Stage-specific regulation of signalling pathways to differentiate pluripotent stem cells to cardiomyocytes with ventricular lineage

Background Pluripotent stem cells (PSCs) can be an ideal source of differentiation of cardiomyocytes in vitro and during transplantation to induce cardiac regeneration. However, differentiation of PSCs into a heterogeneous population is associated with an increased incidence of arrhythmia following transplantation. We aimed to design a protocol to drive PSCs to a ventricular lineage by regulating Wnt and retinoic acid (RA) signalling pathways. Methods Mouse embryonic stem cells were cultured either in monolayers or three-dimensional hanging drop method to form embryonic bodies (EBs) and exposed to different treatments acting on Wnt and retinoic acid signalling. Samples were collected at different time points to analyse cardiomyocyte-specific markers by RT-PCR, flow cytometry and immunofluorescence. Results Treatment of monolayer and EBs with Wnt and RA signalling pathways and ascorbic acid, as a cardiac programming enhancer, resulted in the formation of an immature non-contractile cardiac population that expressed many of the putative markers of cardiac differentiation. The population exhibited upregulation of ventricular specific markers while suppressing the expression of pro-atrial and pro-sinoatrial markers. Differentiation of EBs resulted in early foetal like non-contractile ventricular cardiomyocytes with an inherent propensity to contract when stimulated. Conclusion Our results provide the first evidence of in vitro differentiation that mimics the embryonic morphogenesis towards ventricular specific cardiomyocytes through regulation of Wnt and RA signalling pathways. Supplementary Information The online version contains supplementary material available at 10.1186/s13287-022-02845-9.


Background
Cardiac regeneration via stem cell therapy is a promising field, and stem cell therapy using multipotent adult stem cells have been shown to ameliorate cardiac function. However, the efficacy of multipotent stem cells to differentiate into cardiovascular cells is still in debate, which is also reflected by the inconsistent results from the clinical trials [1][2][3]. Pluripotent stem cells (PSCs) are a potential source of cardiovascular cells due to their ability to differentiate into functional cardiovascular cells in vitro, and hence in replacing the loss of cardiomyocytes as well as vascular cells due to ischemia [4,5]. However, one of the major challenges, and one that has received very little attention, is the need to generate chamber-specific cardiomyocytes. This is critical, as different chambers in the heart have different functions. For instance, the right atrium contains the pacemaker cells and many conduction cells required to initiate and conduct the electrical activity, respectively, whereas the ventricles are the heart's major pumping chambers containing the working cardiomyocytes. Unsurprisingly, our recent data show that the expression profiles of calcium handling proteins, crucial for contraction, are different between both chamber types in humans [6]. In general, PSCs yield a heterogeneous population of cardiomyocytes of which only 30-50% are ventricular cardiomyocytes [7,8]. Since the majority of heart disease occurs in the ventricle, transplantation of a heterogeneous population with distinct functional and electrophysiological properties into the ventricle can contribute to cardiac arrhythmias [9,10]. Recent animal studies have shown that exquisitely coordinated expression of various genes in a spatially and temporally controlled fashion is essential for normal cardiac development, with Wnt signalling, ascorbic acid (AA) and retinoic acid (RA) being key regulators of cardiogenesis [11,12].
Wnt signalling has been reported to play a biphasic role in the differentiation of PSCs to cardiomyocytes where initial induction of Wnt signalling leads to the formation of mesoderm and the inhibition of Wnt signalling at later stages leads to differentiation to cardiomyocytes [13]. AA has been shown to enhance cardiac differentiation by upregulating the expression of transcription factors involved in the proliferation of cardiac progenitors in both mouse and human PSC models [14,15]. On the other hand, RA has a significant role in specifying atrial and ventricular cardiomyocyte lineage in mouse and human PSCs. Positive signals of RA would lead to atrial specific cardiac differentiation, whereas inhibition of RA signalling drives the differentiation towards ventricular lineage [16]. These data therefore show that the individual modulation of Wnt, AA and RA signalling cascades can differentiate PSCs to cardiomyocytes. However, it is unknown whether combinatorial treatment with AA and temporal regulation of Wnt and RA signalling pathways, in a stage-specific manner, would lead to improved differentiation into a more homogeneous population of ventricular cardiomyocytes by suppressing the pacemaker population of cells. Therefore, the rationale of this study is to demonstrate if temporal regulation of Wnt and RA signalling pathway can differentiate PSCs to chamberspecific cardiac lineage.
In the current study, mouse embryonic stem cells (mESCs) were used as an in vitro PSC model. They were driven to cardiac lineage differentiation by regulating Wnt and RA signalling in the presence of AA as a cardiac programming enhancer. Cardiac differentiation in monolayer resulted in an immature non-contractile cardiac population that expressed all the putative markers of cardiac differentiation. They specifically exhibited upregulation of ventricular specific markers while suppressing the expression of pro-atrial and pro-sinoatrial markers.

mESC culture and characterization
A transgenic TNG-A (a clone of transgenic Nanog) mESC strain and its corresponding predecessor control E14 mESC strain were used for all the experiments [17]. The TNG-A strain displays a green fluorescence protein (GFP) reporter tagged to one allele of the Nanog gene, whereby GFP expression corresponds to the expression of pluripotent protein Nanog (Additional file 1: Figure  S1) [18]. The loss of GFP indicates the corresponding loss of pluripotent Nanog expression. The mESCs were seeded and expanded in culture dishes that were precoated with autoclaved and sterile-filtered 0.1% gelatin (Sigma-Aldrich, NZ). The mESCs were supplemented with mESC medium [High glucose Dulbecco's modified eagle's medium (DMEM, Thermofisher, NZ), 15% fetal bovine serum, 1X non-essential amino acids (Sigma-Aldrich, NZ), 100 µM β-mercaptoethanol (Sigma-Aldrich, NZ), 1 μg/ml puromycin, and 10 ng/ml leukemia inhibitory factor (LIF) (Thermofisher, NZ)]. The mESCs were maintained at 37˚C in 5% CO 2 in a humidified normoxic environment. The cells were dissociated with TrypLE (Thermofisher, NZ) and passaged once every 3-4 days when they attained 70% confluency to avoid spontaneous differentiation induced by over confluency. mESCs were characterized for their pluripotency before exposing them to differentiation conditions (Additional file 1: Figure S2).

Cardiac differentiation of mESCs
Monolayer differentiation mESCs (200,000 cells/well) were seeded in a 24-well plate and allowed to grow for 3 days until they reached 80% confluent. To induce cardiac differentiation, the culture medium was replaced with serum-free medium, and cells were treated with AA (Sigma-Aldrich, NZ), IWP2 (Wnt inhibitor (Wi), Sigma-Aldrich, NZ) and BMS-189453 (RA inhibitor (RAi), Sigma-Aldrich, NZ) at different points as shown in Fig. 1A. Cells were harvested on days 3, 5, 9 and 14 for characterization.

Three-dimensional (3D) differentiation by embryoid body (EB) formation
EBs were generated using the three-dimensional hanging drop (HD) technique. For this, mESCs (1000 cells/10 µL) were suspended in the EB differentiation medium and placed as a drop (10 µL) on the inner side of the lid of a petri dish (Sigma-Aldrich, NZ). Around 40-50 drops were evenly placed per dish (Fig. 1B). The base of the petri dish was filled with 10 mL of sterile 1X phosphate buffered saline (PBS) as a source of humidity to the cell cluster in the hanging drops. The lid was then inverted, closed and cultured at 37 °C in a 5% CO 2 incubator. Spherical EBs were visibly formed by day 2. The EBs on day 3 were gently flushed and plated on uncoated Petri dishes in EB differentiation medium supplemented with AA and Wi followed by treatment with RAi as shown in Fig. 1B.

Quantitative real time-PCR (RT-PCR) analysis
Total RNA isolated using TRIzol (Thermofisher) was reverse-transcribed using PrimeScript ™ RT Reagent Kit (Takara Bio, Japan). The resulting cDNA was amplified using SYBR green premix Ex Taq II (Takara Bio, Japan) in StepOne plus PCR machine (Thermo Fisher Scientific, USA) to quantify the expression of genes associated with pluripotency (Nanog, Oct4 and Sox2) and The day 0-mESCs was used as a negative control for the experiment to normalize the transcript levels of differentiated cells. The entire differentiation process, sample processing, and qPCR were independently repeated three times (n = 3). Data are presented as log fold change to day 0-mESCs. The list and sequence of all the primers used in this study are listed in Additional file 1: Table S1.

Flow cytometry analysis
Cells in suspension were fixed with 4% PFA for 20 min at room temperature and probed with antibodies against pluripotency markers (Oct3/4, Sox2) and cardiac differentiation markers (T brachyury (T-Bry), KDR, ISL-1, Nkx2.5, cTnT, IRX4, MLC2V and connexin-43 (Cx43)). Details of antibodies, including the concentration used, are listed in Additional file 1: Table S2. To compensate for the use of multiple fluorescent antibodies, single stain controls comprising of antibody capture beads were used (AbC ™ anti-mouse bead kit, Thermo Fischer Scientific, USA). Cell acquisition was performed using the flow cytometer (Gallios, Beckman Coulter Inc., USA). In total, 5000 to 50,000 events within the gate of interest were acquired depending on the sample type. Flow cytometer voltages were set using an unstained, single stained bead or secondary antibody controls so that the spill-over was minimum in the adjacent channels. Post-acquisition compensation and data analysis were performed using Flowjo 10.6.1 software (TreeStar, USA) [19].

Immunofluorescence analysis
Monolayer cells seeded on glass coverslips in a 24-well plate were fixed with 4% PFA and permeabilized with pre-chilled methanol for intracellular staining. The cardiac cells of EBs were dissociated by collagenase IV/ Trypsin treatment, singularized and seeded onto 0.1% gelatin-coated coverslips before fixation and permeabilization. After blocking with either 10% serum or 1%BSA, cells were incubated with primary antibodies against NkX2.5, Cx43, α-sarcomeric actin, IRX4 and MLC2V (all antibodies and concentrations listed in Additional file 1: Table S2) and incubated overnight at 4 °C. After washing, the cells were probed with secondary antibodies (1:500 dilution) conjugated with Alexa flour. Finally, cells were stained with DAPI (0.1 ug/mL, Thermo Fischer Scientific, USA) to visualize the nuclei and mounted on glass slides using Fluoromount G (Southern Biotech, USA). Images were captured in NikonA1R Resonant Scanner confocal microscope and analysed using ImageJ software, version 1.52 (NIH, USA). A control group for secondary antibodies (with the same labelling procedure without the primary antibody treatment, Additional file 1: Figure S3) was used to rule out the non-specific binding of the secondary antibody.

Transmission electron microscopy (TEM)
The cells were fixed in 2.5% glutaraldehyde (primary fixative) and 1% osmium tetroxide (secondary fixative) dissolved in sodium cacodylate buffer. Post-fixation, Fig. 1 Schematic diagram of the protocols to differentiate mESCs to cardiovascular lineage. A Monolayer differentiation of mESCs to cardiomyocyte lineage. The mESCs were cultured in mESC growth medium (MGM) for three days (day − 3 to 0). On day 0, the medium was replaced with RPMI/ B27-insulin as the serum-free differentiation medium (SFDM). Cells were treated with ascorbic acid (AA) from day 0, which continued until the end of the differentiation protocol (day 14). From days 3-5, cells were treated Wnt inhibitor (Wi), and from days 6-9, they were treated with retinoic acid inhibitor (RAi). B Three-dimensional differentiation of mESCs by EB method. EBs were generated by the hanging drop technique. On day 0, mESCs were dissociated, counted, and pipetted as drops of media on the inner side of a petri dish lid. The lid was inverted, and the petri dish was closed to form hanging drops. On day 2, spherical EBs was formed. On day 3, EBs were flushed with SFDM and cultured in suspension in an uncoated petri dish. On day 5, EBs were transferred to a 24-well plate containing 0.1% gelatin-coated coverslips. From day 5 to 30, EBs were cultured and differentiated in a monolayer. From day 6-9, cells were treated with RAi. the cells were dehydrated by washing with an ethanol series (50/70/90/100% ethanol, 5 min each). The cells were then treated with 100% propylene oxide, resin infiltrated, embedded in resin-filled silicone moulds, and finally were cured at 60 °C for 48 h. Post-curing, ultrathin sections (85 nm) were cut with an ultramicrotome (Leica EM UC7) and transferred to a copper grid (Sigma-Aldrich, NZ). The sections were then stained with 4% uranyl acetate for 25 min, rinsed in Milli-Q water, and transferred to 1% lead citrate for 5 min before rinsing in Milli-Q water and air drying. Images were captured with a Philip CM100 BioTWIN transmission electron microscope (Philips/FEI Corp., Eindhoven, Holland). The following parameters were measured; cell area (µm 2 ), perimeter (µm), and circularity index [20]. The differentiated cells were identified based on the presence of multinucleated cardiac cells, increased cell size (area and perimeter), reduced circularity index, increased mitochondrial number, the presence of lipid droplets and glycogen bodies as some of the putative morphological characteristics observed during cardiac differentiation.

Cytosolic Ca 2+ imaging of differentiated cardiac cells
To measure the cytosolic Ca 2+ , differentiated cells were loaded with fluo-4-AM (2 μM, Thermofisher, NZ), a cell-permeable acetoxymethyl ester (AM) form of Ca 2+ indicator for 10 min at room temperature. Cells were then washed twice with Krebs-Ringer HEPES (KRH) buffer to remove any de-esterified dye. Fluorescence of fluo-4 was detected by imaging the cells using CoolS-NAP HQ2 CCD camera (Nikon). The analysis of cytosolic Ca 2+ imaging was done using NIS-Elements AR Analysis 4.00.03 64-bit software (Coherent Scientific Pty. Ltd, Australia).

Western blot analysis
mESCs at different time points of differentiation were homogenized in ice-cold RIPA lysis buffer supplemented with a 1% protease inhibitor cocktail (Sigma Aldrich, USA) [21]. The total protein was quantified by Bradford protein assay (BioRad, New Zealand). Twenty micrograms of protein were resolved by 15% SDS-PAGE and transferred to polyvinylidene fluoride membrane. After blocking and confirming the successful transfer of proteins using ponceau staining, membranes were probed with primary antibodies against CX43 (1:1000 dilution, Sigma Aldrich), MLC2V (1:1000 dilution, Abcam), and β-actin (1:1000 dilution, Santacruz Biotechnologies) overnight at 4 °C. Goat antirabbit secondary antibody (1:10,000 dilution, Abcam) was used for detection. The density of bands was analysed using ImageJ (NIH, United States) software and normalized to GAPDH.

Experimental repeats and statistical analysis
The experimental repeats (n) were assay dependent. Quantitative real-time PCR, flow cytometry, and calcium imaging assays were performed with three independent repeats (n = 3). For TEM analysis, 10 cells in each sample were used for analysis. Spontaneous Ca 2+ activity in the stimulated EBs was performed with at least 8 EBs. All the statistical analyses were done using Graph Pad Prism 9 software. Statistical significance was determined by (1) Student's t-test to identify the difference between two groups (2) a one-way ANOVA with Dunnett's or Sidak's multiple comparison post hoc test to identify the difference when there were more than two groups, and (3) a two-way ANOVA with Tukey's or Sidak's multiple comparison post hoc test to compare more than two groups with two independent variables. Data are presented as mean ± SEM. P < 0.05 was considered statistically significant.
We further confirmed that mESCs do not "spontaneously" differentiate into cardiac lineage when cultured for 14 days in the absence of signalling regulators (Additional file 1: Figure S4).

The ultrastructure of differentiated mESCs revealed early-stage cardiomyocytes features
Differentiated mESCs (Fig. 4E-H) were larger compared to the small spherical undifferentiated mESCs ( Fig. 4A-D) and displayed heterogeneous shapes (spherical, triangular or cylindrical) (Fig. 4E-H). Quantitative analysis showed an increase in cell area (83.18 ± 5.107 µm 2 vs 600 ± 53.32 µm 2 , P < 0.0001) and perimeter (30.7 ± 1.075 µm vs 119.4 ± 6.596 µm, P < 0.0001), as well as decrease in circularity index (0.916 ± 0.004 vs 0.547 ± 0.05, P < 0.0001) (Fig. 4I-L). The majority of the differentiated cells were multinucleated, which is crucial before neonatal CMs are terminally differentiated. In contrast to the spherical immature mitochondria with little or no cristae in the undifferentiated mESCs (Fig. 4D), the mitochondria in the differentiated cells showed crisp membrane structures and cristae formation (Fig. 4H) and were high in number (Fig. 4L). The large number, size and mature cristae in the mitochondria of the differentiated cells could be attributed to their high demand for oxidative phosphorylation and hence, ATP production.

Cardiac differentiation of mESCs from EBs
Monolayer differentiation did not show any maturation in morphology (Fig. 4) or Ca 2+ activity (Additional file 1: Figure S5), which may be attributed to the monolayer cell culture not representing the in vivo embryonic niche. We next induced cardiac differentiation of mESCs using the EB method, which provides a 3D environmental niche. EB-differentiated mESCs exhibited mature morphology compared to the cells differentiated in monolayers (Additional file 1: Figure S6). This was also confirmed by immunofluorescence analysis of cardiac transcription factors Nkx2.5 (Additional file 1: Figure S7A) and Isl-1 (Additional file 1: Figure S7B), and ventricular lineage markers IRX4 (Additional file 1: Figure S7C), MLC2V (Additional file 1: Figure S7C), Mef2c (Additional file 1: Figure S7D) and Cx43 (Additional file 1: Figure S7E). Flow cytometry analysis also confirmed the expression of IRX4 (Additional file 1: Figure S7F) and structural protein cTnT (Additional file 1: Figure S7F). Using this model, we determined the importance of RA inhibition (RAi) in ventricular specific differentiation of mESCs.

EB-differentiated mESCs displayed spontaneous contractility and cytosolic Ca 2+ activity
In contrast to the cells differentiated in monolayers, EBdifferentiated mESCs which were not treated with RA inhibitor (-RAi) showed spontaneous contractility, while this was not observed in the EB treated with RA inhibitor (+RAi, Fig. 5A). The area under the curve of averaged contracting EBs increased to 565 ± 25.2 in +RAi cells after 20 days of differentiation compared to 26.7 ± 5.8 − RAi cells (P < 0.001). This correlated with the spontaneous Ca 2+ peaks in the −RAi (Fig. 5B) and higher expression of calcium release channel ryanodine receptor 2 (RyR2) (Fig. 5D). Spontaneous Ca 2+ peaks were not observed in +RAi cells (Fig. 5B). However, when electrically stimulated, both −RAi and +RAi EBs demonstrated spontaneous contraction and hence the spontaneous Ca 2+ peaks (Fig. 5C).

RAi downregulated sinoatrial and atrial specific gene/protein expression
Similar to the monolayer differentiation, RT-PCR analysis confirmed the downregulation of pluripotent gene Oct4 (Fig. 6A) and upregulation of cardiac progenitor gene Nkx2.5 (Fig. 6B) and structural gene cTnT (Fig. 6C) in both +RAi and −RAi groups. The +RAi group showed a significant upregulation in the expression of ventricular-specific progenitor gene Tbx5 (Fig. 6D). At the same time, there was no significant difference in the expression of adult ventricular genes IRX4 (Fig. 6E) and MLC2V (Fig. 6F) between the groups. However, RA inhibition prevented the upregulation of pro-atrial (Isl-1 (Fig. 6G) and MLC2a (Fig. 6H)) and sinoatrial genes (T-box transcription factor 3 (Tbx3)) ( Fig. 6I). This is supported by the absence of contracting EBs in +RAi EBs (Fig. 5A). These results suggest that AA and Wnt inhibition is sufficient to drive the differentiation towards ventricular phenotype. However, RA inhibition could act as a determining step to downregulate atrial and nodal subtypes in the differentiated cells, thus driving the cells towards ventricular specific differentiation.

RA inhibition increased mature cardiomyocytes phenotype of differentiated mESCs
Like monolayer-differentiated mESCs, the EB-differentiated mESCs were multinucleated with increased mitochondrial number and length, accumulation of lipid droplet, glycogen granule deposition, the SR calcium stores, and presence of clusters of rough-ER (Fig. 7A-J).

Discussion
We have developed a unique combinatorial protocol to differentiate mESCs to cardiovascular lineage in a stage-specific manner. Although there are a plethora of protocols to differentiate PSCs to heterogeneous cardiomyocyte populations (nodal, atrial, and ventricular), to our knowledge, this is the first study to develop a protocol to differentiate PSCs to cardiomyocytes that are oriented towards ventricular-lineage without automaticity, via sequential modulation of the Wnt and RA signalling pathway in the presence of AA [15,22,23]. In this study, we differentiated the cells by growing them in monolayers and in EBs. In both models, RAi played a crucial role in differentiating the mESCs to a ventricular specific phenotype. Using the monolayer cardiac differentiation protocol, we were able to demonstrate a stage-specific upregulation of cardiovascular precursors (KDR and Mesp-1), cardiac progenitors (Isl-1, Nkx2.5, and Tbx5), and ventricular cardiomyocyte markers (Tbx5, IRX4, and MLC2V) and downregulation of pluripotent markers (Oct-4 and Nanog). Our results showed that AA specifies cardiovascular fate in mESCs, and inhibition of Wnt signalling induces cardiac mesoderm population. This is likely due to the interplay of epigenetic changes in DNA methylation of cardiac genes such as Isl-1 and the integrin signalling system [24]. In support of this, Isl-1 expression was upregulated at both gene and protein level as early as day 3 of differentiation in our study. Myocyte enhancer factor 2C (Mef2C) is a direct transcriptional target of Isl-1, and the expression of Mef2C is a sign of gradual cardiac looping morphogenesis during in vitro cardiac differentiation [25]. Sustained increase in Mef2C protein expression from day 3 to 14 of differentiation in our study signifies that the cardiac cells were undergoing an event of morphogenesis throughout the differentiation. Nkx2-5 is one of the earliest known transcription factors required for cardiac cell specification and maintains the ventricle chamber identity. Nkx2.5 mutants had a reduced number of ventricular cardiomyocytes and excess atrial cells [26]. The double mutants for Nkx2-5 and Heart and Neural Crest Derivatives Expressed 2 (HAND2) were shown to completely abolish the expression of the ventricular-specific homeobox gene Irx4. Nkx2.5 is also reported to be essential in maintaining the ventricle chamber identity [26,27]. This evidence reiterates the significance of Nkx2.5 expression in early cardiac mesoderm specification, proliferation, heart tube morphogenesis, and  [28,29] and pacemaker cells [30], respectively. Upregulation of both Tbx5 and Tbx3 from day 3 differentiation suggested an additional role of AA in specifying the ventricular and pacemaker progenitor pool. While TBX5 gradually upregulated and sustained with RA inhibition, Tbx3 showed a gradual downregulation suggesting that the current treatment protocol does not promote specification or differentiation into pacemaker cells. This could partially explain the absence of automaticity in our differentiated cells. Indeed, EBs differentiated in the absence of RAi exhibited spontaneous contraction. Further, RA inhibition also induced downregulation of pro-atrial gene MLC2a while upregulating pro-ventricular gene MLC2V. Although we did not explore the underlying mechanism by which RAi causes ventricular lineage differentiation, Wang et al. demonstrated that the formation of an inhibitory complex by IRX4 with retinoic X receptor (RXR) downregulated the expression of atrialspecific a unique myosin heavy chain (AHMC1), while promoting differentiation into the ventricular phenotype [31]. Based on this, upregulation of IRX4 in our study following RAi possibly increased its binding to RXR, inducing ventricular lineage differentiation. This is further supported by reduced IRX4 and MLC2V in EB-differentiated cells in which RA was not inhibited.
Monolayer differentiation resulted in cells displaying the features of immature embryonic cardiomyocytes such as deposition of glycogen granules, lipid droplets, and rough ER. Cardiac glycogen is present at high levels during early to mid-gestation before falling to low levels at the time of birth, thereby playing a critical role as an energy source during cardiogenesis [32]. Lipid droplets act as a triglyceride pool for the mitochondria that may serve as a crucial source of signalling intermediates to govern the mitochondrial function and fatty acid oxidation by regulating the expression of peroxisome proliferator-activated receptor-gamma coactivator-1-alpha (PGC-1α) and PGC-1-beta(β) and the activity of Peroxisome proliferator-activated receptor-α (PPAR-α) [33]. Further, despite the expression of cTnT and α-sarcomeric actin, the ultrastructural analysis failed to identify sarcomeres. Asmuda et al. demonstrated the expression of cTnT in the nucleus of neonatal cardiomyocytes, indicating that cardiac troponin and tropomyosin could have an important cellular function(s) beyond Ca 2+ regulation [34]. Although monolayer differentiated cardiomyocytes expressed RyR2 (Additional file 1: Figure S5C), a protein channel responsible for balancing Ca 2+ homeostasis by effluxing excess of Ca 2+ from its Ca 2+ store sarcoplasmic reticulum (SR) upon stimulation [35], they failed to emit any Ca 2+ peaks when stimulated with Ca 2+ infusion. This could be possibly due to the RyR2 channels being not sensitive enough to respond to their external calcium cues. Interestingly, differentiation by EB formation showed mature cardiomyocyte phenotype, suggesting that mimicking embryonic morphogenesis results in the generation of morphologically and functionally mature CMs [36]. However, in spite of the expression of α-sarcomeric actin, these cells still lack sarcomeres, which is a functional unit of the mature cardiomyocytes. Therefore, future studies should aim to focus on molecular pathways that could trigger the formation of sarcomeres required for functional excitation-contraction coupling.

Conclusions
The current study features a novel protocol to differentiate mESCs into cardiomyocytes of ventricular lineage. Our results indicate that stage-specific combinatorial treatment with AA and Wnt and RA inhibition can induce differentiation of mESCs to ventricular lineagespecific cardiomyocytes by suppressing the expression of pro-atrial and pro-sinoatrial lineage. Differentiation via monolayer generated non-contracting ventricularspecific cardiomyocytes. EB generated cardiomyocytes were spontaneously contracting, which was not observed when treated with RA inhibitor, thereby demonstrating the importance of RA inhibition in shifting differentiation of cardiomyocytes towards ventricular lineage. Moreover, the electrically stimulated +RAi cells showed spontaneous contraction and Ca 2+ release similar to −RAiEBs. Next critical step is to determine if in vivo transplantation of the homogenous population of cardiomyocytes that specifically express ventricular lineage markers into the ischemic left ventricle shows superior fusion with the host cardiomyocytes and exhibit better therapeutic efficacy. This will allow for a significant advancement in increasing the efficacy of stem cells therapy in regenerating the diseased heart. Moreover, with the recent use of induced pluripotent stem cells (iPSCs) in regenerative therapy in the clinic and its ability to differentiate into cardiomyocytes, it will be crucial to test the efficacy of our differentiation strategy in iPSCs to determine its translation ability [37].