Calcium signaling in human stem cell-derived cardiomyocytes: Evidence from normal subjects and CPVT afflicted patients

Abstract

Derivation of cardiomyocyte cell lines from human fibroblasts (induced pluripotent stem cells, iPSCs) has made it possible not only to investigate the electrophysiological and Ca²⁺ signaling properties of these cells, but also to determine the altered electrophysiological and Ca²⁺-signaling profiles of such cells lines derived from patients expressing mutation-inducing pathologies. This approach has the potential of generating in vitro human models of cardiovascular diseases where cellular pathology can be investigated in detail and possibly specific pharmacotherapy developed. Although this approach has been applied to a number of mutations in channel proteins that cause arrhythmias, there are only few detailed reports addressing Ca²⁺ signaling pathologies beyond measurements of Ca²⁺ transients in intact non-voltage clamped cells. Unfortunately full understanding of Ca²⁺ signaling pathologies remains elusive, not only because of the plethora of Ca²⁺ signaling proteins defects that cause arrhythmias and cardiomyopathies, but also because detailed functional properties of Ca²⁺ signaling proteins are difficult to obtain. Catecholaminergic polymorphic ventricular tachycardia (CPVT1) is a malignant inherited arrhythmogenic disorder predominantly caused by mutations in the cardiac ryanodine receptor (RyR2). Thus far over 150 mutations in RyR2 have been identified that appear to cause this arrhythmia, a number of which have been expressed and studied in transgenic mice or cell-line models. The development of human iPSC-technology makes it possible to create human heart cell-lines carrying these mutations, making detailed identification of Ca²⁺ signaling defects and its specific pharmacotherapy possible.

In this review we shall first briefly summarize the essential characteristics of the mammalian cardiac Ca²⁺ signaling, then compare them to Ca²⁺ signaling phenotypes of human induced pluripotent stem cell derived cardiomyocytes (hiPSC-CM) and to those of rat neonatal cardiomyocytes, and categorize the possible variance in Ca²⁺ signaling defects caused by different CPVT-inducing mutations as expressed in hiPSC-CMs.

Graphical Abstract

1. Calcium Signaling in adult Cardiomyocytes

Following seminal experiments of Ringer some 150 years ago it became apparent that Ca²⁺ played a central role in initiation and regulation of cardiac contractility, but It was not until the 1960s that the advent of technologies of voltage-clamp for ventricular strips [1, 2], the development of Ca²⁺ sensitive dyes and the whole-cell patch-clamp technique for single cells [3,4] made it possible to explore the intricate details of cardiac Ca²⁺ signaling. The picture that emerged suggested that the entry of Ca²⁺ through the Ca²⁺ channel, during the action potential, triggers much larger releases of Ca²⁺ from the RyR2 receptors [5] located on the dyadic junctions of sarcoplasmic reticulum (SR), setting in motion the activation of contraction (Fig. 1). This process became known as Ca²⁺-induced Ca²⁺-release (CICR, [6]), a mechanism that seems to have evolved specifically in mammalian heart versus a faster voltage-gated Ca²⁺ release (VICR) mechanism in skeletal muscle. The advent confocal Ca²⁺ imaging in line-scan mode made it possible to capture focal rise of Ca²⁺ in the dyadic micro-domains between SR and t-tubular membranes, known as Ca²⁺ spark [7]. Ca²⁺ spark are sporadic events that occur during diastole or are triggered by activation of I_Ca. Cleemann et al. [8] using 2-D rapid (240 f/s) confocal Ca²⁺ imaging in whole cell clamped ventricular myocytes dialyzed with high concentrations of EGTA to limit the diffusion of Ca²⁺-bound Fluo-3, imaged simultaneously over 100s of sparks and found spark durations of ~7ms, and spark amplitudes that were invariable with voltage, but spark frequencies that were dependent on the magnitude of I_Ca. These studies also showed that a Ca²⁺ spark represented the release of ~100,000 calcium ions, resulting most likely from activation of ryanodine receptor clusters in a dyadic junction. Consistent with this idea the sparks occurred at the same cellular location with over 70% consistency from beat to beat (8).

Figure 1.

Schematic of cardiac Ca²⁺ signaling. Top panel show Ca fluxes (green arrows) generated by dyadic junctions (box), Na⁺/Ca²⁺ exchangers (NCX, I_NCX), sarco- and endoplasmic reticular Ca²⁺ ATPase (SERCA), inositol triphosphate receptor (IP₃R), and mitochondria in relation to the cell membrane, the transverse tubular system (t) and the sarcoplasmic reticulum (SR). All are shown superimposed on the contractile filaments to signify the regulation of cellular contraction and relaxation by cytosolic Ca²⁺. The lower diagram shows key proteins within dyadic junctions where clusters of Ca²⁺ release channels (ryanodine receptors, RyR2) in the SR membrane are functionally and structurally associated with Ca²⁺ channels (DHP receptors, L-type Ca²⁺ current, I_Ca) in t-tubules, calsequestrin (CASQ2) in SR and, triadin (TRDN), junction (JNCT), and FKBP (calstabin) in association with RyR2.

It is important to note that although CICR is a ligand-gated process, it has significant voltage-dependence [9], such that the gain (amplification factor) of the process may exceed 100 at −40 to −30 mVs, but it drops to values below 15 at voltages positive to 10 mVs, [10]. The mechanism for this unexpected property of CICR remains somewhat controversial, but experiments on rat atrial myocytes using rapid 2-D (240 f/s) confocal imaging have identified two groups of RyR2: those near the surface membrane, associated with dihydropyridine receptors (DHPRs), showing significant voltage-dependence and those centrally located, not associated with DHPRs, showing no voltage-dependence [11]. These findings imply that the voltage-dependence of gain of CICR reflects possible direct interaction of RyR2 with Ca²⁺ channels. In support of this possibility, cellular introduction of LA-fragments of cloned carboxyl-tail of alpha subunit of Ca²⁺ channel into freshly isolated rat atrial myocytes imparted voltage-dependence to the CICR of centrally located “naked” RyR2. This finding was specific to LA, but not to the K fragment of carboxyl-tail that includes the I-Q motif [11].

Calcium signaling during evolution appears to have evolved from a surface-membrane (Ca²⁺-channel/ NCX) dominated process, as in amphibian & shark hearts [2, 12], to an intracellular regulated system, in mammalian hearts, where the Ca²⁺ channel serves as a trigger to release of Ca²⁺ from an internal Ca²⁺ stores of SR, and where Ca²⁺ is sequestered via SERCA2a/Phospholamban proteins and stored on Ca²⁺-binding protein calsequestrin [13]. In mammalian hearts with morphological changes that accompany growth and differentiation of the heart that includes the development of t-tubules, SR, adrenergic receptors, and myofilaments, there appears also to be a shift from Ca²⁺ entry across the membrane to Ca²⁺ release from intracellular stores.

In addition to the fast I_Ca-gated CICR mechanism, a slower IP₃-gated Ca²⁺ release mechanism appears also to have evolved in the mammalian hearts [14]. This signaling pathway appears to be more dominant in immature developing heart, possibly playing a critical role in nuclear signaling during growth and development [15–18]. The role of IP₃-gated signaling in EC-coupling of healthy adult hearts remains somewhat cloudy, but significant data supports over expression of IP₃-gated signaling pathway in a number of cardiac pathologies [19–21].

Mitochondria also appear to contribute to Ca²⁺ signaling processes of cardiac myocytes. Early studies on isolated mitochondria show that mitochondria sequester and release Ca²⁺ from and to the cytosol [22]. The sequestered Ca²⁺ is thought to stimulate production of nitric oxide (NO), which in turn regulates oxygen consumption, ATP production, and ROS generation [23]. Nevertheless, it remains still unclear whether mitochondria play a critical role in beat-to-beat Ca²⁺ regulation of the myocyte. Since mitochondria make up some 30–40% of the cardiac cell volume, their contribution to cellular Ca²⁺ signaling could be significant [24]. In support of direct role for mitochondria in cellular Ca²⁺ signaling, shear stress pulses (pressurized flow of solution) appeared to trigger Ca²⁺ transients that originate from mitochondrial Ca²⁺ pools, by a mechanism independent of CICR, that had mitochondrial pharmacology [25]. The notion of a rapid (~500 ms) mitochondrial Ca²⁺ release was also confirmed in different cardiac cell types (adult ventricular cells, rat neonatal cardiomyocytes and hiPS-CM), under voltage-clamp conditions, and with genetically targeted mitochondrial probes. This is somewhat surprising, but is supported by key observations that: a) shear stress pulses induce Ca²⁺ release only after Ca²⁺-loading; b) perinuclear and peripheral mitochondrial populations often produce simultaneous Ca²⁺ signals with opposite polarity (uptake vs. release); c) the propensity of mitochondria to release Ca²⁺ increases with Ca²⁺ loading (e.g. by Na⁺ withdrawal); and d) mitochondrial Ca²⁺ release may escape detection in whole-cell measurements, where mitochondrial Ca²⁺ release maybe counter-balanced by uptake signals in different populations of mitochondria [25, 26].

Cardiac Ca²⁺ signaling remains an active area of research as state-of-the-art technologies ranging from molecular, genetic, electrophysiological and imaging approaches become available. Nevertheless, though I_Ca-gated Ca²⁺ release is the dominant pathway for activation of contraction in the heart, other pathways that include transport of Ca²⁺ across the membrane or released from ER and mitochondria may critically modulate cardiac Ca²⁺ signaling depending on pathophysiological or developmental stage of the heart.

2. Ca²⁺ signaling in human fibroblast-derived cardiomyocytes (hiPSC-CM)

It is well documented that iPSC-CMs beating clusters display mixed ventricular, atrial, and pacemaker-like cell populations based on their action potential phenotypes [27–29]. Although research is still ongoing to develop protocols that generate purer populations of a specific type of myocyte, the effort to generate homogenous cell populations of atrial, ventricular or pacemaker myocytes, remains to be fully achieved. Characterization of iPSC-CMs has revealed that while molecular, electrophysiological, metabolic, and Ca²⁺ signaling properties of hiPSC-CM are remarkably similar to those of adult cardiomyocytes, the genetic expression pattern, the cellular morphology of hiPSC-CM, might be more similar to the neonatal instead of adult cardiac myocytes [30, 31]. Studies with human embryonic stem cell-derived cardiomyocytes have revealed that t-tubules are absent in these developing cells [31, 32]. Table 1 compares the molecular and functional properties of a number of hiPSC derived cardiomyocyte cell-lines reported thus far. A cursory survey of the reported data, suggests significant differences between different cell lines, even though detailed analysis of Ca²⁺ signaling parameters that includes: the level of expression of Ca²⁺ channels, I_Ca-gated Ca release and its gain from RyR2s of SR, the magnitude of SR Ca²⁺ store, and the level of functional expression of NCX and SERCA/PLB and their regulation by adrenergic agonists, or the role of mitochondria in hiPSC-CM Ca²⁺ signaling; is often missing.

Table 1.

Calcium signaling characteristic of various hiPSC-CM cell lines.

hiPSC-CM Cell-lines	mRNA	Immunostaining	Action Potential	ICa	SR Ca2+ content	INCX	SERCA	Spontaneous Ca2+ transients	Ca2+ sparks	IP3R
[69]	NKX2.5 MLC2v ACTN2 MYH6 GATA4 a	No data	No data	No data	++	No data	No data	++	No data	No data
[40]	RyR2, IP3R2, SERCA2a Cav1.2 PLB; CSQ	α-actinin & RyR2	No data	No data	++++	No data	Thapsigargin suppress spontaneous and caffeine induced Ca2+ transients	suppressed by nifedipine or ryanodine	No data	2-APB or U73122 suppress Ca2+ transients
[28]	NCX1, RyR2, SERCA2a, MHC	α-actinin & TPM	45% N-CM, 38% V-CM, 17% A-CM	No data	+	No data	No data	U-shape Ca2+ wave; rise of calcium in the periphery faster than center	No data	No data
[41]	No data	RyR2 & CSQ	No data	No data	+	No data	No data	++	No data	No data
[35]	No data	No data	54% V-CM, 22% N-CM, 24% A-CM	ICa, bell shape I-V	No data	No data	No data	No data	No data	No data
[33] [45]	No data	RyR2	No data	ICa, bell shape I-V	++++	++++	Thapsigargin suppress spontaneous Ca-transients	Suppressed by NCX or RyR2 blockers	Sporadic and brief	2-APB or U73122 slightly affect spontaneous beating
[29]	No data	α-actinin & α-MHC	61% V-CM, 17.4% A-CM and 21.6% N-CM,	No data	No data	No data	No data	Ca2+ increase first at periphery then propagate to centre	Stochastic and repetitive	No data
[34]	No data	No data	No data	ICa bell shape I-V	++++	++++	SERCA Ca2+ transport are functional	Slower time to peak and decay rates	No data	No data

CSQ: calsequestrin; PLB Phospholamban; MHC: Myosin heavy chain; RyR2: Ryanodine receptor type 2; NCX: Sodium Calcium exchanger; TPM Tropomyosin

2.1. L-type Ca²⁺ current expression and whole cell Ca²⁺ transients

L-type Ca²⁺ channels are critical in excitation-contraction coupling of both cardiac and skeletal muscles. Human or mouse iPSC-CMs express robust levels of L-type Ca²⁺ channels, the gating properties and pharmacology of which are similar to those reported for mammalian and human cardiomyocytes [33–35]. I_Ca is activated at potentials positive to −40 mV, reached its peak values of ~ 8 pA/pF at 0 mV, reversed direction becoming outward at potentials positive to +60 mVs (e.g. fig. 2A), and could be blocked by conventional L-type Ca²⁺ channel blockers. The developed Ca²⁺ transients [36, 37] had similar bell-shaped voltage-dependencies as that found for I_Ca ; i.e. I_Ca-gated Ca²⁺ release activated at −30 mV, peaked at ~ 0 mV, and triggered little or no release of Ca²⁺ at potentials positive to +60 mVs (Fig. 2A). Repolarization from potentials positive to 60 mVs back to the resting potentials also triggered significant release of Ca²⁺ consistent with the idea that the influx of Ca²⁺ and not the depolarization of membrane triggers the release of Ca²⁺, as proposed for the adult mammalian cardiomyocytes (Fig. 2A, red traces, [38]. Positive depolarizing potentials also activated significant influx of transmembrane Ca²⁺ on the Na⁺/Ca²⁺ exchanger (NCX), which appears to contribute significantly to a slower developing component of Ca²⁺ transients [33]. This possibility is supported by the finding that 1 μM nifedipine while completely blocking I_Ca, suppresses only the rapidly developing component of Ca²⁺ transients, suggestive of considerable contribution of Na⁺/Ca²⁺ exchanger to Ca²⁺ transients [33]. This property of hiPSC-CM is consistent with previous findings in developing and neonatal mammalian cardiomyocytes suggestive of a more significant role for Na⁺/Ca²⁺ exchanger in EC-coupling of developing cardiomyocytes [39]. It should be noted that other groups have reported that in intact, non-patch clamped single or small monolayer clusters of hiPSC-CMs, 1 μM nifedipine completely suppressed the cellular Ca²⁺ transients [40], suggestive of full dependence of Ca²⁺-transients on I_Ca. This finding is not necessarily contradictory to reports in whole cell clamped hiPSC-CM identifying both nifedipine-sensitive and insensitive components Ca²⁺ transients, as in intact cells blocking I_Ca would either suppress excitability fully or shorten the action potential to levels that could not support the longer depolarization periods required for the influx of Ca²⁺ on Na⁺/Ca²⁺ exchanger. Simultaneous recordings of Ca²⁺ transients and contractions from small clusters of iPS-CMs also show similar characteristic as those of adult mammalian cardiomyocytes [41]. Comparison of the Ca²⁺ handling of field-stimulated (0.5Hz) hiPSC-CMs to those of acutely-isolated adult primary ventricular rabbit and mouse cardiomyocytes under identical experimental conditions, also showed that their Ca²⁺ transient amplitudes were not significantly different than those of adult rabbit and mouse cardiomyocytes, whereas time to peak and decay rates of Ca²⁺ transients were significantly slower in hiPSC-CM [34].

Figure 2.

Comparison of I_Ca- and caffeine-induced Ca²⁺ signals in hiPSC-CM from a healthy control subject and a CPVT1-afflicted patient with point mutation F2483I in RyR2. Currents were recorded with a 150ms or 250 ms step depolarizations from holding potential of −40 mV in 10 mV steps to +60 mV. (A) and (B) Left part of panels shows representative I_Ca and fluorescence traces where depolarization to 0mV rapidly activates Ca²⁺ release (black traces) while clamp pulses to +60 or +70 mV only activate rises in Ca²⁺ on repolarization (red traces). The right panel shows the I–V curves for I_Ca and the corresponding Ca²⁺ Fluo-4 signal from control and CPVT hiPSC-CM. (C) Representative caffeine-induced NCX currents and corresponding fluorescence Ca²⁺ signal from control and mutant hiPSC-CM. (D) Average values of caffeine-activated Ca²⁺ signals (top) and I_NCX currents (bottom) in each group. (E) Confocal image of immunofluorescence labeled RyR2 (green) in a small cluster of control hiPS-CM with DAPI-labeled nuclei (blue). Modified from [33]

The quintessential property of the gain of CICR in cardiomyocytes is its voltage-dependence, (See above, [2, 12]). This characteristic of CICR appears also to be preserved in hiPSC-CM, suggesting that iPSC-derived cardiomyocytes have the phenotypic cardiac Ca²⁺ signaling characteristics, in sharp contrast to the neuronal CICR that shows no indication of voltage-dependence [42]. Thus, the available data on characterization of spontaneously- or depolarization-triggered Ca²⁺ transients suggests that hiPS-CMs express the cardiac-type excitation–contraction-coupling that closely resembles that found in the mammalian myocardium.

2.2. Functional SR Ca²⁺ store: SERCA AND NCX

Molecular studies show that Ca²⁺-handling proteins of the mammalian heart: RyR2, IP₃R2, SERCA2a, PLN are all expressed in hiPSC-CM [28, 40, 43]. Immuno-fluorescence studies suggest punctuates clustering of RyR2s near the surface membrane in hiPSC-CM (Fig. 2E), consistent with functional findings that Ca²⁺ influx via the L-type Ca²⁺ channels triggers the release of the SR Ca²⁺ stores [33]. Immuno-cyto-staining studies of RyR2 and sarcomeric alpha-actinin, in small mono-layered clusters of hiPSC-CM, show that sarcomeric alpha-actinin staining has a relatively disorganized striated sarcomeric pattern [44]. In another study, RyR2 expression was also detected throughout the cytosol with some myofilaments co-localization [40].

The magnitude SR Ca²⁺ stores of hiPSC-CM has been quantified using rapid and short applications of caffeine pulses. Caffeine concentrations of 3–5 mM triggered large global Ca²⁺ transients that activated inward I_NCX, of about 2.5 pA/pF (Fig. 2D) as compared to ~1.0 pA/pF in rat neonatal cardiomyocytes, confirming functionally the existence of large SR Ca²⁺ stores in hiPSC-CM [45]. The larger density of I_NCX activated on application of caffeine as compared to values of ~1.0 pA/pF in adult cardiomyocytes of rat, rabbit, mouse, and human hearts suggests either larger stores of Ca²⁺ accessed by caffeine in hiPSC-CM or a greater density of NCX in hiPSC-derived myocytes. Comparison of caffeine triggered Ca²⁺ transients in hiPSC-CM with adult rabbit and mouse cardiomyocytes, in another study [34], suggests equivalent magnitudes, but diverse I_NCX values of only 0.6 pA/pF in hiPSC- CM and 1.3pA/pF in rabbit myocytes. This diversity between the two sets of data may have resulted from either higher Ca²⁺ buffering concentrations, or developmental maturity of iPSC-CMs used. Consistent with this idea the same group also reported that the sequestration rate of the thapsigargin-sensitive caffeine-triggered store in hiPSC-CM was significantly slower than that of rabbit or mouse cardiomyocytes, [34]. Yet others have reported very small caffeine induced Ca²⁺ release stores even when compared to human embryonic stem cell derived myocytes [28]. Given the variation in technics of creating hiPSC-CM cell-lines, the level of their maturity, the degree of Ca²⁺ buffering and concentrations of Ca²⁺-sensitive dyes used in intact or whole-cell clamped myocytes, it is likely that the quantitative variations in Ca²⁺ signaling parameters result from both developmental and electrophysiological approaches used.

2.3. Calcium Sparks and sporadic focal Ca²⁺ releases

Our experiments, using rapid (120–240 f/s) 2-D confocal imaging, show that the majority of sporadic and brief Ca²⁺-sparks and focal releases in hiPSC-CMs had spatiotemporal properties analogous to those of adult rat cardiomyocytes, Ca²⁺ spark durations rarely exceeded 20 ms in control hiPSC-CM [33]. Others using line-scan imaging, have identified stochastic and repetitive Ca²⁺ sparks that re-occurred at the same site in hiPSC-CM. Ca²⁺ sparks appear to be predominately triggered by activation of L-type Ca²⁺ channels, and are enhanced in frequency on elevation of extracellular Ca²⁺. Ca²⁺ diffusion from the center of Ca²⁺ sparks to periphery appears often to be asymmetric [29].

2.4. IP₃-gated Ca²⁺ pools

Although the primary Ca²⁺-signaling pathway in adult mammalian cardiomyocytes occurs through activation of RYR2, in some myocytes and at different stages of their development there appears to be significant contribution from the IP₃-gated signaling pathway [46] [47]. IP₃R-gated Ca²⁺ release is expressed both in endoplasmic reticulum and the nuclear envelop [48]. The functional role of IP₃R in cardiomyocytes remains poorly understood. One study shows that IP₃R mRNA levels were about 50-fold lower than those of RYR2 mRNA in adult cardiac myocytes [14], nevertheless there is abundant expression of this protein in embryonic, neonatal, and adult atrial myocytes [18, 49–52]. In fact, it has been reported that the spontaneous electrical activity of embryonic pacemaker and stem cell-derived cardiomyocytes depends, in part, on IP₃-mediated Ca²⁺ release [53–55]. Since IP₃-gated Ca²⁺ signaling is reported to trigger both sub-sarcolemmal and perinuclear Ca²⁺ releases in developing cardiomyocytes [18], it is more likely that the sub-sarcolemmal Ca²⁺ release would activate the Na⁺/Ca²⁺ exchanger to depolarize the surface membrane, while Ca²⁺ released at the nuclear envelop might have higher likelihood to be sequestered by SERCA2a [56]. Immuno-cyto-staining studies also suggest that IP₃R is mostly distributed around the nucleus and that this pool may contribute to the modulation of Ca²⁺-signaling in hiPSC-CM [57] [40]. Consistent with this idea IP₃R antagonist 2-APB or phospholipase C inhibitor U73122 have been reported to significantly slow the kinetics and decrease the amplitude of whole cell Ca²⁺ transients [40]. We have not been able to confirm this finding in our hiPSC-CM and neonatal cardiomyocytes, and in sharp contrast find that Ca²⁺ release from the SR and mitochondria might be the primary mechanism for the spontaneous pacing activity. In our study U73122 only slightly affected the beating frequency of hiPSC-CM, suggesting that IP₃R Ca²⁺ signaling played a minor role in generation of Ca²⁺ transients and regulation of pacing [33].

2.5. Mitochondrial Ca²⁺ signaling

There is little or no information as yet on mitochondrial Ca²⁺ signaling in hiPSC-CM. The contribution of mitochondrial Ca²⁺ signaling to spontaneous beating activity of hiPSC-CM has been examined recently using genetically encoded mitochondrial probes targeted to mitochondrial subunit VIII of cytochrome C. These studies showed both release and uptake of Ca²⁺ by different mitochondrial populations in hiPSC-CM; generally, though not always, perinuclear mitochondrial population released Ca²⁺ while the peripheral mitochondria took-up Ca²⁺ (Fig. 3, traces and panels 1,2,3). FCCP, a mitochondrial uncoupler, at 50nM suppressed the spontaneous beating of hiPSC-CM and prolonged the relaxation time of caffeine-induced Ca²⁺ transients, suggestive of a direct role for mitochondria in the cycling of cytosolic Ca²⁺ and regulation of pacing, (fig 3, trace 4 and panel 4). Our results therefore support a novel role for rapid release of Ca²⁺ from mitochondria in generation and regulation of spontaneous pacing [33].

Figure 3.

Suppression of oscillatory mitochondrial Ca²⁺ signals in a spontaneously pacing control hiPSC-CM infected with the genetically engineered targeted, Ca²⁺ probe, myticam-E31Q [26]. (A) Tracings showing changes in regional mitochondrial Ca²⁺ signals (top) and I_NCX (bottom) before and 30 s after exposure to 50 nM FCCP. Increasing mitochondrial Ca²⁺ corresponds to downward deflection. (B) Images show baseline fluorescence (F₀) and color-coded regions of interest (ROI, with “n” indicating nuclei) and 4 differential fluorescence images (measured at the times indicated along the traces) showing repeatable mitochondrial Ca²⁺ signals before FCCP application (1, 2, 3) and smaller, more localized responses after (4). Modified from [45]

3. Ca²⁺ signaling profiles of catecholaminergic polymorphic ventricular tachycardia (CPVT1) in hiPSC-CM model: “from mice to men”

Two genetic forms of CPVT have been thus far identified: 1) CPVT1, with point mutations in RyR2, accounting for at least 50% of all cases, resulting from abnormal intracellular Ca²⁺ handling caused by autosomal dominant mutations in the RYR2 gene [58] [59]; and 2) CPVT2, with an autosomal recessive mutations in cardiac calsequestrin, CASQ2, isoform also resulting in Ca²⁺ signaling defect [60].

CPVT1 causing mutations have been extensively studied in transgenic mice models or transformed cell lines. Evidence from a number of labs suggest that CPVT1 results from PKA-mediated phosphorylation of RyR2/FKBP complex resulting in SR diastolic Ca²⁺ leak that produces local depolarizations that trigger DADs via activation of NCX [61]. It has also been proposed that SR Ca²⁺ overload and altered Ca²⁺ sensitivity of SR luminal sites, not related to FKBP modulation, can cause diastolic Ca²⁺ release from RyR2, leading to local depolarizations and arrhythmogenesis [62] These opposing views are in part fueled by multiplicity of point mutations that cause CPVT1, and the variety of genetically engineered mice and animal cell-lines used. The iPSC technology has made it possible to create human cardiomyocyte cell-lines that express CPVT1 causing mutations. The general approach has been to obtain dermal fibroblasts from patients expressing CPVT1 arrhythmias, reprogramming and driving them into hiPSC-CM lineage [27]. To date, over 150 mutations have been identified in RyR2 gene that causes CPVT1 [63] and 11 patient-specific CPVT1 hiPSC-CM cell lines have been generated and evaluated by several labs (Table 2). Since Ca²⁺ signaling pathway defects play critical roles in many cardiac pathologies, accurate knowledge as to the specific details of Ca²⁺ signaling defects in various CPVT1 models is required to provide insights into the mechanisms underlying the particular form of the disease and help design approaches for it in vitro pharmacotherapy, before applying the therapy to patients. Table 2 summarizes the Ca²⁺ signaling phenotypes of CPVT1 causing mutations as expressed in hiPSC-CM cell-lines, as depicted in the cartoon of RyR2 (Fig. 4). Only few studies appear to have attempted simultaneous measurements of membrane currents and Ca²⁺ signaling parameters beyond measurements of Ca²⁺ transients. The parameters required for full analysis of Ca²⁺ signaling must include: the density of I_Ca and I_NCX, the characteristics of I_Ca-gated Ca²⁺ release, the gain of CICR, Ca²⁺ storage capacity of the SR and its fractional release, the unitary properties of Ca²⁺ sparks and their frequency of occurrence, in addition to frequency of spontaneously triggered DADs or EADs, the generation of aberrant Ca²⁺ transients, and the alterations caused by application of isoproterenol. For instance, in hiPSC-CM from a patient expressing point mutation F2483I in RyR2, the caffeine-triggered Ca²⁺ transients and accompanying I_NCX were found to be significantly smaller than in control myocytes (Fig. 2D), consistent with a smaller SR Ca²⁺ store, possibly due to Leaky RyR2 [33]. Interestingly the smaller store of Ca²⁺ appeared to have been in part compensated for by a larger fractional release of Ca²⁺ triggered by I_Ca. Thus, the gain of CICR was higher in F2483I mutant expressing hiPSC-CM. As might be expected rapid electrical pacing, beta-adrenergic stimulation, or higher intracellular Na⁺ exacerbated Ca²⁺ handling abnormalities of F2483I mutant cells and increased the diastolic Ca²⁺ levels, thereby triggering spontaneously occurring Ca²⁺ transients [33]. Confocal Ca²⁺ imaging also showed that the F2483I mutants exhibit larger and more recurrent and spatially wandering sparks compared to sporadic and brief sparks of control iPSC-CM, especially under Ca²⁺-overload conditions. Often, the Ca²⁺-induced Ca²⁺-release events continued after repolarization in CPVT1-CMs suggesting slower decay of Ca²⁺ transients, Fig. 2B, consistent with the continued release of Ca²⁺. A large fraction (59%) of CPVT-CMs also showed slower frequency of spontaneous beating and 34% developed arrhythmia and DAD after Isoproterenol stimulation [27].

Table 2.

Characteristics of 11 hiPSC-CM cell lines expressing CPVT-causing mutations.

CPVT Cell-lines	Content Ca2+ store	Ca2+ sparks	L-type ICa	Beating frequency	Diastolic Ca2+	Aberrant Ca2+ transients & oscillations	β-Adrenergic responses
Exon3-del[66]	No data	No data	No data			+++++	Beating frequency Diastolic Ca2+
S406L [64]	No change	Bigger and longer	No data	No data	Comparable to control, with Iso	+++++	Spark frequency & decay time
R420Q [44]		No data	No data		with Iso	++ Iso: ++++++	59% insensitive to; 16% arrhythmogenic; 25% [Ca2+]i rise associated with Ca2+ transients
E2311D [65]	No data	No data	No data	No data	No data	+++++ Triggered from multiple foci	DAD frequency
P2328S [43] [66]		No data	No data		Comparable to control	+++++	Diastolic Ca2+
F2483I [27] [33]		Longer & wandering	Comparable to control	with Iso	with Iso	+++++	Beating frequency DAD
T2538R[66]	No data	No data	No data		No change	+++++	Beating frequency
M4109R [67]	No data	No data	No data	No data	No data	+++++	DAD Frequency
L4115F [66]	No data	No data	No data			+++++	Beating frequency
Q4201R [66]	No data	No data	No data	26		+++++	Beating frequency
V4653F [66]	No data	No data	No data			+++++	Beating frequency Diastolic Ca2+

Figure 4.

Diagram of RyR2 receptor showing key functional domains in relation to 11 CPVT1-causing mutations that are located from N-terminal to central domain to C-terminal (red circles), which have been studied in hiPSC-CM. Functional domains: Trans-membrane α-helices (M#), phosphorylation sites (P), and binding sites for FKBP, calmodulin (CaM) and protein kinase A (PKA/MKAP) are also indicated.

In the N-terminal segment, two CPVT1 causing point mutations have been reported in human cell lines, fig. 4. Novak et al. [44] reported that R420Q mutation exhibited immature ultra-structural features including poorly developed SR, poorly organized myofibrillar sarcomeric pattern, and higher mitochondrial expression. They also report 60% slower spontaneous beating rates as compared to control hiPSC-CM. Caffeine induced Ca²⁺ transients were smaller and shorter, consistent with smaller SR Ca²⁺ stores but with faster reuptake of intracellular Ca²⁺. Even though isoproterenol increased the amplitude and maximal rate of rise and fall of Ca²⁺ transients in control cells, three types of responses were found in R420Q mutant cells: 1) 59% of CPVT1-CM were irresponsive to isoproterenol; 2) 16% CPVT1-CM showed arrhythmogenic responses; and 3) in 25% of CPVT-CMs isoproterenol decreased Ca²⁺ transients and elevated diastolic Ca²⁺ levels, Table 2. In another N-terminal CPTV1-causing mutation, S406L [64], although cardiomyocytes appeared to have similar diastolic and systolic Ca²⁺ levels and comparable SR Ca²⁺ content under basal condition as control myocytes, in the presence of isoproterenol the diastolic Ca²⁺, the frequency of Ca²⁺ sparks, and the susceptibility to DADs and arrhythmia increased, even though systolic Ca²⁺ and SR Ca²⁺ content remained unaltered [64]. Interestingly, high frequency electrical pacing also increased the percentage of cells with abnormal Ca²⁺ handling. These authors report that dantrolene restored Ca²⁺ sparks frequency in CPVT mutant cells both under basal and adrenergic stimulated conditions, and abolished DAD and triggered arrhythmias [64]. Comparison of Ca²⁺ content of these two point mutations in the N-terminal domain suggests a decrease in R420Q, consistent with RyR2 leak hypothesis, and no change in S406L, possibly supporting increased luminal Ca²⁺ sensitivity of SR, Table 2.

Four hiPSC-CM cell lines have been thus far created carrying CPVT point mutations flanking FKBP domain in the cytosolic region of RyR2, Fig 4. In E2311D cell-line Ca²⁺ transients appeared to initiate from multiple loci and isoproterenol increased the number of sites triggering the spontaneous Ca²⁺ transients. Under basal conditions, 43% CPVT-CMs developed DAD during the diastolic depolarization phase. 12% of the cells also developed DADs even when paced at low frequencies of 0.5Hz, and the DAD frequencies increased significantly after β-adrenergic stimulation. Interestingly, CaMKII inhibitor KN-93 stabilized the Ca²⁺ activity and suppressed isoproterenol-induced DAD in this model [65]. The P2328S CPVT1 causing mutation cell-line [43] also displays aberrant Ca²⁺ cycling, decreased Ca²⁺ content, and a fractional Ca²⁺ release that was significantly higher both in control and isoproterenol-treated myocytes. In this cell-line both EAD and DAD were consistently recorded and their frequency was increased on adrenergic stimulation.

Penttinen et al. [66] also developed 6 CPVT cell lines with different mutations located in different parts of the RyR2 protein from N-terminal, to central, to cytoplasmic, to C-terminal domains (Fig. 4) that included: Exon-3 deletion and point mutations T2538R, P2328S, Q420I, L4115F and V4653F. All cell lines showed abnormalities in Ca²⁺ transients and spontaneous beating frequency both under basal conditions and adrenergic stimulation, but diastolic Ca²⁺ was only increased in P2328S CPVT-CMs. Unfortunately there is no data on Ca²⁺ content of SR or its fractional release in five of six mutant cell-lines, providing no mechanistic support for either of the two CPVT1 hypothesis, Table 2. Nevertheless, pharmacotherapy with dantrolene was effective in suppressing Ca²⁺ cycling abnormalities only in cells having the RyR2 mutations in the N-terminal and central regions, but not in transmembrane domains, consistent with the antiarrhythmic effect of this drug in patients with RyR2 mutations in the N-terminal or central regions of RyR2 protein [66].

Itzhaki et al. also generated a hiPSC-CM heterozygous cell line carrying point mutation at M4109R from a CPVT patient [67]. These myocytes also show abnormal Ca²⁺ transients and appeared to develop store-overload induced Ca²⁺ release (SOICR) at much lower Ca²⁺ concentrations than the healthy control cells. In this cell-line, DADs developed in most cells even when paced at 0.5 to 1 Hz. Adrenergic stimulation significantly increased the frequency and magnitude of DADs and flecainide and thapsigargin significantly suppressed the incidence of DAD [67].

4. Conclusion and Perspectives

Although there are many descriptive electrophysiological studies and measurements of Ca²⁺ transients in hiPSC-CM cell-lines, there are only few reports that have quantified the Ca²⁺ signaling parameters ad their regulation in whole-cell patch clamped myocytes. In few studies where hiPSC-CM cell-lines, created either from normal subjects or CPVT patients, were studied in greater detail using confocal and TIRF imaging in voltage-clamped cells, the data suggest that all the parameters of mammalian cardiac signaling (i.e., phenotypic bell-shaped voltage-dependence of I_Ca-gated Ca²⁺ release and its voltage-dependent gain, spontaneously triggered Ca²⁺ sparks, functional levels of NCX gene, β-adrenergic regulation of I_Ca, and SERCA2a/PLB complex) are expressed in human fibroblast-derived cardiomyocytes. It is, therefore, fair to conclude that hiPSC-CM can serve as a reliable model for human cardiomyocytes, but care must be exercised to characterize fully the various electrophysiological and Ca²⁺ signaling parameters of each new hiPSC-CM cell line before pharmacological and pathophysiological studies are undertaken.

Clearly, there appears to be significant variations on the degree of maturity of cell-lines generated by different labs that may underlie some of the observed differences in the reported Ca²⁺ content of SR, their beta-adrenergic modulation, the systolic and diastolic Ca²⁺ levels, and their spontaneous beating rates. It is not as yet clear whether such variations arise solely from developmental stages of created hiPSC-CM or are related to various cellular biological approaches and in vitro experimental conditions modifying the physiological signaling pathways.

This review also compares the Ca²⁺ signaling parameters of hiPSC-CMs of control subjects to those derived from patients expressing CPVT1, Table 2. The mutations expressed in various hiPSC-CM cell lines are located in the well-known CPVT hotspots of the RyR2 gene. All mutations appear to cause aberrant Ca²⁺ transients and arrhythmias. Ca²⁺ transient abnormalities were somewhat more common in mutations located in transmembrane domain than in central and N-terminal regions, and were preferentially suppressed by dantrolene [64, 66]. Exon 3 deleted hiPSC-CMs had both lower beating frequencies and diastolic Ca²⁺ levels when compared to other mutations, possibly resulting from altered secondary structures critical to folding of the N-terminal domain and its conformational changes [68]. There also appears to be significant variability in myocyte responsiveness to β-adrenergic stimulation in various cell lines expressing CPVT-inducing mutations. For instance, β-adrenergic agonists failed to increase the diastolic Ca²⁺ in Exon-3 deleted cells, as well as in cell lines expressing T2538R, L4115F, Q4201R and V4653F, as compared to S406L, R420Q, P2328S, F2483I, which in sharp contrast have increased diastolic Ca²⁺ levels on exposure to Isoproterenol. Given the inter cellular variability of hiPSC-CMs’ responsiveness to beta agonists it becomes difficult to develop a comprehensive hypothesis as to the mechanisms responsible for development of CPVT on the global whole-heart levels, but this variability may underlie the intriguing finding that the occurrence of CPVT in patients are generally episodic and often unpredictable on exertion or adrenergic stimulation.

CPVT cell lines expressing mutations of different RyR2 loci also show some diversity in the SR Ca²⁺ content. In 4 of the 11 mutant CPVT1 human cell lines, thus far reported, the Ca²⁺ content of SR was smaller, consistent with “leaky” RyR2s, Table 2, but at least in F2483I cell line (located in the vicinity of FKBP binding site) the decreased Ca²⁺ content is accompanied by higher I_Ca-gated fractional Ca²⁺ release. In this respect, it appears to be a common finding that the rate of spontaneous beating is lower in CPVT1 mutant cell lines as compared to control lines, possibly because of the lower SR Ca²⁺ content. This finding is consistent with the hypothesis that Ca²⁺ signaling mechanism, rather than the activation of I_f regulates spontaneous pacing activity in hiPSC-CMs [45].

Given the inter-cellular variability in adrenergic responsiveness and diversity in Ca²⁺ content of SR in different CPVT cell lines, perhaps more sophisticated approaches of simultaneous mapping of both electrical and calcium signals in monolayers of CPVT-expressing cell lines should be combined with single cell studies in identifying and quantifying the possible defects of Ca²⁺ signaling pathways. These are difficult issues to sort out, but they must be done if iPSC-derived human cardiac cell lines are to be used reliably, especially in pharmacotherapeutic studies. Clearly there are differences in the cellular phenotype, the functional pathology, and the pharmacology of different CPVT1 cell-lines. This, we believe, bodes well for the future of this field especially as more detailed characterizations of these cell-lines and their sensitivity to various pharmacotherapies are undertaken.

Highlights.

1. Specific carboxyl-tail fragments of L-type Ca²⁺ channel maybe responsible for the voltage-dependence of cardiac CICR.

2. Ca²⁺ signaling in human fibroblast-derived cardiomyocytes shows similar voltage-dependence of CICR and Ca²⁺ spark properties as those of adult mammalian cardiomyocytes.

3. CPVT mutants myocytes of N-terminal and cytoplasmic-domains show smaller Ca²⁺ stores but longer, wandering, and recurrent Ca²⁺ sparks.

4. Myocytes expressing different point mutations express varied Ca²⁺ signaling phenotypes and pharmacological sensitivity.