PGC-1α activator ZLN005 promotes maturation of cardiomyocytes derived from human embryonic stem cells

Human pluripotent stem cell-derived cardiomyocytes (hPSC-CMs) have great potential in biomedical applications. However, the immature state of cardiomyocytes obtained using existing protocols limits the application of hPSC-CMs. Unlike adult cardiac myocytes, hPSC-CMs generate ATP through an immature metabolic pathway—aerobic glycolysis, instead of mitochondrial oxidative phosphorylation (OXPHOS). Hence, metabolic switching is critical for functional maturation in hPSC-CMs. Peroxisome proliferator-activated receptor gamma coactivator 1α (PGC-1α) is a key regulator of mitochondrial biogenesis and metabolism, which may help promote cardiac maturation during development. In this study, we investigated the effects of PGC-1α and its activator ZLN005 on the maturation of human embryonic stem cell-derived cardiomyocyte (hESC-CM). hESC-CMs were generated using a chemically defined differentiation protocol and supplemented with either ZLN005 or DMSO (control) on differentiating days 10 to 12. Biological assays were then performed around day 30. ZLN005 treatment upregulated the expressions of PGC-1α and mitochondrial function-related genes in hESC-CMs and induced more mature energy metabolism compared with the control group. In addition, ZLN005 treatment increased cell sarcomere length, improved cell calcium handling, and enhanced intercellular connectivity. These findings support an effective approach to promote hESC-CM maturation, which is critical for the application of hESC-CM in disease modeling, drug screening, and engineering cardiac tissue.

In vivo, cardiomyocytes undergo multiple adaptive, structural, functional, and metabolic alterations during maturation, while hPSC-CMs fail to fully experience these developmental processes [14]. The mitochondrial structure and function change drastically during cardiac development [15,16]. A transition in energy metabolism is implicated in cardiogenesis and maturation, including mitochondrial biogenesis, fuel shift, mitochondrial quantity, and morphology [17,18]. Mitochondria in immature cardiomyocytes occupy a small fraction of cell volume, and distribute irregularly in the cytoplasm [19,20], whereas mitochondria in mature cardiomyocytes occupy ~20-40% of the myocyte volume, and have extensive surface area provided by densely packed cristae [19,20]. During development, fetal cardiomyocytes reside in a hypoxic environment and mainly rely on glycolysis to generate ATP. On the contrary, adult cardiomyocytes mostly depend on mitochondrial oxidative phosphorylation for ATP generation, especially through fatty acid oxidation (FAO) [8,21,22]. The adult heart is a high energy consumption organ and always changes its workload, so it is critical for the heart to adjust its energy output to properly support contraction. Hence, strategies promoting metabolic maturation and metaboliccontraction coupling are promising to promote cardiomyocytes' maturation [6,18,23].
Cardiac energy metabolism is regulated by several molecular pathways during cardiac development. The peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1) family consists of three isotypes-PGC-1α, PGC-1β, and PGC-1-related coactivator (PRC)-that function as transcriptional co-activators [24]. PGC-1s activate a variety of coactivated genes, including estrogen-related receptors (ERRs), nuclear respiratory factors (NRFs), and peroxisome-proliferating receptors (PPARs), and they can be regulated at both the transcriptional and post-translational levels [25]. Of all the isoforms, PGC-1α is a key regulator of mitochondrial biogenesis and metabolism, and it is highly expressed in organs with high energy oxidative capacity, such as the heart [26]. By modulating the binding activity of different transcription factors, PGC-1α upregulates the expression of various genes involved in diverse metabolic pathways, including fatty acid oxidation, glycolysis, fatty acid synthesis, and mitochondrial biogenesis [27][28][29]. In addition, it has been reported that PGC-1α is repressed in cardiac hypertrophy, which accompanied a decrease in fatty acid β-oxidation and mitochondrial oxidative capacity [27,30,31]. Conversely, PGC-1α overexpression induced the expression of a number of coactivated genes (ERRs, PPARs, and NRFs), which virtually enhanced all aspects of mitochondrial biogenesis and energy metabolism [24,32,33]. Collectively, PGC-1α can function as a metabolic regulator, which makes it an attractive target to promote the maturation of cardiomyocytes.
In this study, we tested the hypothesis that increasing PGC-1α pathway with the small compound ZLN005 would enhance mitochondrial metabolism, ensure proper coupling of energy production and force generation, and improve the maturation of hESC-CMs. To this end, the early stage hESC-CMs (day 10) were treated with ZLN005 for 48h and then subjected to molecular and functional analysis at day 30. Our results show the underlying mechanism controlling the metabolic maturation of stem cell-derived cardiomyocytes and demonstrate the application of ZLN005 as an effective and convenient approach in enhancing various aspects hPSC-CMs maturation.

Expression profile of PGC-1α during cardiomyocyte differentiation of hESCs
The PGC-1 family consists of a group of transcriptional coregulators that coordinates mitochondrial biogenesis and is highly expressed in cardiac tissue [26]. To investigate which PGC-1 family member may facilitate hPSC-CMs maturation, we examined the expression profile of PGC-1 family members during cardiac differentiation of hESCs. As shown in Figure 1A, the AGING expression of PGC-1α mRNA was significantly upregulated from cardiac differentiation of hPSCs at day 5 and day 6, reached the peak at differentiation day 7, and then sustained high expression during days 15-30. In addition, the protein level of PGC-1α also dramatically increased from differentiation day 7 ( Figure 1B). However, there was no significant change in the expression of PGC-1β mRNA during CM differentiation and maturation (Supplementary Figure  1B). Hence, manipulating PGC-1α might be an effective approach to promote hPSC-CM maturation.

ZLN005 treatment upregulated the levels of PGC-1α mRNA and protein in hESC-CMs
It has been shown that treatment with 10 μM ZLN005 for 48h can effectively increase the expression of PGC-1α in skeletal and cardiac myocytes [35,37,38,41]. In this study, to test whether the biological effects of ZLN005 were relevant with PGC-1α expression in hESC-CMs, we generated hPSC-CMs and treated hESC-CMs with ZLN005 (10 μM) from day 10 to day 12, as depicted in Figure 1C. In brief, early hESC-CMs (day 10) were treated with DMSO (control) or ZLN005 at a final concentration of 10 μM for 48h, and then subjected to biological and functional analyses at day 30. As shown in Figure 1D, 10 μM ZLN005 caused a 1.7 fold increase of PGC-1α mRNA levels over the control after 48h of treatment. PGC-1α protein expression was significantly elevated in the ZLN005treated groups compared with control groups, as evident in Figure 1E. These results demonstrate that ZLN005 increases the expression of PGC-1α in hESC-CMs at both the transcript and protein levels.

ZLN005 improved mitochondrial maturation of hESC-CMs
As a coactivator of the PGC-1 family, the estrogen related receptors (ERRs) play a key role in regulating genes involved in biogenesis, respiratory function, and mitochondrial fatty acid oxidation [42]. We therefore examined the expression levels of PGC1 family and ER family. After treatment with ZLN005, the expression of PGC-1α and ERRγ mRNA was significantly upregulated ( Figure 1D). In addition, qRT-PCR analysis revealed that gene expression for mitochondrial biogenesis (GABPA, TFAM), mitochondrial OXPHOS (NRF1, CS (TCA cycle), NDUFA4 (i), SDHB (II), COX5a (IV), ATP5b (V)), and fatty acid β-oxidation (CPT1b, CPT2, SLC25a20, ACADM, ESCHS1, HADHA) markers were significantly upregulated in the hESC-CMs following ZLN0025 treatment relative to the control (Figure 2A, 2B). To determine whether the ZLN005 treatment had effects on the mitochondrial DNA content, we examined the mtDNA copy number. qPCR analysis revealed the mtDNA copy number was 1.5-fold higher in the ZLN005-treated hESC-CMs than in the control ( Figure 2C). Transmission electron microscopy analysis showed that the mitochondrial number in the ZLN005-treated hESC-CMs was significantly increased ( Figure 2D) and ZLN005-treated hESC-CMs contained more mitochondrial cristae formation than those in the controls ( Figure 2D). To characterize mitochondrial energy metabolism, we further analyzed the oxygen consumption rate (OCR) values and ATP production between ZLN0025 treatment and control hESC-CMs. As shown in Figure 2E-2G, the cardiomyocytes treated with ZLN005 had significantly higher basal OCR and levels of ATP per cell compared with the controls.

ZLN005 improved structural maturation of hESC-CMs
Immunostaining results of α-actinin showed clear sarcomere structures in both ZLN005-treated and the control groups ( Figure 3A); however, the hESC-CMs treated with ZLN005 had longer sarcomeres

ZLN005-treated hESC-CMs show improved electrical activity and synchronicity
Gap junction morphology is an important parameter for evaluating the maturation state of hPSC-CMs. Connexin 43 (CX43), a major component of gap junctions, has fundamental roles in excitable tissues by facilitating rapid transmission of action potentials between adjacent cells. We therefore tested whether ZLN005 treatment affected CX43 expression. As shown in the supplementary movies (Supplementary Movie 1-4), ZLN005-treated hESC-CMs exhibited more homogeneous, spontaneous beating rhythm and electrical propagation. Immunostaining results showed that ZLN005 treatment upregulated the density of CX43 ( Figure 4A). The enhanced levels of CX43 in the ZLN005-treated hESC-CMs were confirmed by qRT-PCR ( Figure 4B) and Western blot ( Figure 4C). Furthermore, electrical propagation and activity of CMs were evaluated by measuring surface potentials on Microelectrode arrays (MEAs). The color map represents a snapshot of the electrical propagation in the monolayer where the direction of the arrow indicates the direction of local electrical propagation, suggesting that hESC-CMs developed a more homogeneity electrical propagation due to ZLN005 treatment ( Figure 4D). Representative sample traces of spontaneous beating in control (black) and ZLN005 treated (red) CMs are shown in Figure 4D, 4E, respectively. Detailed analyses revealed that the ZLN005-treated CMs exhibited greater field potential amplitude (FPA) ( Figure 4G) than those in the control cells. Further analyses showed that ZLN005 treatment significantly increased the frequency of spontaneous beatings (1.11 ± 0.21 Hz vs 1.61 ± 0.12 Hz) ( Figure 4H) and decreased the inter-spike interval time (1122.1 ± 205.1 ms vs 598.8 ± 68.42 ms). In addition, the electrical activity exhibited higher regularity and reduced variability in both parameters as evident by the AGING reduced coefficient of variation, indicating a high degree of temporally synchronized depolarization and strong electrical coupling across the cell monolayer. These data demonstrate that ZLN005 treatment significantly improves electrical activity and synchronicity of hESC-CMs.

ZLN005-treated hESC-CMs exhibited more negative resting membrane potential
To further determine the maturity of electrical properties of the hESC-CMs, we performed electrophysiological assessments (representative spontaneous action potential traces, Figure 5A) and found that ZLN005-treated hESC-CMs had a more negative resting membrane potential compared with cardiomyocytes from the control group (−58.83 ± 1.28 vs. −63.47 ± 1.14 mV, P=0.01, Figure 5B) and no significant differences in peak amplitude ( Figure 5C). The action potential duration at 90% repolarization (APD90) and the maximum upstroke velocity were significantly higher in the ZLN005 treated CMs ( Figure 5D, 5E).

ZLN005-treated hESC-CMs exhibited more mature calcium transient kinetics
Since Ca 2+ transients are essential for the regulation of excitation-contraction coupling in CMs, we next determined the effects of ZLN005 treatment on the Ca 2+ transients. qRT-PCR results showed that ZLN005-treated hESC-CMs exhibited higher expression of genes encoding for proteins involved in Ca 2+ handling (ATP2A2, RYR2, PLN) ( Figure 6A). To determine whether the changes in these genes affected hESC-CMs Ca 2+ transients, spontaneous intracellular calcium ([Ca 2+ ]i) transients in the hESC-CMs were recorded with Fura-2AM dye. Representative traces of Ca 2+ influx for the two groups are shown in Figure 6B.  Figure 6D, 6F). Taken together, these results provide evidence for the enhanced Ca 2+ transient kinetics in the ZLN005-treated cells.

DISCUSSION
hPSC-CMs are promising candidates for drug development, human disease modeling, and cardiac toxicity screening in vitro. In order to promote the practical application of hPSC-CMs, numerous studies have aimed to enhance hPSC-CMs' maturation using chemical, genetic, and biomechanical approaches. Here, we focus on the bioenergetics of hESC-derived cardiomyocytes and functional impact of its manipulation. We demonstrated that increasing PGC-1α strongly stimulated mitochondrial biogenesis and promoted cardiomyocyte maturation. For the first time, our results revealed that the small molecule ZLN005 helped facilitate hPSC-CMs maturation by: 1) promoting maturation of mitochondrial biogenesis and energy metabolism; 2) enhancing structural maturation, including anisotropy, and sarcomere length; 3) increasing expression of CX43 and improving electrical activity; and 4) improving Ca 2+ handling and electrophysiological properties. The application of ZLN005 both increased the electronic excitation and metabolic output, which led to better coupling of the energy supply with the electrical and contractile demand.
Increasing evidence has emphasized the importance of mitochondrial oxidative metabolism in the functional maturation of hPSC-CMs [6,17,18,23]. PGC1α is a transcriptional co-regulator that regulates mitochondrial oxidative metabolism by activating genes involved in the OXPHOS and mitochondrial biosynthesis [14]. In addition, the perinatal period is a critical window for maturation of CMs [14,43]. Our data show that the expression of PGC-1α was greatly upregulated at day 7 (spontaneous contraction began) of cardiac induction ( Figure 1A), consistent with previous reports [44]. Multiple studies have shown that relatively mature hPSC-CMs exhibited a higher level of PGC-1α mRNA, which accompanied a higher mtDNA content and mitochondrial mass [18,43]. Overexpression of PGC-1α in cardiomyocytes would induce hundreds of genes that encode for key enzymes in mitochondrial biogenesis and energy metabolism [24].   Since early-stage hPSC-CMs has been shown to display significant plasticity [24], we postulated that PGC-1α would regulate mitochondrial biogenesis program during the early maturation stage of hPSC-CMs. To efficiently and economically manipulate PGC-1α activity, ZLN005 was applied during the early time window. Consistent with previous studies, treatment of ZLN005 induced the upregulation of the PGC-1α gene and protein levels in hESC-CMs. As reported previously, ZLN005 increases the expression of PGC-1α and its downstream genes by promoting its transcriptional activity, which is dependent on the myocyte enhancer factor 2 (MEF2) binding site at the promoter region of the PGC-1α gene [37]. Moreover, we found that ZLN005 treatment also increased expression of the coactivated gene, estrogen-related receptor γ (ERRγ) ( Figure 1D), which regulates mitochondrial metabolism, such as the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS) [45]. This observation is complementary with our functional data, which showed that the resting OCR and ATP production are elevated in ZLN005-treated cells. It has been identified that the PCG-1α expression stimulated by ZLN005 in L6 myotubes is dependent on the AMPK pathway [41]. Moreover, activated AMPK enhances glucose/fatty uptake and mitochondrial metabolism by phosphorylating GLUT4/CD36 and PGC-1α/SIRT, respectively, which ultimately increases ATP production in skeletal muscle cells [46][47][48]. Certainly, this possibility should be further investigated in hPSC-CMs.
Although glycolysis is a major source of energy for cardiomyocytes during early cardiac development [21,49], fatty acid oxidation becomes the primary source for fully matured cardiomyocytes [22,50]. Indeed, dramatic shifts in energy metabolism during the cardiac fetal-to-adult transition have been reported [15]. The switch from glycolysis to fatty acid oxidation in cardiac metabolism is mainly mediated by the PGC-1α axis [51]. Similarly, the majority of energy production occurs by glycolysis in hPSC-CMs, while adult CMs use mostly oxidative metabolism [16,52]. Upregulation of PGC-1α in cultured cardiomyocytes induced expression of mitochondrial genes involved in the multiple mitochondrial energy production and energy transduction pathways [50]. In this study, after treatment with ZLN005, several downstream transcription factors (TFAM, NRF1) of PGC-1α that promote mitochondrial biogenesis and metabolism were increased. PGC-1α transactivates NRF1, which subsequently activates TFAM, and consequently regulates mtDNA transcription and replication [53,54]. As expected, the mtDNA copy number in ZLN005-treated CMs was higher than in controls. Moreover, a previous study reported that ZLN005-treatment increased the protein expression levels of OXPHOS complexes I, II, III and IV in bovine hepatocytes [55]. Consistent with these prior findings, our data show that the expression of several genes that encode for respiratory chain subunits and other genes that are responsible for mitochondrial biogenesis were significantly upregulated in ZLN005treated CMs.

AGING
Additionally, a recent report indicates that shifting cultured hPSC-CMs' metabolism could improve its metabolic and functional maturation [17]. Here, an increase in resting OCR and ATP productions were observed in ZLN005-treated CMs, which was most likely a result of increased mitochondrial content or increased mitochondrial activity [56]. Mitochondrial bioenergetics and ATP production provide key insights into the maturity of cardiomyocytes, since proper mitochondrial function is more crucial in adult and mature cardiomyocytes compared to neonatal and immature cardiomyocytes [16,43]. Furthermore, our transmission electron microscopy results showed that ZLN005-treated hESC-CMs contained more mitochondria and exhibited better mitochondrial morphology and mitochondrial content compared with the controls. Hence, changes in mitochondrial content are critical components of cardiomyocyte maturation during development [8,15,21], which might contribute to an overall increase in mitochondrial oxidative phosphorylation. These data suggest that ZLN005 treatment increased PGC-1α and downstream gene expression and improved mitochondrial biogenesis and metabolic maturation in hESC-CMs. Compared to adult cardiomyocytes, immature cardiomyocytes exhibited smaller Ca 2+ transient amplitudes and slower rises and decay kinetics. The electromechanical activity of cardiomyocytes is highly dependent on ATP produced by oxidative phosphorylation [57,58].
Recent studies reported [11,17,18] that shifting cardiomyocyte metabolism to oxidative pathways promoted cellular maturation, which affected sarcomere lengths and Ca 2+ transient kinetics. In particular, PGC-1α plays a key role in the glucocorticoid-induced maturation of fetal cardiomyocyte structure, oxygen consumption, ion channel expression, and Ca 2+ handling [59,60]. In line with this, we observed that ZLN005 treatment significantly increased the expression of representative of structural genes (MYL3, MYL4, MYH7, TNNT2) in hESC-CMs. Consistent with the increased gene expression, the sarcomere lengths and circularity were improved in ZLN-treated CMs. These findings suggest that ZLN005 treatment promoted hPSC-CMs' morphology maturation, most likely due to a PGC-1α-driven increase in mitochondrial capacity [59]. Furthermore, Ca 2+ imaging analysis showed that ZLN005-treated hESC-CMs exhibited significant increases in the upstroke rate and decay rates compared with controls, indicating more effective Ca 2+ handling. In addition, ZLN005-treated hESC-CM exhibited a more negative resting membrane potential compared to the control group. Connexin 43 (CX43), as the major connexin protein in ventricular cardiomyocytes, brings about electrical excitability and mechanical coupling between cardiac cells [61][62][63]. It has been identified that substantial deficiencies in cell-cell coupling of immature hPSC-CMs may result from low CX43 membrane expression levels [16,64]. The present study is the first to demonstrate that the expression of CX43 is significantly upregulated in the ZLN005-treated hESC-CMs. Importantly, hESC-CMs that were treated with ZLN005 exhibited more homogeneous, spontaneous beating rhythm and electrical propagation (supplementary movies). During cardiomyocyte maturation, CX43 becomes concentrated toward the cell-cell junction, which contributes to the final cell coupling and conduction velocity. A recent report demonstrated that increased CX43 expression improved cell coupling and excitability of iPSC-CMs [65]. Indeed, our data shows that ZLN005-treated CMs exhibited more uniform electrical propagation and higher electrical excitability, consistent with the increased expression of CX43 gap junctions. The results indicated that ZLN005 treatment improves cell-cell communication in hESC-CMs. During cardiac development, PGC-1α plays a key role in controlling the complex regulatory network in cardiac energetics. We sought to determine whether ZLN005 (a smallmolecule transcriptional regulator of PGC-1α) promotes the maturation of hESC-CMs. As expected, ZLN005 significantly increased the expression of PGC-1α and improved mitochondrial biogenesis and metabolic maturation. In addition to stimulating mitochondrial development, ZLN005 treatment also promoted multiple aspects of hPSC-CM maturation.
In summary, we have shown that the small molecule compound ZLN005 is a valuable candidate for hPSC-CM maturation. ZLN005 treatment not only increased the expression of PGC-1α in hESC-CM, but also induced a multitude of target genes that could potentially promote maturation. Importantly, compared with the control hESC-CMs, ZLN005-treated hESC-CMs exhibited a more mature phenotype in various aspects than the control group, including mitochondrial oxidative metabolism, cellular structural, intercellular mechanical coupling, and Ca 2+ handling. This study provides a simple and useful approach that will help improve the applications for cardiac disease modeling, drug screening, and regenerative medicine.

Immunostaining
The cells were fixed in 4% paraformaldehyde for 10 minutes and then permeabilized in 0.4% Triton X-100 for 6 minutes. The fixed cells were blocked with 5% normal goat serum for 1 hour at room temperature. Samples were incubated overnight at 4°C with following primary antibodies: mouse polyclonal anti-α-ACTININ (1:400, Sigma), rabbit polyclonal anticonnexin-43 (0.5μg/ml, Boster Biological). The samples were rinsed with PBS and stained for 1 hour with the following secondary antibodies: goat anti-mouse IgG Alexa Fluor488, goat anti-rabbit IgG Alexa Fluor594 (all from Jackson Immuno Research Laboratories, West Grove, PA, 1:1,000). All samples were counterstained with Hoechst 3342 (1μg/ml, Sigma-Aldrich) for 3 min. A Zeiss Observer microscope was used for slide observation and image capture. Sarcomere lengths were measured using Image J software. We analyzed the hESC-CM sarcomere length (measured by the distance between Z-disks) and the cell circularity index (Circularity = 4π·Area / Perimeter 2 ) using previously published methods [43,70,71].

Calcium (Ca 2+ ) imaging
Cardiomyocytes were loaded with Fura-2 AM (Molecular Probes Inc, Eugene, OR) as previously described [43]. Briefly, cells were incubated with 5 μM Fura 2 AM in Tyrodes Solution (140mM NaCl, 5mM KCl, 1mM MgCl2, 1.8mM CaCl2, 10mM glucose and 10mM HEPES, adjust pH 7.4 with NaOH) for 5 min at 37°C and then washed three times with PBS. The spontaneous calcium transient of single cardiomyocyte was recorded with a 40× Olympus objective using a Nikon Eclipse T5100 fitted with an IonOptix video microscopy system (Ionoptix Corporation, Milton, MA, USA). The experiments were done with culturing medium containing 15 mM HEPES at 35°C.

Quantitative PCR and mitochondrial DNA content
Total RNA was extracted with RNAiso Plus Kit (TAKARA) and quantified using NanoDrop (Thermo Scientific). Reverse transcription was performed using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific) following the manufacturer's instructions. Quantitative PCR was performed on a Bio-Rad iCycler iQ (Bio-Rad) where each well contained a reaction mix totaling 20 μl, with 10μl of SYBR Premix Ex Taq (TAKARA), 5μl of cDNA (10 ng), and 5μl of primers (400 nmol/l each). Each assay was repeated three times, and the relative mRNA levels were normalized using GAPDH as the housekeeping gene.
The mitochondrial DNA Content was quantified as described previously [18]. Genomic and mitochondrial DNA was extracted using Genomic DNA Extraction Kit (TaKaRa). Relative mtDNA content was measured by quantitative real-time PCR (Applied Biosystems 7900) using primers for mitochondrially encoded NADH dehydrogenase (mt-ND1 or mt-ND2); values were normalized to the nuclear gene ACTB. Table 1 shows the primers used.

ATP assay
Cellular ATP content was determined using an Enhanced ATP Assay Kit (Beyotime) according to the manufacturer's instructions. Briefly, cells were lysed with ATP lysis buffer and centrifuged at 12,000 g for 5 min at 4°C. The supernatant was collected and mixed with the assay buffer at room temperature. Luminance (RLU) was assayed in the CLARIOstar microplate reader (BMG LABTECH, Germany). The luminescence data were normalized by the sample protein concentrations.

Transmission electron microscopy (TEM)
For TEM studies, cells were fixed with 4% glutaraldehyde solution (pH 7.2) for 2 hours and further fixed with 1% OsO4 in Sorenson's buffer for 1 hour. After dehydration, the cells were embedded in Epon 812. Ultrathin sections (60 nm) were poststained with uranyl acetate and lead citrate. Sections were examined on a transmission electron microscope (TECNAL G2 Spirit Biotwin) operating at a voltage of 100 kV. Images were acquired at random locations throughout the samples.

Measurement of cellular respiration
Oxygen consumption rate experiments were performed by the CLARIOstar microplate reader (BMG LABTECH, Germany) as previously described [72,73]. In brief, CMs were plated in 96-well assay plates with a density of 30,000 per well in Culture medium. Replace seeding medium with 100 μl of culture media containing MitoXpress Intra probe at 10μg/ml after 24 hours culture. Then 100μL of pre-warmed mineral oil was overlaid in each well to create a seal. The plates were immediately read on the microplate reader according to manufacture protocol for 2 hours. A ratiometric analysis was performed to determine changes in the Lifetime fluorescence signal. Lifetime slopes were calculated using each Lifetime measurement over time. These conversions were performed using the MARS data analysis software.

Multi electrode array (MEA)
Standard 60-electrode MEA probes (Multi Channel Systems, Germany) were plasma-cleaned and coated with Matrigel for 1 h at 37°C. Monolayer hESC-CMs were dissociated and re-plated on MEA probes. Extracellular recording was performed using a MEA2100-System (Multi Channel Systems, Germany). Output signals were digitized at 10 kHz. Recordings were performed between day 14 and day 21 after the cardiac differentiation. During recordings, temperature was kept at 37°C. Data were recorded using Cardio2D (Multi Channel Systems, Germany) and analyzed off-line with Cardio2D+ (Multi Channel Systems, Germany).

Statistical analysis
All of the data were shown as mean ± SEM and statistical analyses were performed using Microsoft GraphPad Prism software (LA Jolla, CA). Student's t test was applied for the comparisons between ZLN005 treatment and control group. One-way analysis of variance (ANOVA) followed with Bonferroni's multiple comparisons was applied for the PGC-1α and PGC-1β expression during cardiomyocyte differentiation day 0 to day 30. Two-way ANOVA was applied with Tukey's multiple comparison for the quantitative analysis of PGC1α protein expression levels at differentiation day 15 and day 30 after ZLN005 treatment. P < 0.05 was regarded as statistically significant.

ACKNOWLEDGMENTS
We thank WiCell Research Institute for providing the H7 hESCs.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.
AGING FUNDING