SEARCH

SEARCH BY CITATION

Keywords:

  • In vitro differentiation;
  • Embryonic stem cells;
  • Stroma cells;
  • Stem/progenitor cells;
  • Differentiation;
  • Tissue regeneration

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Regeneration of cardiac pacemakers is an important target of cardiac regeneration. Previously, we developed a novel embryonic stem (ES) cell differentiation system that could trace cardiovascular differentiation processes at the cellular level. In the present study, we examine expressions and functions of ion channels in ES cell-derived cardiomyocytes during their differentiation and identify ion channels that confer their automaticity. ES cell-derived Flk1+ mesoderm cells give rise to spontaneously beating cardiomyocytes on OP9 stroma cells. Spontaneously beating colonies observed at day 9.5 of Flk1+ cell culture (Flk-d9.5) were significantly decreased at Flk-d23.5. Expressions of ion channels in pacemaker cells hyperpolarization-activated cyclic nucleotide-gated (HCN)1 and -4 and voltage-gated calcium channel (Cav)3.1 and -3.2 were significantly decreased in purified cardiomyocytes at Flk-d23.5 compared with at Flk-d9.5, whereas expression of an atrial and ventricular ion channel, inward rectifier potassium channel (Kir)2.1, did not change. Blockade of HCNs and Cav ion channels significantly inhibited beating rates of cardiomyocyte colonies. Electrophysiological studies demonstrated that spontaneously beating cardiomyocytes at Flk-d9.5 showed almost similar features to those of the native mouse sinoatrial node except for relatively deep maximal diastolic potential and faster maximal upstroke velocity. Although ∼60% of myocytes at Flk-d23.5 revealed almost the same properties as those at Flk-d9.5, ∼40% of myocytes showed loss of HCN and decreased Cav3 currents and ceased spontaneous beating, with no remarkable increase of Kir2.1. Thus, HCN and Cav3 ion channels should be responsible for the maintenance of automaticity in ES cell-derived cardiomyocytes. Controlled regulation of these ion channels should be required to generate complete biological pacemakers.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Cardiac regeneration is one of the largest targets of regenerative medicine. There have been numerous trials using various cell sources; embryonic stem (ES) cells [1, [2], [3]4], bone marrow cells [5, 6], peripheral blood cells [7, 8], mesenchymal stem cells [9], and so on [10, 11] are being made for cardiac regeneration. Nevertheless, those studies are mainly aiming for the recovery of cardiac pump function by regenerating cardiac muscle walls. Regeneration of cardiac pacemakers is another important target of cardiac regeneration therapy [12]. Sick sinus syndrome (SSS) is a disorder and/or a loss of sinoatrial (SA) node pacemakers, which causes symptomatic bradycardia including syncope and even sudden death [13]. Permanent cardiac pacing is still the only effective treatment for SSS. Approximately half of SSS patients are the primary indication for pacemaker implantation in the U.S. [14]. Recent progress of stem cell biology has facilitated the development of curative treatments for SSS by regeneration of biological pacemakers.

SA-node pacemaker cells in mouse were reported to express various kinds of ion channels that generate various electrical currents [15, [16]17]. Pacemaker potential in the SA-node pacemaker cells seems to be generated by a balance between the activation of inward currents (Ih/If, ICa-T, ICa-L, Ist) and the deactivation of outward currents (IKr and IK,slow) [15]. Generation of cardiac pacemaker cells mimicking the SA-node has been reported using cardiac tissue cells, mesenchymal stem cells, and ES cells, including human cells. That is, depletion of inward rectifier potassium channel (kir)2 gene [18] or expression of the hyperpolarization-activated cyclic nucleotide-gated (hcn) gene family [19, [20]21] induced automaticity in ventricular or atrial cells. Early cardiomyocytes induced from ES cells showed pacemaker properties [22, 23], and human ES cell-derived beating embryoid bodies [24] could function as an ectopic pacemaker after transplantation to the heart [25]. Nevertheless, complete SA-node pacemaker cells have not been generated to date. In many cases, simple genetic modifications on a single ion channel could not completely reproduce the SA-node pacemaker properties. As for ES cells, the majority of ES-derived cardiomyocytes lost their automaticity and ceased spontaneous beating during long-term culture [26]. In spite of promising potentials of ES cells as a cellular source of biological pacemakers, this decline of automaticity during differentiation has been hampering development of pacemaker regeneration therapy.

Previously, we established a novel ES cell differentiation system that can reproduce the early process of cardiovascular development in vitro [27]. Vascular endothelial growth factor receptor-2 (also designated as Flk1) is the earliest differentiation marker for endothelial cells (ECs) and blood cells and a marker for lateral plate mesoderm [28, 29]. We induced Flk1+ cells from ES cells, purified them by fluorescence-activated cell sorting (FACS), and recultured the purified cells [30]. We succeeded in inducing cardiovascular cells, such as vascular ECs, mural cells (pericytes and vascular smooth muscle cells) [31, 32], and cardiomyocytes [22] from common progenitors, Flk1+ cells. When Flk1+ cells are cultured on OP9 stroma cells [30, 33], spontaneously beating cardiomyocytes are induced on two-dimensional culture conditions even from a single Flk1+ cell. Induced cardiomyocytes can be purified by FACS using α-myosin heavy chain (MHC) promoter-driven green fluorescent protein (GFP) expression, which includes myosin-positive ventricular cells, connexin 40+ conduction system cells [34], and HCN 4+ pacemaker cells [35, 36]. Using this system, we can trace and evaluate the differentiation process of cardiomyocytes including pacemaker cells at the cellular level. In the present study, to generate complete SA-node pacemaker cells from ES cells and develop a biological pacemaker cell therapy, we examined ion channel expressions and functions in ES cell-derived cardiomyocytes during their differentiation and tried to identify ion channels that confer the automaticity of the ES cell-derived cardiomyocytes.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Antibodies

Monoclonal antibodies (MoAbs) for murine E-cadherin (ECCD2) [37] and murine Flk1 (AVAS12) [29] were prepared and labeled in our laboratory as described previously [30, 31]. Antibodies for HCN1 and HCN4 were purchased from Chemicon (Temecula, CA, http://www.chemicon.com). Anti-Kir2.1, voltage-gated calcium channel (Cav)3.1, and Kv11.1 antibodies were from Alomone (Jerusalem, http://www.alomone.com). Antibodies for Cav3.2, Kir3.1, and Kir3.4 were from Santa Cruz Biotechnology (Santa Cruz, CA, http://www.scbt.com). MoAb for cardiac troponin T (cTnT) was from NeoMarkers (Fremont, CA, http://www.labvision.com).

Reagents

To block HCN channel and Kir2.1 channel, cesium chloride (CsCl) (Nacalai Tesque Inc., Kyoto, Japan, http://www.nacalai.co.jp/en) and barium chloride (BaCl2) (Nacalai Tesque) were used, respectively. For L- and T-type Ca2+ channel blocker, we used efonidipine hydrochloride (efonidipine) (Nissan Chemical Industries, Tokyo, http://www.nissanchem.co.jp/english).

Cell Culture

EMG7, a subline derived from the EB5 ES cell line, carries the MHC promoter-driven enhanced green fluorescent protein gene. EMG7 cells were maintained as described [22]. OP9 stroma cells, established from mouse calvaria, were maintained as described [33]. Induction of ES cell differentiation was performed as described [30, 31] using differentiation medium (α-minimum essential medium [Gibco, Grand Island, NY, http://www.invitrogen.com] supplemented with 10% fetal calf serum and 5 × 10−5mol/l 2-mercaptoethanol).

FACS Sorting

FACS of ES cells was performed as previously described [22, 31]. After 96–108 hours of ES cell differentiation, cultured cells were harvested and stained with allophycocyanin-conjugated AVAS12 and fluorescein isothiocyanate-conjugated ECCD2. Viable Flk1+/E-cadherin cells, excluding propidium iodide (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), were sorted by FACSVantage (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com) and plated on OP9. After 9.5 or 23.5 days of culture of Flk1+ cells on OP9 cells (Flk-d9.5 or -d23.5), they were dissociated with 0.25% trypsin/EDTA (Gibco), and GFP+ cells were sorted as differentiated cardiomyocytes. Purified cardiomyocytes were then subjected to Western blotting.

Immunostaining of Cardiomyocytes

Immunostaining of cardiomyocytes was performed as described previously [22]. Cardiomyocytes were fixed with 4% paraformaldehyde. First, antibodies for HCN1 (1:200), HCN4 (1:200), Kir2.1 (1:200), Cav3.1 (1:200), Cav3.2 (1:200), Kir3.1 (1:200), or Kir3.4 (1:200) together with cTnT (1:1,000) were followed by secondary antibodies (1:500) Alexa Fluor 488-conjugated anti-rabbit or anti-goat Ig (Molecular Probes, Eugene, OR, http://www.probes.invitrogen.com) and 567-conjugated anti-mouse Ig (Molecular Probes). Stained cells were photographed with inverted fluorescent microscopy, Eclipse TE2000-U (Nikon, Tokyo, http://www.nikon.com), and digital camera system AxioCam HRc (Carl Zeiss, Jena, Germany, http://www.zeiss.com). Images were imported into Adobe Photoshop as JPEG files for figure assembly.

Western Blotting

Extracts from cardiomyocytes at Flk-d9.5 or -d23.5 were run on SDS/polyacrylamide gel electrophoresis using gradient gel (Daiichi Pure Chemicals Co. Ltd., Tokyo, http://www.daiichichem.jp/english) followed by electrophoretic transfer onto nitrocellulose membranes. The blots were incubated for 1 hour in blocking agents Blocking One (Nacalai Tesque). Afterward, the membranes were incubated overnight with either of the first antibodies (1:200) for HCN1, HCN4, Kir2.1, Cav3.1, Cav3.2, Kir3.1, Kir3.4, or Kv11.1. Secondary antibodies (1:500) were horseradish peroxidase (HRP)-conjugated anti-goat antibody (Zymed, San Francisco, http://www.invitrogen.com) or HRP-conjugated anti-rabbit antibody (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Can Get Signal Immunoreaction Enhancer Solution kit (Toyobo, Osaka, Japan, http://www.toyobo.co.jp/e) was used. Immunoreactivity was detected with the enhanced chemiluminescence kit Chemi-Lumi One (Nacalai Tesque). β-Actin staining with anti-mouse β-actin antibody (Sigma; 1:1,000) and HRP-conjugated anti-mouse (Zymed) was used as control. Signal intensity was calculated with NIH Image software (the Research Service Branch of the National Institutes of Health, Bethesda, MD).

Electrophysiological Studies

GFP-positive cells were collected by FACS sorting and seeded on collagen-coated coverslips. The myocytes were cultured for 2–5 days under this condition before use. The coverslips were then transferred to a patch clamp recording chamber, and electrophysiological measurements were carried out using Axopatch200B amplifier and Digidata 1320 interface (Axon Instruments/Molecular Devices Corp., Union City, CA, http://www.moleculardevices.com).

Composition of Solutions.

Physiological bathing solution contained (in mM) 140 NaCl, 5.4 KCl, 0.33 NaH2PO4, 0.5 MgCl2, 1.8 CaCl2, and 5 HEPES (pH = 7.4 with NaOH). The Na+-free bathing solution contained 110 N-methyl-d-glucamine (NMDG), 40 TEA, 5 4-aminopyridine, and 5 CsCl (pH = 7.4 with HCl). Standard high K+ pipette solution contained 110 aspartic acid, 30 KCl, 5 MgATP, 5 Na2 creatine phosphate, 0.1 Na2GTP, 2 EGTA, and 10 HEPES (pH = 7.2 with KOH). Cs+-rich pipette solution contained 100 CsCl, 50 NMDG, 10 TEA, 5 MgATP, 10 EGTA, and 5 HEPES (pH = 7.2 with HCl). The electrode resistance was 4–6 MOhm. All of the experiments were carried out at 32°C–34°C.

Recording of Supplemental Online Movies

Supplemental online movies are real-time videos showing cultured spontaneously beating colonies induced from ES cells recorded by the 3CCD color video camera DXC-390 (Sony, Tokyo, http://www.sony.com) and digital videocassette recorder DSR-45 (Sony). Digital video data were converted into Windows media video files (.wmv) by digital video capture software DV Gate Motion version 2.2 (Sony) and Windows Media Encoder 9 Series (Microsoft, Redmond, WA, http://www.microsoft.com).

Statistics

All results were expressed as mean ± SD unless indicated. Statistical analysis of the data was performed using analysis of variance or Student's t test (for electrophysiological study); p < .05 was considered significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Flk1+ cells were induced from ES cells and purified as previously described [31]. In brief, undifferentiated ES cells (E-cadherin+) maintained on gelatin-coated dishes with leukemia inhibitory factor (LIF) were plated onto type IV collagen-coated dishes and cultured with differentiation medium (Materials and Methods) in the absence of LIF to induce differentiation. After 96–108 hours of differentiation, induced Flk1+ (E-cadherin) cells were sorted and purified by FACS and recultured. When purified Flk1+ cells were plated onto OP9 stroma cells, which were established from the calvaria of op/op mice [33], spontaneously beating cells started to be observed from day 4 or 5 of Flk1+ cell culture [22] (Flk-d4 or -d5; supplemental online Fig. 1).

First, we evaluated the number of beating colonies during cardiomyocyte differentiation. We cultured Flk1+ cells on OP9 at a low cell density (2,500 cells per cm2) and counted beating colony number at very early (Flk-d5.5), early (Flk-d9.5), and late (Flk-d23.5) stages of cardiomyocyte differentiation. Spontaneously beating colonies were significantly decreased at Flk-d23.5 to approximately 40% of those at Flk-d5.5. That is, colony numbers at Flk-d5.5, -d9.5, and -d23.5 that appeared from 5,000 Flk1+ cells were 19.9 ± 2.8 (n = 7), 16.7 ± 3.4 (n = 7, p = .09 vs. Flk-d5.5, p < .001 vs. Flk-d23.5), and 7.7 ± 1.5 (n = 7, p < .001 vs. Flk-d5.5, -d9.5), respectively.

Next, we examined expressions of ion channels as functional markers of various cardiac tissues [15, [16]17, 38] in ES cell-derived cardiomyocytes with immunocytochemical staining. HCN1 and -4 that generate hyperpolarization activated, cyclic nucleotide sensitive cation current (Ih/If) and Cav3.1 and -3.2 that generate T-type Ca2+ current (ICa-T) were stained as SA-node specific markers. In native cardiac tissue, Kir2.1 is expressed in ventricle and atrium and generates inward-rectifier K+ current (IK1). These molecules were stained together with a cardiomyocyte marker, cTnT, at Flk-d9.5 and Flk-d23.5 (Fig. 1). At Flk-d9.5, staining of all these ion channels largely overlapped in cTnT+ cardiomyocytes. However, expression of HCN1 and -4 and Cav3.1 and -3.2 was decreased in cardiomyocytes at Flk-d23.5. On the other hand, Kir2.1 expression showed no apparent difference.

thumbnail image

Figure Figure 1.. Ion channel expressions in cardiomyocytes at Flk-d9.5 and -d23.5. Double immunofluorescent staining for cTnT (red) and ion channels (green): (A) HCN1, (B) HCN4, (C) Cav3.1, (D) Cav3.2, (E) Kir2.1, (F) Kir3.1, and (G) Kir3.4 in induced cardiomyocytes. Left panels: Flk-d9.5; right panels: Flk-d23.5. Scale bars = 100 μm. Abbreviations: Cav, voltage-gated calcium channel; cTnT, cardiac troponin T; Flk-d9.5, day 9.5 of Flk1+ cell culture; Flk-d23.5, day 23.5 of Flk1+ cell culture; HCN, hyperpolarization-activated cyclic nucleotide-gated channel; Kir, inward rectifier potassium channel.

Download figure to PowerPoint

When cardiomyocyte differentiation was induced using EMG7 ES cells that carry the MHC promoter-driven GFP gene, bright GFP-positive beating cells were observed from Flk-d4 to -d5 [22]. We purified GFP+ cardiomyocytes at Flk-d9.5 and -d23.5 and quantitatively evaluated ion channel expressions in FACS-purified cardiomyocytes. As shown in Figure 2, expressions of HCN1 and -4 and Cav3.1 and -3.2 at Flk-d23.5 were significantly decreased to approximately 66% ± 3.3% (n = 3, p < .001 vs. Flk-d9.5), 30% ± 17% (n = 3, p < .01 vs. Flk-d9.5), 46% ± 16% (n = 3, p < .05 vs. Flk-d9.5), and 24% ± 3.3% (n = 3, p < .01 vs. Flk-d9.5) of those at Flk-d9.5, respectively. Kir2.1 expression remained the same during differentiation. Expression of mouse ether-a-go-go-related gene (ERG) potassium channel (Kv11.1) [38] was also not different between Flk-d9.5 and -d23.5 (supplemental online Fig. 2A). These results strongly suggest that decrease of HCN1 and -4 and Cav3.1 and -3.2 expressions in the later stage cardiomyocytes should be correlated with the decrease of beating colonies of ES cell-derived cardiomyocytes.

thumbnail image

Figure Figure 2.. Western blot analyses for ion channel expressions in cardiomyocytes. (A): Representative results of Western blotting for HCN1, -4, Cav3.1, -3.2, Kir2.1, Kir3.1, and -3.4 in purified cardiomyocytes at Flk-d9.5 and -d23.5. (B–H): Quantitative evaluation of ion channel expressions in cardiomyocytes. Relative intensities normalized with β-actin expressions are shown. (B): HCN1; (C): HCN4; (D): Cav3.1; (E): Cav3.2; (F): Kir2.1; (G): Kir3.1; (H): Kir3.4; *, p < .05; **, p < .01 versus Flk-d9.5 (n = 3). Abbreviations: Cav, voltage-gated calcium channel; d23.5, day 23.5 of Flk1+ cell culture; Flk-d9.5, day 9.5 of Flk1+ cell culture; HCN, hyperpolarization-activated cyclic nucleotide-gated; Kir, inward rectifier potassium channel.

Download figure to PowerPoint

To examine the functional significance of HCNs and Cav3s in spontaneous beating of the cardiomyocytes, we examined the effects of inhibitors of HCN (CsCl) and T- and L-type Ca2+ channel inhibitor (efonidipine) on beating rates of the colonies. Addition of CsCl (1 mmol/l) or efonidipine (10 μmol/l) [39] significantly decreased beating rates of cardiomyocyte colonies at Flk-d9.5. Simultaneous administration of CsCl and efonidipine additionally suppressed the beating rates to approximately 50% of control (supplemental online Fig. 3, supplemental online videos 1–3), indicating that HCNs and Cav3s are functionally contributing to the spontaneous beating property of induced cardiomyocytes.

Based on the above findings, we evaluated the electrophysiological properties of ES cell-derived cardiac myocytes at Flk-d9.5 and Flk-d23.5. As shown in Figure 3A, FACS-sorted GFP-positive myocytes at Flk-d9.5 possessed spontaneous action potential when recorded with the ruptured whole cell patch method. The rate of slow diastolic potential was −0.151 ± 0.04 V/second (n = 12), and the beating rate was 482.8 ± 134.1/minute (n = 12). The maximal diastolic potential (MDP; −68.2 ± 5.2 mV, n = 12) appeared more hyperpolarized than that of native mouse pacemaker cells (−56.7 ± 7.4 mV). The maximal upstroke velocity (dV/dtmax;; −69.1 ± 55.1 V/second, n = 12) was also larger than dV/dtmax of native pacemaker cells (−14.5 ± 4.4 V/second) [15] (see also summary in Fig. 7). In the voltage clamp experiment (Fig. 3B), robust activation of Na+ current (INa) was observed in all cells examined, which probably underlies faster dV/dtmax. The INa was insensitive to tetrodotoxin (TTX; data not shown). Ih was also consistently activated upon hyperpolarization. E4031-sensitive ERG current was activated by depolarization (supplemental online Fig. 2B–2D). Current-voltage relationships are indicated in Figure 3C. Similar to native pacemaker cells, spontaneously beating myocytes at Flk-d9.5 also retained the response to parasympathetic nerve transmitter. As shown in Figure 3D, 3 μM of acetylcholine (ACh) completely suppressed spontaneous action potential and induced remarkable hyperpolarization. ACh-induced current showed typical inward rectification, and reversal potential was close to the equilibrium potential for K+ (Fig. 3E). The presence of muscarinic K+ current (IK-Ach) was in good agreement with the existence of Kir3.1 and -3.4, which generate IK-Ach [38], in cardiomyocytes at Flk-d9.5 (Fig. 1F, 1G). Significant decreases in Kir3.1 and -3.4 expressions at Flk-d23.5 (Figs. 1F, 1G, 2G, 2H) should reflect the decrease in spontaneously beating colonies at Flk-d23.5.

thumbnail image

Figure Figure 3.. A representative action potential and membrane current of spontaneously beating myocytes at day 9.5 of Flk1+ cell culture. (A): Action potential was recorded with current clamp mode. The dashed line indicates 0 mV level. (B): Current traces in the voltage clamp experiments. The holding potential was −50 mV, and test pulses were applied for 500 msec. The dotted lines indicate zero current levels. The membrane potentials during test pulses are shown on the traces; ○ indicates initial current at the onset of hyperpolarization or the peak of inward current during depolarization, and • indicates current amplitude measured at the end of test pulses. The arrows indicate Na+ current. (C): Current-voltage (IV) relationship. The amplitude was measured at the position indicated by the same symbols in (B). (D): Effect of 3 μM acetylcholine on spontaneous action potential. The dotted line indicates 0 mV level. ACh terminated the spontaneous action potential, and the membrane potential was hyperpolarized to −76.8 mV. (E): IV-relationship was obtained from descending limb of ramp pulses (holding potential = −50 mV; dV/dt = 0.2 V/second). The intersection of two IV curves was at −82.5 mV. Abbreviations: ACh, acetylcholine; msec, millisecond; pA, picoampere; sec, second.

Download figure to PowerPoint

After 2 days of culture of FACS-sorted GFP+ cardiomyocytes at Flk-d23.5, 7 out of 13 myocytes examined showed spontaneous beating. The action potential configuration and membrane currents of this cell type were essentially the same as those of myocytes at Flk-d9.5. MDP was −66.8 ± 5.8 mV and dV/dtmax was −61.5 ± 48.9 V/second. The rate of slow diastolic potential was −0.149 ± 0.07 V/second, and beating rate was 417 ± 166 per minute (n = 7). On the other hand, 6 out of 13 cells were quiescent. We then recorded the membrane potential of the quiescent myocytes. As shown in Figure 4A, quiescent myocytes at Flk-d23.5 possessed a resting membrane potential (RMP) ranging from −82 mV to −71 mV (−75.4 ± 4.1 mV, n = 6). The action potential was induced by current injection, although its configuration appeared distinct from triangular action potential of adult mouse ventricular myocytes, suggesting that these cells were not completely mature as native ventricular cells. dV/dtmax of quiescent myocytes was −137.8 ± 52.7 V/second (n = 6). In the voltage clamp experiments, no obvious Ih current was observed during hyperpolarization. Noninactivating outward current was elicited by depolarization, although its molecular entity is unclear in the present study (Fig. 4B).

thumbnail image

Figure Figure 4.. A representative action potential and membrane current of quiescent myocytes at day 23.5 of Flk1+ cell culture. (A): Action potential induced by current injection. The resting membrane potential was −78.3 mV. (B): A family of current traces. The holding potential was −50 mV, and test pulses were 500 msec; •, current amplitude measured at the end of test pulses. (C): Current-voltage relationship obtained from the current traces in (B). Abbreviations: msec, millisecond; pA, picoampere; sec, second.

Download figure to PowerPoint

Ih (i.e., HCNs) and ICa-T (i.e., Cav3s) have been suggested to underlie pacemaker depolarization of spontaneous action potential in many excitable cell types [40, [41], [42]43]. In order to address the question of why spontaneous action potential was lost during the course of prolonged cell culture, we compared the amplitude of Ih and ICa-T in cardiomyocytes at Flk-d9.5 with nonbeating myocytes at Flk-d23.5. In Figure 5A and 5C, the amplitude of Ih was measured after inhibiting IK1 by the application of 1 mM BaCl2. The amplitude of Ih (○ in Fig. 5B, 5D) was the difference between the amplitudes measured at the onset and at the end of hyperpolarizing pulses. The amplitudes of the Ba2+-sensitive component (i.e., IK1) were measured at the end of pulses, and indicated by •. It is evident from Figure 5 that, although clear Ih current was observed in beating cells at Flk-d9.5 (−6.3 ± 1.1 picoampere [pA]/picofarad [pF] at −140 mV, n = 6), Ih was almost completely lost in the nonbeating myocytes at Flk-d23.5 (−0.11 ± 0.03 pA/pF at −140 mV, n = 6, p < .05). The amplitude of IK1 was slightly larger in individual nonbeating cells at Flk-d23.5 (−5.8 ± 2.9 pA/pF at −140 mV) than that of Flk-d9.5 (−3.1 ± 0.8 pA/pF at −140 mV, n = 6), although the difference was not statistically significant (Fig. 5B, 5D). In line with this finding, Western blotting failed to detect the increase of protein levels in the bulk of cardiomyocytes (Fig. 2F). In addition, it should be noted that density of the IK1 in Flk-d23.5 was far too small when compared with adult ventricular myocytes (−20.9 ± 5.3 pA/pF at −140 mV, data not shown).

thumbnail image

Figure Figure 5.. Changes of Ih and IK1 during the differentiation of ESC-derived myocytes. (A): The membrane current elicited by hyperpolarization at day 9.5 of Flk1+ cell culture (Flk-d9.5). The pulse protocol is shown at the top; left, in control bathing solution; right, in the presence of 1 mM BaCl2. Note that the initial current jump at the onset of test pulse became smaller. (B): •, Current-voltage (IV) curve of IK1 at Flk-d9.5. The amplitude of the Ba2+-sensitive component was measured at the end of test pulse; ○, IV-curve of Ih at Flk-d9.5. After the inhibition of IK1, the amplitude of the time-dependent component was measured by subtracting the initial current at the onset of test pulse and the steady-state current at the end of test pulse. The capacitance of the cell was calculated by integrating the capacitive current elicited by a 10-mV step pulse. (C): The membrane current elicited by hyperpolarization at Flk-d23.5. Same as in (A). (D): Symbols • and ○, IV curves of IK1 and Ih at Flk-d23.5, respectively. The amplitudes of Ih from −140 to −100 mV were reduced more in cells at Flk-d23.5 than in cells at Flk-d9.5 (n = 6; *, p < .05 vs. Flk-d9.5). Abbreviations: msec, millisecond; pA, picoampere; pF, picofarad.

Download figure to PowerPoint

In the experiments shown in Figure 6, ICa-T was recorded in Na+-free bathing solution. Voltage-gated K+ current was suppressed using Cs+-rich pipette solution. As shown in Figure 6A, ICa-T was robustly activated at −45 mV in the myocyte at Flk-d9.5. Subsequently, L-type Ca2+ current (ICa-L) was activated at −5 mV. The amplitudes of ICa-T at the potentials more positive than −45 mV were measured by subtracting the current traces with different conditioning pulses at −85 mV and −45 mV (Fig. 6B). Using the same pulse protocol, the amplitude of ICa-T in nonbeating myocytes at Flk-d23.5 (−5.9 ± 1.0 pA/pF at −40 mV, n = 5) was found to be significantly smaller than that in beating cardiomyocytes at Flk-d9.5 (−9.7 ± 1.1 pA/pF at −40 mV, n = 5, p < .05) (Fig. 6C, 6D). In contrast, the amplitude of ICa-L was not significantly different between Flk-d9.5 and Flk-d23.5.

thumbnail image

Figure Figure 6.. Changes of ICa-T and ICa-L during differentiation. (A): Ca2+ currents recorded in Na+-free bathing solution at day 9.5 of Flk1+ cell culture (Flk-d9.5). The holding potential was −85 mV. Note that during the first voltage step to −45 mV, ICa-T was completely inactivated after robust activation. The second voltage pulse to 5 mV activated ICa-L. The dotted line indicates zero current level. (B): The test pulse (duration = 500 msec) was applied from −80 mV to −40 mV with 5-mV increments without conditioning pulse; •, the amplitude of ICa-T was defined as the difference between peak inward current and the steady-state current at the end of test pulse. From −35 mV to +5 mV, the test pulse was changed with 10-mV increments, and the amplitude of ICa-T + ICa-L was measured. Then, the conditioning pulse was set to −45 mV, and the test pulse was applied from −35 mV to +35 mV with 10-mV increments; ○, the amplitude of ICa-L was measured under this condition. The amplitude of ICa-T at the potential between −35 mV and +5mV was measured by subtraction. (C): Ca2+ currents recorded in Na+-free bathing solution at Flk-d23.5. (D): Current-voltage curve of ICa-T (•) and ICa-L (○) at Flk-d23.5. The pulse protocol was the same as in Flk-d9.5. The amplitude of ICa-T from −45 to −15 mV in Flk-d23.5 was significantly smaller than those in Flk-d9.5 (n = 5, *, p < .05 vs. Flk-d9.5). The amplitude of ICa-L showed no significant difference between Flk-d9.5 and Flk-d23.5. Abbreviations: msec, millisecond; pA, picoampere; pF, picofarad.

Download figure to PowerPoint

Finally, we summarized the electrophysiological findings of ES cell-derived cardiomyocytes at different stages. At Flk-d9.5, 12 out of 13 myocytes showed spontaneous action potential as indicated by ○ in Figure 7A. Their MDPs are shown on the abscissa and dV/dtmax on the ordinate. For comparison, MDP (or RMP) and dV/dtmax of adult mouse pacemaker cells (▾) and ventricular cells (▴) are indicated. Only one myocyte was quiescent, but the action potential could be induced by current injection. At Flk-d23.5, the number of quiescent myocytes (•, n = 6) clearly increased, and, as a consequence, the number of beating myocytes with spontaneous action potential (○, n = 7) decreased, as shown in Figure 7B. At Flk-d23.5, quiescent cells showed significantly deeper RMP (−75.4 ± 4.1 mV; •, n = 6; p < .05 vs. beating cells) and faster dV/dtmax (−137.8 ± 52.7 V/second; •, n = 6; p < .05 vs. beating cells) than those of beating cells (MDP = −66.8 ± 5.8 mV and dV/dtmax = −61.5 ± 48.9 V/second; ○, n = 7), respectively. These cells may represent an intermediate differentiation stage of cardiomyocytes diversifying into two populations, pacemaker cells and nonbeating ventricular or atrial cells. Whereas cells with relatively deeper value of MDP (−75 to −70 mV) at Flk-d9.5, which retained Ih and ICa-T expression, maintained spontaneous beating (Fig. 7A), quiescent cells at Flk-23.5 with similar value of MDP (−75 to −70 mV) ceased spontaneous beating (Fig. 7B), accompanied by loss of Ih and significant decrease in ICa-T with no significant increase in IK1 (Figs. 5, 6). These results indicate that the loss of spontaneous beating appears mainly to be due to the decline in Ih and ICa-T rather than the increase in IK1.

thumbnail image

Figure Figure 7.. Summary of (A, B) action potential parameters and (C) differentiation and diversification process of cardiomyocytes. (A): Flk-d9.5. Each symbol represents the data from individual myocytes: ○, spontaneously beating myocytes; •, quiescent myocytes; ▴, adult mouse ventricular myocyte; ▾, adult mouse pacemaker cells. Dotted lines indicate −70 mV of maximal diastolic potential (MDP) (or resting membrane potential [RMP]) and −90 V/sec dV/dt max. (B): Flk-d23.5. (C): Blue to red color gradient represents MDP (or RMP) and dV/dt max. Blue box: spontaneously beating cells. Red box: nonbeating cells. Corresponding periods of ES cell differentiation are indicated. Abbreviations: Cav, voltage-gated calcium channel; Flk-d4–5, days 4–5 of Flk1+ cell culture; Flk-d9.5, day 9.5 of Flk1+ cell culture; Flk-d23.5, day 23.5 of Flk1+ cell culture; HCN, hyperpolarization-activated cyclic nucleotide-gated; Kir, inward rectifier potassium channel; Na Ch., Na+ channel; sec, second.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Here, we described expression patterns and functional features of ion channels in ES cell-derived cardiomyocytes during their differentiation process and succeeded in demonstrating the functionally responsible ion channels that are required to maintain the automaticity of cardiomyocytes during differentiation. Although attenuation of spontaneous beating has been observed in long-term culture of ES cell-derived cardiomyocytes [26], responsible ion channels for the maintenance of the spontaneous beating have not been well described. One of the reasons may be a limitation in differentiation methods using embryoid bodies in which it is difficult to trace and analyze the cellular and molecular process of differentiation at the cellular level [20, 32]. Functional assessment of individual differentiating cardiomyocytes using our two-dimensional differentiation system successfully reproduced the early differentiation and diversification processes of cardiomyocytes and demonstrated that loss of spontaneous beating was caused by the decline of HCN and Cav3 channel expressions.

Our present study suggests the endogenous cardiac developmental processes as summarized in Figure 7C. That is, cardiomyocytes first appear as a homogenous spontaneously beating population resembling the primordial heart tube, and then embryonic cardiomyocytes are functionally diversified into pacemakers and nonbeating ventricular or atrial cells. The loss of spontaneous beating in differentiating cardiomyocytes occurs as the initial phenotypic change of cardiomyocyte diversification. The significant decrease in HCN and Cav3 expression with subtle increase in Kir2.1 induces the loss of spontaneous beating in cardiomyocytes with relatively deep MDP and faster dV/dtmax. Nonbeating cells at this stage are, however, still immature as ventricular myocytes. Whereas, during endogenous development of the heart, robust increase of IK1 (i.e., Kir2.1) is indispensable for the electrophysiological maturation of dV/dtmax ventricular myocytes [44], protein levels of Kir2.1 did not significantly increase at Flk-d23.5 (Figs. 1, 2), and the density of IK1 in Flk-d23.5 was much smaller than adult ventricular myocytes (Fig. 5). Transient outward current, which is highly expressed in matured mouse ventricular myocytes, was not present in quiescent myocytes at Flk-d23.5 (Fig. 4B). Further maturation process for ventricular myocytes would be required accompanied by robust increase in Kir2.1. On the other hand, cardiomyocytes that retain HCN and Cav3 expression maintain spontaneous beating. Nevertheless, spontaneously beating cells at this stage are also immature as pacemaker cells. Electrophysiological properties of ES-derived spontaneously beating myocytes showed deeper MDP and faster dV/dtmax than those of native mouse pacemaker cells (Figs. 3, 7A) [15, 38]. These differences appeared due to the presence of IK1 and TTX insensitive INa (Figs. 3, 5). IK1 is not present in rabbit and guinea pig pacemaker cells but is present at low levels in rat and mouse pacemaker cells. INa is not present in rabbit, guinea pig, and rat pacemaker cells but is present only in mouse pacemaker cells, although it is TTX sensitive [16, 17]. Reduction of IK1 (i.e., Kir2.1) and INa (Na+ channels) would be required for the maturation of spontaneously beating cells as complete pacemaker cells. Thus, the loss of spontaneous beating is first determined by the decline in HCN and Cav3 ion channels followed by further maturation processes for pacemaker cells or ventricular myocytes. ES cell-derived cardiomyocytes were diversified but not completely matured as pacemaker cells or ventricular and atrial myocytes in this culture condition. Further culture prolongation or additional factors should be required to complete maturation.

Balanced regulation of ion channel expressions, especially HCNs and Cav3s in cardiomyocytes, should be important to constitute the spontaneously beating properties in ES cell-derived cardiomyocytes and for the generation of complete biological pacemakers. However, little has been reported concerning regulatory mechanisms of HCN and Cav3 ion channel expression during cardiomyocyte differentiation. Recently, we reported that overexpression of the dominant negative form of neuron-restrictive silencer factor (NRSF) transcription factor in cardiomyocytes led to increase of HCN2 and -4 and Cav3.2 mRNA expressions [45], and NRSF regulated HCN4 transcription in vitro [46]. Recently, a T-box transcriptional repressor, Tbx3, was reported to be involved in SA-node development and to control various ion channel expressions [47]. Manipulation of HCN and Cav3 expression may be possible by controlling these regulatory machineries.

Our two-dimensional differentiation systems, thus, should provide novel insights into the mechanisms of cardiomyocyte differentiation and diversification. Combined and coordinated regulation of multiple ion channels should be required to generate complete biological pacemakers and working cardiac muscles.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

We thank Dr. H. Niwa for EB5 ES cells. We thank Dr. M. Takahashi for critical reading of the manuscript and Dr. K. Ono and many of our colleagues for helpful suggestions and discussion. J.K.Y. was supported by grants from the Ministry of Education, Science, Sports and Culture of Japan, the New Energy and Industrial Development Organization (NEDO) of Japan, Takeda Science Foundation, the Cell Science Research Foundation, the Naito Foundation, NOVARTIS Foundation for the Promotion of Science, Terumo Life Science Foundation, Tanabe Medical Frontier Conference, and PRESTO JST. M.T. was supported by Takeda Science Foundation, the Vehicle Racing Commemorative Foundation, and the Japan Medical Association.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosure of Potential Conflicts of Interest
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
sc-06-0388Suppl_Fig1.pdf737KSupplemental Figure 1
sc-06-0388Suppl_Fig2.pdf166KSupplemental Figure 2
sc-06-0388Suppl_Fig3.pdf495KSupplemental Figure 3
Supp_video_on_line_1_before_inhibitor_admin.wmv3029KSupplemental Video 1
on_line_3_10min_after_washout_the_reagents.wmv3240KSupplemental Video 2
supple-3.wmv3201KSupplemental Video 3

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.