SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • Wnt3A;
  • Directed differentiation;
  • Mesendoderm;
  • Cardiomyogenesis;
  • Preclinical safety;
  • Cell pharmacology

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

In vitro differentiation of human embryonic stem cells (hESCs) into pure human cardiomyocytes (hESCMs) would present a powerful tool to further the creation of cell models designed to advance preclinical drug development. Here, we report a novel differentiation method to substantially increase hESCM yield. Upon early and transient treatment of hESCs with Wnt3a, embryoid body and mesendoderm formation is enhanced, leading to greater differentiation toward cardiomyocytes. Moreover, the generated beating clusters are highly enriched with cardiomyocytes (50%) and express genes characteristic of cardiac cells, providing evidence that these hESCMs are competent to develop in vitro into functional and physiologically relevant cardiomyocytes. In summary, this protocol not only has the potential to guarantee a renewable supply of enriched cardiomyocyte populations for developing novel and more predictive cell models, but it also should provide valuable insights into pathways critical for cardiac regeneration. STEM CELLS 2009;27:1869–1878


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

Because terminally differentiated cardiomyocytes have been shown to possess a limited potential to proliferate, human embryonic stem cell-derived cardiomyocytes (hESCMs) would represent a powerful tool, not only for advancing our understanding of early human cardiogenesis and cardiac pathology, but also for developing cell-based assays for assessing preclinical drug safety and efficacy. Additionally, hESCMs could open opportunities for identifying pathways critical to cardiac regeneration and ultimately lead to clinical applications supporting stem-cell-based therapy.

The development of the mammalian heart involves complex networks of signaling pathways that temporally and spatially regulate cell growth, proliferation, specification, and differentiation toward cardiac cell lineages. Conventional methods for differentiating human embryonic stem cells (hESCs) into cardiomyocytes result in low and inconsistent yields, making it difficult to establish cellular models [1, 2]. Developing superior differentiation protocols by inducing defined, developmental transitions would offset such limitations [3–7]. Mesendoderm, an early precursor cell lineage, gives rise to mesoderm (specifying cardiac, blood, and bone cells) and endoderm (specifying pancreatic, hepatic, and lung cells). Increasing mesendoderm formation, therefore, would be a first step toward significantly improving hESCM yield. As shown by Laflamme et al. [8], early treatment of hESCs with bone morphogenic protein 4 (BMP4) and activin A enhances hESC cardiomyogenesis by mesendoderm induction. In addition, cell–cell interactions mediated by cocultivation of hESCs with END2 (mouse endoderm-like cells) or forced aggregation during embryoid body (EB) formation increase hESC cardiac differentiation [3, 6]. Although these methods show improvements over conventional protocols, they often only provide subtle enhancements and, in some cases, are hESC line dependent.

To systematically and consistently improve hESC differentiation toward the cardiac lineage, we established a stepwise protocol using Wnt3a, an activator of the canonical Wnt/β-catenin signaling pathway for early induction of mesendoderm. Wnt3a has been shown to play a critical role in the formation of mouse primitive streak, a structure where cells later differentiate into the mesoderm and endoderm lineages [9–14]. It also exerts multiple, cellular effects such as increasing cell viability, cell proliferation, and cardiac differentiation from mouse ESCs [15–18]. Its effect in driving hESCs toward the cardiac lineage, however, has not been shown [19–21].

In addition to inducing mesendoderm formation by transiently treating hESCs with Wnt3a, we identified that the generation of proper-sized EBs and the gradual reduction of serum and insulin levels in differentiation medium are critical steps for inducing cardiomyogenesis in hESCs. By combining these parameters, we significantly improved the yield of hESC-derived cardiomyocytes to (a) develop novel cell models for investigating mechanisms underlying cardiovascular pathology, (b) develop assays for preclinical drug discovery, and (c) assess cardiac toxicity earlier during drug development.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

Materials

Recombinant mouse Wnt3a, human Wnt7a, and human Dkk1 were purchased from R&D Systems, Inc. (Minneapolis, MN, http://www.rndsystems.com). Recombinant human Wnt1 was purchased from Fitzgerald Industries International (Concord, MA, http://www.fitzgerald-fii.com) and Wnt2 was purchased from Novus Biologicals (Littleton, CO, http://www.novusbio.com). (2′Z,3′E)-6-bromoindirubin-3′-oxime (BIO) was purchased from Calbiochem (Gibbstown, NJ, http://www.emdbiosciences.com).

Cultivation and Differentiation of hESCs

H1, H7, and H9 hESCs were obtained from WiCell Research Institute (Madison, WI, http://www.wicell.org). Cells (passages 35-60) were maintained on irradiated CF1 feeder cells with complete growth medium (CGM) containing 20% KO-serum replacement, 0.5% glutamine, 1% nonessential amino acids, 0.1 mM β-mercaptoethanol, and 10 ng/ml basic fibroblast growth factor in Dulbecco's modified Eagle's medium (DMEM)/F12 medium. Cells were passaged as follows: incubated with collagenase (1 mg/ml) for 10 minutes in 5% CO2, 37°C incubator, scraped and neutralized with CGM, triturated into small aggregates, centrifuged at room temperature and 1,000 rpm for 1 minute, washed twice with basal DMEM/F12, reconstituted with CGM, and distributed onto irradiated CF1 feeder cells.

For differentiation of hESCs, cells were treated as described above with some modifications (Fig. 1A). Briefly, cells were triturated into larger aggregates and reconstituted with differentiation medium (DM) containing 15% fetal bovine serum, 1% glutamine, 1% nonessential amino acids, and 0.1 mM β-mercaptoethanol in KO-DMEM. Cells were distributed onto ultralow-binding six-well plates and treated with varying factors for 48 hours. Cells were maintained in DM for 6 days in suspension as they formed cell aggregates, termed EBs. On day 6, EBs were attached onto 0.1% gelatin-coated six-well plates with DM containing 5% serum. From day 8 up until harvest, cells were fed with serum-free medium containing 0.2% bovine serum albumin (BSA), 5 mM taurine, 2 mM carnitine, and 5 mM creatine in high-glucose DMEM [22].

thumbnail image

Figure 1. Cardiomyogenic induction of hESCs by BMP4 or Wnt3a. (A): A schematic diagram of hESC cardiac differentiation. (B): hESCs were treated with vehicle, BMP4 (100 ng/ml), or recombinant Wnt3a (100 ng/ml) for 48 hours. Cells were harvested and subjected to qRT-PCR analysis. Data represent mean ± SD of a triplicate experiment. #p < .01, treatment versus control; *p < .05, treatment versus control. (C): Cells were treated as in (B). On day 6, cells were visualized under the microscope. EB numbers were tabulated and EB sizes were measured. *p < .05, treatment versus control, n = 3; ¥p < .05, BMP4 versus Wnt3a; #p < .0001, treatment versus control. (D): Cells were treated as in (B) and cultivated as described in Materials and Methods. Cells were harvested on day 12 after differentiation and subjected to qRT-PCR analysis. Data represent mean ± SD, n = 5; *p < .05; #p < .01. Abbreviations: αMHC, α-myosin heavy chain; ANF, atrial natriuretic factor; BMP4, bone morphogenic protein 4; EB, embryoid body; hESC, human embryonic stem cell; MLC2v, myosin light chain 2v; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; SD, standard deviation.

Download figure to PowerPoint

Immunohistochemistry

Microdissected beating clusters were dissociated into single cells with 0.05% trypsin/EDTA solution and cultivated on 0.1% gelatin-coated four-well chamber slides with 5% serum-containing DM for 3 days. Cells were fixed with 4% paraformaldehyde for 30 minutes and permeabilized with 0.2% Triton X-100 for 1 hour. To block nonspecific binding, cells were incubated with 1% BSA for 1 hour at room temperature. Cells were coimmunostained with polyclonal antibody against cardiac troponin I (cTnI) (Cell Signaling Technology Inc., Danvers, MA, http://www.cellsignal.com), Nkx2.5, or Mef2a (Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) and monoclonal antibody against α-actinin (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) or cTnI (Meridian Life Science Inc., Saco, ME, http://meridianlifescience.com) overnight at 4°C. Cells were then incubated with Texas Red-conjugated anti-mouse IgG and fluorescein isothiocyanate (FITC)-conjugated anti-rabbit IgG (GE Healthcare Life Science, Pittsburgh, PA, http://www.gehealthcare.com) for 2 hours. Cells were mounted with Prolong Gold Antifade reagent with 4′,6-diamidino-2-phenylindole (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) and visualized under a confocal fluorescent microscope.

Taqman Real-Time Polymerase Chain Reaction Analysis

Total RNA was isolated using the RNeasy mini kit (Qiagen, Valencia, CA, http://www1.qiagen.com) according to the manufacturer's protocol. The concentration was quantified using a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE, http://www.nanodrop.com). First-strand cDNA was synthesized using Taqman reverse transcription reagents (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Real-time polymerase chain reaction (PCR) was performed using an Applied Biosystems 7900 fast real-time (RT) PCR system. All the Taqman probes were optimized by Applied Biosystems (supporting information Table 1). The reactions were run in a 384-well optical plate with the following cycle settings: stage 1, 2 minutes at 50°C; stage 2, 10 minutes at 95°C; and stage 3, 15 seconds at 95°C, and 1 minute at 60°C for 40 cycles. Fam fluorescent dye was used to detect the gene of interest and Vic dye was used for the internal control (18S rRNA). Relative gene expression was calculated as 2−ΔCt where ΔCt = (CtFam − CtVic) and Ct is the threshold cycle.

SYBR Green RT-PCR

Total RNA was isolated using the RNeasy mini kit (Qiagen). First-strand cDNAs were generated using Taqman reverse transcription reagents (Applied Biosystems). cDNAs were diluted 1:5-1:10 depending on the starting amount of RNA. Amplicons were amplified using the 7900HT FAST Real Time PCR System (Applied Biosystems) with 3 μl per diluted cDNA reaction, 0.9 μM sense and antisense primers (supporting information Table 2), and SYBR Green PCR master mix (Applied Biosystems). The PCRs were set in three stages: stage 1, 2 minutes at 50°C; stage 2, 10 minutes at 95°C; stage 3, 15 seconds at 95°C and 1 minute at 60°C for 40 cycles. β-actin was used as the loading control. Gene expression signals were analyzed as following: Expression = 2−ΔCt, where ΔCt = (Ctgene − Ctβ-actin) and Ct is the threshold cycle.

Immunoblot Analysis

Cells were lysed with RIPA lysis and extraction buffer (Thermo Scientific, Rockford, IL, http://www.thermofisher.com) containing 1% protease inhibitor and 1% phosphatase inhibitor (EMD Chemicals Inc., Gibbstown, NJ, http://www.emdbiosciences.com) and homogenized using QIAshredder (Qiagen). Proteins (15 μg per well) were resolved through premade NuPAGE Novex 4-12% Bis-Tris Gel (Invitrogen) for 1 hour at 200V. Proteins were then transferred onto 0.2-μm nitrocellulose membrane using the iBlot dry blotting system with iBlot mini gel transfer stacks (Invitrogen). The membrane was incubated with 5% BSA in 0.01% Tween-20 containing Tris-buffered saline (Fisher Scientific) for 2 hour at room temperature with primary antibody against total β-catenin (BD Biosciences, San Jose, CA, http://www.bdbiosciences.com), active β-catenin (Alexis Biochemicals, San Diego, CA, http://www.alexis-corp.com), proliferating cell nuclear antigen (PCNA) (Cell Signaling Technology), or glyceraldehyde-3-phosphate dehydrogenase (Novus Biologicals Inc., Littleton, CO, http://www.novusbio.com) overnight at 4°C and with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG (Cell Signaling) for 2 hours. The membrane was exposed to ECL Plus (GE Healthcare) for two minutes and developed using autoradiographic film.

Microarray Analysis

Undifferentiated hESCs, cells treated with or without Wnt3a (day 2), and beating clusters (day 16 or day 23) were collected. RNA was isolated using the RNeasy mini kit. Global gene expression was profiled using the Affymetrix Human Genome U133 Plus v2.0 array. The raw intensities from Affymetrix cell files were normalized and summarized into gene expression indices at the probe-set level using the Probe Logarithmic Intensity Error (PLIER) method (http://www.affymetrix.com) implemented in the PLIER package with quantile normalization available at Bioconductor (http://www.bioconductor.org) [23]. A constant number was added to summarized data before final log2 transformations were performed [24]. MAS5 present/absent calls were also generated for gene filtering purposes. National Center for Biotechnology Information Entrez Gene (genome build 36.46, September 11, 2006) based custom Chip Definition File (version 10.2, February 2008) published by the Microarray Lab of Neuroscience Institute at the University of Michigan was used to summarize the probe-set level signals (http://brainarray.mhri.med.umich.edu/Brainarray/) [25]. In total, 17,589 Entrez Gene probe set expression values were generated. Probe sets that were absent across all conditions were eliminated from further statistical analyses. To identify differentially expressed genes, the moderated F-test implemented in the “limma” package was used to compare expression values across different groups within the contrast matrix [26]. Two group comparison analyses specified in the contrast matrix were performed using moderated t-tests. p-values generated by these tests were adjusted for multiple-hypothesis testing using the method of Benjamini and Hochberg [27]. Hierarchical clustering was performed in R (http://www.r-project.org) using 1 − Pearson correlation distance and Ward linkage methods to generate the dendrogram.

Statistical Analysis

Paired, two-tailed, Student's t-tests were performed to determine the significant difference among various conditions. A p-value < .05 was considered statistically significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

Inducing Mesendoderm Improves Cardiomyogenesis from hESCs

To determine factors that induce mesendoderm formation in hESCs and subsequently lead to cardiac differentiation, we treated cells with control vehicle, BMP4, or Wnt3a for 48 hours and subjected them to gene expression analysis. As shown in Figure 1B, BMP4 and Wnt3a substantially increased cell lineage-specific gene expression toward mesendoderm (eomesodermin homolog (EOMES), 33.7 ± 4.5 or 12.1 ± 2.1 versus 1.94 ± 0.6 for BMP4 or Wnt3a versus control, respectively; p < .01), mesoderm (mesoderm posterior 2 homolog (MESP2), 30.2 ± 9.0 or 1.83 ± 0.05 versus 0.47 ± 0.1 for BMP4 or Wnt3a versus control, respectively; p < .05), and endoderm (SOX17, 3.05 ± 0.3 or 3.03 ± 0.2 versus 1.25 ± 0.2 for BMP4 or Wnt3a versus control, respectively; p < .01 and Forkhead box A2 (FOXA2), 1.97 ± 0.1 versus 0.96 ± 0.04 for Wnt3a versus control, respectively; p < .01), excluding ectoderm (orthodenticle homeobox 2 (OTX2) and SOX1). By activating these two pathways, we also significantly enhanced the formation of EBs (Fig. 1C), both in size (234 μm ± 82 μm or 263 μm ± 70 μm versus 80 μm ± 28 μm for BMP4 or Wnt3a versus control, respectively; p < .0001) and in number (209 ± 31 or 330 ± 53 versus 38 ± 12 for BMP4 or Wnt3a versus control, respectively; p < .05). To determine whether this greater EB formation was a result of elevated cell survival or enhanced cell proliferation, Wnt3a-induced hESCs were subjected to immunoblot analysis probing for PCNA, a protein marker for cell proliferation. In the absence of Wnt3a, PCNA levels declined, whereas hESCs exposed to Wnt3a for 48 hours showed stabilized PCNA levels for up to 6 days (data not shown). This suggests that Wnt3a promotes cell survival of hESCs during EB formation.

Even though BMP4 and Wnt3a exerted equivalent effects on hESCs toward mesendoderm and EB formation, Wnt3a is more effective at stimulating cardiomyogenesis from hESCs (Fig. 1D). This is demonstrated by induced gene expression of atrial natriuretic factor (ANF) (87.6 ± 68.7 versus 3.4 ± 2.2 for Wnt3a versus BMP4, respectively), Nkx2.5 (63.4 ± 17.2 versus 18.7 ± 1.8, respectively; p < .01), α-myosin heavy chain (αMHC) (187 ± 30 versus 6.1 ± 9.4, respectively; p < .01), and myosin light chain 2v (MLC2v) (143 ± 86.6 versus 7.3 ± 3.4, respectively; p < .05), markers specific for cardiomyocytes. As shown in Figure 1C, activating hESC differentiation in the absence of BMP4 or Wnt3a leads to progressive cell death, resulting in a low number of EBs after 6 days of differentiation. These findings demonstrate that stimulating mesendoderm formation during early hESC differentiation is important for inducing cardiomyocyte differentiation in vitro.

Cell–Cell Interaction Enhances hESC Cardiomyogenesis

We next sought to determine whether providing an optimal environment for cell–cell interaction plays an essential role in improving hESC cardiomyogenesis. To do this, we compared the effects of initiating the differentiation process from monolayers versus EB suspensions. Both conditions were similarly treated with Wnt3a for 48 hours, fed with medium containing reduced serum, and collected for gene expression analysis. As shown in Figure 2A and 2B, we found that, in comparison with cardiomyocyte differentiation from monolayers, EB differentiation exhibited a better expression profile for mesendoderm and cardiac cell lineages (average gene expression: EOMES, 409 versus 13.4; MESP2, 10.3 versus 2.1; ANF, 145 versus 1.1; Nkx2.5, 148 versus 2.43; αMHC, 181 versus 2.43; MLC2v, 200 versus 2.6 for EB versus monolayer, respectively). These findings imply that providing an optimal environment for cell–cell interaction could be crucial for differentiating hESCs toward cardiomyocytes in vitro.

thumbnail image

Figure 2. Effects of cell–cell interaction, serum, and insulin on hESC cardiomyogenesis. (A, B): hESCs were treated with or without Wnt3a for 48 hours as cell aggregates called EBs or as a monolayer (Mono) and collected on day 2 (A) or day 12 (B). Total RNA was isolated and subjected to qRT-PCR analysis. Data represent means of two independent experiments. (C, D): Cells were treated with or without Wnt3a for 48 hours with increasing levels of serum concentration, collected on day 2, and subjected to qRT-PCR analysis. Data represent means of two independent experiments. (E, F): Cells were treated with or without Wnt3a for 48 hours with increasing levels of insulin, collected on day 2, and subjected to qRT-PCR analysis. Data represent means of two independent experiments. Abbreviations: αMHC, α-myosin heavy chain; ANF, atrial natriuretic factor; BMP4, bone morphogenic protein 4; EB, embryoid body; EOMES, eomesodermin homolog; hESC, human embryonic stem cell; MESP2, mesoderm posterior 2 homolog; MLC2v, myosin light chain 2v; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; SD, standard deviation SOX17, sex determining region y-box.

Download figure to PowerPoint

To determine whether the timing and duration of factor treatment would affect hESC cardiomyogenesis, cells were incubated with control vehicle, BMP4, or Wnt3a either at the initiation of EB formation or 48 hours later. Analysis on day 6 after EB formation showed that early treatment with BMP4 or Wnt3a resulted in greater EB formation (Fig. 1C), whereas late treatment led to little or no improvement (data not shown). In addition, we found that prolonged treatment of cells with Wnt3a was no more effective than transient (48-hour) treatment (data not shown).

Serum and Insulin Negatively Regulate hESC Cardiomyogenesis

Even though serum and insulin have been shown to negatively affect differentiation of cardiomyocytes from hESCs, similar inhibitory effects using Wnt3a-mediated cardiac differentiation were unproven. We therefore treated hESCs with or without Wnt3a in medium containing increasing levels of serum or insulin and analyzed the samples by gene expression. As shown in Figure 2C–2F, we found that serum and insulin consistently induced an approximately two- to threefold reduction in the gene expression level of the mesoderm (MESP2, 28.9 versus 14.3 for 0% versus 20% serum or 32.2 versus 8.5 for 0% versus 1% insulin, respectively) and endoderm (sex determining region y-box (SOX17), 228.6 versus 123.2 for 0% versus 20% serum or 4,617 versus 2,163 for 0% versus 1% insulin, respectively) lineages. Furthermore, they invariably reduced Wnt3a-induced cardiomyogenesis (data not shown). Interestingly, cells treated with serum-free medium and Wnt3a exhibited the highest level of mesendoderm gene expression, but showed poor expression of cardiac genes. In the absence of serum, however, we observed that cells formed smaller-sized EBs and were less likely to attach to gelatin-coated culture dishes; this effect, consequently, led to less than optimal differentiation conditions.

Specific Effects of Wnt3a on hESC Cardiomyogenesis

In order to leverage the conditions that enhance cardiomyogenesis from hESCs, we developed a stepwise approach to improve the differentiation process to derive cardiomyocytes from hESCs. We treated H9 hESC cell aggregates with Wnt3a for the first 48 hours, attached the generated EBs on day 6 in medium containing 5% serum, and subsequently cultivated them in medium without serum or insulin for up to 60 days (Fig. 1A). By using this novel differentiation protocol, we found that beating clusters usually emerged on day 8 and progressively increased in size and number. There was little or no additional increase in beating clusters observed after day 12 of differentiation. These beating clusters could be maintained in culture for up to 60 days. Moreover, Wnt3a treatment substantially increased, in a dose-dependent manner, the number of beating clusters during differentiation (Fig. 3A) (39.5 versus 0 mean beating clusters for 30 versus 0 ng/ml rWnt3a, respectively). Inhibition of the Wnt canonical signaling pathway by Dkk1 caused a dose-dependent blockade of Wnt3a-induced cardiac differentiation (Fig. 3B) (31.5 versus 0 mean beating clusters for 0 versus 0.3 μg/ml Dkk1, respectively). Gene expression analysis revealed that Wnt3a mediates a time-dependent upregulation of genes that are characteristic for the mesendoderm (Eomes) and cardiac (ANF, αMHC, Nkx2.5, and Mlc2v) lineages. At the same time, the stem cell marker Oct3/4 was downregulated during differentiation (data not shown). When alternative hESC lines (H1, H7) were used, we showed comparable Wnt3a-mediated effects on gene expression, EB formation, and cardiomyogenesis (supporting information Fig. 1A–1F), suggesting a cell lineage-independent effect.

thumbnail image

Figure 3. Wnt3a improves hESC cardiomyogenesis. (A): hESCs were treated with increasing concentrations of rWnt3a for 48 hours and cultivated as described in Materials and Methods. BCs were quantified on day 12 of differentiation. (B): Cells were preincubated with increasing concentrations of Dkk1 for 1 hour, treated with control or Wnt3a CM for 48 hours, and cultivated as described in (A). BCs were tabulated on day 12 of differentiation. Data are shown as means of two independent experiments. (C): Cells were treated with vehicle, rWnt1 (300 ng/ml), rWnt2 (300 ng/ml), rWnt3a (100 ng/ml), rWnt7a (300 ng/ml), or BIO (300 nM) for 48 hours and subjected to qRT-PCR analysis. Data represent means of two independent experiments. (D): Cells were treated with control, Wnt3a, or Wnt5a CM for 48 hours and subjected to qRT-PCR analysis. Data represent means of two independent experiments. Abbreviations: BC, beating cluster; BIO, (2′Z,3′E)-6-bromoindirubin-3′-oxime; CM, conditioned medium; hESC, human embryonic stem cell; qRT-PCR, quantitative reverse transcription-polymerase chain reaction.

Download figure to PowerPoint

Mechanisms of Action Underlying Wnt3a-Induced hESC Cardiomyogenesis

To better understand the mechanism of the canonical Wnt/β-catenin signaling pathway that mediates cardiomyogenesis in hESCs, we treated hESCs with various Wnt isoforms (1, 2, 3a, or 7a), insulin-like growth factor one (IGF1), phorbol myristate acetate (PMA), isoproterenol (ISO), or BIO for 48 hours. We found that none of these factors showed any of the previously demonstrated Wnt3a-mediated effects on hESC differentiation (Fig. 3C, supporting information Fig. 2).

In an attempt to understand the origins of this particular outcome, we found that Wnt3a induced the nuclear translocation of β-catenin, an effect indicative of β-catenin activation (supporting information Fig. 3A). Furthermore, immunoblot analysis supported these results, showing that Wnt3a stabilized and activated β-catenin (supporting information Fig. 3B, 3C). However, all other factors that were analyzed did not activate and only slightly stabilized β-catenin (supporting information Fig. 3B–3D). These findings suggest that β-catenin activation is required for inducing hESC-mediated mesendoderm formation and subsequent cardiomyogenesis.

Wnt11, a member of the noncanonical Wnt signaling pathway, has been shown to enhance cardiomyogenesis of human endothelial circulating progenitor cells, mouse ESCs, and mouse bone marrow mononuclear cells. We thus sought to determine whether Wnt5a, also a member of the noncanonical Wnt signaling pathway, would improve hESC cardiac differentiation using our novel approach. We subsequently discovered that cells treated with Wnt5a exhibited little or no difference in mesendoderm and EB formation (Fig. 3D and data not shown).

Characterization of Beating Clusters

To determine the percentage of cardiomyocytes within hESC-derived beating clusters and to confirm the cells' cardiac attributes, we microdissected Wnt3a-induced beating clusters on day 16 and day 30 of differentiation and analyzed them for cardiac-specific protein and gene expression. As shown in Figure 4A, beating clusters contained cells expressing cardiac-related proteins such as cTnI, α-actinin, Nkx2.5, and Mef2a. Quantitative scoring showed that there were approximately 50% cardiomyocytes (α-actinin+ cells) within the beating clusters (Fig. 4B, 4C). There were also cells positively stained with smooth muscle actin, vimentin, and von Willebrand factor VIII (supporting information Fig. 7D and data not shown), suggesting that the remaining 50% of cells within the beating cluster consisted of smooth muscle, fibroblast, and endothelial cells. In addition, these beating clusters showed a highly induced expression of cardiac-related genes such as αMHC (∼100,000-fold induction; p < .01) and Nkx2.5 (∼3.000-fold induction; p < .05) (Fig. 4D, 4E).

thumbnail image

Figure 4. Beating clusters express genes and proteins characteristic of cardiomyocytes. (A): hESCs were treated with Wnt3a CM for 48 hours and cultivated as described in Materials and Methods. Beating clusters (day 16-30) were microdissected, dissociated into single cells, and cultivated on 0.1% gelatin-coated slides for another 3 days. Cells were fixed and immunostained with antibodies against cTnI, Nkx2.5, α-actinin, or Mef2 as indicated above. Images were taken using a Zeiss confocal microscope with 63× objective lens. (B, C): Day 16-30 beating clusters were microdissected, dissociated into single cells, and cultivated on gelatin-coated chamber slides for another 3 days. Cells were fixed and immunostained with antibody against α-actinin (B). The percentage of actinin+ cells was tabulated (C). (D, E): Cells treated with control or Wnt3a CM were collected at day 2 and beating clusters were microdissected at day 16 or day 23. Total RNA was isolated and subjected to RT-PCR analysis probing for Nkx2.5(D) and αMHC(E) gene expression. Data are shown as mean ± SD, n = 4; #p < .01; *p < .05. (F, G): Beating clusters were microdissected and treated with increasing concentrations of isoproterenol. Data represent mean ± SD, n = 4 beating clusters. Abbreviations: αMHC, α-myosin heavy chain; CM, conditioned medium; cTnI, cardiac troponin I; DAPI, 4′,6-diamidino-2-phenylindole; hESC, human embryonic stem cell; RT-PCR, reverse transcription-polymerase chain reaction; SD, standard deviation.

Download figure to PowerPoint

We also determined the functional property of these beating clusters by analyzing their response to isoproterenol, a β-adrenergic agonist. As shown in Figure 4F, the beating clusters increased their beating frequency with increasing concentrations of isoproterenol. Furthermore, as the beating clusters aged, their responses also increased (Fig. 4G). We attributed such effects to the time-dependent maturation of the cardiomyocytes and consequently induction of β-adrenergic receptor expression.

Genome-Wide Analysis of Wnt3a-Induced Cardiomyogenesis

We next tried to better understand the effects of Wnt3a-mediated hESC differentiation by carrying out a global gene-expression analysis. We collected (a) undifferentiated hESC, (b) hESCs differentiated in suspension culture with or without Wnt3a supplementation on day 2, and (c) Wnt3a-induced beating clusters at differentiation day 16 or day 23, and then subjected these samples to a genome-wide microarray analysis. A heat map analysis of 1,050 filtered probes exhibited distinct expression patterns for each group (Fig. 5A). As expected, we confirmed that putative stem cell marker genes, such as Nanog, CYP26A1, Lin28, and PRMD14, markedly decreased as differentiation progressed (supporting information Fig. 4). We also noticed that genes implicated in cell proliferation and viability (Bmp2, Bmp4, Tgf3β, Bmpr2, and Smad6) were induced by Wnt3a on day 2 and increased progressively thereafter (supporting information Fig. 5).

thumbnail image

Figure 5. Global gene expression analysis of Wnt3a effects. (A): Heat map of gene expression for the five different conditions. Probe sets of 1,035 filtered genes (∼10% of total) were clustered relative to detection signals. (B): Genes shown are derived from global gene-expression analysis data. Fold changes in detection signals of day 2, control, or Wnt3a treatment against undifferentiated hESCs are shown for genes implicated to play important roles in mesendoderm and mesoderm formation. Data are shown as mean ± SD, n = 4; #p < .01. (C): Genes derived from microarray analysis are presented. Fold changes in detection signals of differentiated against undifferentiated hESCs are shown for cardiac-specific or cardiac-related genes including those encoding transcription factors and structural proteins. Data are shown as mean ± SD, n = 4; #p < .01. Abbreviations: BC, beating cluster; hESC, human embryonic stem cell; SD, standard deviation.

Download figure to PowerPoint

Consistent with our previous observation, on day 2 of differentiation, Wnt3a highly induced lineage-specific genes characteristic of mesendoderm (Wnt3, Gata4, Gsc, Mixl1, T, Eomes, Gata6), mesoderm (Tbx6, Kdr, Cdx2, Cdh11, Mesp1, Pdgfra, Mesp2), and endoderm (Sox7, Pthr1, Cxcr4, Hhex, Lhx1, Foxa2, Cer1, Sox17) (Fig. 6B and data not shown). At the same time, Wnt3a did not induce the expression of ectoderm lineage-specific marker genes (Pax6, Zic1, Sox1, Sox3, En2, Otx2, Ncam1). Furthermore, by analyzing the expression patterns of beating clusters isolated on differentiation day 16 and day 23, we identified strong enrichment of cardiac-specific genes (Fig. 6C and data not shown), including those encoding transcription factors (Mef2a, Mef2c, Nkx2.5, Hand1, Hand2), structural proteins (Tnni3, Tnnt2, Myh6, Myh7, Myl2, Myl3), calcium homeostatic proteins (Pln, ATP2A2, RYR2), and cardiac-related channels/receptors (Cav1.2, NCX1, Adrb2). Moreover, these beating clusters, with their highly specific cardiac gene signature, exhibited little or no increase in lineage marker gene expression for liver, pancreas, lung (endoderm), blood (mesoderm), or neural cells (ectoderm) (supporting information Fig. 6). These beating clusters, however, also expressed genes characteristic of endothelial, epithelial, fibroblast, and smooth muscle cells (supporting information Fig. 7 and data not shown). These findings are consistent with those observed through immunohistochemical analysis.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

We have described a protocol that substantially improves the differentiation of hESCs into cardiomyocytes by implementing the following critical steps: (a) the enzymatic and physical disruption of undifferentiated hESCs to generate EBs, (b) the early and transient treatment of hESCs with Wnt3a, and (c) the gradual reduction of serum and insulin levels.

The first step of our protocol, the enzymatic and physical disruption of undifferentiated hESCs into small-cell aggregates for developing suitably sized EBs, is critical for improving hESC cardiomyogenesis. We witnessed that EB size significantly influenced differentiation efficacy, an observation consistent with that reported by Burridge et al. [3]. Because the mechanism behind this effect is unknown, we speculate that defined EB size mediates lineage-specific cell interactions important for proper differentiation. Mesoderm and endoderm cell–cell interactions are especially important and have been shown to play an essential role in the specification of cardiac, hepatic, and pancreatic lineages [28, 29]. For example, Sox17, an endodermal marker gene, is shown to be essential for cardiomyogenesis of mouse ESCs through a nonautonomous cellular mechanism [28]. In addition, the development of liver and pancreas (endoderm lineage) requires the interaction of cardiac mesodermal and septum transversum mesenchymal cells (mesoderm lineage), a mechanism possibly mediated through BMP and fibroblast growth factor signaling pathways [29]. Separating mesodermal and endodermal cells during early hESC differentiation will, therefore, further our understanding of early precursor cell interactions and their ability to specify commitment to the cardiac, pancreatic, or hepatic lineages.

In the second step of our protocol, we substantially improved our differentiation procedure by optimizing the exposure window of Wnt3a. We found that early and brief exposure with Wnt3a during the first 48 hours of differentiation was critical for preferentially reprogramming hESCs toward mesoderm and endoderm, excluding ectoderm. This effect could not be reproduced using other Wnt isoforms (1, 2, or 7a), IGF1, PMA, ISO, or BIO, but can be explained through our results showing that these factors merely stabilize β-catenin. Wnt3a, on the other hand, not only increased the protein level of active β-catenin but also induced its nuclear translocation, suggesting that the activation of β-catenin is critical for inducing mesendoderm in hESCs. Because both emerin, an inner nuclear membrane protein, and c-Jun N-terminal kinase have been shown to regulate the nuclear accumulation of β-catenin [30, 31], it is possible that these factors could be responsible for preventing β-catenin activation by other Wnt isoforms (1, 2, or 7a), BIO, IGF1, PMA, or isoproterenol.

In addition to mesendoderm induction, Wnt3a significantly improved the formation of EBs, by increasing either cell viability or cell proliferation. Earlier studies have shown that Wnt3a prevents apoptosis and promotes cell proliferation [20, 21]. Our global gene-expression analysis showed a Wnt3a-dependent upregulation of genes known to enhance cell survival (Bmp2, Bmp4, Tgf3β, and Bmpr2) [32–34] on day 2 of hESC differentiation. To validate these results, we added BMP4 during early hESC differentiation and showed a dose-dependent increase in EB formation (data not shown). Furthermore, the PCNA immunoblot analysis of cells harvested at day 2 of differentiation indicated that Wnt3a had little effect on cell proliferation. This suggests that the observed Wnt3a-mediated superior EB formation is caused mainly by enhanced cell survival and only to a lesser extent by cell proliferation.

During the third step of our differentiation process, we enhanced cardiomyogenesis from hESCs even further by gradually reducing serum and insulin levels. Under these conditions, we were able to generate significant quantities of beating clusters that were substantially enriched in cardiomyocytes (∼50% enrichment) and could be cultivated under serum-free conditions for up to 60 days. We found that serum and insulin exerted inhibitory effects on hESC mesendoderm formation and cardiomyogenesis. These findings are consistent with recent studies demonstrating the inhibitory effects of serum and insulin on hESC cardiac differentiation [6, 35]. By immunohistochemical staining for cardiac-specific transcription factors (Nkx2.5 and Mef2a) and structural proteins (α-actinin and cTnI), we confirmed that cells within a beating cluster exhibit markers characteristic of cardiomyocytes. These results, which depict well-organized myofibrillar structures within cardiac cells, were strengthened using microarray analysis to show a highly specific gene-expression pattern characteristic of cardiac cells, but not lung, pancreatic, hepatic, blood, or neural cells. This gene signature could be classified into cardiac-specific transcription factors, calcium regulatory proteins, ion channels, and receptors, and, consequently, furnished evidence that hESC-derived cardiomyocytes develop the significant functional features of human cardiomyocytes. We also found that these beating clusters dose dependently responded to a β-adrenergic agonist, indicating a functionality resembling that of native cardiomyocytes. These important observations should, therefore, provide a strong foundation for designing assays that can detect cardiac-specific, pharmacological effects on calcium flux, sarcomere shortening, beating frequency, and action potentials [23, 36–41].

The identification of superior differentiation protocols for generating pure lineage-specific human cell types becomes even more valuable in the light of the recent discovery that human fibroblasts can be genetically reprogrammed to form induced pluripotent stem (iPS) cells [42, 43]. A recent study using mouse iPS cells showed the feasibility of efficiently differentiating these cells into cardiac myocytes at percentages comparable with those of mouse ES cells [44]. And, by using our Wnt3a-mediated differentiation protocol, it may now be possible to generate patient-specific, highly enriched human iPS-derived cardiac myocytes that can ultimately unravel the underlying mechanisms of cardiovascular defects in humans.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

We have developed a hESC-line independent, Wnt3a-mediated protocol to efficiently generate cardiomyocytes from hESCs by inducing mesendoderm formation and cell viability during the early stages of differentiation. More importantly, these human cardiomyocytes possess cardiac-specific markers, suggesting the feasibility of developing novel and more predictive preclinical cellular models for drug development. Functional analysis will further validate these cardiomyocytes regarding their ability to produce physiologically relevant human cell models; and lineage-specific selection should provide greater enrichment toward our goal of generating large amounts of pure human cardiomyocytes.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES
  10. Supporting Information

Additional supporting information available online.

FilenameFormatSizeDescription
STEM_95_sm_suppinfofigure1.tif2124KSupporting Information Figure 1. Wnt3a induces of hESC mesendoderm formation, cell viability, and cardiomyogenesis in cell line-independent manner. H1 (A-C) or H7 (D-F) ESC were treated with control or Wnt3a CM for 48 hours and analyzed for mesendoderm formation (A & D), EB formation (B & E), and cardiomyogenesis (C & F).
STEM_95_sm_suppinfofigure2.tif1771KSupporting Information Figure 2. A. Human ESC were treated with vehicle, rWnt1 (300 ng/ml), rWnt2 (300 ng/ml), rWnt3a (100 ng/ml), rWnt7a (300 ng/ml), or BIO (300 nM) for 48 hours, harvested and subjected to RT-PCR analysis detecting Brachyury gene expression. B. Cells were treated with vehicle, IGF1 (100 ng/ml), PMA (200 nM), ISO (5 mM), or rWnt3a (30 ng/ml) for 48 hours, harvested and subjected to RT-PCR analysis detecting Brachyury gene expression. C. Cells were treated as in 2A-B and cultivated as described in Materials and Methods. On day 6, cells were visualized under the microscope.
STEM_95_sm_suppinfofigure3.tif2026KSupporting Information Figure 3. Effects of β-catenin/canonical signaling pathway on hESC EB formation, mesendoderm formation, and cardiomyogenesis. A. Human ESC were treated with control or Wnt3a CM for 48 hours, fixed, and immunostained for Brachyury (red) and β-catenin (green). B-C. Cells were treated as described in 3A, harvested, and subjected to immunoblot analysis probing for total or active β-catenin. Fold changes of normalized signals of total and active β-catenin are shown as means of two distinct experiments. D. Cells were treated with vehicle, IGF1 (100 ng/ml), PMA (200 nM), ISO (5 mM), or rWnt3a (30 ng/ml) for 48 hours, harvested, and subjected to immunoblot analysis probing for total β-catenin.
STEM_95_sm_suppinfofigure4.tif1949KSupporting Information Figure 4. Suppression of stem cell genes. Signals were derived from microarray analysis data. Fold changes in detection signals of day-23 beating clusters against undifferentiated hESC are shown for genes solely expressed in stem cells. Data are shown as mean±SD, n=4, #p<0.01.
STEM_95_sm_suppinfofigure5.tif1351KSupporting Information Figure 5. Wnt3a enhanced expression of activating but not inhibiting TGFβ family related genes. Signals for these genes were obtained from microarray analysis. Fold changes in detection signals of day-2, control- or Wnt3a-treated against undifferentiated hESC are shown for genes implicated in activating or inhibiting TGFβ signaling pathway. Data are shown as mean±SD; n=4; *p<0.05.
STEM_95_sm_suppinfofigure6.tif2047KSupporting Information Figure 6. Beating clusters exhibited little or no increase in none-cardiac related genes. Signals shown represent data derived from microarray analysis. Fold changes in detection signals of day-23 differentiated against undifferentiated hESC are shown for genes specific to liver (A), pancreatic (B), lung (C), blood (D), or neural cells (E). Data are shown as mean±SD, n=4. Broken line represents no change (1-fold change).
STEM_95_sm_suppinfofigure7.tif1919KSupporting Information Figure 7. Beating clusters exhibited an increase in genes characteristic of other cells. A-C. Signals shown represent data derived from microarray analysis. Fold changes in detection signals of day-23 differentiated against undifferentiated hESC are shown for genes specific to endothelial (A), epithelial (B), or fibroblast (C) cells. Data are shown as mean±SD, n=4. D. Human ESC were treated with Wnt3a CM for 48 hours and cultivated as described in Materials and Methods. Beating clusters (day 16-30) were microdissected, dissociated into single cells and cultivated on 0.1% gelatin-coated slide for another 3 days. Cells were fixed and immunostained with antibodies against Nkx2.5 and Vimentin or smooth muscle Actin (SMA). Images were taken using fluorescent microscope with 20X objective lens.
STEM_95_sm_suppinfotable1.tif605KSupporting Information Table 1. Primers for Taqman real time PCR analysis.
STEM_95_sm_suppinfotable2.tif1618KSupporting Information Table 2. Primers for SYBR Green real time PCR analysis.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.