High-Resolution Electrophysiological Assessment of Human Embryonic Stem Cell-Derived Cardiomyocytes
A Novel In Vitro Model for the Study of Conduction
The goal of the present report was to establish a new in vitro model for the study of impulse propagation in human cardiac tissue. By using the human embryonic stem cell differentiating system, spontaneously contracting areas were generated in three-dimensional differentiating cell aggregates (embryoid bodies). Morphological analysis revealed an isotropic tissue of early-stage cardiac phenotype. Gap junctions, assessed by immunostaining of connexin43 and connexin45, were distributed along the cell borders. High-resolution activation maps demonstrated the presence of a functional syncytium with stable focal activation and conduction properties. Conduction was significantly slower in narrow bands of contracting tissue compared with broad cardiomyocyte regions. Establishment of this unique in vitro human model may be used for the assessment of long-term structure-function relationships, for pharmacological studies, for tissue engineering, and may permit the study of genetically modified cardiomyocytes.
Impulse propagation in cardiac tissue is a complex process governed by the excitable properties of the tissue and its macroscopic and microscopic architecture. Although the spread of electrical excitation has been studied extensively in the native heart, only a few in vitro models are currently available for studying structure-function relationships. These models include aggregates of spherical chick embryonic cardiomyocytes1 and two-dimensional isotropic, anisotropic, and patterned neonatal rat and mouse cardiomyocyte monolayers.2–4⇓⇓ Although important experimental findings were gathered from these models, they may be hampered by their relatively short life span, by the inability to genetically manipulate the cells directly, and by the lack of a human model.
Human embryonic stem (hES) cells are pluripotent cell lines recently derived from human blastocysts.5 Cultured as three-dimensional embryoid bodies (EBs), hES were demonstrated to differentiate into derivatives of all three germ layers including spontaneously contracting tissue.6 We have recently demonstrated that dispersed cells isolated from these beating areas have ultrastructural, gene expression, and functional properties of early-stage cardiac phenotype.6
In this report, we combined the hES differentiating system with a recently described microelectrode array (MEA) mapping technique.7,8⇓ Our results demonstrate that hES cells can differentiate to generate a functional syncytium with synchronized action potential propagation. We also characterized the structural and functional properties of this new experimental system and demonstrated that it could serve as a unique in vitro model for studying conduction in human cardiac tissue
Materials and Methods
For the complete Materials and Methods section, see the online data supplement, available at http://www.circresaha.org.
Results and Discussion
Cell dimensions, orientation, and connections are key determinants of electrical load distribution and of the conduction properties of cardiomyocyte cell networks. Therefore, as a first step in characterizing this new model, we attempted to define its structural properties.
Spontaneously contracting areas were identified at the outgrowth of the EBs and dissected (at a stage of 24±9 days after plating). The diameter of these contracting foci varied between 0.3 to 2 mm and their thickness was 0.04 to 0.1 mm (4 to 10 cell layers). Immunostaining with anti-cardiac troponin I (cTnI) antibodies (Figure 1a) identified an isotropic tissue with the cardiomyocytes arranged in various orientations. The cells were relatively small and round-, triangular-, or rod-shaped. Average cell length and width were 44.2±10.9 and 16.0±4.6 μm respectively, with a mean length/width ratio of 2.9±0.9 (n=67). Noncardiomyocytes were also identified in the contracting areas (Figure 1a) mainly at the periphery and accounted for 41±6% of all cells.
We next sought to determine the presence and properties of gap junctions within the contracting areas because their number, size, and distribution are important determinants of conduction during physiological and pathological conditions.9–11⇓⇓ Gap junctions were identified by the positive punctate immunostaining of connexin43 (Cx43) and Cx45 (Figures 1b through 1d). In contrast, Cx40 was not identified in the cardiomyocytes and was seen, rarely, only in nonmyocytes. Based on Cx43 immunostaining, the average gap junction size was 0.58±0.08 μm2. The number of gap junctions per 100-μm2 tissue area was 0.45±0.18 and the proportion of tissue area occupied by high-intensity immunoreactive signal was 0.27±0.13%. By using the Cx45 immunosignal, the same parameters were 0.57±0.12 μm2, 0.88±0.33, and 0.48±0.13%, respectively. Double-labeling experiments demonstrated that most of the Cx43 and Cx45 immunosignals were colocalized to the exact same spots (same gap junctions).
These results present some interesting findings. Gap junctions were relatively small and distributed homogeneously along the cell circumference with no preferential polar orientation. This pattern is similar to the one observed by Peters et al12 in human fetal and neonatal tissue. The significance of Cx45 in this model is not surprising. Although almost absent in adult ventricular myocardium, Cx45 has been shown to play a major role in early cardiac embryonic development.13
Spontaneously contracting areas were plated on MEA plates (Figure 2a). This allowed long-term, high-resolution electrophysiological recordings from all 60 electrodes (Figure 2b). After determination of the local activation time at each electrode, detailed activation maps and conduction-velocity vectorial plots were constructed (Figure 2c). These maps demonstrated the development of a functional syncytium with synchronized action potential propagation. Interestingly, both the site of earliest focal activation and the conduction properties within each EB were relatively reproducible. The average standard deviation of the location of earliest activation was minimal during short-term (3 hours) and long-term (10±5 days) recordings, measuring 103±100 μm and 264±146 μm, respectively. Similarly, the average deviation in total activation time, global velocity, and the mean magnitude of the local velocity vector were 8±3%, 8±5%, and 6±3%, respectively, during short-term recordings and 18±13%, 20±8%, 21±14%, respectively, during long-term recordings.
In contrast to the relatively reproducible measurements within each EB, the conduction properties between different EBs were more heterogeneous. In general, two conduction types were noted. In the first type (n=6), a single, relatively broad cardiomyocyte area was present, resulting in relatively fast conduction (Figure 2c). Total activation time in these EBs averaged 10.3±3.8 ms, and the average magnitude of the local conduction-velocity vector was 14.2±9.5 cm/s. The presence of notches in some of the waveforms in Figure 2b, however, may suggest the presence of discontinuities at the microscopic scale, even at these relatively fast conducting areas.
The second type of conduction was observed in EBs, in which a narrow strand of conducting tissue interconnected two contracting areas (Figure 3a). Activation and isochronal maps constructed in these EBs (Figures 3b and 3c) demonstrated relatively fast conduction within the two broad contracting areas and significant conduction delay at the narrow connecting strand. This type of conduction pattern resulted in significantly longer activation times (30.6±18.9 ms, n=6) than in the first group (P<0.05) and in a lower magnitude of the average conduction-velocity vector (4.4±2.9 cm/s). The unipolar electrogram morphology also showed an interesting spatial distribution with a QS pattern at the site of earliest focal activity, a large-amplitude deflection in the broad conducting areas, and double potentials at sites of anatomical block (Figures 3a and 3b).
These findings stress the important effect of the tissue microarchitecture on conduction. The slow conduction recorded at the narrow conducting strands is most likely due to sink-source mismatches.3 A similar pattern of narrow strands connecting different cardiomyocyte regions within the EB was also described in the murine model. These structures resulted in complex activity patterns with intermittent conduction blocks,7 a phenomenon not observed in the human model.
The conduction-velocity values observed in the present report are lower than those reported in the intact human heart and in the neonatal rat and mouse monolayers.4,9⇓ This slower conduction may stem from the small cell dimensions, the isotropic nature of the tissue, the heterogeneous distribution and presence of nonmyocytes within the cell network (which may electrically couple with cardiomyocytes and thereby slow conduction2), the lower gap junction size and density, the significant presence of Cx45, and possibly by the less developed ionic channel machinery in early-stage cardiomyocytes.
Implications of the Model
To our knowledge, this is the first long-term in vitro model for human cardiac tissue. This may pave the way for a variety of physiological, pathophysiological, and pharmacological studies that have been hampered by the lack of a suitable human model. Second, because the hES cell-derived cardiomyocytes can be assessed for prolonged periods, the model provides a unique opportunity to study short- and long-term remodeling phenomena. Moreover, the ability to assess structure and function at a relatively high resolution may enhance the investigation of their interrelations at the microscopic level.
In addition, because the cardiomyocytes in this model are generated from clonally derived ES cells, genetic manipulation may be performed with relative ease. This may allow assessing the consequences of deletion or overexpression of genes that are important in depolarization, repolarization, and cell-to-cell coupling including the role of lethal mutations that cannot be explored in transgenic animals.
Limitations of the Model
The new model also holds a number of limitations. The generated tissue is isotropic and differs from that of the adult heart. In addition, the geometry of the contracting tissues within the EB is unpredictable, ranging in size and shape. This drawback is counterbalanced by the ability to study the same tissue for prolonged periods and under different experimental settings, using each EB as its own control. Another limitation stems from the constraint of the two-dimensional mapping technique used for the assessment of the three-dimensional EB. In these respects, the present model does not replace the existing models but complements them.
This research was supported in part by the Israel Science Foundation, the Israel Ministry of Health, and the Johnson & Johnson Focused Giving grant. We thank Dr Ofer Shenker and the unit of intradepartmental equipment for their technical help.
Original received January 16, 2002; resubmission received August 5, 2002; revised resubmission received September 12, 2002; accepted September 16, 2002.
- ↵Clapham DE, Shrier A, DeHaan RL. Junctional resistance and action potential delay between embryonic heart cell aggregates. J Gen Physiol. 1980; 75: 633–654.
- ↵Fast VG, Darrow BJ, Saffitz JE, Kleber AG. Anisotropic activation spread in heart cell monolayers assessed by high-resolution optical mapping: role of tissue discontinuities. Circ Res. 1996; 79: 115–127.
- ↵Thomas SP, Bircher-Lehmann L, Thomas SA, Zhuang J, Saffitz JE, Kleber AG. Synthetic strands of neonatal mouse cardiac myocytes: structural and electrophysiological properties. Circ Res. 2000; 87: 467–473.
- ↵Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, Jones JM. Embryonic stem cell lines derived from human blastocysts. Science. 1998; 282: 1145–1147.
- ↵Zhuang J, Yamada KA, Saffitz JE, Kleber AG. Pulsatile stretch remodels cell-to-cell communication in cultured myocytes. Circ Res. 2000; 87: 316–322.
- ↵Peters NS, Wit AL. Myocardial architecture and ventricular arrhythmogenesis. Circulation. 1998; 97: 1746–1754.
- ↵Saffitz JE, Laing JG, Yamada KA. Connexin expression and turnover: implications for cardiac excitability. Circ Res. 2000; 86: 723–728.
- ↵Peters NS, Severs NJ, Rothery SM, Lincoln C, Yacoub MH, Green CR. Spatiotemporal relation between gap junctions and fascia adherens junctions during postnatal development of human ventricular myocardium. Circulation. 1994; 90: 713–725.
- ↵Alcoléa S, Théveniau-Ruissy M, Jarry-Guichard T, Marics I, Tzouanacou E, Chauvin JP, Briand JP, Moorman AF, Lamers WH, Gros DB. Downregulation of connexin 45 gene products during mouse heart development. Circ Res. 1999; 84: 1365–1379.