Absence of Transverse Tubules Contributes to Non-Uniform Ca²⁺ Wavefronts in Mouse and Human Embryonic Stem Cell–Derived Cardiomyocytes

Abstract

Mouse (m) and human embryonic stem cell-derived cardiomyocytes (hESC-CMs) are known to exhibit immature Ca²⁺ dynamics such as small whole-cell peak amplitude and slower kinetics relative to those of adult. In this study, we examined the maturity and efficiency of Ca²⁺-induced Ca²⁺ release in m and hESC-CMs, the presence of transverse (t) tubules and its effects on the regional Ca²⁺ dynamics. In m and hESC-CMs, fluorescent staining and atomic force microscopy (AFM) were used to detect the presence of t-tubules, caveolin-3, amphiphysin-2 and colocalization of dihydropyridine receptors (DHPRs) and ryanodine receptors (RyRs). To avoid ambiguities, regional electrically-stimulated Ca²⁺ dynamics of single ESC-CMs, rather than spontaneously beating clusters, were measured using confocal microscopy. m and hESC-CMs showed absence of dyads, with neither t-tubules nor colocalization of DHPRs and RyRs. Caveolin-3 and amphiphysin-2, crucial for the biogenesis of t-tubules with robust expression in adult CMs, were also absent. Single m and hESC-CMs displayed non-uniform Ca²⁺ dynamics across the cell that is typical of CMs deficient of t-tubules. Local Ca²⁺ transients exhibited greater peak amplitude at the peripheral than at the central region for m (3.50 ± 0.42 vs. 3.05 ± 0.38) and hESC-CMs (2.96 ± 0.25 vs. 2.72 ± 0.25). Kinetically, both the rates of rise to peak amplitude and transient decay were faster for the peripheral relative to the central region. Immature m and hESC-CMs display unsynchronized Ca²⁺ transients due to the absence of t-tubules and gene products crucial for their biogenesis. Our results provide insights for driving the maturation of ESC-CMs.

Introduction

Adult cardiomyocytes (CMs) are unable to regenerate cells lost through disease or old age due to their inability to proliferate. Cell replacement therapy is a promising treatment, but is greatly hindered by limited cell source. Human (h) embryonic stem cells (ESCs) that are pluripotent and exhibit high proliferative capacity can be an unlimited source for CMs. However, human embryonic stem cell-derived cardiomyocytes (hESC-CMs) are physiologically different from their adult counterpart. They are smaller in size and electrophysiologically unstable with inefficient excitation–contraction coupling due to a relatively underdeveloped sarcoplasmic reticulum (SR), the intracellular Ca²⁺ store [1–3]. These immaturities and heterogeneous population of CMs at various differentiation stages or inclusion of different cell types can be substrates of arrhythmias. A better understanding of the basic biology of hESC-CMs is crucial for developing strategies for facilitated maturation to improve both the clinical efficacy and safety.

The main function of ventricular CMs is to generate efficient contractions for circulation. This is highly dependent upon the intracellular Ca²⁺ concentration, [Ca²⁺]_i, that is available to bind to troponin C allowing actin–myosin cross-bridge cycling to generate contractions. This fast [Ca²⁺]_i increase depends on the efficiency of Ca²⁺-induced Ca²⁺ release (CICR) that involves an initial Ca²⁺ influx through the activated L-type Ca²⁺ channels or dihydropyridine receptors (DHPRs) followed by Ca²⁺-induced activation of the ryanodine receptors (RyRs) that results in Ca²⁺ release from the SR [4]. The efficiency of this positive feedback mechanism is dependent upon the diffusion distance between DHPRs and RyRs. Adult ventricular CMs circumvent this diffusion-limited problem by developing transverse (t) tubules, invaginations in the sarcolemmal membrane that concentrates DHPRs and brings them spatially close to RyRs residing on the SR membrane located deeper in the cytoplasm [5,6]. By physically minimizing the diffusion distance, RyRs in CMs even with large cross-sectional area can participate in CICR without a lag. The result is a synchronized, faster, and greater transient [Ca²⁺]_i increase from the peripheries to the center, creating a uniform Ca²⁺ wavefront across the transverse section with simultaneous recruitment of all SR, contrasting with a U-shaped Ca²⁺ wave propagation seen in a de-tubulated ventricular or atrial CMs [5], due to a time delay that is proportional to the diffusion distance squared in recruiting the Ca²⁺ stores at the cell center [7]. Fast and synchronized activation of RyRs translates into a greater Ca²⁺ transient amplitude, recruitment of more actin–myosin cross-bridge cycling, and generation of greater contractile force [5,8].

In this study, we examined the presence and the consequence of t-tubules on the colocalization of DHPRs with RyRs, and the resulting transient [Ca²⁺]_i spatiotemporal gradient in mouse (m) and hESC-CMs. Unlike several previous studies [1,9–11], electrical stimulation of isolated, single ESC-CMs rather than spontaneously beating clusters were studied to prevent gap junction-mediated electrical communications with neighboring cells that might contaminate genuine CICR mediated by DHPRs. Furthermore, while lipophilic staining is effective for identifying highly ordered arrangement of t-tubules in adult cells, debris or artifacts often show up as bright fluorescent spots that can be difficult to distinguish from genuine t-tubules if they are absent or have disorganized arrangements. Therefore, the highly sensitive technique atomic force microscopy (AFM) was employed to accurately probe the physical membrane topology of m and hESC-CMs. With AFM, debris would show up as elevations rather than invaginations in the membrane. We found that t-tubules were absent in m and hESC-CMs. These immature cells also did not exhibit colocalization of DHPRs and RyRs. Crucial proteins for t-tubule biogenesis were also absent. The resulting [Ca²⁺]_i change was a faster and higher magnitude of increase at the cell periphery than the central region similar to the immature fetal or neonatal CMs lacking t-tubules.

Materials and Methods

mESC culture and differentiation into mESC-CMs

D3 mESCs were cultured as previously described [12]. Briefly, mESCs were co-cultured with mouse embryonic fibroblasts (MEF) in Dulbecco's modified Eagle's medium (DMEM) (Cat. #11965; Invitrogen, Carlsbad, CA) supplemented with 15% fetal bovine serum (Hyclone), 1% nonessential amino acids (Invitrogen), 1 mmol/L l-glutamine (Invitrogen), 0.1 mmol/L β-mercaptoethanol (Invitrogen), 0.5 U/mL penicillin (Invitrogen), 0.5 µg/mL streptomycin (Invitrogen), and 10 ng/mL leukemia inhibitory factor (Sigma) at 37°C with 5% CO₂. Approximately 800 cells per 20 µL medium of hanging drops were allowed to incubate for 2 days to form embryoid bodies, and then grown in suspension for an additional 5 days before plating on 0.1% gelatin-coated Petri dishes. After 16–20 days post-differentiation, mESC-CMs were harvested and digested into single cells with collagenase II prior to plating for assessment of t-tubule formation, DHPRs colocalization with RyRs, presence of t-tubule biogenesis proteins, and local Ca²⁺ transients.

hESC culture and differentiation into hESC-CMs

hESCs, H1 line (WiCell) were cultured on MEF in DMEM/F12 (Invitrogen) supplemented (all from Invitrogen) with 15% knockout serum, 1% nonessential amino acids, 1 mmol/L l-glutamine, 0.1 mmol/L β-mercaptoethanol, 0.5 U/mL penicillin, 0.5 µg/mL streptomycin, and 4 ng/mL basic fibroblast growth factor at 37°C with 5% CO₂. To initiate differentiation into hESC-CMs, hESC colonies were grown in suspension in low-attachment plates for 7 days in differentiation medium with the same composition as the mESC medium without β-mercaptoethanol. After 7 days of suspension, hESC embryoid bodies were plated in 0.1% gelatin-coated Petri dishes. After 40–50 days post-differentiation, hESC-CMs were identified by beating clusters and manually dissected out, then plated after digestion into single cells with collagenase II for assessment of t-tubule formation, DHPRs colocalization with RyRs, presence of t-tubule biogenesis proteins, and Ca²⁺ transients.

t-tubule fluorescent staining

For fluorescent t-tubule staining, m and hESC-CMs were incubated in 10 µM Di-8-ANEPPS (Invitrogen) for 10 min at room temperature, then washed for 10 min with phosphate-buffered saline (PBS). Images at the midplane of the cell height were taken with a Nikon laser scanning confocal microscope with a 60× oil immersion objective.

t-tubule detection with AFM

To probe the membrane topology, a Bioscope II (Veeco Metrology, Santa Barbara, CA) AFM with Microlever silicon nitride probe (MLCT-AUNM) tip D (spring constant [k] of 0.03 N/m) was used in contact mode to raster scan the surface of 0.05% glutaraldehyde-fixed m and hESC-CMs at room temperature in an aqueous environment at a scan speed of 0.5–0.7 Hz with 512 × 512 spatial resolution for a 10 µm region of interest. Images were generated either in 2D with cell topology variation shown in a gray scale or as a rendered 3D surface map using Nanoscope® Software.

Immunostaining of mESC-CMs and hESC-CMs

mESC-CMs and hESC-CMs were fixed with 4% paraformaldehyde for 15 min at room temperature and permeabilized with 0.2% saponin for 10 min at room temperature. Prior to overnight primary antibody incubation at 4°C, the cells were blocked in 1% BSA and 10% goat serum for 1 h at room temperature. DHPRs were labeled with rabbit Cav1.2 antibody (Sigma, St. Louis, MO) at 1:100. RyRs were labeled with mouse anti-RyR (Affinity Bioreagent) at 1:100. Caveolin-3 (Santa Cruz Biotechnology, Santa Cruz, CA) and amphiphysin-2 (Santa Cruz Biotechnology) antibody concentrations were both 1:50. For DHPRs and RyRs staining, Alexa fluor 488 goat anti-rabbit IgG and Alexa fluor 546 goat anti-mouse IgG (Invitrogen, Carlsbad, CA) were used, respectively. Caveolin-3 and amphiphysin-2 slides were incubated in Alexa fluor 488 goat anti-mouse IgG (Invitrogen) for 1 h at room temperature. Finally, all nuclei were counterstained with Hoechst 33342 (Invitrogen). Slides were mounted using Prolong Gold Anti-fade reagent (Invitrogen).

Confocal Ca²⁺ imaging

To detect fast local Ca²⁺ changes in ESC-CMs, a spinning disk confocal microscope (Yokogawa CSU10) with high sampling frequency was used. Measurements using confocal microscopy allows for temporal as well as spatial [Ca²⁺]_i changes of m and hESC-CMs rather than global whole-cell [Ca²⁺]_i changes. Single m or hESC-CMs were incubated for 30 min at 37°C in 5 µM Fluo-4 (Invitrogen) with 0.02% pluronic (Invitrogen) in Hank's Buffered Salt Solution (Invitrogen). Single-cell Ca²⁺ transients of m or hESC-CMs at ∼180 frames per second with a 40× microscope objective, stimulated by electric field of 40 V with 90 ms pulse duration at 0.2 Hz, were recorded at room temperature in Tyrode's solution consisting of (mM): 140 NaCl, 5 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES at pH 7.4. The Ca²⁺ transient changes were quantified by the background subtracted fluorescent intensity changes normalized to the background subtracted baseline fluorescence using Image J. Specifically, the transient Ca²⁺ increases at the periphery and center of the cell due to electrical stimulations were compared. Paired Student's t-test analysis yielding P < 0.05 was considered to be statistically significant.

Results

mESC-CMs and hESC-CMs lack organized t-tubules that are present in adult CMs

To detect t-tubules, we initially used a fluorescent lipophilic dye, Di-8-ANEPPS, to stain three separately differentiated batches of >30 non-permeabilized m and hESC-CMs, and then imaged at their midplane in the z-axis using a laser scanning confocal microscope. Both m and hESC-CMs did not display brightly labeled spots in the central region of the cells (Fig. 1A and C), unlike our positive control of mature guinea pig ventricular CMs displaying an organized array of bright spots in the cellular midplane that are typical of CMs with t-tubules (Fig. 1E). For m and hESC-CMs, only their cellular periphery displayed bright fluorescence, indicative that the dye did bind to the cell membrane and that there was no presence of invaginated lipid membrane in their central region, thus, no organized t-tubules were present. To confirm these observations, we next physically scanned the three batches of >15 m and hESC-CM surface topography with AFM. Indeed, the surfaces of m and hESC-CMs were smooth with occasional appearance of the cytoskeleton beneath the sarcolemma in the AFM images (Fig. 1B and D). In contrast, adult guinea pig ventricular CMs known to have t-tubules exhibited invaginations that appeared periodically every 2 µm and coincided with the z-lines (Fig. 1F).

FIG. 1.

T-tubule imaging of a mouse embryonic stem cell-derived cardiomyocyte (mESC-CM), a human ESC-CM (hESC-CM), and a mature ventricular CM. Di-8-ANEPPS confocal microscopic images of an mESC-CM (A) and an hESC-CM (C) did not show intracellular fluorescent spots like those in an adult guinea pig ventricular CM (E), suggesting the absence of t-tubules. The absence of t-tubules in ESC-CMs was further confirmed by AFM imaging of an adult guinea pig ventricular cardiomyocyte (F), showing regularly spaced pores in the sarcolemma that coincide with the z-lines, while mESC-CM (B) and hESC-CM (D) surface showed comparatively smoother topology with no presence of invaginations that are indicative of t-tubules.

DHPRs and RyRs do not colocalize in mESC-CMs and hESC-CMs

Double-staining of m (Fig. 2A) and hESC-CMs (Fig. 2B) for DHPRs and RyRs showed that the two crucial Ca²⁺-handling proteins did not colocalize. Clearly, dyads that are the signature of adult ventricular CMs were not observed. DHPRs were scattered diffusely throughout the derived CMs without any distinct pattern. RyRs were cytoplasmic and tended to localize around the periphery of the nuclei. This lack of colocalization was consistent with our findings that t-tubules were absent in m and hESC-CMs.

FIG. 2.

Double immunostaining of DHPRs and RyRs in mouse and human embryonic stem cell-derived cardiomyocytes (mESC-CMs and hESC-CMs). DHPRs and RyRs showed lack of colocalization in mESC-CMs (A) and hESC-CMs (B). All cells were nuclear counterstained with Hoechst 33342.

Caveolin-3 and amphiphysin-2 are absent in mESC-CMs and hESC-CMs

Since t-tubules were largely absent in both m and hESC-CMs, we next investigated the expression of two gene products, caveolin-3 and amphiphysin-2, that have been implicated to play crucial roles in t-tubule biogenesis. Figure 3 shows that neither caveolin-3 nor amphiphysin-2 could be detected in m and hESC-CMs. In stark contrast, adult ventricular CMs consistently showed intense positive staining for both.

FIG. 3.

Immunostaining of caveolin-3 and amphiphysin-2 in mouse embryonic stem cell-derived cardiomyocytes (mESC-CMs), human ESC-CMs (hESC-CMs), and mature adult CMs. Caveolin-3 was absent in both mESC-CMs (A) and hESC-CMs (B). Adult ventricular CMs were stained as positive control (C). Amphiphysin-2 was also deficient in mESC-CMs (D) and hESC-CMs (E). Amphiphysin-2 was also stained in adult ventricular CMs as positive control (F). All cell nuclei were counterstained with Hoechst 33342.

mESC-CMs and hESC-CMs show non-uniform electrically-induced Ca²⁺ wavefronts

To explore the functional consequences due to the absence of t-tubules, single-cell local transient intracellular Ca²⁺ changes due to electrical stimuli were recorded by spinning disk confocal microscopy in Fluo-4-loaded m and hESC-CMs. Electrical stimulations of single ESC-CMs avoided gap junction-mediated electrical communications in spontaneously beating clusters. Time-lapsed images acquired at the z-axis midplane of a cell by confocal microscope were analyzed by converting the fluorescence from cell border to border into a spatial linescan versus time pseudo-colored for Ca²⁺ level to illustrate the Ca²⁺ wavefront propagation.

Both m and hESC-CMs exhibited a U-shaped Ca²⁺ wavefront typical of t-tubule-deficient CMs with a delayed increase in the central region of the cell relative to the peripheral region (Figs. 4A and 6A). These increases in [Ca²⁺]_i in m and hESC-CMs were lower in the central relative to the peripheral region as depicted by the color scale. Such U-shaped Ca²⁺ wavefronts were clearly different from the uniform wavefronts measured in adult ventricular CMs (Fig. 5A). This phenomenon of non-uniform wavefronts in immature mESC-CMs was also observed in coupled mESC-CMs (Fig. 5B). To quantify the regional differences in the fluorescence changes due to temporal Ca²⁺ fluxes, the peak amplitude, the maximum rate of intracellular Ca²⁺ increase, and the decay constant fitted to an exponential function for the Ca²⁺ transients were analyzed and compared (Figs. 4B,C and 6B,C). For mESC-CMs, their averaged peak Ca²⁺ transient amplitude was greater for the periphery relative to the center at 3.50 ± 0.42 and 3.05 ± 0.38, respectively (P < 0.01; n = 12). Their maximum transient rate of rise (V _rise) was faster at the peripheral (159 ± 39 s⁻¹) than the central region (130 ± 32 s⁻¹; P < 0.01). Their transient decay was also faster at the periphery as shown by their decay time constant (τ) at 0.88 ± 0.08 s⁻¹ and 1.13 ± 0.10 s⁻¹ for the central region (P < 0.05). Similar to mESC-CMs, the peak Ca²⁺ transient amplitude for the hESC-CMs at the periphery and the center were 2.96 ± 0.25 and 2.72 ± 0.25, respectively (n = 11; P < 0.05). Their transient rate of rise (V _rise) was also faster at the periphery (118 ± 18 s⁻¹) than the central region (96 ± 16 s⁻¹; P < 0.05). The transient decay was noticeably faster for the periphery (τ =1.34 ± 0.20 s⁻¹ vs. 1.84 ± 0.48 s⁻¹ for the central region), although this difference did not reach statistic significance (P = 0.23).

FIG. 4.

Electrically-induced transient Ca²⁺ increase in mouse embryonic stem cell-derived cardiomyocytes (mESC-CMs). (A) Time progression linescans of pseudo-colored transient increase in intracellular Ca²⁺ across the midplane of a mESC-CM showed a U-shaped wavefront typical of CMs lacking t-tubules. (B) Quantified Ca²⁺ transient of linescans in A. The arrow indicates time of pulsed electrical stimulus. (C) The peak amplitude increase was statistically greater for the periphery relative to the center (n = 12). The transient rate of rise (V _rise) and the transient decay as shown by the decay time constant (τ) were both statistically faster at the peripheral than the central region. Data shown as mean ± SEM.

FIG. 6.

Electrically-induced transient Ca²⁺ increase in human embryonic stem cell-derived cardiomyocytes (hESC-CMs). (A) Time progression linescans of pseudo-colored transient increase in intracellular Ca²⁺ across the midplane of a hESC-CM showed a U-shaped wavefront typical of CMs lacking t-tubules. (B) Quantified Ca²⁺ transient of linescans in A. The arrow indicates time of pulsed electrical stimulus. (C) The peak amplitude increase was statistically greater for the periphery relative to the center (n = 11). The transient rate of rise (V _rise) was statistically faster at the periphery than the central region. The transient decay was faster for the periphery as shown by their decay time constant (τ), but the difference did not reach statistical significance. Data shown as mean ± SEM.

FIG. 5.

Electrically-induced transient Ca²⁺ increase in adult guinea pig ventricular cardiomyocytes (CMs) and coupled cluster of mouse embryonic stem cell-derived CMs (mESC-CMs). (A) Time progression linescans show uniform Ca²⁺ wavefronts in mature adult ventricular CMs typical of CMs with t-tubules. (B) Time progression linescans of mESC-CMs in coupled cluster exhibited the same U-shaped Ca²⁺ wavefronts as the single mESC-CMs. The arrows indicate time of electrical stimulation.

Discussion

Self-renewal and pluripotent ESCs that can differentiate into all cell types of the embryo proper is a potential unlimited source to derive CMs for cell replacement therapy; however, ESC-CMs are phenotypically and functionally immature relative to their adult counterparts [2,3,9,11,13,14]. We have previously shown at the whole-cell level that the CICR mechanism is not as efficient as that of adults due to the immature SR unable to fully contribute to the overall elevation of [Ca²⁺ ]_i [1]. In this study, we focused on t-tubules and local regional differences in Ca²⁺ transients.

t-tubules are invaginations about 200–300 nm in diameter occurring every ∼2 µm at the sarcomere z-lines in the sarcolemmal membrane of adult mammalian ventricular CMs [5,6]. These structures concentrate the DHPRs and bring them spatially closer to the RyRs embedded in the SR membrane. Using fluorescent staining and AFM, we showed that t-tubules were absent from both m and hESC-CMs. Although some t-tubules have been detected in mESC-CM 28 days post-differentiation [15] and some or none in hESC-CMs 30 to 40 days post-appearance of beating areas by TEM [10,16], it is nonetheless clear that the frequency and the pattern of their occurrence do not compare to the extent seen in adult ventricular CMs. Therefore, generally the t-tubule development is still in its early stage like those of fetal or neonatal CMs [17,18]. Of note, any comparison of m and hESC-CMs should take into consideration any differences in their absolute ages and the relative gestational periods between the species. The time point that we have chosen for our mESC-CMs is 16–20 days, which is within the range of what are considered to be late stage in the literatures [13,19]. For hESC-CMs, the time point that we have chosen of 40–50 days was motivated by data of different stages in the literature and our own observations. Early (15–30 days) and late (55–110 days) stage hESC-CMs have been compared by Sartiani et al. [2]. Based on their data, there was not a significant change in the electrophysiological profile between days 57 and 110. Even between 25 and 57 days post-differentiation, various ion channel expression levels were already pretty similar. This means that an additional 50 days of culturing time did not promote significant improvements in hESC-CM maturity. Moreover, based on our observation, appearance of MLC2v-positive or ventricular CMs, which is the only CM type with presence of t-tubules, increased about 35 days post-differentiation in the beating areas. Hence, we have chosen a sampling time after appearance of ventricular hESC-CMs. Finally, we have decided to study hESC-CMs that have been in culture for a reasonable amount of time, but not too long for practicality of clinical application. If more mature hESC-CMs are more suitable for clinical practice, but these cells cannot reach the necessary maturity in a reasonable amount of culturing time, then facilitated induction of maturation may be a direction that needs to be explored.

Our stainings of m and hESC-CMs demonstrated no colocalization of DHPRs with RyRs, showing intracellular and periphery DHPRs and cytoplasmic RyRs that preferentially localize around the cell nuclei. This perinuclear pattern of RyRs was also observed by Sauer et al. in mESC-CMs at 5 to 11 days post-differentiation that progressed to the entire cytoplasmic compartment by the late stage [11]. This is consistent with our findings that t-tubules are absent in both m and hESC-CMs. Without t-tubules, there is no structure present to confine the DHPRs in a specific region and bring them spatially close to the RyRs residing on the SR membrane, leading to the lack of localization. Double-staining of DHPRs and RyRs in fetal rat CMs similarly exhibited no colocalization, with intracellular DHPRs that have been theorized to be in route to the subsarcolemma [17].

Caveolin has long been implicated to play a role in the formation of caveolae. T-tubule formation was thought to begin with the formation of caveolae chains that are unable to undergo fission [20,21]. Caveolin-3, an isoform unique to muscle cells, is an integral membrane protein with a hairpin structure that can form oligomers and has the ability to bind to cholesterol to form highly ordered structures for spatially confining integral proteins [20,21]. Amphiphysin-2 localized in the t-tubules is another membrane-associated protein with a BAR domain at the N-terminus, which is named for its presence in the vertebrate Bin1 and Amphiphysins and in the Saccharomyces cerevisiae Rvs proteins [22], that can sense the membrane curvature, binds and evaginates the membrane even in nonmuscle cells [22–24], suggesting that amphiphysin-2 may be an initiator in t-tubule formation. Neither caveolin-3 nor amphiphysin-2 was expressed in m and hESC-CMs; our laboratory is currently investigating whether their forced expression may drive structural maturation to facilitate Ca²⁺ handling.

For m and hESC-CMs, the periphery displayed statistically greater peak amplitude increase and faster rise time and exponential decay constant relative to the center. We and Satin et al. have also observed similar regional differences in small clusters of m and hESC-CMs [10]. Similarly, asynchronous Ca²⁺ wavefronts have been observed in immature fetal or newborn ventricular CMs [18]. The observed slower rate of rise and smaller amplitude increase are due to the time needed for the Ca²⁺ influx through DHPRs to diffuse from the border to recruit the RyRs residing on the SR membrane located deep inside the cytosol to activate the CICR mechanism. Although statistical significance has been reached, our regional differences are less than those seen in fetal or neonatal, atrial or de-tubulated adult ventricular CMs [17,18,25–27]. The smaller differences may be due to the intrinsically smaller cell size of ESC-CMs. This small cross-sectional area or shorter diffusion distance from the cell border to the center is further diminished by the 2D culture condition that is routinely in practice. Planer culture condition has been shown to result in cellular polarity, such that the colocalization pattern of HSP90 and eNOS differs between 2D and 3D culture environments due to the free cellular surface lacking cell–cell interaction in the 2D condition [28]. We chose to detect t-tubules in single cells to match the state of ESC-CMs in our Ca²⁺ imaging experiments. With ESC-CMs randomly oriented and not exhibiting rod-shape morphology like the fetal CMs in vivo, we had typically selected lines straight across a cell in the XY-plane that had the shortest distance; however, the diffusion distance in these immature cells with small cross-sectional area is furthered diminished by the 2D culture environment. Their shortest diffusion distance was actually in the z-axis or the cell height with typical distance of ∼6 µm relative to our linescan recording across the cells at 18.0 ± 1.4 and 16.6 ± 1.6 µm for mESC-CM and hESC-CMs, respectively. Therefore, although how we chose the region of interest is important, the more critical point is that the Ca²⁺ in the central region most likely diffused in from the top of the cells rather than from the sides. Since the diffusion time is proportional to the diffusion distance squared [7], the diffusion-limited effect may not be as pronounced at the current diffusion distance for ESC-CMs with a small cross-sectional area compared to that of more mature CMs in vivo with cross-sectional area of 10 µm or more. Thus, further investigations of the effect of 3D culturing are warranted, and may lead to optimized culturing methods to facilitate maturation for clinical applications. Another experimental limitation is that all Ca²⁺ transients were elicited with a field stimulator that may excite multiple sites rather than point stimulation with a micropipette; however, our conclusion remains valid since the spread of depolarization depends on the intrinsically faster Na⁺ channels with a lower activation threshold than the DHPRs, followed by the activation of DHPRs that triggers the CICR process.

Although hESC-CMs have been shown to functionally integrate with resident CMs in vitro and in vivo [29], the transplanted ESC-CMs with asynchronous Ca²⁺ wavefronts or weaker contractile force may not augment cardiac contraction enough to improve blood circulation. Moreover, the weak contractile force may create a mechanical strain imbalance for the resident CMs at the interface between the native and the newly introduced ESC-CMs, which could result in the advancement of pathological hypertrophy in the heart. Therefore, improvements in the excitation–contraction coupling such as t-tubules and the contractile force in ESC-CMs are desirable.

Interestingly, CMs with smaller cross-sectional areas such as mature avian or fish and immature fetal or neonatal mammalian CMs do not possess t-tubules [30]. Even those atrial cells that do possess irregular internal transverse-axial tubular system, a modified t-tubule system, tend to have a greater mean diameter than those that do not [31]. The lack of t-tubules in these CMs may be dependent on the cell cross-sectional area and/or the diffusion distance. Therefore, whether t-tubules development is triggered by an increase in cell size to compensate for the increase in diffusion distance or initiation of t-tubule development allows for cellular hypertrophy is an interesting issue that may shed important insights for deriving more functionally viable, mature CMs for cell transplantation or drug screening.

Acknowledgments

This work was supported by the National Institutes of Health (R01-HL72857 to R.A.L.), the California Institute of Regenerative Medicine (to R.A.L.), the Stem Cell Program of the University of California, Davis (to R.A.L.), and Research Grant Council and CC Wong Stem Cell Fund (to R.A.L. and H.F.T.). D.K.L. was supported by a fellowship from the NSF Center for Biophotonics Science & Technology and a fellowship from Shriners Hospital for Children. The Center for Biophotonics, an NSF Science and Technology Center, is managed by the University of California, Davis, under Cooperative Agreement No. PHY 0120999. C.W.S. and A.A.K. were supported by a Croucher Fellowship and a T32 training grant (5TL1RR024145-02), respectively, during the tenure of this project.

Author Disclosure Statement

No competing financial interests exist.