SEARCH

SEARCH BY CITATION

Keywords:

  • Human embryonic stem cells;
  • Myogenesis;
  • Fluorescence microscope;
  • Embryoid body

Abstract

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

Cell replacement strategies are promising interventions aiming to improve myocardial performance. Yet, the electrophysiological impact of these approaches has not been elucidated. We assessed the electrophysiological consequences of grafting of two candidate cell types, that is, skeletal myoblasts and human embryonic stem cell-derived cardiomyocytes (hESC-CMs). The fluorescently labeled (DiO) candidate cells were grafted into the rat's left ventricular myocardium. Two weeks later, optical mapping was performed using the Langendorff-perfused rat heart preparation. Images were obtained with appropriate filters to delineate the heart's anatomy, to identify the DiO-labeled cells, and to associate this information with the voltage-mapping data (using the voltage-sensitive dye PGH-I). Histological examination revealed the lack of gap junctions between grafted skeletal myotubes and host cardiomyocytes. In contrast, positive Cx43 immunostaining was observed between donor and host cardiomyocytes in the hESC-CMs-transplanted hearts. Optical mapping demonstrated either normal conduction (four of six) or minimal conduction slowing (two of six) at the hESC-CMs engraftment sites. In contrast, marked slowing of conduction or conduction block was seen (seven of eight) at the myoblast transplantation sites. Ventricular arrhythmias could not be induced in the hESC-CM hearts following programmed electrical stimulation but were inducible in 50% of the myoblast-engrafted hearts. In summary, a unique method for assessment of the electrophysiological impact of myocardial cell therapy is presented. Our results demonstrate the ability of hESC-CMs to functionally integrate with host tissue. In contrast, transplantation of cells that do not form gap junctions (skeletal myoblats) led to localized conduction disturbances and to the generation of a proarrhythmogenic substrate. STEM CELLS 2010;28:2151–2161


INTRODUCTION

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

Myocardial cell replacement strategies are emerging as exciting therapeutic approaches for assisting the failing myocardium [1–4]. The rationale underlying these strategies is that the myocardial function may be improved by repopulating the diseased areas with a new pool of functional cells. Based on this assumption, a variety of cell types have been suggested as a potential sources for tissue grafting with some (autologous skeletal myoblasts [5, 6] and bone marrow-derived progenitor cells [7, 8]) already entering the clinical arena in ongoing phase I and II clinical trials.

The enormous enthusiasm, surrounding cardiovascular regenerative medicine, has led to numerous studies focusing on identifying a variety of stem cell sources, on finding means for directing stem cells to differentiate into the cardiomyocyte or vascular lineages, on designing myocardial cell delivery or mobilization techniques, and on the assessment of the structural (using detailed histological techniques) and functional (by assessing global ventricular function) outcomes of cell grafting. Although the experiential evidence on these important areas suggest great promise for this emerging therapy, studies on the electrophysiologic impact of cell therapy are lacking. In particular, little data exist about whether particular cell types electrophysiologically integrate with host myocardium and/or whether the engraftment is proarrhythmic.

In this study, we used a unique optical mapping technique, capable of high spatial and temporal analysis, to assess the regional and global effects of cell transplantation on the myocardial electrophysiological substrate. Specifically, we assessed the effects of grafting of two candidate cell types, that is, skeletal myoblasts [5, 6, 9] and human embryonic stem cell-derived cardiomyocytes (hESC-CMs) [10–14]. Our results demonstrate that although hESC-CMs seem to integrate electrically following cell grafting, skeletal myoblasts do not and their engraftment results in localized conduction delays, conduction blocks, and this abnormal substrate forms the basis for initiation of sustained re-entrant ventricular tachycardias.

MATERIALS AND METHODS

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

Derivation of hESC-CMs

Undifferentiated hESCs of the H9.2 clone [15] were grown on top of mouse embryonic fibroblast feeder layer as previously described [16]. The culture medium consisted of 20% fetal bovine serum (FBS; HyClone, ThermoScientific, Pittsburg, PA, https://www.thermoscientific.com) and 80% knockout Dulbecco's modified Eagle's medium (DMEM) supplemented with 1 mM L-glutamine, 0.1 mM mercaptoethanol, and 1% nonessential amino acids (all from Life Technologies).

To induce differentiation, the hESCs were dispersed to small clumps using collagenase IV (1 mg/ml, Worthington, Lakewood, NJ, http://www.worthington-biochem.com) and then cultured in suspension for 7–10 days where they formed embryoid bodies (EBs). The EBs were plated on gelatin-coated culture dishes. Spontaneously beating areas were noted in some of the EBs after 5–20 days of plating. For the transplantation studies, the contracting areas within the EBs were microdissected with a curved 23-gauge needle after 30–45 days of in vitro differentiation. The contracting areas were then dissociated into small cell clusters (20–100 cells) by incubation with 1 mg/ml of collagenase B (Roche, Mannheim, Germany, https://www.roche-applied-science.com) for 45 minutes [17]. To identify the cells following in vivo grafting, they were labeled with the fluorescent cell tracer DiO (2.5 μg/ml for 45-minute incubation; Molecular Probes, Eugene, Oregon).

Skeletal Myoblasts Isolation and Cell Culture

Myoblasts were isolated from the limb muscle of neonatal Sprague-Dawley rats as previously described [18]. In brief, after rinsing the limbs with 70% ethanol the muscle was dissected, minced, incubated at 37°C, and digested with a solution containing collagenase, dispase, and CaCl2 until the mixture became fine slurry. The mixture was centrifuged for 5 minutes and the pellet resuspended in 10 ml of F-10 DMEM (Gibco; Invitrogene, Carlsbad, CA, http://www.invitrogen.com) and primary myoblast growth medium (containing 2.5 ng/ml basic fibroblast growth factor [bFGF]). The cells were plated in 60-mm collagen-coated culture dishes and incubated in media containing 20% fetal bovine serum and F-10-DMEM. Eight to ten passages were performed to eliminate fibroblasts from the culture.

To confirm the ability of the skeletal myoblasts to form myotubes, they were initially cultured to 70% confluence (fifth generation) using the myoblast culture medium (F-10-DMEM). The medium was then changed to a myotube culture medium (5% horse serum, DMEM). Myotubes appeared after 3 days of culturing. The cells were then fixed on glass coverslips with 4% paraformaldehyde for immunostaining. Briefly, after 1 hour in blocking solution (5% normal goat serum, 0.1% Triton), the samples were incubated overnight with a monoclonal antibody anti-myogenin (Santa Cruz; Santa Cruz, CA, http://www.scbt.com), followed by 1 hour with Alexa goat anti-mouse IgG (Invitrogene, Carlsbad, CA, http://www.invitrogen.com) secondary antibody.

Cell Transplantation

All animal studies were approved by the Institutional Animal Care and Use Committee of the University of California, San Francisco and the Technion's Faculty of Medicine. Male Sprague-Dawley rats weighing 200–250 g were anesthetized using Isoflurane and mechanically ventilated using a Harvard small animal mechanical respirator. Animals were randomized to be engrafted with DiO-labeled cell clusters of either beating EBs or myoblasts. Using a left thoracotomy approach, the cells were injected to a single left ventricular (LV) myocardial site (at anterior wall) using a 28-gauge needle. Approximately 1.5–2 × 106 cells were injected in each animal. Following the procedure, the animals were treated with daily injections of cyclosporine-A (15 mg/kg) and methylprednisolone (2 mg/kg) to prevent immune rejection.

Optical Mapping of the Cell Grafts in the Langendorff-Perfused Rat Heart

A charge-coupled device (CCD)-based system (Scimedia; Costa Mesa, CA, http://www.scimedia.com) using two high-speed CCD cameras that enable imaging with a resolution as high as 384 × 256 pixels on each camera and can image at a temporal resolution as fast as 1.3 milliseconds (at 48 × 32 pixels) was used in this study. We have successfully adapted this system for use with the isolated Langendorff-perfused rat preparations (Supporting Information Fig. 1). Excitation light was performed via a 1,000 W tungsten-halogen source and a light guide to direct the light on the specimen. To block cardiac contractility, the heart was immobilized with the excitation-contraction uncoupler 2,3-butanedione monoxime (15 mM). Using a custom-made chamber, which allows for mechanical immobilization, flat imaging surface, and temperature control, we were able to obtain stable optical action potentials for >5 hours. The setup also allows pacing from multiple sites in the ventricles.

To identify the injected cells, they were prelabeled (prior to cell injection into the rat's ventricular myocardium) with DiO (Molecular Probes), an inert green fluorescent lipophilic tracer. This dye was chosen because its green emission (excitation 485 nm and emission 530 nm) has no overlap with the red emission wavelength of the voltage-sensitive dye PGH-I (excitation 530 nm and emission >630 nm) used for optical mapping. At follow-up, rats underwent optical mapping using a Langendorff aorta-perfused preparation. Following positioning of the rat heart in the optical mapping chamber, the heart was loaded with PGH-I (10 μl of 5 mM stock solution) via the perfusate. Images were obtained using the CCD cameras with white light illumination for delineation of anatomy, then with filters for DiO to detect the green wavelength emission to identify the transplanted cells, and then with filters to detect PGH-I fluorescence for voltage mapping. As the heart is stationary for all these images, image overlay could be obtained to relate anatomy and DiO fluorescence with the voltage maps.

Programmed Electrical Stimulation

Programmed electrical stimulation was performed to allow optical mapping at various conditions and to evaluate the arrhythmogenic risk of the cell transplantation procedure [19, 20]. Ventricular epicardial pacing was performed at double-threshold amplitude from up to four different ventricular sites and at varying pacing cycle lengths (CLs) ranging from 400 milliseconds to the fastest CL still allowing 1:1 conduction. Next, a programmed electrical stimulation protocol, consisting of a train of 20 stimuli delivered at a constant CL of 250 milliseconds followed by up to three extrastimuli (S2–S4) delivered at decreasing coupling intervals (10 milliseconds decrements) until loss of capture, was also performed. Finally, trains of 30 stimuli from 90 to 60 milliseconds (decremented by 2 milliseconds) were performed with the aim of inducing ventricular arrhythmias.

Data Analysis

Optical mapping data were analyzed using custom software (courtesy Bum-Rak Choi and Guy Salama). Raw fluorescence data, obtained from optical imaging of the voltage-sensitive dye was viewed initially as a movie depicting changes in the normalized fluorescence intensity at each pixel. This movie revealed the global activation pattern within the field of view. Quantitative data was then obtained from the optically derived action potentials for each of the imaged pixel. Activation times were calculated at the maximum rate of rise of the fluorescent action potential signal (dF/dt) [21, 22] using previously described automated algorithms. Isochronal maps were then constructed using this information. These spatial contour maps of impulse propagation were then projected onto the imaged epicardial surface of the heart to correlate the electrophysiological changes with the location of the grafted cells. Conduction velocity (CV) vector plots were constructed using a method previously described [23]. Linear CV was measured at the site of cell grafting (S) and at two control sites in each heart, that is, (a) just beyond the transplanted site in the same direction (C2) and (b) just adjacent to the transplanted site but in a direction of propagation that is at 90° to the grafted site (C1). Linear CV was measured by dividing the distance along the propagation path by the propagation time.

Immunostaining

Hearts were harvested, frozen in liquid nitrogen, and cryo-sectioned (8 μm). Sections were permeabilized with 1% Triton and blocked with 3% normal goat serum (NGS). Immunostaining was performed using mouse anticardiac troponin I (cTnI), rabbit anti-Cx43 antibodies (all from Chemicon; Millipore, Billerica, MA, http://www.millipore.com), antimyogenin (Santa Cruz), and antiskeletal muscle myosin (MY-32; Sigma, https://www.sigma-aldrich.com). Preparations were incubated with secondary antibodies at 1:100 dilutions and analyzed by a Nikon inverted fluorescent microscope or by confocal microscopy (Nikon and Bio-Rad scanning system).

In some of the hearts, to compare graft size between the two cell types, the myoblasts or hESC-CMs were prelabeled with the fluorescent cell tracker Q-tracker585 (Invitrogene, Carlsbad, CA, http://www.invitrogen.com). Graft area was then quantified using the ImageJ-1.41 software (NIH) in at least five slides for each heart.

Statistical Analysis

All results are expressed as mean ± SEM. One-way analysis of variance was used to compare the CV measurement at the site of cell transplantation to the two control sites in the same hearts. If it was found to be significant, then post hoc analysis using Tukey's test was performed. Unpaired t test was used to compare the CV measurements at the sites of myoblast transplantation to that measured at the engraftment sites in the hESC-CM-treated animals.

RESULTS

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

Identification of the Grafted Cells

A prerequisite for the assessment of the impact of cell therapy on the myocardial electrophysiological substrate is the identification of the transplanted cells and the precise spatial association with the electrophysiological outcome. To this end, we labeled the grafted cells with a fluorescent cell tracer (DiO), just prior to in vivo delivery, and assessed the ability of this technique to allow accurate identification of the location of the grafted cells during the electrophysiological follow-up study performed 2 weeks later.

In the follow-up experiments, the transplanted rat hearts were studied using the Langendorff, aorta-perfused, preparation. The hearts were placed in a specially designed imaging chamber and were studied using a CCD camera (Supporting Information Fig. 1). Images were obtained with unfiltered light illumination for delineation of the heart's anatomy (Fig. 1A, left panel), then with filters for DiO to detect the green wavelength emission for identification of the transplanted cells (Fig. 1A, middle panel), and finally with filters to detect PGH-I fluorescence for voltage mapping. The location of the DiO-labeled cell grafts could be readily identified in all the animals studied as a bright green wavelength fluorescent signal (Fig. 1A). Histological examination performed at the end of the optical mapping studies confirmed the presence of the grafted cells at the transplanted DiO-labeled sites in both the myoblast (Fig. 1B–1D) and hESC-CMs (Fig. 1E–1F) transplantation studies.

thumbnail image

Figure 1. Identification of the transplanted cells. (A): Identification of the DiO-labeled-grafted skeletal myocytes. Shown are images obtained with unfiltered light illumination for delineation of the rat heart's anatomy (top), with filters for DiO for identification of the transplanted cells (middle), and superposition of the two images (bottom). (B–E): Histological identification of the transplanted skeletal myocytes. (B): Coimmunostaining of the transplanted site using antibodies for both cardiac (TnI, blue) and skeletal muscle (skeletal muscle myosin [MY32], green) markers. Note the presence of the transplanted cells, their skeletal myocyte phenotype, and the formation of myotubes. (C): High-magnification image of the transplanted myocytes with anti-skeletal muscle myosin antibodies (red cells). Note the typical striated pattern of the generated myotubes. (D): Immunostaining of the myoblast transplanted site. Notice the DiO-labeled (green)-grafted skeletal myocytes and host rat cardiomyocytes (TnI immunosignal, blue). Cells are shown in the short-axis view (transverse section). Also note the presence of gap junctions (red punctuate Cx43 immunostaining) between host cardiomyocytes and the absence of such structures among the DiO-expressing skeletal myocytes. (E): Identification of the transplanted human embryonic stem cell-derived cardiomyocytes (hESC-CMs; identified by their green fluorescence, top panel) within the rat myocardium. Note that the grafted cells as well as host rat cardiomyocytes are stained positively for the cardiac-specific marker TnI (red, middle panel). Fused image displaying both the green and red fluorescence signals (bottom panel). Arrows mark examples of the transplanted hESC-CMs. (F): Positive Cx43 immunostaining (green) at the interface (arrow) between donor human (H) and rat (R) cardiomyocytes. Blue, nuclear DAPI staining. Abbreviations: CMs, cardiomyocytes; Myo, skeletal myocytes.

Download figure to PowerPoint

Myoblast Transplantation

To assess the electrophysiological impact of transplantations of cells that do not form gap junction with host cardiomyocytes, we studied hearts that were engrafted with skeletal myoblasts. These cells tend to lose their ability to form gap junctions when forming myotubes [24]. Our preliminary studies confirmed the ability of the myoblasts used for these transplantation studies to form myotubes ex vivo (Supporting Information Fig. 2). As described earlier and shown in Figures 1A and 2A, precise localization of the area of cell grafting was possible in all animals by the detection of the DiO signal.

thumbnail image

Figure 2. Skeletal myocyte transplantation results in localized conduction block. (A): Identification of the location of the transplanted DiO-labeled skeletal myocytes. (B): Representative optically derived action potentials. (C): Sequential frames from the optical mapping movie depicting ventricular activation of the myoblast-transplanted rat heart (pacing at 250 milliseconds). White indicates increased PGH-I fluorescence (depolarization). Notice the presence of localized conduction block at the site of cell grafting. Top panel: Pacing from top-left corner. Bottom panel: Pacing from top right corner. (D, E): Isochronal activation map of the myocyte-transplanted heart. Note in the merged image (E) the spatial correlation between the site of conduction block (isochronal crowding) and the location of the grafted cells. (F): Vectorial map depicting the spatial distribution of the conduction velocity vectors. Note the presence of conduction block at the site of cell grafting (red box) as identified by opposite directions of the vectors at opposing sides of the grafted site. (G): Isochronal activation map of the same heart when paced at a rapid rate (CL-100 milliseconds).

Download figure to PowerPoint

Histological examination confirmed the presence of transplanted skeletal myoblasts within the rat myocardium (Fig. 1B–1D). These coimmunostaining studies targeting both cardiac (troponin I) and skeletal muscle (skeletal muscle myosin, MY32) markers confirmed the presence of the transplanted myocytes within the myocardial tissue, their skeletal muscle phenotype (note the development of a typical striated pattern in the high-magnification image in Fig. 1C) and their ability to form myotubes in vivo (Fig. 1B, 1C). These studies also revealed the absence of gap junction formation by the grafted skeletal myocvtes among themselves or with neighboring host cardiomyocytes. Figure 1D shows the presence of the Cx43 immunosignal (pink) only between host cardiomyocytes and its absence among the transplanted DiO-expressing myocytes. Finally, we also did not note any significant inflammatory response at the site of cell grafting in these immunosupressed animals.

To determine the electrophysiological impact of cell grafting, we performed optical mapping studies in the isolated perfused rat heart at 2 weeks following transplantation. An example of an optically derived action potential is shown in Figure 2B, and it is similar to that previously described in this isolated rat heart model by other groups [25]. The changes in fluorescence intensity with time, in all imaged pixels, were viewed initially as a dynamic display (Fig. 2C and Supporting Information Movie 1), depicting activation spread across the ventricle. Interestingly, when viewing the movies of the paced myoblast-transplanted hearts, we could identify the presence of discrete areas of conduction disturbances. These conduction disturbances were localized to the sites of cell transplantation and ranged from areas of complete conduction block (Supporting Information Movie 1) to varying degrees of slow conduction (Supporting Information Movie 2).

We next quantified the electrophysiological changes induced by the grafted cells. The optical signal-derived action potentials (Fig. 2B), which were obtained throughout the ventricle, were used to construct high-resolution activation maps depicting impulse propagation. Activation times were calculated at each imaged pixel as the timing of the maximum rate of rise of the fluorescent (dF/dt) signal. This information was used to generate detailed isochronal activation maps (Fig. 2D) that could be superimposed on the merged anatomical images of the epicardial surface (Fig. 2E) depicting the location of the grafted cells. As can be seen in the isochronal maps (Fig. 2D, 2E), myoblast transplantation in some of animals resulted in the formation of discrete areas of conduction block. Evidence for complete conduction block can also be appreciated in the local CV vector plot (Fig. 2F). The generated conduction blocks were nonfunctional as they persisted at both slow- and fast-pacing CLs (Fig. 2D, 2G).

Clear conduction disturbances induced by myoblast transplantation were identified in seven of the eight hearts studied. These abnormalities were characterized by complete localized conduction block in three hearts, as defined by the presence of opposite orientations of the CV vectors on opposing sides of grafted area and by an activation time difference across the transplanted site that was >50% of total ventricular activation time (Fig. 2). In the other four hearts, myoblast transplantation resulted in varying degrees of slow conduction, as identified by the localized crowding of isochrones at the site of cell transplantation (Fig. 3), ranging from significant conduction slowing (Fig. 3A–3C and Supporting Information Movie 2) to milder degrees of slow conduction (Fig. 3D–3F).

thumbnail image

Figure 3. Varying degrees of conduction slowing at sites of myocyte transplantation. (A–C): A combined DiO and light image depicting the skeletal myocyte transplanted site (A), the corresponding isochronal activation map while pacing at 250 milliseconds (B), and superposition of the two images. Note the presence of significant conduction slowing (crowding of isochrones, 0.36 mm/millisecond) at the site of cell grafting. (D–F): Similar images depicting the presence of a lower degree of conduction slowing (0.56 mm/millisecond) at the site of myocyte transplantation. Also shown in (F) is an illustration of the sites where conduction velocity values were measured across the site of cell transplantation (arrow marked by “S”) and in two control sites (C1 and C2).

Download figure to PowerPoint

Figure 4 summarizes the CV values as measured at the skeletal myoblasts transplantation sites (the site of measurement is identified by the “S” arrow in the schematic example in Fig. 3F) and in the two control sites in the same hearts (“C1” and “C2” arrows in Fig. 3F). Average CV across the site of cell grafting (n = 8) was 0.47 ± 0.55 mm/millisecond and was significantly slower (*, p < .05) than that measured at both control sites (0.78 ± 0.07 and 0.91 ± 0.11 mm/millisecond, respectively).

thumbnail image

Figure 4. Summary of the conduction changes induced by the transplanted cells. Summary of the average CV values as measured at the cell transplantation sites and at two control sites in the same hearts. Note the significant slowing of conduction at the skeletal myocyte transplantation sites (n = 8) but not at the hESC-CM-grafted sites (n = 6). *, p < .05 when compared with the control sites. **, p < .05 when compared with the hESC-CM-transplanted sites. Abbreviations: CV, conduction velocity; hESC-CMs, human embryonic stem cell-derived cardiomyocytes.

Download figure to PowerPoint

Myoblast Transplantation Is Arrhythmogenic

We next proceeded to assess the arrhythmogenic risk of the myoblast transplantation procedure. To this end, we performed ventricular-programmed electrical stimulation in the transplanted animals. Interestingly, in 50% of the animals (four of eight), we were able to reproducibly induce sustained monomorphic ventricular tachycardias. These stable ventricular tachycardias could be induced either by programmed electrical stimulation using up to three extrastimuli or by rapid ventricular pacing. Optical mapping of the ventricular tachycardia episodes demonstrated that in all cases they were re-entrant in nature and involved the DiO-labeled area of cell engraftment as a zone of slow conduction or as an obstacle around which re-entry occurred. Figure 5 and Supporting Information Movie 3 show an example of such a tachycardia. Note the development of a stable re-entry circuit rotating around the site of myoblast cell grafting (identified as the gray area in Fig. 5C). The ventricular arrhythmias induced in these animals were sustained and could be entrained as well as pace-terminated; further confirming that these rhythm disorders were due to re-entry.

thumbnail image

Figure 5. Development of ventricular tachycardia in the myoblast-transplanted heart. (A): A combined DiO and light image depicting the myoblast transplantation site. (B): Sequential images from the dynamic movie depicting the propagation of the re-entrant activation wave front. (C): The corresponding isochronal activation map showing the re-entry circuit traveling around the area of anatomical conduction block at the site of cell grafting (identified in the map as the gray area).

Download figure to PowerPoint

hESC-CM Transplantation

We next continued to assess the electrophysiological impact associated with the transplantation of a different type of candidate cell, that is, hESC-CMs. hESC-CMs represent a prototype of cells that have both excitable properties as well as the ability to generate electrical connections through gap junction with host cardiac tissue [11–14, 26–29]. Similar to the myoblast transplantation experiments, the hESC-CMs (obtained by microdissection of spontaneously contracting areas in the differentiating EBs) were prelabeled with DiO and grafted to a site in the rat LV myocardium (Fig. 6A). This allowed accurate identification of the grafted area for the follow-up optical mapping studies. Histological examination confirmed the presence of the grafted hESC-CMs in all animals studied. Note the immunostaining example in Figure 1E that the grafted cells (identified by green fluorescence, top panel) were also positively stained for a cardiac-specific marker (troponin I, TnI, red in middle and bottom panels). Gap junctions (positive immunostaining for Cx43, green) could be identified between the transplanted hESC-CMs and host cardiomyocytes (Fig. 1F; arrow). The pattern of gap junction distribution among the hESC-CMs differed from that of the adult pattern; they were smaller, they had a lower density, and they were arranged isotropically.

thumbnail image

Figure 6. Electrophysiological effects of human embryonic stem cell-derived cardiomyocyte (hESC-CM) transplantation. (A): A merged DiO and light image depicting the transplanted hESC-CMs site. (B, C): The resulting isochronal activation maps (with or without the overlay of the DiO image) while pacing at 250 milliseconds. Note the normal conduction pattern (and lack of conduction slowing) at the site of hESC-CMs grafting. (D): The corresponding conduction velocity vectorial map.

Download figure to PowerPoint

To verify that graft size in the hESC-CMs-transplanted hearts was of similar magnitude to that of the myoblast graft area, we prelabeled the grafted cells (either myoblasts or hESC-CMs) in some of the animals with the fluorescent cell-tracker (Q-tracker585). Quantitative analysis of the fluorescent area (Supporting Information Fig. 3) performed 2 weeks following transplantation demonstrated that the extent of the hESC-CMs graft area was at a similar range (although slightly smaller, by 33%) when compared with the skeletal myocyte graft area. This difference, interestingly, did not reach statistical significance (p = .11).

Optical mapping of the hearts transplanted with the hESC-CMs revealed a different pattern than hearts grafted with the skeletal myoblasts. Figure 6 and Supporting Information Movie 4 show typical results obtained in one of the animals studied. As can be seen in the dynamic display, a normal conduction pattern was noted throughout the LV without any obvious conduction disturbances at the site of cell transplantation. A similar pattern could be noted in the corresponding isochronal (Fig. 6B, 6C) and CV vector maps (Fig. 6D). Normal conduction at the site of cell grafting was found in four of the six hESC-CMs-transplanted sites. In two of the six hESC-CMs-transplanted sites, we found evidence for minimal conduction slowing at the site of cell grafting (Fig. 7), but in none could we find conduction block.

thumbnail image

Figure 7. Slight conduction slowing at a human embryonic stem cell-derived cardiomyocytes (hESC-CMs)-transplantation site. (A): A merged DiO and light image depicting the transplanted hESC-CMs site. (B): Frame sequences from the optical mapping movie of the hESC-CMs-transplanted heart while pacing at 250 milliseconds. Note the slight conduction delay at the site of hESC-CMs grafting. (C, D): The resulting isochronal activation maps (with or without the overlay of the DiO image). Note the slight conduction delay at the site of hESC-CMs grafting.

Download figure to PowerPoint

Figure 4 summarizes the CV values in the hESC-CMs-transplanted sites (n = 6). CV at the site of hESC-CMs transplantation was 0.83 ± 0.08 mm/millisecond. This value did not differ significantly from the conduction values measured at two control sites in the same hearts (0.87 ± 0.07 and 0.96 ± 0.10 mm/millisecond, respectively) but was significantly greater (p < .05) than the average CV value measured at the site of myoblast transplantation (0.47 ± 0.55 mm/millisecond).

Most importantly, in marked contrast to the skeletal myoblasts-treated animals, we could not induce sustained ventricular arrhythmias in any of the hESC-CMs-transplanted rats during both programmed electrical stimulation (with up to three extrastimuli) and rapid ventricular pacing.

DISCUSSION

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

Myocardial cell therapy procedures are emerging as novel strategies for myocardial repair [1–4]. In this study, we addressed an important but frequently overlooked issue associated with these procedures, that is, the electrophysiological impact of cell transplantation. Using an optical mapping technique and fluorescently labeled cell grafts, we were able to perform high-resolution electrophysiological mapping of the isolated perfused rat heart and to assess the localized and global electrophysiological effects of cell transplantation. The main findings of this study include: (a) Transplantation of skeletal myoblasts, which do not form gap junction with host cardiomyocytes, results in the generation of significant localized conduction slowing and conduction block. (b) These conduction disturbances form the necessary substrate for the initiation of stable ventricular re-entrant arrhythmias. (c) In contrast, hESC-CMs, which were demonstrated to form gap junctions with neighboring rat cardiomyocytes, seemed to electrically integrate with host cardiac tissue and did not form any significant conduction disturbances. Consequentially, ventricular arrhythmias could not be induced in these transplanted hearts.

Electrical Integration of the Grafted Cells

The first important electrophysiological consideration associated with myocardial cell therapy is related to potential mechanism by which stem cell engraftment may improve myocardial performance. The ultimate goal of cardiovascular regenerative medicine is to generate functional cardiac tissue that will become well integrated with host myocardium and will restore the myocardial electromechanical properties. Although a variety of cell types were demonstrated to improve ventricular performance in animal models of myocardial infarction (including skeletal myoblasts [9] and mouse and hESC-CMs [17, 30–32]), the mechanism underlying this functional improvement is not clear. Although cell grafting can improve myocardial performance through a variety of indirect mechanisms affecting the postinfarction remodeling process (such as changes in the structural properties of the scar, amplification of an endogenous repair mechanism, or induction of angiogenesis), true systolic augmentation through the contraction of the transplanted cells would require the electrophysiological integration of these cells with host tissue. For such integration to occur, currents generated in host cells, passing through gap junctions, must be sufficient to depolarize the transplanted cells, which in turn should have the appropriate excitable and excitation-contraction coupling properties.

Despite the importance of this issue, only a handful of studies assessed the ability of the grafted cells to electrically integrate with host tissue. Kehat et al. [27], using multielectrode mapping, demonstrated long-term synchronous electromechanical activity and action potential propagation between neonatal rat cardiomyocytes and hESC-CMs in coculturing studies. In vivo integration of the hESC-CMs in that study was manifested by their ability to function as a biological pacemaker and pace the ventricle in the pig model of complete atrioventricular (AV) block [27]. Similar results were also obtained by Xue et al. demonstrating the ability of enhanced green fluorescent protein (eGFP)-expressing hESC-CMs to pace the isolated guinea pig heart [29]. In contrast, Abraham et al. [33], using optical mapping, showed the absence of such functional integration in cocultures of rat cardiomyocytes and skeletal myoblasts.

To assess the ability of grafted cells to integrate in vivo at the single-cell level, Rubart et al. [34] used two-photon microscopy for imaging of intracellular calcium transients in the isolated Langendorff-perfused heart. By grafting cells that were labeled with a transgenic fluorescent marker (eGFP), the investigators were able to document electrical integration (presence of synchronous calcium transients) between individual transplanted mouse fetal cardiomyocytes and host cells [34]. A similar study by the same group demonstrated the lack of such integration when skeletal myoblasts were used [35]. In a more recent study, electrical integration between donor and host cardiomyocytes was assessed in a unique cardiac slice preparation by direct intracellular recording from eGFP-expressing mouse fetal cardiomyocytes transplanted into the cryoinfarct mouse heart model [36]. Synchronous electrical activity was shown in grafted cells that were in direct contact with host cardiomyocytes but not in those which were embedded within scar tissue.

The current work adds to the aforementioned studies, which assessed electrical integration either in vitro or at the single-cell level in vivo, by assessing for in vivo donor-host electrical integration at the level of the entire cell graft. We initially assessed for such electrical integration using a prototype of a candidate cell that does not generate gap junctions with host cardiomyocyte. To this end, we transplanted skeletal myoblasts that do not form gap junctions when becoming myotubes. Clear evidence is provided in this study for the lack of electrical integration of these transplanted cells at the whole tissue level as manifested by the presence of localized conduction block at the area of cell grafting. These conduction disturbances were nonfunctional in nature, as they persisted at both slow- and fast-pacing rates.

We next continued to assess the electrical integration of a candidate cell type that has both excitable properties (can generate action potentials) and the ability to electromechanically couple with host cardiomyocytes through the formation of gap junctions [11–14, 26–29]. To this end, we transplanted hESC derivatives that were initially differentiated ex vivo to generate cardiomyocytes (hESC-CMs). In marked contrast to the skeletal myoblasts experiments, the hESC-CMs seemed to have integrated with host tissue as manifested by the presence of relatively smooth conduction through the transplanted area. Interestingly, although in the majority of the transplanted sites, we noted rapid conduction that was comparable with the normal rat myocardium in two of the six animals, we observed slight slowing of conduction at the area of cell transplantation. This conduction slowing, which was by far much lower than that observed in the myoblast transplantation sites, may have resulted from a number of reasons including: (a) the smaller size of the grafted hESC-CMs, (b) potential difference in gap junction number (lower density), size (smaller), composition (hESC-CMs express both Cx43 and Cx45), and arrangement (isotropic) when compared with the native myocardium [26], (c) the presence of nonmyocytes within the cell graft, and (d) change or disruption of fiber orientation.

An additional explanation may be a potential mismatch between the maximum firing rate of rat and human cardiomyocytes, with the human cells incapable of following at 1:1 conduction at extremely fast rates. In this respect, it is worth mentioning that human cardiomyocytes can potentially generate action potentials at extremely fast rates (such as occurs, for example, during clinical atrial and ventricular fibrillation) and that in preliminary studies, we were able to pace EBs-containing hESC-CMs ex vivo at a rate approaching 3.5 Hz. Unfortunately, we cannot provide data whether the transplanted hESC-CMs can follow such high rates also in vivo as our optical mapping system does not provide single-cell resolution. Nevertheless, it is possible that the smooth conduction, observed at the site of cell grafting, can still occur even if some fractions of the transplanted cells do not conduct each beat. This may be the result of adequate conduction by the rest of the transplanted cells (and potentially also by intermixed host cardiomyocytes) or merely through electrotonic coupling (even without generation of action potentials) by the gap junction-forming cells, as elegantly shown both in vitro [33] and in vivo [37]. However, if such conduction disturbances do occur, they are likely of minimal significance, as no arrhythmias could be induced in these hearts.

Finally, our study only evaluated the effect of hESC-CMs transplantation in the normal heart and at a single time point (2 weeks following engraftment). Data from previous studies [17, 27, 38] have shown robust engraftment of hESC-CMs in this experimental setting. Our histological quantification data further support these findings and demonstrate that the size of the hESC-CM cell grafts was comparable with that of the skeletal myoblast-derived cell grafts and, hence, that the electrophysiological differences observed cannot be explained merely by differences in cell survival between the two cell types. Further studies will have to evaluate these interactions, however, at different time points and in the diseased (infarcted) heart setting.

Arrhythmogenic Risk of Cell Therapy

The second important electrophysiological consideration relates to the effect of cell therapy on the myocardial electrophysiological substrate and to the possible arrhythmogenic risk of these procedures. The clinical significance of this issue could already be appreciated by the disturbingly high incidence of ventricular arrhythmias observed in the initial skeletal myoblast transplantation trials [5, 6]. Although direct causal relationship is hard to prove in these initial, noncontrolled trials, given the expected high incidence of ventricular arrhythmias in this patient population, this potentially life-threatening side effect warrants further considerations.

In the case of skeletal myoblasts, the generated myotubes have completely different physiological properties than host myocytes. Moreover, because of their lack of gap junctions, these myotubes are completely uncoupled to the surrounding ventricular myocytes and may therefore act as anatomical obstacles, increasing tissue inhomogeneity, slowing conduction, and increasing the likelihood for the formation of re-entrant arrhythmias. This hypothesis was recently tested experimentally. Coculturing of skeletal myoblasts with primary rat cardiomyocyte cultures resulted in the formation of slow conduction zones that led to the generation of spiral (re-entrant) wave in this in vitro model [33]. Interestingly, genetic modification of the myoblasts to express the major gap junction protein, Cx43, improved conduction and decreased the tendency for arrhythmias in these cocultures, presenting a potential new method for increasing the safety of these procedures [33].

Recent in vivo studies, performed in the rat and mouse chronic infarction models, strengthened the observations in the coculturing studies by demonstrating increased propensity for ventricular arrhythmias following skeletal myoblast transplantation [37, 39]. Mills et al. [40], for example, demonstrated that injection of skeletal myoblasts to the rat infarct results in an increased risk of ventricular arrhythmias, whereas transplantation of a similar number of mesenchymal stem cells (MSCs) improved electrical activity and reduced arrhythmogenicity. Interestingly, genetically modifying the wild-type skeletal myoblasts to overexpress Cx43 in one of these studies resulted not only in the elimination of this proarrhythmic effect but also in a pronounced antiarrhythmic outcome [37].

Our study strengthens the aforementioned observations and provides important mechanistic information to the nature of this proarrhythmic effect. Transplantation of the skeletal myocytes resulted in the generation of significant conduction disturbances at the site of cell grafting. These conduction abnormalities ranged from varying degrees of slow conduction to complete conduction block (which was nonfunctional in nature as it persisted at both slow- and fast-pacing CLs). Remarkably, these conduction abnormalities were sufficient to allow the development of sustained ventricular tachycardias following programmed electrical stimulation in 50% of the animals, even in the current experimental setting that consisted of a healthy ventricle in a small animal model. Optical mapping of these tachyarrhythmias confirmed their re-entrant nature and highlighted the importance of the conduction disturbances generated at the site of cell transplantation for both the initiation and maintenance of these ventricular rhythm disorders.

In contrast to the skeletal myocyte transplantation group, we could not induce any ventricular arrhythmias in the hESC-CMs-transplanted hearts. The inability to induce arrhythmias was noted in both animals that displayed rapid conduction at the site of cell transplantation and in the minority of the hearts that displayed minimal conduction slowing at the grafted area. Although these results are encouraging, future studies should be performed to further assess the arrhythmogenic risk of hESC-CMs transplantation in the setting of diseased hearts (chronic myocardial infarction), large animal models (in which arrhythmia induction may be easier), more extensive areas of cell grafting (potentially resulting in larger area of conduction slowing), and in animal models with slower heart rates (to assess their arrhythmogenic potential via abnormal automaticity or triggered activity mechanisms).

CONCLUSION

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

In this study, we present a unique approach to assess the electrophysiological consequences of cell replacement strategies both from the mechanistic point of view and from the therapeutic and potential adverse effects point of view. By linking spatial information regarding the location of the transplanted cells with high-resolution optical mapping imaging, we were able to study the electrophysiological impact of transplantation of two candidate cell types for myocardial cell therapy strategies, that is, skeletal myoblasts and hESC-CMs. Our results demonstrate the lack of functional integration of skeletal myoblasts with host myocardium and provide supporting evidence for the presence of such integration in the case of hESC-CMs transplantation. Moreover, our studies also demonstrate the presence of significant conduction disturbances at the site of myoblast transplantation. The development of this abnormal electrophysiological substrate resulted in increased propensity for induction of ventricular tachycardias and may explain the reported high incidence of ventricular arrhythmias in the early human clinical trials of skeletal myoblast transplantation.

Acknowledgements

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

This study was supported in part by the Nancy and Stephen Grand Philanthropic Fund (L.G, J.E.O.), by the Yad Hanadiv scholarship (L.G.), and by the Israel Science Foundation (#1781/07; L.G).

REFERENCES

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

Supporting Information

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

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
STEM_545_sm_suppinfofigure1.tif109KSupplement Figure 1: The experimental setup used for the optical mapping studies.
STEM_545_sm_suppinfofigure2.tif330KSupplement Figure 2: Formation of myotubes (ex-vivo) by the skeletal myoblasts. [A] Initial immonostaining with myogenin antibodies (green) showing the cultured skeletal myoblasts prior to the formation of myotubes. [B-C] Formation of myotubes as identified by phase contrast microscopy (B) and nuclear immunostaining for myogenin (C- left) and myogenin and DAPI (C-right).
STEM_545_sm_suppinfofigure3.tif64KSupplement Figure 3: Low magnification fluorescent image of the skeletal-myocyte transplanted area within the rat myocardium used for graft-size quantification. The transplanted cells were pre-labeled with the fluorescent cell tracker Q-tracker 585.
STEM_545_sm_suppinfomovie1.wmv2465KSupplement Movie 1: Dynamic display showing ventricular activation spread in a myoblast-transplanted heart. Notice the presence of localized conduction block at the site of cell grafting.
STEM_545_sm_suppinfomovie2.wmv2561KSupplement Movie 2: Dynamic display showing ventricular activation spread in a myoblast-transplanted heart. Notice the presence of localized conduction slowing at the site of cell grafting.
STEM_545_sm_suppinfomovie3.wmv5121KSupplement Movie 3: Optical mapping movie displaying a ventricular tachycardia circuit in a myoblast-transplanted heart. Notice the reentrant activation wavefront circulating around the site of cell grafting.
STEM_545_sm_suppinfomovie4.wmv3137KSupplement Movie 4: Dynamic display showing ventricular activation spread in a hESC-CMs-engrafted heart. Notice the absence of any conduction disturbances at the site of cell grafting.
STEM_545_sm_suppinfomovie5.wmv6209KSupplement movie 5: Dynamic display showing ventricular activation spread in a hESC-CMs-transplanted heart. Notice the presence of slight conduction slowing at the site of cell grafting.

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.