Calcium signalling of human pluripotent stem cell-derived cardiomyocytes

Abstract

Loss of cardiomyocytes (CMs), which lack the innate ability to regenerate, due to ageing or pathophysiological conditions (e.g. myocardial infarction or MI) is generally considered irreversible, and can lead to conditions from cardiac arrhythmias to heart failure. Human (h) pluripotent stem cells (PSCs), including embryonic stem cells (ESC) and induced pluripotent stem cells (iPSCs), can self-renew while maintaining their pluripotency to differentiate into all cell types, including CMs. Therefore, hPSCs provide a potential unlimited ex vivo source of human CMs for disease modelling, drug discovery, cardiotoxicity screening and cell-based heart therapies. As a fundamental property of working CMs, Ca²⁺ signalling and its role in excitation–contraction coupling are well described. However, the biology of these processes in hPSC-CMs is just becoming understood. Here we review what is known about the immature Ca²⁺-handling properties of hPSC-CMs, at the levels of global transients and sparks, and the underlying molecular basis in relation to the development of various in vitro approaches to drive their maturation.

Since non- or lowly regenerative adult cardiomyocytes (CMs) lack an innate clinically relevant ability to regenerate, their significant loss due to ageing or pathophysiological conditions (e.g. myocardial infarction or MI) can have lethal consequences by hastening the progression of heart failure (HF, primarily a disease of the ventricle) and/or predisposing to conduction abnormalities and arrhythmias. Current therapeutic regimes are palliative in nature, and in the case of end-stage HF, heart transplantation remains the last and only resort. Since this option is severely limited by the number of available donor organs, cell replacement therapy presents a laudable alternative for myocardial repair. Unfortunately, however, it is also limited by the availability of transplantable human CMs (e.g. human fetal CMs) due to practical and ethical considerations. As a result, transplantation of non-cardiac cells such as skeletal muscle myoblasts (SkMs), smooth muscle cells and bone marrow-derived mesenchymal stem cells (MSCs) has been sought as a potentially viable alternative. However, the non-cardiac identity of these cell sources has presented major limitations. In the case of SkMs, their lack of electrical integration after transplantation into the myocardium has been shown to underlie the generation of malignant ventricular arrhythmias, leading to the premature termination of their clinical trials. As for bone marrow stem cells, it is now well established that they lack the capacity to transdifferentiate into cardiac muscle (Murry et al. 2004), limiting their utility for myocardial repair. Indeed, various cardiac and non-cardiac lineages, as well as embryonic and adult stem cell populations, have been investigated as potential sources, with their pros and cons extensively reviewed elsewhere (Menasche et al. 2003; Smits et al. 2003; Murry et al. 2004; Sil et al. 2004; Kong et al. 2010; Poon et al. 2011). This review focuses on human (h) pluripotent stem cells (PSCs) that have been shown to generate genuine human CMs, with an emphasis on their Ca²⁺-handling properties.

Human pluripotent stem cells – embryonic and induced pluripotent stem cells

Upon fertilization of an oocyte by sperm, the resultant zygote, which possesses the total potential (i.e. totipotency) to develop into all cell types including those necessary for embryonic development (such as extra-embryonic tissues), undergoes several rounds of cell division to become a compact ball of totipotent cells known as the morula. As the morula continues to grow (∼4 days after fertilization), its cells migrate to form a more specialized hollow, fluid-filled structure known as the blastocyst consisting of an outer cell layer, the trophectoderm, and an inner cluster of cells collectively known as the inner cell mass (ICM). While the trophectoderm is committed to developing into extra-embryonic structures for supporting fetal development, the ICM that retains the ability to form any cell of the body except the placental tissues (i.e. pluripotency) will give rise to the embryo. Embryonic stem cells (ESCs) are isolated from the ICM. ESCs possess the ability to remain undifferentiated and propagate in vitro while maintaining their normal karyotype and pluripotency to differentiate into all the three embryonic germ layers (i.e. endoderm, mesoderm and ectoderm) as well as their lineage derivatives including brain, blood, pancreatic, heart and other muscle cells. Pluripotent mammalian ESC lines were first derived from rodent blastocysts 30 years ago (Evans & Kaufman, 1981; Martin, 1981), leading to the generation of the first transgenic animal and thereby revolutionizing genetics and disease modelling; the human counterpart was first successfully isolated about a quarter century later (Thomson et al. 1998). As an alternative, direct reprogramming of adult somatic cells to become hES-like induced pluripotent stem cells (iPSCs) has been developed. Forced expression of four pluripotency genes, Oct3/4, Sox2, c-Myc, and Klf4 (Takahashi & Yamanaka, 2006; Meissner et al. 2007; Takahashi et al. 2007), or Lin28 (Yu et al. 2007) suffices to reprogramme fibroblasts into iPSCs. Recent studies have further demonstrated the successful use of fewer pluripotency factors (Huangfu et al. 2008; Kim et al. 2008; Nakagawa et al. 2008) and non-viral methods (e.g. with synthetic modified RNA; Warren et al. 2010). Although concerns such as induced somatic coding mutations (Gore et al. 2011) and immunogenicity (Zhao et al. 2011) have yet to be fully addressed, human (h) iPSC largely resemble hESCs in terms of their pluripotency, surface markers, global transcriptomic profile, and can likewise be differentiated into CMs. Therefore, hESC/iPSC-CMs serve well as a potential unlimited ex vivo source of human CMs for disease modelling, drug discovery, cardiotoxicity screening and cell-based heart therapies.

Ca²⁺ cycling and excitation–contraction (EC) coupling as a fundamental property of working CMs

During an action potential (AP) of adult ventricular (V) CMs, membrane depolarization leads to the opening of sarcolemmal voltage gated L-type Ca²⁺ (I_Ca,L) channels (a.k.a. dihydropyridine receptors, DHPR), which are localized at the invaginations of the T-tubular network, in close spatial proximity to the sarcoplasmic reticulum (SR), in which a large internal pool of Ca²⁺ is stored. Ca²⁺ entry through I_Ca,L channels activates a positive feedback process termed as Ca²⁺-induced-Ca²⁺ release (CICR; Bers, 2002) that triggers the release of SR Ca²⁺ via the ryanodine receptors (RyRs), escalating the cytosolic Ca²⁺ and resulting in tropomyosin translocation and myofilament contraction. For relaxation, elevated Ca²⁺ gets pumped back into the SR by the sarcoplasmic–endoplasmic reticulum Ca²⁺-ATPase (SERCA) and extruded to the extracellular space by the electrogenic Na⁺–Ca²⁺ exchanger (NCX) to return to the resting Ca²⁺ level. Such a rise and subsequent decay of Ca²⁺, known as the Ca²⁺ transient, modulates both the contractile force (inotropy) and frequency (chronotropy) of CM contraction. Indeed, Ca²⁺-handling abnormalities due to malfunction of EC coupling proteins can be arrhythmogenic. As examples, SERCA2a down-regulation, hyper-phosphorylation and enhanced leak of RYR2, NCX up-regulation all lead to Ca²⁺ cycling defects. The process of EC coupling is schematically summarized in Fig. 1.

Figure 1. Schematic comparison of Ca²⁺ signalling pathways in adult and hPSC-derived cardiomyocytes.

In both adult (A) and hPSC-CMs (B), Ca²⁺ entry via I_CaL triggers Ca²⁺ release from SR via RyRs, leading to the rise of Ca²⁺ transient; the subsequent decay is similarly accomplished by Ca²⁺ reuptake and extrusion via SERCA and NCX, respectively. The smaller amplitude and slower kinetics and null inotropic response of hPSC-CMs compared to adult can be attributed to the following differences: (1) lack of junction (JCTN) and triadin (TRDN) to facilitate RyR function in hPSC-CMs; (2) lack of calsequestrin (CSQ) for SR Ca buffering (rather, calreticulin (CALR) is expressed in hPSC-CMs); (3) lack of phospholamban (PLB) for sarcoplasmic–endoplasmic reticulum Ca²⁺-ATPase (SERCA) modulation; (4) lower SERCA and RyR expression in hPSC-CMs; (5) lack of T-tubules which contributes to a U-shape of Ca²⁺ propagation wavefront.

In addition to I_Ca,L, RyR, SERCA and NCX, Ca²⁺ homeostasis also depends on an array of accessory Ca²⁺-handling proteins. RyRs are coupled to triadin (TRDN), junctin (JCTN) and calsequestrin (CSQ) at the luminal SR surface (Zhang et al. 1997). As the most abundant, high-capacity but low-infinity Ca-binding protein in the SR, the cardiac isoform CSQ2 can store up to 20 mm Ca²⁺ while buffering the free SR [Ca²⁺] at ∼1 mm (Beard et al. 2004), allowing repetitive muscle contractions without run-down. CSQ2 also coordinates the rates of SR Ca²⁺ release and loading by modulating RyR activities. In fact, the SR Ca²⁺ content affects the amount of Ca²⁺ released via CICR (Bassani et al. 1995; Shannon et al. 2000). For a given I_Ca,L trigger, a high SR Ca²⁺ load enhances the open probability of RyRs while directly providing more Ca²⁺ available for release (Lukyanenko et al. 1996). By contrast, I_Ca,L can no longer cause CICR when the SR Ca²⁺ content is sufficiently low. Mechanistically, CSQ2 senses the levels of luminal Ca²⁺ and effects RyRs via TRDN and JCTN. For instance, when SR Ca²⁺ declines (during Ca²⁺ release), an increase of Ca²⁺-free CSQ2 deactivates RyRs by binding via JCTN and TRDN; alternatively, SR Ca²⁺ reload (upon relaxation when CICR terminates) relieves the CSQ2-mediated inhibition of RyRs (Beard et al. 2004; Gyorke et al. 2004). Thus, CSQ2 is an important determinant of the SR load. For β-adrenergic signalling, the cascade involves such components as β-adrenergic receptors (β-ARs), the Gs protein–adenylyl cyclase (AC) and protein kinase A (PKA). PKA phosphorylates such substrates as I_Ca,L, RyR, PLB, troponin I, myosin binding protein-C (MyBP-C) and protein phosphatase inhibitor-1 to increase Ca²⁺ influx, Ca²⁺ cycling and myofilament Ca²⁺ sensitivity, thereby augmenting contractility (Zhao et al. 1994; Sulakhe & Vo, 1995; Gerhardstein et al. 1999; Kunst et al. 2000; Marx et al. 2000; Zhang et al. 2002; MacLennan & Kranias, 2003).

In mature ventricular CMs (VCMs), CICR is also optimized by the presence of T-tubules, invaginations in the sarcolemmal membrane that concentrates I_Ca,L channels in close spatial proximity to RyRs (Brette & Orchard, 2003, 2007). With a minimized Ca²⁺ diffusion distance between I_Ca,L and RyRs, SR deep in VCMs with large cross-sectional area can participate in CICR without significant time lags. This increased efficiency is demonstrated by a uniform increase in cytosolic Ca²⁺ across the transverse section of the cell (with simultaneous recruitment of all SRs). Such a uniform Ca²⁺ wave, with peripheral and central Ca²⁺ transients having similar amplitudes and kinetics, starkly contrasts the U-shaped Ca²⁺ wave propagation in de-tubulated CMs (Brette & Orchard, 2003). The U-shaped waves result from a time delay that is proportional to the diffusion distance squared in recruiting the Ca²⁺ stores at the cell centre (Song et al. 2005). Fast and synchronized activation of RyR translates into a larger Ca²⁺ transient amplitude, recruitment of more actin–myosin cross-bridge cycling, and generation of greater contractile force (Brette & Orchard, 2003; Louch et al. 2004).

Taken collectively, proper Ca²⁺-handling properties of hESC/iPSC-CMs are therefore crucial for their clinical and other applications. Transplantation of cells with improper Ca²⁺-handling and electrophysiological properties could introduce sources of Ca²⁺-mediated afterdepolarizations and/or automaticity which in turn result in various forms of potentially lethal arrhythmias.

Cardiac differentiation of hPSCs and heterogeneity

The natural development of the heart is one of the earliest and most essential steps during vertebrate embryonic development (Musunuru et al. 2010). For in vitro differentiation, PSCs can spontaneously form 3-D aggregates called embryoid bodies (EBs) with a variety of specialized cell types including CMs (Kehat et al. 2001; Xu et al. 2002; He et al. 2003; Xue et al. 2005). Different PSC lines display distinct cardiogenic potentials to become early ventricular-, atrial- and pacemaker-like derivatives as gauged by their signature AP profiles (Moore et al. 2008). For instance, compared to H1, HES2 cells have a higher likelihood of differentiating into ventricular-like hESC-CMs (Fig. 2). The cardiogenicity of PSCs is also highly influenced by a number of factors such as the EB size, media composition (e.g. serum, cytokines), the time and duration of differentiation, etc. Recently, various directed cardiac differentiation protocols (Yang et al. 2008; Zhu et al. 2010; Kattman et al. 2011; Ren et al. 2011; Lian et al. 2012; Minami et al. 2012; Zhang et al. 2012) have been developed to enable the derivation of PSC-CMs in large quantities with yields orders of magnitude higher than the traditional method of EB formation (Kehat et al. 2001). Despite the improved yields, the resultant cell population continues to be heterogeneous. Various purification methods including Percoll gradient centrifugation (Xu et al. 2002), optical signatures (Chan et al. 2009), second harmonic generation (Awasthi et al. 2012) and genetic selection based on the expression of a reporter protein under the transcriptional control of a cardiac-restricted promoter (e.g. α–MHC: Anderson et al. 2007; MLC2v: Huber et al. 2007; Fu et al. 2010) have been developed to generate purer CM preparations.

Figure 2.

Action potentials of ventricular, atrial and pacemaker cardiomyocytes derived from HES2 (A) and H1 hESCs (C). The corresponding pie charts (B and D). Adapted from Moore et al. (2008).

Functional but immature Ca²⁺ handling of hPSC-CMs

In murine (m) ESC-CMs, both the SR load and RyR are essential for regulating contractions even at very early developmental stages (Fu et al. 2006a,b). Similarly, Ca²⁺ transients have been recorded from hESC-CMs as beating clusters (Dolnikov et al. 2006; Liu et al. 2007) or single cells (Fu et al. 2010). In brief, Ca²⁺-handling properties of hESC-CMs resemble those of human fetal left ventricular (LV) CMs (16–18 weeks) but significantly differ from adult LVCMs. Upon electrical stimulation, hESC-CMs elicit Ca²⁺ transients with much smaller amplitudes and slower kinetics compared to adult (Fig. 3 and Table 1). Pharmacologically, ∼40% are responsive to caffeine (Liu et al. 2007). The responsiveness of hESC-CMs to caffeine are also reported in other studies (Satin et al. 2008; Zhu et al. 2009). Ryanodine reduces the electrically evoked Ca²⁺ transient amplitudes and slows the upstroke of caffeine-responsive but not -insensitive hESC-CMs. Thapsigargin, a SERCA inhibitor, similarly reduces the amplitude and slows the decay of only caffeine-responsive hESC-CMs(Liu et al. 2007). NCX is functional (Fu et al. 2010), with its expression highest in hESC-CMs but intermediate in fetal and lowest in adult LVCMs. Although SERCA2a expression is most robust in adult LVCMs, it is already substantially and comparably expressed in hESC- and fetal LVCMs. RyRs are expressed in hESC-CMs and fetal LVCMs, but lack the organized pattern seen in adult due to the lack of T-tubules. On the far end, the regulatory proteins junctin, triadin, and calsequestrin (CSQ) are robustly expressed in adult LVCMs but completely absent in hESC-CMs. Consistent with the lack of T-tubules, the Ca²⁺ wavefront of hESC-CMs is U shaped (Fig. 4), indicating a time delay in activation of RyRs at the centre but faster and greater magnitude of Ca²⁺ transient increase at the periphery. Furthermore, a negative force–frequency relationship that is different from adult but typical of immature CMs has been reported (Dolnikov et al. 2006; Turnbull et al. 2013). These key differences from adult CMs are summarized in Fig 1 and Table 1.

Figure 3. Functional yet immature Ca²⁺ handling in hESC-CMs compared with fetal ventricular CMs (FLV-CMs) and adult ventricular CMs (ALV-CMs).

A, caffeine-induced Ca²⁺ transient in HES2-CMs, H1-CMs and FLV-CMs. B, percentage of respective CMs that are sensitive to caffeine treatment. C, Ca²⁺-handling protein expression profile in HES2-CMs, H1-CMs, FLV-CMs and ALV-CMs. D, immunostaining of ryanodine receptors in HES2-CMs, H1-CMs and FLV-CMs. Adapted from Liu et al. (2007).

Table 1.

Expression levels of Ca²⁺-handling proteins, and Ca²⁺ transient properties in hESC-CMs, fetal LVCMs and adult LVCMs

		hESC-CMs	Fetal LVCMs	Adult LVCMs
Expression levels of Ca2+-handling proteins	RyR	++	++	++++
	SERCA	+++	+++	++++
	Phospholamban	–	++	++++
	CSQ/TRDN/JCTN	–	+	++++
	NCX	+++	++++	+
Ca2+ transient properties	Basal [Ca2+]i	++	+++	++++
	Amplitude	++	++	++++
	Decay	++	++	++++
	Upstroke	++	++	++++

Adapted from Kong et al. (2010).

Figure 4. Absence of T-tubules in hESC-CMs.

Di-8-ANEPPS staining shows no intracellular fluorescent spots in hESC-CMs (A) compared with adult cardiomyocytes (C). B and D, Atomic force microscopy imaging reveals no T-tubules, present as regularly spaced pores, in hESC-CMs. E, lack of T-tubules results in a non-uniform, U-shaped Ca²⁺ transient. Adapted from Lieu et al. (2009).

Immature Ca²⁺ transient properties of hESC-CMs can be attributed to the differential developmental expression profiles of specific Ca²⁺-handling proteins in hESC-CMs compared to adult. Gene transfer of CSQ that is otherwise absent in hESC-CMs significantly increases the transient amplitude, upstroke and decay velocities, as well as both the SR Ca²⁺ load and elevated basal cytosolic Ca²⁺ (Fig. 5), but without altering I_Ca,L, suggesting that the improved transient is not simply due to a higher Ca²⁺ influx for CICR.

Figure 5. CSQ overexpression increases the amplitude of caffeine-induced Ca²⁺ transient.

A, representative caffeine-induced Ca²⁺ transient tracings for Ad-GFP, Ad-CSQ and Ad-CSQΔ transduced hESC-CMs. B, bar graphs of amplitude. *P < 0.05. Adapted from Liu et al. (2009).

Ca²⁺ sparks in hPSC-CMs

Global Ca²⁺ transient is crucial for CM contraction; but, Ca²⁺ signals can be highly localized, with Ca²⁺ release units (CRU) being clusters of Ca²⁺ release channels (RyRs, inositol-1,4,5-trisphosphate receptors (IP₃Rs) or a mixture) located on the SR membrane (Cheng & Lederer, 2008). Discrete> local Ca²⁺ release through RyRs has been termed ‘Ca²⁺ sparks’. Ca²⁺ waves form when Ca²⁺ release from CRU is regenerative, typically reflecting the high sensitivity of CICR (Keizer et al. 1998; Izu et al. 2001; Cheng & Lederer, 2008). Ca²⁺ sparks are the elementary events of EC coupling: electrical stimulation triggers thousands of Ca²⁺ sparks that spatio-temporally summate to increase the level of cytoplasmic Ca²⁺ level, thereby causing contractions (Cheng et al. 1993, 1996). In murine embryonic CMs, sparks are largely limited to area around the nucleus (Janowski et al. 2006). By contrast, those of adult ventricular and atrial cells are near the sarcolemma (Janowski et al. 2006). In mESC-CMs, it has been reported that Ca²⁺ sparks and spontaneous activity are essentially dependent on RyR2 (Itzhaki et al. 2006). Ca²⁺ sparks in mESC-CMs are absent in very early cardiac progenitors, but markedly increase in their frequency along with increased RyR expression at later stages (Sauer et al. 2001). In hESC-CMs (unsorted), both RyR2 and IP₃R have been shown to contribute to the generation of Ca²⁺ sparks (Satin et al. 2008; Sedan et al. 2008; Itzhaki et al. 2011). In general, spontaneous Ca²⁺ sparks with smaller peak amplitude, slower kinetics and lower frequency compared to fetal LVCMs have been reported to localize to the surface membrane (Guan et al. 2007; Zhu et al. 2009). In our laboratory, hESC-VCMs, selected on the basis of dual expression of a fluorescent reporter (e.g. tdTomato) and zeocin resistance under the transcriptional regulation of the ventricle-specific promoter MLC2v and derived using our ventricular specification protocol, have similar Ca²⁺ spark parameters, although sub-cellular localization to around the nucleus or near the sarcolemma is not observed; rather, Ca²⁺ sparks occur readily throughout the cytoplasm (Awasthi et al. 2012).

β-Adrenergic responses in hESC-CMs

β-Adrenergic signalling figures prominently in the modulation of cardiac function. Upon its stimulation via β-ARs by their agonists, the heart rate, contractile force and AP conduction velocity (dromotropy) increase thereby escalating the total cardiac output. Several reports show that β-adrenergic stimulation by isoproterenol (Iso) hastens the spontaneous contraction or AP firing and conduction of hESC-CM (Xue et al. 2005; Sedan et al. 2008; Burridge et al. 2011; Wang et al. 2013). When stimulated by Iso, however, engineered human cardiac tissues strips made of hESC-VCMs exhibit a positive chronotropic response by increasing their spontaneous contraction frequency, but without an inotropic effect (i.e. no increase in contractile force; Turnbull et al. 2013). Consistently, Iso increases AP firing but not the transient amplitude in single hESC-VCMs (G. Chen & R. A. Li, unpublished data). Similar observations have been made using unsorted multi-cellular hESC-CM clusters (Pillekamp et al. 2012). Phospholamban (PLB) is an inhibitor of SERCA in its unphosphorylated form, and a critical determinant of contractility and inotropic responses (MacLennan & Kranias, 2003). Once its primary site Ser-16 gets phosphorylated by PKA, its inhibitory effect on SERCA is reversed, leading to acceleration of Ca²⁺ sequestration into the SR and hastening cardiac relaxation. Interestingly, overexpression of PLN which is poorly expressed in hESC-CMs restores the positive ionotropic response of hESC-VCMs to Iso (G. Chen & R. A. Li, unpublished data).

iPSC-based modelling of Ca²⁺-handling defects

Patient-specific iPSCs have been generated to model various forms of Ca²⁺-handling genetic defects. To give a few examples, hiPSC-CMs derived from catecholaminergic polymorphic ventricular tachycardia (CPVT) patients with a missense mutation D307H in CSQ2 have been shown to display arrhythmogenicity in response to β-adrenergic stimulation (Novak et al. 2012). Similarly, CPVT-iPSC-CMs that carry the RYR2 mutations S406L (Jung et al. 2012), M4109R (Itzhaki et al. 2012) or P2328S (Kujala et al. 2012) exhibit increased susceptibility to arrhythmogenic events such as delayed after-depolarizations (DADs) when subjected to catecholaminergic stress as a result of their changes in Ca²⁺ spark and transient properties. As for iPSCs derived from a familial dilated cardiomyopathy (DCM) patient with the cardiac troponin T point mutation R173W, their CMs have altered Ca²⁺ homeostasis, decreased contractility, and abnormal distribution of sarcomeric α-actinin but functional improvements are seen after beta-blocker treatment or Serca2a overexpression (Sun et al. 2012).

Driven maturation by in vitro electrical conditioning

Other than Ca²⁺ handling, the electrophysiological properties of PSC-CMs are also immature (Kehat et al. 2002; Sartiani et al. 2007; Poon et al. 2011). Adult LVCMs are normally electrically silent yet excitable upon stimulation. By contrast, hESC-VCMs display significant automaticity, depolarized resting membrane potential (RMP), phase 4 depolarization and delayed after-depolarization (DAD) that are not seen in adult VCMs (Lieu et al. 2012; Fig. 6A). Long-term (>120 days) culturing leads to some but still limited structural and functional maturation (Lundy et al. 2013). While we and others have shown that when transplanted such AP-firing hESC-CMs can serve as an epicardial source of automaticity and cause potential arrhythmias (Kehat et al. 2004; Xue et al. 2005), it has been recently reported that their transplantation in injured hearts can suppress electrical disturbances (Shiba et al. 2012). For transplantation of murine fetal cardiomyocytes, physiological but immature AP properties are found in transplanted cells surrounded by cryoinjured tissue (Halbach et al. 2007). Among the panoply of sarcolemmal ionic currents investigated (I_Na⁺/I_CaL⁺/I_Kr⁺/I_NCX⁺/I_f⁺/I_to⁺/I_K1⁻/I_Ks⁻), we recently pinpointed the lack of the Kir2.1-encoded inwardly rectifying K⁺ current (I_K1) as the single mechanistic contributor to the immature electrophysiological properties in hESC-CMs (Lieu et al. 2012). Forced expression of Kir2.1 in hESC-CMs completely ablates all the pro-arrhythmic AP traits, rendering the electrophysiological phenotype indistinguishable from adult (Fig. 6B). Despite electrical maturation, Ca²⁺-handling properties remain immature with smaller transient amplitudes and slow kinetics. Indeed, the expression levels of sarcomeric genes, such as MHCα, MHCβ, MLC2a and MLC2v, of Kir2.1-silenced cells even deteriorate due to the lack of spontaneous contractions after electrical silencing (Fig. 6C). Based on these results, we have further developed a bio-mimetic culturing strategy for enhancing maturation. In vitro electrical conditioning of hESC-CMs promotes electrophysiological maturation (Fig. 6D); pacing-induced regular contractions likewise facilitate maturation of Ca²⁺-handling and contractile properties with augmented Ca²⁺ transient and SR Ca²⁺ load (Fig. 6E). Consistently, the expression levels of CSQ, JCTN, and TRDN (Liu et al. 2007), as well as the T-tubule biogenesis proteins caveolin-3 (Cav3) and amphiphysin-2 (Amp2), which are typically absent or barely expressed in hESC-CMs, all increase. We conclude that key environmental cues are missing in conventional culturing methods, leading to the immaturity of hESC-CMs.

Figure 6. Adenovirus-mediated Kir2.1 overexpression led to hESC-CMs maturation.

A, representative tracings of human embryonic stem cell-derived cardiomyocytes produced by directed differentiation protocol (ddhESC-CMs), showing action potentials (APs) of spontaneously firing (left) and quiescent (right) ventricular CMs. Arrow indicates phase 4-like depolarization. B, APs of Ad-Kir2.1-transduced ventricular ddhESC-CMs. I_k1 in inset. The phase 4-like depolarization was eliminated (arrow) by Ad-Kir2.1 transduction (left). The percentage of quiescent ventricular ddhESC-CMs increased significantly to 100% after Ad-Kir2.1 transduction (middle). Resting membrane potentials (RMPs) of Ad-Kir2.1-transduced ventricular ddhESC-CMs became significantly hyperpolarized relative to control (right). C, after Ad-Kir2.1 transduction, the mRNA expression of contractile elements is significantly reduced relative to control hESC-CMs. D, both electrically conditioned atrial and ventricular ddhESC-CMs harboured action potentials without phase 4-depolarization (left). More hyperpolarized resting membrane potential (RMP) was shown (right). E, caffeine-elicited Ca²⁺ transients from control (blue) and electrically conditioned (red) ddhESC-CMs, and their averaged peak amplitude were also presented in bar graph. Adapted from Lieu et al. (2012).

Conclusion

The field of regenerative medicine is hampered by an incomplete understanding of mechanisms that underlie the proper maturation of CMs derived from hPSC. As fundamental properties of working CMs, the electrophysiological and Ca²⁺-handling properties of hESC/iPSC-CMs are just becoming understood. Such knowledge will lead to better tools (e.g. for accurate disease modelling, drug discovery and cardiotoxicity screening) and novel effective approaches for cell-based therapies by improving both the efficacy and survival of hPSC-VCMs after transplantation.

Additional information

Competing interests

None declared.

Funding

This work was supported by the Research Grant Council (T13-706/11), SCRMC and Faculty Cores of HKU.