelatin-coated dishes.

EB morphology, cell concentration, and kinetics

For EB morphological analysis, 5 mL of the EB-containingsuspension was sampled, transferred into culture dishes, andanalyzed microscopically. The average diameter of randomlyselected 20 EBs in each system was measured. To determineEB cell concentration, 3 mL of homogeneously suspendedEBs from the SF or three Petri plates containing the HD werecollected. To obtain single cells, the EBs were mechanicaldissociated and then enzymatically dispersed using 0.25%trypsin–EDTA. Cell viability was determined by Trypan-blue staining (Biological Industries). Growth curves were fitto exponential growth model and the cell doubling time wascalculated by logarithmic integration as indicated by theslope of the line on the accelerated growth phase of eachcurve, as described by Nie et al.11

Apoptosis detection

In situ cell death detection kit (Fluorescein, Roche) andconfocal microscopy was used to detect apoptosis in 8 daysEBs derived by both methods (n¼ 10).

Flow cytometry

EBs were dissociated by mechanical disruption and incu-bation with type-I collagenase (Sigma) and trypsin–EDTA.Cell viability was tested by Fixable-Viability Dye eFluor�450(eBioscience) for 30 min at 48C. Cells were then washed withIscove’s modified Dulbecco’s medium (IMDM) with 5% fetalbovine serum and fixed with 4% paraformaldehyde. Next, thecells were permeabilized for 10 min with 0.5% saponin inphosphate-buffered saline, and incubated with 1:100 dilution ofanti-cardiac troponin T (cTnT) monoclonal antibodies (Thermo)for 30 min. Negative control cells were incubated with isotype-IgG control antibodies (1:100, Thermo). After washing withphosphate-buffered saline, secondary fluorescein isothiocyanate(FICT)-conjugated anti-mouse IgG antibodies were added(1:100). Cells were analyzed with a flow cytometer (CyAn ADP;Beckman Coulter), using Summit software (Beckman Coulter).

Immunostaining studies

Undifferentiated iPS colonies, whole EBs, and single cellsfrom beating EBs were fixed with 4%-paraformaldehyde,permeabilized with 1%-Triton-X-100 (Sigma), blocked with5%-horse serum, and incubated overnight at 48C with primaryantibodies targeting: Oct-4 at 1:150 dilution, connexin-43(Cx43, 1:10) (all from Santa Cruz), Nanog (1:60, Abcam), SSEA-1(1:100, R&D), cardiac-troponin-I (cTnI, 1:200, Chemicon),cTnT (1:150 R&D), and sarcomeric-a-actinin (1:200, Sigma).Preparations were incubated with secondary antibodies( Jackson) at 1:200 dilution for 1 h. Nuclei were counterstainedwith DAPI (1:500, Sigma). Preparations were examined using alaser scanning confocal microscope (Zeiss LSM-510-PASCAL).

Reverse-transcription polymerase chain reactionand quantitative real-time polymerase chain reaction

Undifferentiated iPS (day-0) and differentiating EBs (3, 7,11, and 14 days) were frozen in liquid nitrogen. RNA wasisolated using the RNeasy-plus mini kit (Qiagen-Gm). Reversetranscription into cDNA was conducted using Superscript-IIfirst-strand synthesis system (Invitrogen). The polymerasechain reaction (PCR)-related primers and reaction conditionsare detailed in Table 1. Briefly, each reverse-transcription PCR(RT-PCR) included the following PCR program: 938C for3 min, 938C for 30 s, 608C for 30 s, and 728C for 30 s. cDNAwas synthesized using Verso RT PCR Kit (Thermo Scientific).About 1.6 ng of cDNA was used.

SYBR-Green quantitative real-time PCR studies wereperformed in triplicates using Absolute blue QPCR-SYBR-mix-ROX (Thermo Fisher). Samples were cycled for 40 timesusing an ABI-7000 Sequence Detector (Applied Biosystems) asfollows: 508C for 2 min, 958C for 15 min, followed by 40 cyclesof 958C for 15 s, 608C for 30 s, and 728C for 30 s. Thresholdcycle (CT) was calculated under default settings for RT se-quence detection software (Applied Biosystems). Expressionlevels of target genes were normalized to b-tubulin transcriptlevels. Data are provided as mean� standard error of mean ofnormalized gene expression levels from three differentiationexperiments.

Multielectrode array recordings

To characterize the electrophysiological properties of theiPS-derived cardiomyocytes, a multielectrode array (MEA)

1028 ZWI-DANTSIS ET AL.

recording system (Multichannel Systems)8,12,13 was used.The contracting EBs were microdissected and plated onfibronectin-coated MEA plates. The MEA system allows si-multaneous recordings from 60 titanium-nitride-coated goldelectrodes (30mm) at a high spatial (200 mm) and temporal (15kHz) resolutions.

To assess the effects of different drugs on the electro-physiological properties, 10mL of tested drug stock solutionwas added to 1 mL of the culture medium. The tested drugsincluded isoproterenol hydrochloride, carbamylcholine, andheptanol (all from Sigma). Extracellular recording wereperformed for 60 s at baseline and at 5 min following drugapplication.

Statistical analysis

Results are given as mean� standard error of mean. Datawere analyzed using unpaired (for comparison of EB sizesbetween the two culture methods) or paired (for the MEAdrug studies) Student’s t-test. Values with p< 0.05 wereconsidered statistically significant.

Results

Differentiation of c-Myc-free iPS cellsusing the HD EB method

In this study we utilized a murine iPS cell line generatedby the retroviral delivery of three factors (OCT4, SOX2, andKLF4) without the oncogene c-Myc.9 The undifferentiatediPS cells were grown on the MEF feeder layer and formedsimilar colonies to mouse ESC (Fig. 1A). These undifferen-tiated colonies were positively stained for the pluripotentmarkers SSEA-1, NANOG, and OCT4 (Fig. 1B–D).

Differentiation of the c-Myc-free iPS cells into cardio-myocytes was initially performed using the well-establishedHD method, traditionally used for differentiation of mouseESC. The iPS cells were cultivated in 20 mL drops (eachdrop containing 500 cells) for 2 days. During this period,three-dimensional differentiating cell aggregates (EBs)were formed (Fig. 2A). Subsequently, the EBs weretransferred to bacterial-grade dishes for another five daysin suspension, where they formed larger EBs (Fig. 2B). TheEBs were then plated on gelatin-covered plates. Sponta-neously beating areas appeared in EBs beginning fromdays 2 to 4 postplating.

Scalable production of cardiomyocytesfrom iPS using SFs

Next, we designed a strategy for scaling-up of the differ-entiation procedure using stirred-suspension cultures. Un-differentiated iPS cells were removed from the MEF feeder,trypsinized into single cells, and cultivated in SFs at a con-centration of 1�105 cells/mL (in a final volume of 100 mL).The cultures were stirred at 60 revolutions per minute (rpm)for 7 days and then plated on gelatin-covered plates.

In contrast to previous studies in which ESC cells seededinto SFs quickly agglomerate into single pellet,14,15 by em-ploying SFs equipped with bulb shaped impellers, homo-geneous EBs were formed 24–48 h after seeding. These EBscontinued to grow during the stirred-suspension culture (Fig.2E, F). To provide some quantitative information regardingthe efficiency of this process, we counted the number ofgenerated EBs. We found that the average number of EBsthat can be obtained in a single SF is 1016� 508 (n¼ 4) for aninitial dose of 1�107 cells.

In a similar manner to the HD method, following platingof the EBs on gelatin-coated plates most of the EBs did notstay in their round three-dimensional shape, but ratherspread out on the culture plate. Beating areas appearedmainly in the outgrowths of the aggregates but also at thecenter of the mass. The percentage of plated EBs containingbeating areas continued to increase with time until reachinga maximal number at 10–14 days postplating.

EB morphology, growth kinetics, and apoptosis

When comparing the size of the EBs generated in bothmethods, the SF technique resulted in significantly larger EBs(394� 66mm in SF vs. 280� 70mm in HD at 7 days of culti-vation, Fig. 2G) when the initial cell dosage of the drop (HDmethod) was 500 cells/20 mL. However, increasing the initialcell number and volume of the drop (1000 cells/30 mL) re-sulted in the derivation of aggregates with comparable sizesin both methods (Fig. 2C, D, H). Thus, at 3 days of differ-entiation the mean EB diameters were 246� 24 mm and239� 40mm in the HD and SF, respectively, and after 7 days,394� 66mm and 421� 65 mm, respectively.

We next compared the viable cell concentrations in theaggregates and their doubling rates in both the static and thestirred systems. EB samples were trypsinized into single cellsand the viable cell concentration was evaluated at days 2, 4,

Table 1. List of Primers Used in Reverse-Transcription Polymerase Chain Reaction

and Quantitative Polymerase Chain Reaction

Name Forward Reverse

OCT4 GTGGAGTCTGGAGACCATGTTTC TGGCGCCGGTTACAGAACNANOG AGCAGAAGATGCGGACTGTGT TCAGGTTCAGAATGGAGGAGAGTTNKX2-5 CTTCAAGCAACAGCGGTACCT CGCTGTCGCTTGCACTTGTABRACHYUTY CTCCAACCTATGCGGACAATTC GGCTGGCGTTATGACTCACAGATA4 TCGTTTTTTGATCTCCGGTTGT CATCGCCCCAAGCTCTGAMYL7 TCATGACCCAGGCAGACAAG CCCCGTGGGTGATGATGTAGMYH6 TCCACCGGGAAAATCTGAAC CCGGAAGTCCCCATAGAGAATMYH7 GAATGGCAAGACGGTGACTGT AAGAGGCCCGAGTAGGTATAGATCATNNI GGACTCGTTGCCAGCCTTT GGACTCGTTGCCAGCCTTTACTC1 CCAAGGCCAACCGTGAGA CAATGCCTGTGGTTCTTCCAb-TUBULIN GCAAGTATGTCCCTCGAGCTATC GCCCCAGACTGACCGAAA

IPS-DERIVED CARDIOMYOCYTES 1029

6, and 8 of differentiation (Fig. 2I). After 8 days, the viablecell concentration in the HD method was 23.8�103 cells/EB,with a doubling rate of 1.44� 0.1/day. In the stirred SFmethod, the viable cell concentration reached 19.7�103 cells/EB, and the doubling rate was 1.66� 0.7/day (n¼ 3 inde-pendent experiments). In both systems, cell viability re-mained above 90% during the entire culture period (Fig. 2J).

Since the trypan blue staining may be somewhat limited inestimating cell survival due to the possible confounding ef-fects of cell damage during enzymatic dissociation, we alsoevaluated the magnitude of cell apoptosis in intact EBs. Tothis end, at 8 days of cultivation, 10 EBs from each systemwere immunostained for detection of apoptosis. As appre-ciated in Figure 2K, both types of EBs showed minimal ap-optosis (green). Moreover, the location of the apoptotic cellswithin the EBs was completely random, indicating no criticallimitation in mass transfer.16

Gene expression patterns during differentiation

Using quantitative real-time PCR, we evaluated the geneexpression pattern associated with the in vitro cardiomyocytedifferentiation process in both the HD and SF methods (Fig.3). The expression of the pluripotency genes OCT4 andNANOG was greatly decreased during differentiation. Thefirst event related to cardiomyogenesis was the upregulationof the mesoderm marker BRACHYURY at 3 days of differ-entiation, and its subsequent downregulation. This was fol-

lowed by the expression of the cardiac-related transcriptionfactor GATA4 on day 7. The last marker to be expressed wasthe cardiac-specific structural gene MYH6 encoding the sar-comeric protein a-myosin heavy chain.

Structural and molecular characterizationof the iPS-derived cardiomyocytes

We next compared the cardiomyocyte differentiationcapacity using both culturing methods. To estimate thepercentage of the beating EBs, 372 HD-derived EBs and540 SF-derived EBs (from three separate differentiation ex-periments) were included in the analysis. A very similar ef-ficiency was noted. Thus, at 14 days of differentiation,63%� 4% of the differentiating EBs in the HD method de-veloped spontaneously contracting areas and a similar per-centage of beating EBs (59%� 8%) was also observed usingthe SF technique (Fig. 4A).

We next evaluated the number of viable cardiomyocytesin the differentiating EBs derived by both methods by flowcytometric analysis using antibodies targeting the cardiacmarker, cTnT. In these studies we also took advantage ofthe fact that all iPS cell derivatives constitutively expressDsRed.9 Representative flow cytometry dot plots of theEB-derived cells (obtained at 21 days of differentiation) areshown in Figure 4B. The percentage of double-positive cells(DsRedþ and cTnTþ) was 2.98% of the total number ofEB-derived cells (DsRedþ) in this example using the HD

FIG. 1. Undifferentiated iPS cells. (A) Undifferentiated iPS colonies cultured on mouse embryonic fibroblasts (MEF) feederlayer. Scale bar: 20 mm. (B–D) Immunostaining of the iPS colonies with anti-SSEA-1 (B), anti-NANOG (C), and anti-OCT4 (D)antibodies. Scale bars: 500 mm. iPS, induced pluripotent stem. Color images available online at www.liebertonline.com/tea.

1030 ZWI-DANTSIS ET AL.

method. A slightly higher percentage of cardiomyocytes wasfound in the stirred SF method (in this example 4.14%). Deadcells were distinguished by using eFlour-450 and omittedfrom the analysis. These values are comparable to those re-ported for the spontaneous cardiomyocytes differentiationof mouse ESC without the addition of cardiac enrichmentfactors.

To characterize the structural properties of the iPS cell-derived cardiomyocytes, we first performed immunostainingwith antibodies for cardiac-specific proteins using enzymat-ically dispersed single beating cells or small clusters (21 daysof differentiation). As depicted in Figure 5, in both the HDand SF differentiation cultures, the iPS-derived myocyteswere stained positively for a-sarcomeric actinin (present at

FIG. 2. EBs’ size distribution, cell con-centration, and apoptosis. (A–F) Lightmicroscopy analysis of 3 and 7 days dif-ferentiating EBs derived using the HDmethod (using an initial drop cell dose of500 [A, B] or 1000 [C, D] cells) or usingthe SF method (E, F). Scale bars: 500 mm.(G, H) Graphs comparing the mean EBsizes at the 3 and 7 days EBs between theHD and SF systems. Note that EBs’ sizewas smaller in the HD system when using500 cells to create the drops (G, *p< 0.05)but was not significantly different whenusing 1000 cells (H). (I) Viable cell num-bers in the EBs derived using bothmethods. ( J) Cell viability during the 8days of cultivation. (K) Apoptosis loca-tions in whole 8 days EBs as detected byimmunofluorescence staining of entireEBs in both systems. Scale bars: 100mm.EBs, embryoid bodies; HD, hangingdrops; SF, spinner flasks. Color imagesavailable online at www.liebertonline.com/tea.

IPS-DERIVED CARDIOMYOCYTES 1031

the Z-line of the sarcomere), cardiac troponin I (a subunit ofthe troponin complex), and cardiac troponin T (a cardiacmyofilament protein). Gap junctions were identified by thepositive punctuate immunostaining of connexin43 (red, Fig.5G, H) that seemed to be localized to the junctions betweenthe iPS-derived cardiomyocytes.

We next assessed the expression of a variety of plur-ipotency and cardiac-related genes in undifferentiated iPScells and in microdissected contracting EBs derived fromboth differentiation methods. As shown in Figure 5I, thedifferentiating myocytes expressed the cardiac transcriptionfactors NKX2-5 and GATA4 as well as the cardiac-specificgenes cardiac troponin I (TNNI), a-myosin heavy chain(MYH6), b-myosin heavy chain (MYH7), a-actin (ACTC),and myosin light chain-2 atrial isoform (MYL7). The gene

expression pattern was quite similar in cardiomyocytesderived from both HD and SF. The expression of thepluripotency genes OCT4 and NANOG was high in theundifferentiated iPS cells (which did not express the cardiac-specific genes), whereas it greatly decreased during differ-entiation.

iPS cell-derived cardiomyocytes display functionalelectrophysiological properties

We next sought to evaluate the functional properties of theiPS-derived cardiomyocytes derived from both systems.Contracting areas were cultured on MEA plates (Fig. 6A),and the extracellular electrograms recorded from all elec-trodes underlying the beating EB were analyzed. These

FIG. 3. Gene expressionpatterns during iPS cardio-myogenesis. Real-time poly-merase chain reaction studiesshowing that cardiomyocytedifferentiation in both the HDand SF systems was charac-terized by a continuous de-crease in the expression ofpluripotent markers (OCT4and NANOG). This was cou-pled with an initial increase inmesoderm marker (BRACHY-URY) followed by the expres-sion of the cardiac-relatedtranscription factors (GATA4)and finally by cardiac-specificstructural genes (MYH6).

1032 ZWI-DANTSIS ET AL.

studies demonstrated the presence of intact electrical activityin the beating EBs (Fig. 6B).

We next evaluated the ability of the iPS cell-derived car-diomyocytes to respond appropriately to neurohormonaltriggers by assessing their ability to respond to adrenergicand muscarinic stimulation. Specifically, we evaluated thepossible chronotropic changes induced by administration ofthe b-agonist isoproteranol and the muscarinic agonist car-bamylcholine. Isoproterenol administration (10 mM, n¼ 7)resulted in a significant increase in the beating rate by112%� 17% and 83%� 18% in cardiomyocytes derived fromthe HD and SF methods, respectively (Fig. 6B). In contrast,carbamylcholine administration (5mM, n¼ 7) resulted in asignificant decrease in the beating rate by 27%� 10% and19%� 7% from baseline values in the HD and SF cardio-myocytes, respectively (Fig. 6C).

We also evaluated the effect on the conduction propertiesof the iPS cell-derived cardiac tissue by administration of thegap junction uncoupler heptanol. Application of 1 mM ofheptanol resulted in complete cessation of conduction inboth systems (n¼ 7), which was reversible after washout(Fig. 6D).

Discussion

The recent groundbreaking technology allowing the deri-vation of iPS lines by reprogramming of adult somatic cells(such as fibroblasts) with four transcription factors has gen-erated significant excitement among the scientific commu-nity.2–4 This technology holds great promise for the emergingdisciplines of personalized and regenerative medicine; be-cause of the potential to derive patient-specific in vitro dis-ease models,17 due to the prospect for tailored drugdiscovery and individualized drug testing,18 and given thepotential of autologous iPS cell derivatives to elude the im-mune system for cell replacement and tissue engineeringstrategies.19 The ability to generate functional cardiomyo-cytes, as reported for both the mouse5,6 and human7,8 iPS cellsystems, may open the way for the development of similar

applications in the cardiovascular arena. Nevertheless, suc-cessful implementation of this novel approach in these fieldswould require the development of robust, reproducible, andsafe approaches to produce cardiomyocytes in a scalablemanner.

In the present study we aimed to meet these requirementsby generating a scalable cardiomyocyte differentiating sys-tem from mouse iPS cell lines created without the oncogenec-Myc. The differentiating system utilized a scalable stirred-suspension bioreactor (SF) and the differentiating cardio-myocytes were characterized for their molecular, structural,and functional properties.

Our major findings show (a) that a reproducible cardio-myocyte differentiating system can be established from iPSgenerated with only three reprogramming factors (OCT4,SOX2, and KLF4) giving rise, in a similar manner to the ESCsystem,15,20–22 to myocytes with the appropriate molecular,structural, and functional properties; (b) that this dif-ferentiating system can be efficiently scaled up using astirred-suspension bioreactor technique; and (c) that thecardiomyocyte differentiation efficacy, gene expression pat-terns during differentiation, and the phenotypic properties ofthe iPS cell-derived cardiomyocytes are comparable betweenthe static- and stirred-based differentiation systems.

The robust ability to differentiate iPS cells into cardio-myocytes, independently of the reprogramming factor c-Myc,is an important finding of the current study. First, c-Myc-freeiPS lines are expected to reduce the potential dysregulatedgene expression networks and tumorigenic load associatedwith this oncogene.9,23 For example, when c-Myc-iPS lineswere used to produce germline-competent chimeras, in-creased tumorigenicity in the chimeras and progeny mice wasobserved because of reactivation of the c-Myc retrovirus. Incontrast, when c-Myc-free iPS lines were used, no tumorswere found in the offspring.

Second, the c-Myc pathway was previously suggestedto play a role in cardiac lineage differentiation, for example,by contributing to cardiomyocyte hyperplasia during de-velopment24 and to the re-expression of fetal genes under

FIG. 4. (A) Comparison of thepercentage of beating EBs ob-tained in both differentiating cul-tures. NS, non significant. (B)Representative flow cytometricanalysis of the EB-derived cells at21 from both methods. Y-axisrepresents fluorescence intensityof the cells constitutively expres-sing DsRed. X-axis representsfluorescence intensity of the cellsstained with cTnT antibodies.Cells in the right upper quadrate(cTnTþ, DsResþ) are viable cardi-omyocytes. Color images avail-able online at www.liebertonline.com/tea.

IPS-DERIVED CARDIOMYOCYTES 1033

conditions of stress in the adult myocardium.25 Since themajority of studies reporting on cardiomyocyte differentia-tion from iPS cells utilized lines that were created with thetraditional ‘‘4-factors’’ approach including c-Myc,5,6,8 it wasnot clear whether similar differentiation capacity can be es-tablished also in c-Myc-free lines. Our study is in agreementwith the recent report of Martinez-Fernandez et al.26 byshowing that robust cardiomyocyte differentiation (60% of thedifferentiating EBs) can be achieved also when using a c-Myc-free line. Moreover, the generated cardiomyocytes displayedappropriate molecular, structural, and functional properties.

The second important advancement achieved in this studyis the establishment of a scalable method for cardiomyocyte

differentiation using a stirred-suspension bioreactor (SF).Both the cardiogenic differentiation process and the pheno-type of the differentiating cardiomyocytes were found to becomparable in the static- and stirred-based differentiationsystems. These studies demonstrated that the developmentalstages involved in cardiac-lineage commitment and differ-entiation of the iPS cells recapitulates events in previouslydescribed in vitro and in vivo cardiac developmental models.It includes the initial downregulation of pluripotent genescoupled with an upregulation of mesodermal-specificmarkers. This was followed by expression of cardiac-specifictranscription factors and finally by cardiac-specific struc-tural genes. Immunostaining studies confirmed that the

FIG. 5. Structural and molecularcharacterization of the iPS-derivedcardiomyocytes. (A–H) Im-munostaining of cardiomyocytesderived by the HD (A–C, G) and SFsystems (D–F, H). Cells werestained positively for sarcomeric-a-actinin (A, D, G, H), cTnI (B, E),cTnT (C, F), and connexin 43 (redpunctuate immunosignal, G, H).Nuclei were counterstained withDAPI (blue). (I) Semi-quantitativereverse transcription polymerasechain reaction studies evaluatingthe expression of pluripotencygenes (OCT4 and NANOG),cardiac-specific transcriptionalfactors (GATA4, MYL7, and NKX2-5), and sarcomeric proteins (TNNI,MYH6, MYH7, and ACTC) by theiPS cell-derived cardiomyocytes(CMs) derived by the HD andSF systems as well as by theundifferentiated iPS cells. Colorimages available online atwww.liebertonline.com/tea.

1034 ZWI-DANTSIS ET AL.

differentiated myocytes spatially expressed cardiomyocyte-specific proteins (sarcomeric and gap junction proteins).Finally, electrophysiological measurements by means of anMEA system demonstrated that SF-derived cardiomyocyteswere functionally similar to those generated by the HDtechnique, and displayed adequate responses to adrenergicand cholinergic signals.

Besides the phenotype of the generated cardiomyocytes,very similar efficiency of cardiogenesis was also observed inboth systems. This finding is important since cardiomyocytedifferentiation can be affected by mechanical forces.27,28 Theformation of EBs plays an important role in cardiomyocytesdifferentiation. The HD method has been the most widely

used strategy to generate differentiating EBs in both themouse ESC and iPS systems.5,29,30 However, this technique issuitable to produce only a small number of EBs and cannotmeet the demands for therapeutic applications due to itscomplexity and difficult manageability. Stirred-suspensionculture using SF may be ideal for production of clinicallyrelevant numbers iPS-derived cardiomyocytes as it is ame-nable to scale up, facilitate process control, and simplify cellproduction processes.

For example, our findings demonstrate the ability togenerate *1000 EBs from a starting number of 107 cells using100 mL of a single SF run. Although this number may beslightly smaller than the number of EBs generated using the

FIG. 6. Multielectrode array andpharmacological studies. (A) Beat-ing areas were placed on top ofmultielectrode array plates. Thisallowed recording of extracellularelectrograms from all electrodesunderlying the contracting EBs(scale bar: 200 mm). (B, C) Positiveand negative chronotropic re-sponses induced by the adrenergicagonist isoproterenol (B) and themuscarinic agonist carbamylcho-line (C), respectively. Experimentswere performed using cardiomyo-cytes derived from both culturesystems. *p< 0.05 versus baselinevalues. (D) Administration of 1 mMof the gap junction uncoupler hep-tanol caused complete cessation ofconduction. This effect was revers-ible following drug washout.

IPS-DERIVED CARDIOMYOCYTES 1035

HD technique the two methods are comparable in terms ofthe cardiomyogenesis efficiency, in the number of viable cellswithin the EBs, and in the average size of the generated EBs(when using 1000 cells in 30 mL to generate the drops). Moreimportantly, the SF process is dramatically more efficientwith regard to the time and efforts required for the entiredifferentiation process. For example, to obtain the samenumber of EBs (*1000) using a single SF, we would need toprepare more than 20 Petri dishes of drops (�50 EBs aregenerated per plate). It takes no more than 15 min to set up atleast two SFs simultaneously. In contrast, more than 1.5 h ofwork is required to prepare the 20 dishes of drops in the HDmethod (*12-fold increase in the work time). This meansthat to derive *10,000 EBs, for example, we will need morethan 200 plates of drops (15 h of work), but only 10 SFs and75 min of work.

Another finding of the present study is the successfulhomogeneous EBs formation into the SFs, without the for-mation of large cell clumps and/or aggregation betweenEBs. This is in contrast to previous observations showing thatESCs added directly into the stirred cultures clumped to-gether.14,15 One possible explanation is the impeller typeused. SFs equipped with paddle-type impeller frequentlyresults in cell clumping. On the other hand, employing SFsequipped with bulb-shaped impellers, as in our system, re-sults in the formation of homogeneous EBs.21

Importantly, extension of the up scaling method describedhere for mouse iPS to the human system is not a straight

forward conclusion. Human ES or iPS cells are more difficultto handle as single cells and usually require an initiationnucleus of a small aggregate to create EBs. The common wayto overcome this obstacle is to allow EBs formation in staticculture for 24 h before placing them in spin cultures. Thisstep was found to be a critical step for human EB formationin stirred systems in contrast to mouse EBs.16,31 Nevertheless,the potential of SF as up scaling method to derive cardio-myocytes from hiPS cells for clinical application remains tobe proven.

In summary, our study presents, for the first time, a proce-dure for scalable generation of cardiomyocytes from c-Myc-free iPS cell lines. The generated cardiomyocytes displayedstructural, molecular, and functional properties that werecomparable to similar cells derived by the widely used HDstatic method. The establishment of this cultivation systemprovides a powerful research and clinical tool for severalresearch fields including for customized toxicological andpharmacological testing, disease modeling, and for futurecell transplantation and cardiac tissue engineering strategies.

Acknowledgments

This study was supported in part by the Israel ScienceFoundation (#1225/09) and by the Grand Family researchfund. We thank Dr. Edith Suss-Toby, Dr. Ofer Shenker,Yaakov Sakoury, and the interdisciplinary unit for their as-sistance with the confocal imaging and flow cytometry.

Disclosure Statement

All authors declare that no competing financial interestsexist.

References

1. Laflamme, M.A., and Murry, C.E. Regenerating the heart.Nat Biotechnol 23, 845, 2005.

2. Takahashi, K., and Yamanaka, S. Induction of pluripotentstem cells from mouse embryonic and adult fibroblast cul-tures by defined factors. Cell 126, 663, 2006.

3. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka,T., Tomoda, K., et al. Induction of pluripotent stem cells fromadult human fibroblasts by defined factors. Cell 131, 861,2007.

4. Yu, J., Vodyanik, M.A., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J.L., Tian, S., et al. Induced pluripotentstem cell lines derived from human somatic cells. Science318, 1917, 2007.

5. Mauritz, C., Schwanke, K., Reppel, M., Neef, S., Katsirntaki,K., Maier, L.S., et al. Generation of functional murine cardiacmyocytes from induced pluripotent stem cells. Circulation118, 507, 2008.

6. Narazaki, G., Uosaki, H., Teranishi, M., Okita, K., Kim, B.,Matsuoka, S., et al. Directed and systematic differentiation ofcardiovascular cells from mouse induced pluripotent stemcells. Circulation 118, 498, 2008.

7. Zhang, J., Wilson, G.F., Soerens, A.G., Koonce, C.H., Yu, J.,Palecek, S.P., et al. Functional cardiomyocytes derived fromhuman induced pluripotent stem cells. Circ Res 104, e30,2009.

8. Zwi, L., Caspi, O., Arbel, G., Huber, I., Gepstein, A., Park,I.H., et al. Cardiomyocyte differentiation of human inducedpluripotent stem cells. Circulation 120, 1513, 2009.

FIG. 6. (Continued).

1036 ZWI-DANTSIS ET AL.

9. Nakagawa, M., Koyanagi, M., Tanabe, K., Takahashi, K.,Ichisaka, T., Aoi, T., et al. Generation of induced pluripotentstem cells without Myc from mouse and human fibroblasts.Nat Biotechnol 26, 101, 2008.

10. Mogi, A., Ichikawa, H., Matsumoto, C., Hieda, T., To-motsune, D., Sakaki, S., et al. The method of mouse embry-oid body establishment affects structure and developmentalgene expression. Tissue Cell 41, 79, 2009.

11. Nie, Y., Bergendahl, V., Hei, D.J., Jones, J.M., and Palecek,S.P. Scalable culture and cryopreservation of human embry-onic stem cells on microcarriers. Biotechnol Prog 25, 20, 2009.

12. Igelmund, P., Fleischmann, B.K., Fischer, I.R., Soest, J.,Gryshchenko, O., Bohm-Pinger, M.M., et al. Action potentialpropagation failures in long-term recordings from embry-onic stem cell-derived cardiomyocytes in tissue culture.Pflugers Arch 437, 669, 1999.

13. Kehat, I., Gepstein, A., Spira, A., Itskovitz-Eldor, J., andGepstein, L. High-resolution electrophysiological assess-ment of human embryonic stem cell-derived cardiomyo-cytes: a novel in vitro model for the study of conduction. CircRes 91, 659, 2002.

14. Dang, S.M., Kyba, M., Perlingeiro, R., Daley, G.Q., andZandstra, P.W. Efficiency of embryoid body formation andhematopoietic development from embryonic stem cells indifferent culture systems. Biotechnol Bioeng 78, 442, 2002.

15. Zandstra, P.W., Bauwens, C., Yin, T., Liu, Q., Schiller, H.,Zweigerdt, R., et al. Scalable production of embryonic stemcell-derived cardiomyocytes. Tissue Eng 9, 767, 2003.

16. Yirme, G., Amit, M., Laevsky, I., Osenberg, S., and Itskovitz-Eldor, J. Establishing a dynamic process for the formation,propagation, and differentiation of human embryoid bodies.Stem Cells Dev 17, 1227, 2008.

17. Lee, G., and Studer, L. Induced pluripotent stem cell tech-nology for the study of human disease. Nat Methods 7, 25,2010.

18. Ebert, A.D., and Svendsen, C.N. Human stem cells and drugscreening: opportunities and challenges. Nat Rev DrugDiscov 9, 367, 2010.

19. Kiskinis, E., and Eggan, K. Progress toward the clinical ap-plication of patient-specific pluripotent stem cells. J ClinInvest 120, 51, 2010.

20. Maltsev, V.A., Wobus, A.M., Rohwedel, J., Bader, M., andHescheler, J. Cardiomyocytes differentiated in vitro fromembryonic stem cells developmentally express cardiac-specific genes and ionic currents. Circ Res 75, 233, 1994.

21. Schroeder, M., Niebruegge, S., Werner, A., Willbold, E.,Burg, M., Ruediger, M., et al. Differentiation and lineageselection of mouse embryonic stem cells in a stirred benchscale bioreactor with automated process control. BiotechnolBioeng 92, 920, 2005.

22. Niebruegge, S., Nehring, A., Bar, H., Schroeder, M.,Zweigerdt, R., and Lehmann, J. Cardiomyocyte production

in mass suspension culture: embryonic stem cells as a sourcefor great amounts of functional cardiomyocytes. Tissue EngPart A 14, 1591, 2008.

23. Wernig, M., Meissner, A., Cassady, J.P., and Jaenisch, R. c-Myc is dispensable for direct reprogramming of mouse fi-broblasts. Cell Stem Cell 2, 10, 2008.

24. Jackson, T., Allard, M.F., Sreenan, C.M., Doss, L.K., Bishop,S.P., and Swain, J.L. The c-myc proto-oncogene regulatescardiac development in transgenic mice. Mol Cell Biol 10,

3709, 1990.25. Izumo, S., Nadal-Ginard, B., and Mahdavi, V. Proto-

oncogene induction and reprogramming of cardiac geneexpression produced by pressure overload. Proc Natl AcadSci USA 85, 339, 1988.

26. Martinez-Fernandez, A., Nelson, T.J., Yamada, S., Reyes, S.,Alekseev, A.E., Perez-Terzic, C., et al. iPS programmedwithout c-MYC yield proficient cardiogenesis for functionalheart chimerism. Circ Res 105, 648, 2009.

27. Illi, B., Scopece, A., Nanni, S., Farsetti, A., Morgante, L.,Biglioli, P., et al. Epigenetic histone modification and car-diovascular lineage programming in mouse embryonic stemcells exposed to laminar shear stress. Circ Res 96, 501, 2005.

28. Schmelter, M., Ateghang, B., Helmig, S., Wartenberg, M.,and Sauer, H. Embryonic stem cells utilize reactive oxygenspecies as transducers of mechanical strain-induced cardio-vascular differentiation. FASEB J 20, 1182, 2006.

29. Wobus, A.M., Wallukat, G., and Hescheler, J. Pluripotentmouse embryonic stem cells are able to differentiate intocardiomyocytes expressing chronotropic responses to ad-renergic and cholinergic agents and Ca2þ channel blockers.Differentiation 48, 173, 1991.

30. Takahashi, T., Lord, B., Schulze, P.C., Fryer, R.M., Sarang,S.S., Gullans, S.R., et al. Ascorbic acid enhances differentia-tion of embryonic stem cells into cardiac myocytes. Circu-lation 107, 1912, 2003.

31. Cameron, C.M., Hu, W.S., and Kaufman, D.S. Improveddevelopment of human embryonic stem cell-derived em-bryoid bodies by stirred vessel cultivation. BiotechnolBioeng 94, 938, 2006.

Address correspondence to:Lior Gepstein, M.D., Ph.D.

Technion’s Faculty of Medicine2 Efron St., POB 9649

Haifa 31096Israel

E-mail: mdlior@tx.technion.ac.il

Received: April 18, 2010Accepted: November 18, 2010

Online Publication Date: December 27, 2010

IPS-DERIVED CARDIOMYOCYTES 1037

This article has been cited by:

1. T Jadczyk, A Faulkner, P Madeddu. 2013. Stem cell therapy for cardiovascular disease: the demise of alchemy and rise ofpharmacology. British Journal of Pharmacology 169:2, 247-268. [CrossRef]

2. Limor Zwi-Dantsis, Lior Gepstein. 2012. Induced pluripotent stem cells for cardiac repair. Cellular and Molecular Life Sciences69:19, 3285-3299. [CrossRef]

3. Katharina S. Volz, Erik Miljan, Amanda Khoo, John P. Cooke. 2012. Development of pluripotent stem cells for vascular therapy.Vascular Pharmacology 56:5-6, 288-296. [CrossRef]

4. Udi Sarig, Marcelle Machluf. 2011. Engineering cell platforms for myocardial regeneration. Expert Opinion on Biological Therapy11:8, 1055-1077. [CrossRef]

5. Carlos A.V. Rodrigues, Tiago G. Fernandes, Maria Margarida Diogo, Cláudia Lobato da Silva, Joaquim M.S. Cabral. 2011. Stemcell cultivation in bioreactors. Biotechnology Advances . [CrossRef]