The Role of Metabolism in Directed Differentiation versus Trans-differentiation of Cardiomyocytes

Abstract

The advent of induced pluripotent stem cells (iPSCs) and identification of transcription factors for cardiac reprogramming have raised hope to cure heart disease, the leading cause of death in the world. Our knowledge in heart development and molecular barriers of cardiac reprogramming is advancing, but many hurdles are yet to be overcome for clinical translation. Importantly, we lack a full understanding of molecular mechanisms governing cell fate conversion toward cardiomyocytes. In this review, we will discuss the role of metabolism in directed differentiation versus trans-differentiation of cardiomyocytes. Cardiomyocytes exhibit a unique metabolic feature distinct from PSCs and cardiac fibroblasts, and there are multiple overlapping molecular mechanisms underlying metabolic reprogramming during cardiomyogenesis. We will discuss key metabolic changes occurring during cardiomyocytes differentiation from PSCs and cardiac fibroblasts, and the potential role of metabolic reprogramming in the enhancement strategies for cardiomyogenesis. Only when such details are discovered will more effective strategies to enhance the de novo production of cardiomyocytes be possible.

Graphical abstract

1. Introduction

Cardiovascular disease (CVD) is the leading cause of global disease burden, and ischemic heart disease accounts for a half of CVD-caused deaths [1, 2]. An insufficient supply of oxygen or nutrients to the myocardium can result in an irreversible loss of cardiomyocytes (CMs) that are replaced with fibrotic scaring, eventually leading to possible heart failure. There are a number of pharmacological interventions targeting sympathetic activation, pressure overload, fibrosis pathways, which can slow down the progress of heart failure, but the contractile function cannot be restored unless a net loss of CMs is recovered [3]. CMs can renew at a very low rate and about 50% of CMs are exchanged throughout the lifetime [4, 5]. However, an endogenous regeneration of CMs is insufficient to recover a loss of ~1 billion CMs that occur after myocardial infraction, thus the field of heart regeneration has been primarily focusing on remuscularization of the myocardium by exogeneous supplementation of CMs [6].

The discovery of induced pluripotent stem cells (iPSCs) has provided various unlimited sources of differentiated cells without ethical concerns. Pluripotent stem cells (PSCs), which consist of iPSCs and human embryonic stem cells (hESCs), can proliferate indefinitely and differentiate into any desired cell types including CMs [7, 8]. Intramyocardial injection of PSC-derived cardiomyocytes (PSC-CMs) has been shown to improve contractile function in animal models of myocardial infarction [9]. In addition to stem cell-based therapy, direct cardiac reprogramming has been increasingly recognized as an alternative strategy to remuscularize the myocardium. An ectopic expression of defined transcription factors (TFs) can convert cardiac fibroblasts (FBs) to CMs [10]. The delivery of cardiogenic transcription factors in a mouse model of myocardial infarction successfully restored the contractile function and decreased the infarct size [11]. Direct reprogramming holds many advantages over stem cell-based therapy, though the efficiency of transdifferentiation is low due to multiple molecular barriers like an inherent metabolic feature of FBs. In the current review, we will discuss the role of metabolism in cardiomyogenesis and changes in metabolism during directed differentiation and trans-differentiation into CMs.

2. De novo generation of cardiomyocytes from non-cardiomyocytes

2.1. Directed differentiation

The advent of iPSCs has provided an unlimited source of PSCs and PSC-derivatives like CMs for replacement therapy. A successful translation to the clinic setting depends on our capacity to accurately direct PSCs to commit to becoming CMs. In the early era of stem cell biology, we used poorly defined culture media containing fetal bovine serum to generated CMs and the efficiency was extremely low [12]. Advances in our knowledge in cardiovascular development have led to multiple serum-free, chemically defined protocols to finely guide PSCs to CMs. Investigators can now generate a contractile sheet of ~95% of CMs from PSCs. A detailed evolution and comparison of CM differentiation protocol can be found elsewhere [13].

CM differentiation is directed by the signaling pathways known to regulate early developmental processes such as Nodal, Wnt, and BMP. Similar to developmental events occurring in vivo, Wnt signaling is critical to initiate CM differentiation – mesoderm specification [14, 15]. A fine titration of BMP and activin A/Nodal signaling with the timed delivery of Wnt inhibition can generate CMs from PSCs [16]. The most widely used protocol, which does not require recombinant BMP and Nodal/activin A, uses a temporal modulation of Wnt signaling with small molecules [17]. Briefly, Wnt signaling is induced by a small molecule glycogen synthase kinase-3 inhibitor that elevates the endogenous BMP and Nodal pathways for mesoderm specification. Wnt signaling is eventually terminated as an extended activation of Wnt signaling inhibits the later cardiac developmental stage [18]. This small molecule-based protocol is further improved with chemically defined medium components and re-formulated as a highly reproducible and efficient protocol [19].

PSC-CMs can be used for human development and disease modeling, drug screening, and cell replacement therapy. With more advanced differentiation protocols (e.g., generating CM subtypes), we can recapitulate the key developmental events identified in other model organisms and understand the influence of genetic mutations on onset and development of human cardiovascular diseases [20, 21]. Over the last decade, the clinical representation of human heart diseases caused by genetic mutations has been recapitulated in a dish with patient iPSC-CMs [22–26]. Surprisingly, iPSC-CMs can recapitulate individuals’ diverse inherent susceptibilities or responses to drugs, allowing a targeted intervention via personalized drug assessment or development [27–30]. Not only that, PSC-CMs can be directly used for cell therapy. PSC-CMs can be successfully engrafted and remuscularize the myocardium to provide a meaningful recovery of contractile function in multiple animal models of myocardial infarction [31–35]. A scalable production of PSC-CMs is feasible and the first clinical trial using iPSC-CMs is being conducted [36].

2.2. Trans-differentiation

The plasticity of cell identity or cell fate was demonstrated by a seminar work of Davis et al., which found that the ectopic expression of a master regulator of one cell type could covert one cell fate to another [37]. This raises the possibility that somatic cells like fibroblasts can be directly converted into desired cell types without undergoing a transient pluripotent state. A direct reprogramming or transdifferentiation into CMs was then first demonstrated by Ieda et al., who screened 14 TFs related to cardiac development and identified the core cardiac reprogramming TFs: GATA4, Mef2C, and TBX5 (a combination referred to as GMT) [10]. An ectopic expression of GMT successfully converted mouse dermal or cardiac fibroblasts into CM-like cells [10]. These CMs, so-called induced CMs (iCMs), spontaneously beat and exhibited a similar electrophysiology property as CMs. As cardiac FBs exist abundantly in heart and proliferate following injuries, resident cardiac FBs are an ideal source of new CMs. Two seminal works published in 2012 further demonstrated that in situ transdifferentiation of endogenous cardiac fibroblasts into iCMs is feasible in a mouse model of myocardial infarction [11, 38]. A local delivery of GMT or GMTH (addition of Hand2) in a retrovirus system transdifferentiated resident cardiac FBs or non-CMs into iCMs [11, 38]. Intriguingly, the overall efficiency of reprogramming in vivo was higher and iCMs exhibited structurally and functionally mature phenotypes of CMs. This suggests that the physiologically relevant microenvironment (e.g., extracellular matrix, secreted factors, and tissue stiffness) may affect the reprogramming efficiency similar to bioengineered heart tissue in vitro [39]. Both studies observed a functional improvement with a reduction in infarct size after myocardial infarction [11, 38]. Since then, numerous factors influencing the efficiency of reprogramming have been identified. The detailed overview of transdifferentiation methodologies and enhancement strategies can be found in other reviews [40, 41].

2.3. PSC-CMs vs i-CMs

Directed differentiation and trans-differentiation have their own advantages versus disadvantages for therapeutic application [42]. For instance, the efficiency of PSC-derived CMs is very high (90–95%) and PSCs can be cultured indefinitely, so batches of 1~10 billion pharmaceutical-grade CMs can be generated at high purity (>90%) [43]. The engrafted CMs from PSCs can persist for at least 3 months, and they progressively synchronized with the host myocardium to significantly improve the ventricular function [35]. However, the transplantation of xenografts can provoke a severe immune response that requires the use of significant amounts of immunosuppressant. The feasibility of autologous stem cell therapies (e.g., iPSC-CMs) was previously validated by Shiba et al. [44] They generated iPSCs from non-human primates, which underwent myocardial infarction and were administered iPSC-CMs. The engrafted CMS survived for 3 months with no evidence of immune rejection and became electrically coupled with the host myocardium [44]. However, the long process (several months) of generating a therapeutic dose of iPSC-CMs presents a logistical challenge for clinical application. Moreover, the heterogeneity and immature phenotypes of PSC-CMs are major challenges for CM cell therapy. An immaturity of PSC-derived CMs, distinct metabolic and electrophysiological properties from adult CMS, can contributed to ventricular arrhythmias [45]. All studies using large animal models of myocardial infarction and PSC-CMs consistently observed ventricular arrhythmias [33, 35, 44]. Electrophysiological studies found that these arrhythmias originated from contaminating pacemaker cells in the transplant due to imperfect CM differentiation protocols.

iCMs, on the other hand, show relatively more mature phenotypes than PSC-CMs. Zhou et al. compared the transcriptomic profile between iPSC-CMs and iCMs from the same origin at day 28 post-differentiation [46]. They observed highly expressed genes from iCMs being involved in oxidative phosphorylation (OXPHOS) and fatty acid oxidation, whereas suppressed genes were involved in cell cycle. On the other hand, highly expressed genes from iPSC-CMs were involved in glycolysis and cell cycle [46]. Structural and functional parameters also supported the idea that iCMs exhibit more mature phenotypes than iPSC-CMs [46]. Although the authors did not fully take into account strategies to improve the maturity of iPSC-CMs, this study provided the first comparative study between iPSC-CMs and iCMs and suggested a relatively rapid acquisition of maturation state in direct reprogramming. However, a low efficiency of reprogramming with poor reproducibility, complex gene delivery method, and lack of evidence on human models are major drawbacks for clinical translation of cardiac reprogramming [47, 48]. This has led extensive efforts to explore the strategies (e.g. addition of TFs or small molecules) to improve the cardiac reprogramming [40, 41].

3. Metabolic signatures of PSCs, CMs, and FBs

Cardiomyogenesis from directed differentiation or transdifferentiation involves the same principle of “cell fate conversion,” and CMs exhibit a unique metabolic phenotype compared to iPSCs or somatic FBs. To understand the role of metabolism occurred during a cell fate transition, we now discuss the metabolic signatures of each cell types: PSCs, CMs, and FBs (Fig 1).

Figure 1.

Distinct metabolic states of pluripotent stem cells (PSCs), cardiomyocytes (CMs), quiescent and activated fibroblast (FBs). PSCs maintain a highly proliferative capacity via glycolysis, pentose phosphate pathway (PPP) and glutaminolysis. PPP and glutaminolysis enhance an antioxidant capacity of PSCs to alleviate oxidative stress. CMs undergo a metabolic reprogramming from glycolysis to oxidative phosphorylation (OXPHOS) throughout development. Fetal or postnatal CMs rely on glycolysis and lactate oxidation as main energy source, while adult CMs predominantly rely on fatty acid and glucose oxidation for OXPHOS. Quiescent FBs focus the metabolism on PPP to alleviate oxidative stress from OXPHOS. Activated FBs switch to glycolysis and glutaminolysis to maintain a highly proliferative capacity. Increased mitochondrial fission and reactive oxygen species (ROS) elevate oxidative stress and accelerate myofibroblast transition.

3.1. Induced pluripotent stem cells (iPSCs)

PSCs maintain “stemness” by balancing energetic and biosynthetic demands via a tight metabolic regulation of glycolysis and OXPHOS. PSCs exhibit a high glycolysis flux while maintaining low OXPHOS flux for energy production [49, 50]. Glycolysis is a conversion of glucose to pyruvate, which becomes acetyl CoA by pyruvate dehydrogenase (PDH) to feed in tricarboxylic acid (TCA) cycle. PSCs showed a low level of PDH activity while maintaining a high level of hexokinase (HK) activity, a rate limiting pathway governing glycolysis [49]. A chemical inhibition of HK disrupts the metabolic feature of PSCs and results in the loss of pluripotency [51]. During iPSC reprograming, somatic cells undergo metabolic repatterning by upregulating glycolytic genes (Glut1, Hxk2, Pfkm, and Ldha) that precedes the induction of pluripotent markers [52]. The maintenance of gene expression (Hxk2 and PkM2) involved in glycolysis is directly controlled by a reprogramming TF OCT4 [53].

An accumulation of glycolytic metabolites (e.g., glucose-6-phosphate) potentiates the pentose phosphate pathway (PPP), which generate ribose-5-phosphoate that reduces NADPH-generating power for lipid biosynthesis and antioxidant capacity [54]. An enhanced antioxidant capacity is critical to maintain stemness because reactive oxygen species (ROS) from mitochondrial OXPHOS is known to disrupt the pluripotency of PSCs [55]. During the reprograming, somatic mitochondria undergo regression from tubular/cristae-rich mitochondria to sparse/spherical/cristae-poor structures associated with downregulation of nuclear factors for mitochondrial biogenesis and mtDNA content [52].

Although PSCs maintain the low energetic contribution by OXPHOS, mitochondria still play an important role in generating metabolic intermediates for epigenetic modulation of PSCs. PSCs utilize a high rate of glutamine to feed the TCA cycle to maintain alpha ketoglutarate (α-KG) level. An elevated α-KG to succinate ratio promotes the epigenetic modification (histone/DNA demethylation) contributing to maintenance of the pluripotency [56]. PSCs express high level of genes involved in glutaminolysis (SLC1A5, GLS2, GLUL, and GPT2) and glutamate, a metabolite intermediate of glutaminolysis, further increasing the antioxidative capacity to prevent the degradation of pluripotency marker OCT4 [57]. There are multiple other mechanisms of maintenance of pluripotency via metabolites intermediates, which can be found elsewhere [58].

2.2. Cardiomyocytes (CMs)

Cardiomyocytes are one of the most energy-demanding cell types with a very limited energy reserve but continuously produce energy for cardiac function (e.g., contractility, calcium handling, and electrophysiology). Cardiomyocytes harbor the second most numerous mitochondria in cells, and OXPHOS generate more than 90% of energy for the heart [59]. Cardiomyocytes undergo a dramatic metabolic remodeling throughout development and differentiation. Cardiomyocytes in fetal state are still capable of proliferation by relying on glycolysis as a major energy source [59]. A successful transition from PSCs to CMs requires a switch from glycolysis to OXPHOS, but maintaining glycolysis is essential to keep the proliferative capacity of cardiomyocytes. A high dependence on glycolysis is maintained even at the early postnatal developmental period and any injuries occurring during these periods are recoverable by replacing dead cardiomyocytes with new cardiomyocytes [60, 61]. Hif-1α is the master regulator that governs the glycolytic metabolism in the fetal heart by enhancing expression of glycolysis-related genes (Glut1, Hk1, Ldha, and Pdk1). Throughout the development of cardiomyocytes, the oxidative metabolic capacity exponentially increases with an increased reliance on oxidative metabolism and a sharp decline in the regenerative capacity. This transition is related to environmental oxygen level in that an exposure to oxygen-rich coincides with a metabolic transition from glycolysis to OXPHOS, which results in increases reactive oxygen species (ROS) and inhibits the proliferative capacity [62]. Conversely, conditioning the hypoxic environment extends the proliferative capacity even in adult mice by maintaining a high glycolysis rate [63]. During a postnatal development, an increased free fatty acid serum concentration (< 0.1 mM in fetal versus 0.2–0.4 mM in adult cells) in circulation switches a major energy source from glucose to fatty acid. Feeding pups with fatty acid deficient milk preserves the glycolytic metabolism in heart and enhances the proliferative capacity and the regenerative window of heart [64]. Mechanistically, an increased activity of PDK4 driven by free fatty acid mediate PPARα activation restricts glucose oxidation yet enhances fatty acid oxidation [65]. There are various other nuclear receptor proteins (e.g., PPARβ/γ, PGC1α, etc.) regulating the metabolic maturation during the postnatal period and thoroughly discussed elsewhere [59, 66, 67].

2.3. Fibroblasts (FBs)

Fibroblasts are a major non-CM population residing in the heart which contribute to maintenance of normal heart physiology via paracrine signaling, electrical coupling, vascular maintenance, and stress surveillance [68]. Under the normal physiological condition, cardiac fibroblasts remain quiescent or resting, whereas they are activated and differentiated into myofibroblasts upon pro-fibrotic signaling or external stressors like acute injury, hemodynamic overload, and sympathetic nervous activation [68]. Contrary to the usual belief that quiescent cells are low in metabolic activity, quiescent FBs are highly metabolic and emphasize metabolic pathways to PPP [69]. Focused PPP pathway regenerates NADPH and raises antioxidant capacity to alleviate oxidative stress by ROS which triggers a myofibroblast transition via mitochondrial remodeling [70, 71]. In rat cardiac FBs, isoproterenol treatment resulted in ROS generation to increase the proliferation and activate myofibroblast differentiation, which is reversed by antioxidant treatment [72]. Jain et al. reported that profibrotic TGF-β signaling downstream pathways leading to myofibroblast transition requires mitochondrial ROS [71]. Huang et al. reported that an increase in mitochondrial fission with elevation of mitochondrial ROS stimulates a myofibroblast transition upon hypertensive insults [73].

FBs, upon activation by either stressor or pro-fibrotic signaling pathways, become highly proliferative by increasing the expression of cell cycle proteins and glycolysis related genes. Patients with cystic fibrosis have increased expression of glycolysis related genes (Pfk1, Pkm, and Ldha), and a glycolytic switch from OXPHOS is consistently reported in activated FBs from different organs [74]. Conditioning FBs with high glucose is sufficient to stimulate pro-fibrotic TGF-β signaling, corroborating a strong association between cardiac fibrosis and diabetic cardiomyopathy [75, 76]. Profibrotic TGF-β signaling promotes the Warburg effect by redirecting glucose into non-oxidative pathways (e.g. decreased expression of Pkm2) [77]. Recently, Lombardi et al. observed that TGF-β signaling restricted the mitochondrial calcium uptake, which elevated aerobic glycolysis and glutaminolysis to yield a greater α-KG bioavailability that resulted in the epigenetic modulation of myofibroblast gene expression [78]. Increased glutaminolysis is mediated by upregulation of glutaminase 1 (Gls1) via TGF-β signaling [79, 80]. An increase in glutaminolysis is required to activate fibroblasts as pharmacological or genetic disruption or glutamine influx prevents myofibroblast transition. Collectively, these studies suggest that inefficient glucose oxidation with increased glycolysis and glutaminolysis contributes to myofibroblast transition from quiescent FBs.

4. Metabolic shift during cardiomyogenesis

Cell fate is regulated at multiple decision points, and a conversion from one fate to another is mediated by dynamic changes in transcriptomic profiles often associated with the metabolic switch. The metabolic state is determined by an interplay of complex factors such as cell signaling, substrate availability, as well as energetic and biomass demands. Both directed- and trans-differentiation of CMs involve a complex transcriptomic remodeling which contributes to the metabolic switch from glycolysis to OXPHOS.

4.1. Directed differentiation

CM differentiation from PSCs involves a stepwise specification of mesoderm and cardiac progenitor cells via signaling pathways involved in early development. Glycolysis is a key determinant for pluripotency, and a reduction in glycolysis inhibits self-renewing and proliferative capacity of PSCs [81]. PSCs move away from glycolysis by downregulating transcriptional activity of Myc which result in increased OXHPOS and induce mesoderm differentiation [81, 82]. Mesodermal cells have a high oxidative metabolism than other germ layers (endodermal or ectodermal), and extracellular pyruvate supplementation, which feeds the TCA cycle, increases the oxygen consumption rate and potentiates the mesoderm specification by modulating the mTOR signaling pathway [83, 84]. An upstream signaling of mTOR, or insulin signaling, inhibits the mesodermal and cardiac progenitor differentiation from PSCs by inhibiting the expression of mesodermal markers (Brachyury T) and cardiogenesis markers (GATA4 and Nkk2–5) [85]. Insulin stimulates glucose metabolism by enhancing glucose uptake via Akt signaling, and thus most CM differentiation protocols use insulin-depleted culture media to induce CM differentiation [17, 86]. Another important signaling pathway is canonical Wnt/β-catenin signalling. Wnt signaling pathway regulates the amount of the transcriptional co-activator β-catenin, which controls key gene programs involved in cell proliferation, polarity, and cell fate determination during embryonic development [87]. The mesoderm specification of PSCs requires an activation of canonical Wnt/β-catenin signaling and Glycogen synthase kinase 3β (GSK3β) inhibition robustly induces mesoderm differentiation by activating canonical Wnt/β-catenin signaling [20]. GSK3β is known to inhibit glycogen synthesis, which in turn activates glycolysis; therefore, GSK3β inhibition is involved in modulating glucose metabolism in PSCs to initiate CM differentiation [88]. The evidence of GSK3β inhibition results in a metabolic shift from glycolysis to OXPHOS can be found in other different cell types [89].

Although mesodermal induction alone results in a robust metabolic shift from glycolysis to OXPHOS, the glucose dependence of cardiac progenitors or PSC-CMs at early stage of differentiation is still high. This reflects the fact that an early cardiac development where embryonic or fetal heart exists in a hypoxic uterus where Hif-1α, a master regulator of glycolysis, is activated [90]. Ablation of cardiac Hif-1α during an embryonic or a fetal period causes developmental defects, but a decrease in cardiac Hif-1α signaling during a postnatal period promotes a metabolic maturation to OXPHOS [91]. This reflects that timed inactivation of glycolysis or glucose metabolism is critical to acquire the metabolic maturation of cardiomyocytes. When hPSC-CMs are cultured in a glucose-rich medium, glucose interrupts the metabolic and functional maturation of PSC-derived CMs by promoting glycolysis and PPP pathways [92]. The glucose-rich medium aberrantly activates Hif-1α expression and its downstream glycolysis-related genes (Hk2, Glut1, Ldha) [93]. Chemical inhibition of Hif-1α transcriptional activity or glycolysis improves the metabolic and functional parameters of PSC-derived CMs by enhancing mitochondrial content, oxygen consumption, sarcomere length, and contractility [92, 93]. Therefore, the blockade of glucose metabolism is a key determinant in late developmental or differentiation process to achieve mature PSC-CMs.

4.2. Transdifferentiation

The low efficiency or cardiac reprogramming of heterogenous resultant cells has been a major hurdle to reveal the mechanistic insights of iCM transdifferentiation. Advances in single cell technology enables to decipher the detailed molecular remodeling occurring within each cell thus to identify how FBs with different molecular signatures respond differently to GMT delivery. It takes about a month to identify spontaneously beating iCMs in vitro, though single cell RNA-sequencing (scRNA-seq) analysis showed that the cell fate transition is primed within the first 72 hours [94]. Using a trajectory analysis and RNA-velocity technology, Liu et al. identified that only a subset of FBs undergoes cardiac reprogramming and show a dynamic remodelling in a stepwise manner. Unbiased clustering and principal component analysis revealed that cell cycle-related genes (e.g., Cdk1, CCnb1, and Top2a) account for a significant portion of the variability in the data and the predominant FBs undergoing cardiac reprogramming are not proliferative. These FBs progressively express genes involved in mitochondrial respiratory chain complexes and mitochondrial biogenesis throughout the cell fate conversion, whereas genes involved in glycolysis (Hif1a, PFK1) decreased throughout reprogramming [94]. The authors validated that synchronizing the cell cycle or interfering proliferation in FBs enhance cardiac reprogramming, whereas while accelerating proliferation suppress cardiac reprogramming [94]. A follow-up study with human fibroblast reprogramming also found evidence of metabolic shift from glycolysis to OXPHOS during iCM transdifferentiation [95]. Here, authors observed that the expression of genes related to glycolysis and one-carbon metabolism is inversely correlated with the changes with TBX5 downstream target gens. One carbon metabolism is an important regulator to maintains the low oxidative state in proliferative cells [96]. A separate study provided the mechanistic insight [97] by that certain TF sequence motifs are associated with changes in transcriptome in FBs that had undergone iCM reprogramming. The study identified 48 motifs changed significantly throughout cardiac reprogramming and performed knock-down experiments on 18 TFs to modulate the reprogramming efficiency. FBs undergone a successful cardiac reprogramming show a significant decrease in Hif-1 motif and knocking down the expression of Hif-1α in FBs significantly improved the reprogramming efficiency. As mentioned earlier, Hif-1α is a master regulator of glycolytic metabolism by upregulating glycolysis-related genes (Glut1, Ldha, and Mct4) while simultaneously suppressing mitochondrial OXPHOS [90]. Collectively, a cell fate conversion toward cardiomyocyte involves in a metabolic remodeling from glycolysis to OXPHOS.

5. Enhancement strategy to generate metabolically mature CMs

5.1. Metabolites conditioning

A successful translation of PSC-CMs to the clinic is dependent on how close we can mimic the normal developmental process to direct CMs to become an adult-like state. In a conventional 2D monolayer culturing system, various strategies like extending culturing period or adopting bioengineering techniques (e.g., mechanical loading, electrical stimulation, and micropatterning) have been relatively successful in inducing physical and functional maturation [39, 98]. However, the resultant CMs are still far from completely resembling adult-like cardiomyocytes especially metabolic phenotype (e.g., mitochondrial oxygen consumption, substrate utilization, and mitochondrial morphology). Nevertheless, PSC-CMs are flexible in taking up available carbon substrates that influence the relative levels of glycolysis and OXPHOS [99]. This has led to a hypothesis that alterations in substrate availability in culture media that may drive metabolic maturation.

Exogeneous supplementation of fatty acids in the culture media can enhance the functional and metabolic phenotypes of PSC-CMs. For instance, supplementing palmitate/oleate for 3 days significantly enhanced OXPHOS and ATP production [100]. The inclusion of insulin with exogenous supplementation of fatty acids for 3 or 7 days improved the metabolic phenotype (increased oxygen consumption, mitochondrial number, and fatty acid utilization) with structural phenotypes (including features such as sarcomere length, calcium transient, and electrophysiology) [101, 102]. Extending culturing with exogeneous lipid supplementation for 14 days further potentiated the physical characteristics (e.g., increased cell surface area, decreased circularity, increased sarcomere length) while generating a greater contractile force and oxygen consumption rate [103]. Recently, an addition of nuclear receptor agonists (PPARα, PGC1α) has robustly improved structural, functional, and metabolic parameters of PSC-CMs [104]. Metabolites-driven maturation of iPSC-CMs improved the electrophysiological and functional parameter to recapitulate the disease state of patients [105]. However, an exaggerated fatty acid supplementation can lead to lipotoxicity, and adult cardiomyocytes still rely on glucose metabolism to generate a significant portion of energy [67]. A recent thorough metabolomic analysis of mice and human heart confirmed the significance of non-FA source contribution to the adult heart metabolism [106]. Lactate oxidation, in which lactate converts back to pyruvate and feeds the TCA cycle, also contribute to cardiac energetics. A prolonged incubation with galactose/lactate for up to 21 days alone improve the metabolic parameters (e.g. increased oxygen consumption, ATP production and mitochondrial respiratory complex expression) [99]. Intriguingly, lactate supplementation in culture medium is critical in preventing cell death by lipotoxicity due to prolonging fatty acid supplementation [107]. Therefore, a fine balance between lipid and non-lipid metabolism is critical to acquire a true metabolic maturation of cardiomyocytes (Fig 2A).

Figure 2. The role of metabolism in cardiogenesis from PSCs and FBs.

(A) PSCs are differentiated into CMs by modulating BMP/Nodal/Wnt signaling. The resultant PSC-CMs are metabolically immature, exhibiting a high glycolytic and proliferative rate, and Hif1α governs the transcriptomic landscape of immature PSC-CMs. Conditioning PSC-CMs with high lipid and low glucose medium enhance the metabolic maturation by decreasing glucose and lactate metabolism and increasing mitochondrial biogenesis and fatty acid metabolism. (B) FBs can be transdifferentiated into CMs via an ectopic expression of cardiogenic transcription factors GMT (GATA4, Mef2c, TBX5). Additional transcriptional factors (Hand2), micoRNAs and/or small molecules can improve the reprogramming efficiency by enhancing chromatin accessibility to GMT and increasing the expression of genes related to heart development, fatty acid metabolism, and mitochondrial biogenesis. Multiple mechanisms (Fibrosis, Stiffness, Signaling, Cell cycle) underlie repressed or enhanced cardiac reprogramming and various pathways have been identified: repressor – TGFβ, EGF, Rho, Integrin, enhancer – mTOR, Foxo3a, AKT, PI3K.

5.2. TGF-β

A cell fate conversion is governed by cell-type specific signaling events, and TGF-β is the primary signaling governing FB to prevent cardiac reprogramming (Fig 2B). Ilfokovits et al. tested small molecules known to enhance cardiomyogenesis, and found that SB31542, a potent TGF-β signaling inhibitor, increase the reprogramming efficiency up to 5-fold [108]. Inhibition of TGF-β signaling during the early reprogramming process increase the expression of genes involved in cardiac muscle development including PGC-1α, a master regulator of mitochondrial biogenesis [109]. In fact, TGF-β signalling is an important downstream signaling of Nodal/Activin signaling in early cardiac progenitor cell differentiation from PSCs. A persistent activation of TGF-β stimulates non-CMs (smooth muscle cells, endothelial cells) differentiation, whereas an inactivation of TGF-β during the late phase of cardiomyogenesis during PSC-CM differentiation [110]. iPSC-CMs from patients suffering from congenital cardiomyopathy exhibit an aberrant TGF-β signaling pathway which contribute to pathological phenotypes [111]. Therefore, tuning TGF-β signalling is crucial in normal cardiomyogenesis. The efficacy of the enhancement strategy using TGF-β inhibitor has been replicated by other groups and they reported that genes involved in fatty acid metabolism, mitochondrial proteome, and OXPHOS significantly increased in iCMs reprogrammed with GMT and a TGF-β inhibitor compared to iCMS reprogrammed with GMT alone [112, 113]. Recently, Riching et al. showed that TGF-β signaling disrupts the ability of GATA4 to bind to target genes by impairing chromatin accessibility of cardiogenesis gene promoters [114]. GATA4 is a critical TF to regulate a metabolic switch from glycolysis to OXPHOS and a point mutation on GATA4 impairs mitochondrial biogenesis and produce energetic defects in heart [96, 115].

5.3. AKT

A pioneering work by Zhou et al. found that Akt or protein kinase B robustly accelerated and promoted cardiac reprogramming from FBs [116]. FBs infected with GMT with Hand2 and Akt generated more mature CMs (e.g., showing enhanced poly-nucleation, surface area, and Myh6 to Myh7 ratio) and increased expression of genes involved in mitochondrial biogenesis and OXPHOS [116]. They identified that IGF1/PI3K signaling is an upstream signaling and mTOR is a downstream signaling pathway to promote cardiac remodeling. Predictably, pharmacological inhibition of PI3K or mTOR signaling abolished the enhancement of reprogramming efficiency by Akt. A follow-up study adopted multiple ChIP-seq of GMT in different reprogramming experiments (GMT, GMT+Hand2, GMT+Hand2+Akt) and identified how Akt enhances cardiac reprogramming at epigenetic level [117]. They revealed that Akt significantly enhanced the number of GATA4 binding sites and co-binding sites of GATA4-TBX5, which is critical for mitochondrial biogenesis [96]. After a vigorous cross-comparison with in vivo data on genomic occupancy of GATA4 and TBX5, they identified the gene regulatory network by annotating the genes found near the ChIP-seq TF peaks. They identified that genes related to the cell cycle, extracellular matrix, and inflammation pathways were negatively associated with iCM reprogramming. Additionally, they discovered that genes involved in EGF receptor (EGFR) signaling pathways were negatively associated with iCM reprogramming via GMTH + Akt (Fig 2B). EGFR signaling pathway is often upregulated in highly proliferative cells like cancer cells and strongly related to cell proliferation [118, 119]. Although the exact change of the metabolic state in FBs undergoing iCM reprogramming with Akt should be further validated, we can infer they are likely non-glycolytic since the proliferative state of FBs is inversely associated with the efficiency of transdifferentiation.

5.4. Stiffness

The efficiency of in vivo reprogramming is consistently reported to be higher than in vitro reprogramming of FBs [11, 38]. It has been postulated that the relevant cardiac microenvironment can affect the reprograming efficiency. For example, dermal FBs are much harder to be converted into iCMs than if cardiac FBs [120]. In addition, the injection of GMT into the mouse skin did not transdifferentiate iCMs [121]. These studies suggest that a relevant cell origin or microenvironment affect cardiac reprogramming. Among multiple factors contributing to physiologically relevant microenvironment, Kurotsu et al. investigated the role of stiffness in mediating cardiac transdifferentiation [122]. They reported that cardiac reprogramming of FBs cultured on Matrigel-based hydrogels with elasticities relevant to the myocardium (10–20 kPa) significantly improved the reprogramming efficiency and the maturity of iCMs (e.g., in terms of Ca²⁺ transient and contractility), compared to iCMs cultured on the conventional rigid polystyrene dish (~ GPa). RNA-seq analysis showed that gene expression involved in fatty acid metabolism is highly expressed in iCMs cultured on soft materials (8 kPa) compared to iCMs cultured on hard materials (126 kPa). RNA-seq analysis also showed that the expression of YAP/TAZ-related genes were significantly downregulated in iCMs on soft materials [122]. YAP/TAZ are transcriptional co-activators that can sense physical and mechanical cues like extracellular matrix stiffness to control growth, differentiation, and survival of cells [123]. Kurotsu et al. validated that chemical inhibition of mechano-transduction signaling pathways upstream of YAP/TAZ was sufficient to increase the reprogramming efficiency [122]. Although the exact mechanism underlying enhanced metabolic and functional maturation of iCMs on soft materials needs further investigation, emerging evidence suggests that mechanotransduction via YAP/TAZ signaling is closely associated with cell fate conversion [124].

6. Conclusion

Advance in knowledge of normal heart development has evolved the protocol to generate a high quality of PSC-CMs, and the results of preclinical trials in large animal models of cell therapy are very encouraging. However, immature phenotypes of PSC-CMs and multiple technical/logistic problems are hindering the clinical translation. In the current review, we reviewed an underappreciated aspect, metabolic reprogramming, occurred during directed- and transdifferentiation of CMs. We discussed how metabolic guidance can improve and refine the cardiomyogenesis in both methods to acquire mature adult-like cardiomyocytes. Cardiomyocytes become non-proliferative which concomitantly occur with changes in metabolic and redox state. Intriguingly, such changes are observed in cardiac FBs undergone for cardiac reprogramming. Although no studies specifically aimed to manipulate the metabolic state of FBs for improving the efficiency of reprogramming, many existing enhancement strategies are strongly associated with metabolism of cardiac FBs. Cardiac FBs in infarcted hearts are likely in activated state with highly glycolytic and proliferative state, which is known to hinder the reprogramming efficiency. Understanding the metabolic and transcriptional interplay in cell fate conversion will improve the strategies of de novo generation of cardiomyocytes from non-cardiomyocytes and cure cure for heart disease.