Human induced pluripotent stem cell-derived closed-loop cardiac tissue for drug assessment

Summary Human iPSC-derived cardiomyocytes (hiPSC-CMs) exhibit functional immaturity, potentially impacting their suitability for assessing drug proarrhythmic potential. We previously devised a traveling wave (TW) system to promote maturation in 3D cardiac tissue. To align with current drug assessment paradigms (CiPA and JiCSA), necessitating a 2D monolayer cardiac tissue, we integrated the TW system with a multi-electrode array. This gave rise to a hiPSC-derived closed-loop cardiac tissue (iCT), enabling spontaneous TW initiation and swift pacing of cardiomyocytes from various cell lines. The TW-paced cardiomyocytes demonstrated heightened sarcomeric and functional maturation, exhibiting enhanced response to isoproterenol. Moreover, these cells showcased diminished sensitivity to verapamil and maintained low arrhythmia rates with ranolazine—two drugs associated with a low risk of torsades de pointes (TdP). Notably, the TW group displayed increased arrhythmia rates with high and intermediate risk TdP drugs (quinidine and pimozide), underscoring the potential utility of this system in drug assessment applications.


INTRODUCTION
A precise evaluation of the safety and efficacy of newly developed medications is of vital importance for drug discovery.As a promising candidate for drug assessment, human iPSC-derived cardiomyocytes (hiPSC-CMs) have been extensively discussed for the development of a predictable in vitro drug cardiotoxicity screening assay [1][2][3] ; the utility of hiPSC-CMs in detecting drug-induced proarrhythmic effects and their potential to evolve a new paradigm for in vitro proarrhythmic assays were previously demonstrated. 3However, iPSC-CM-created tissues exhibit significant differences in cellular features, such as morphology, contractility, and electrophysiology, 4 when compared to native cardiac tissue, which could lead to variations in drug response and a higher susceptibility to arrhythmia. 1,3urrently, multiple approaches exist to induce the functional and morphological maturation of hiPSC-CMs.8][29] These maturation methodologies can significantly improve the maturation of hiPSC-CMs in terms of sarcomere structure, calcium-handling properties, and electrophysiology, which dramatically affects the response of the cells to drugs.Specifically, long-term cultured hiPSC-CMs 10 and three-dimensionally engineered tissues 22 demonstrated less sensitivity to the calcium and hERG blocker verapamil.Electrical stimulation of mature cardiac tissue showed a physiologically relevant response to the beta agonist isoproterenol. 28hiPSC-CM tissues cultured on soft scaffolds showed a modest response to the IKr blocker E4031, which resembled the response of the adult myocardium. 26e previously developed a spontaneously originating, traveling wave-based platform capable of rapidly pacing and promoting the maturation of three-dimensional hiPSC-CM tissue rings. 30,31The pacing by traveling wave avoided the side effects of electrical stimulation such as heavy metal poisoning, electrolysis, pH shift, and the generation of reactive oxygen species (ROS) 16,32 and there is no need for an external

Traveling waves regulate cardiac-related gene expression
To investigate how TWs regulate gene expression in iCTs, we used RNA sequencing to compare the gene expression profiles among different groups.The hierarchical clustering of the Spearman's correlation heatmap (Figure 2A) and the principal components analysis (PCA; Figure 2B) indicated that there were closer correlations between iCTs without TW training (Control) and compared with iCTs with TW training.Next, we performed gene ontology (GO) analysis of the groups with and without TW.There were 943 upregulated genes out of a total of 26255 genes (adjusted p < 0.05; fold change >1.5) in the TW group compared with the control group.Several GO-enriched terms were associated with maturation, such as muscle structure development, actin filament-based processes, muscle system processes, and cardiac muscle tissue development (Figures 2C; Figures S2 and S3).The TW group (Figure 2D; Figure S2) showed regulated expression of genes associated with conduction (increased GJA1, KCND3, KCNJ2, KCNJ4, KCNJ11, SCN1B, and SCN5A expression), ultrastructure (increased MYH7, JPH2, CDH2, TNNI3 expression and decreased MYH6, TNNI1 expression), energetics (increased NPPA, PPARGC1A, CYCS, MB, and ADRB1 expression), and calcium handling (increased AMPH and ATP1A2 expression).The RNA-sequencing data were also compared with those of fetal and adult heart cells in a previous report. 36Compared with the control group, the TW group showed a change toward adult heart cells (Figure 2D).
The ion channel-related genes are of great importance for studying the drug response of hiPSC-CMs, we found the enhanced expression of several potassium ion channels (KCND3, KCNJ2, and KCNJ4) and sodium ion channels (SCN1B and SCN5A) in the TW group (Figure 2D; Figure S2).In addition, as mentioned in a previous report 37 by the FDA on evaluating the usability of iPSC-CMs for drug assessment, the hiPSC-CM tended to have a higher expression of the calcium ion channel (CACNA1C) compared with those in the adults.We found decreased the expression of CACNA1C in the TW group (Figure 2D), which is closer to the adult group.Similarly, in the commercial Cor.4U, KCNH2 also showed higher expression compared to that in the adult group. 37In the present work, the TW group showed reduced expression of KCNH2, which is closer to the levels in the adult group (Figure 2D) than those in the control group.
Next, we performed immunostaining for cardiac-specific markers; the TW group showed a significantly increased expression of the gap junction marker Cx43 (encoded by GJA1) and b-myosin heavy chain (b-MHC, encoded by MYH7), a cardiac maturity marker associated with muscle contraction (Figures 3A and 3B).Moreover, electron microscopy indicated that the CMs in the TW group had larger sarcomere bundles and well-defined Z disks, I-bands, T-tubules, and myofibrils than those in the control group (Figure 3C).There is no significant difference in the sarcomere length between both groups (Figure 3D).We then confirmed the high expression of proteins such as b-MHC (encoded by MYH7), Cx43 (encoded by GJA1), and N-cadherin (encoded by CDH2) in the TW group by western blotting (Figure 3B).These data indicated that the CMs in the TW group were more mature than those in the control group in terms of sarcomere structure, gap junction, and other maturation-related marker expression.

Traveling waves improved the functional maturation of CMs
To evaluate the functional maturation of CMs, we performed electrophysiology recording and motion analysis on the two groups.Both the TW and control groups showed the homogeneous propagation of contractions; however, the conduction velocity of the TW group was significantly higher than that of the control group, while the Max delay of the TW group was smaller than that of the control group (Figures 4A-4D), which agrees well with the enhanced expression of the gap junction gene GJA1 (Figure 2D) and the protein Cx43 (Figure 3) in the TW group.Both the control and TW groups demonstrated QT intervals at approximately 300 ms, which is much shorter than those ($600 ms) of the LQT2 patient. 38Furthermore, the motion analysis revealed that, after training with TW, the CMs in the TW group showed significantly higher contractility than those in the control group (Figures 4C-4I), including higher contraction velocity (58.68 G 17.49 mm/s vs. 36.74G 12.91 mm/s; p = 0.004), relaxation velocity (39 G 7.87 mm/s vs. 30.19G 10.96 mm/s; p = 0.044), as well as acceleration (2359.54G 844.68 mm/s vs. 1200.29G 361.64 mm/s 2 ; p = 0.0008).
An extracellular flux analyzer was then used to evaluate the mitochondrial function of the CMs in both groups.Mitochondrial ATP synthesis was inhibited by the addition of oligomycin.A proton-gradient discharger (FCCP) was added to evaluate the maximum mitochondrial respiration (Figures 5A-5F).There was no significant difference in the basal respiration (27.96

Traveling waves trained human induced pluripotent stem cell-CMs for drug assessment
We further evaluated the drug response of iCT with two types of cardiomyocytes (253G1-derived cardiomycytes and commercially available iCell 2 ).Rather than directly culturing iCTs on expensive MEA chips, hiPSC-CMs were cultured on the device with a layer of permeable fiber scaffold attached to the bottom.After culturing and training in a normal 24-well plate (Figure 1A), the iCT was directly transferred onto an MEA for drug tests (Figures S4A and S4B), as well as for culturing after the drug tests.Moreover, the MEA chips could be immediately reused for the next iCT recording after washup.The TW-paced iCT beat significantly faster than the control group (Figures S4C and S4D) and showed enhanced sarcomere structure (Figure S4E), which is similar to the results obtained after using the device without a fiber scaffold.The iCTs were then used for drug response tests (Figures 6 and 7; Table 1; Figures S5-S7), including adrenoreceptor agonist (isoproterenol), sodium/potassium current blocker (ranolazine), calcium blocker (verapamil), sodium channel blocker (mexiletine), potassium blocker (E4031), and a negative control (aspirin).Drugs with high (quinidine) and intermediate (pimozide) torsades de pointes (TdP) risks were also assessed to evaluate the system.
Upon exposure of iCT to b-adrenergic stimulation with isoproterenol, the trained tissue from the 253G1 line showed a chronotropic effect at all the tested concentrations (from 0 mM to 10 mM, 141.24% G 16.43%, p < 0.001).The control group demonstrated a chronotropic effect only between 0 mM and 0.1 mM (134.76%G 23.1%, p < 0.01) and the beat rate decreased between 0.1 mM and 10 mM (Figure 6A).The EC 50 values of the TW and control groups were 95.2 nM and 6.7 nM, respectively, similar to those reported in a previous study on the electrical stimulation of the matured cardiac tissue. 28Identically, the iCT with TW made using iCell demonstrated a lower EC 50 value but an enhanced reaction to isoproterenol at all tested concentrations (Figure S5).This result agrees well with that of a previous report 39 in which the chemical factor-matured cardiac tissue showed a chronotropic effect between 0 mM and 10 mM, whereas the control group had a maximum beat rate between 0.01 and 0.1 mM.In addition, the improved expression of the b-adrenergic receptor (ADRB1) may be correlated with the improved reaction of iCT to b-adrenergic stimulation (Figure 2D).We next evaluated the response of both groups from the 253G1 line-derived cardiomycytes to the calcium blocker verapamil (Figure 6B), which is also a hERG blocker with no reports of QT prolongation or TdP in humans. 40Both groups showed a shortened QT interval with the addition of verapamil; 4 out of 6 samples in the TW group stopped beating at 10 mM, while 1 and 2 out of 6 samples in the control group stopped beating at 1 and 10 mM, respectively, similar to a previous report. 10In addition, the TW group showed lower data variation than the control group.The IC 50 value of the control group (0.05 mM) was close to the free effective therapeutic plasma concentration (ETPCunbound), 10,40 while the IC 50 of the TW trained group (1.1 mM) resulted in a higher safety margin (IC 50 /ETPCunbound) of 20, which is close to the safety margins (20-30 22 ) typically required by pharmaceutical companies for the development of new drugs.This improvement was also achieved previously using long-term cultured iPSC-derived cardiac tissue 10 and 3D engineered iPSC-derived cardiac tissue. 22In the iCell-made iCT (Figure S5), 1 in 4 TW samples stopped beating, and 0 of the 6 control samples stopped beating at a concentration of 10 mM.The IC 50 of the two groups was 2.59 and 2.73 mM, respectively.The inter-cell line variation between 253G1-derived CM and iCell may contribute to the different responses to Verapamil.
The sodium blocker ranolazine is a known hERG blocker that prolongs the QT interval, but has a low risk of TdP.It was previously reported that ranolazine causes early afterdepolarization (EAD) in hiPSC-CMs at clinically relevant concentrations (C max , 2-6 mM). 1 Another multisite study showed that arrhythmia or cessation could occur in approximately 40% of samples at 100 mM (>50-fold C max ). 3 In the present study, we observed EAD in 2 out of 6 control samples and 1 out of 6 TW samples at 100 mM.In addition, the control group showed a downward trend in the cQT interval at concentrations beyond 30 mM, while 2 out of 6 samples showed cessation of beating at 300 mM.Meanwhile, the TW group showed a continuous increase in the cQT interval, and none of the 6 samples showed cessation of beating at 300 mM (Figure 6C).Similarly, in the iCT made by iCell, 3 in 8 control samples showed EAD at 100 mM, while 1 in 9 TW samples showed EAD (Figure S5C).When the drug concentration was increased to 300 mM, 2 in 8 control samples stopped beating compared to the 1 in 9 TW samples.The cessation or increase in arrhythmia in the control group could have resulted from the low maturity of the hiPSC-derived cardiomyocytes, as discussed in previous reports. 1,41,42he potassium blocker E4031 was also evaluated, and both the control and TW groups showed QT prolongation after E4031 treatment; no arrhythmic activities were observed in either group at all tested concentrations (Figures S6 and S7).Mexiletine, a sodium and hERG channel blocker, was also added to the two groups.QT prolongation was observed in both groups.In 253G1-derived cardiomycyte samples,   arrhythmic activities occurred in one each of the control (nine samples) and TW samples (nine samples) at 100 mM (>40-fold C max 3 ).In iCell samples, 2 of the 8 control samples showed arrhythmic activities, while none of the 9 TW samples showed abnormal activities.The negative control, aspirin, showed no significant difference between the two groups.
Since most of the above drugs have low or no TdP risk, high (quinidine) and intermediate (pimozide) TdP risk drugs were also applied to the iCT tissues (Figure 7; Figure S8).Quinidine is a known high TdP risk drug that could induce arrhythmia-like events at concentrations close to clinical C max (3 mM). 3 Both groups showed the prolongation of QT intervals after quinidine treatment.However, 1 in 5 control samples showed arrythmia-like events at 10 mM, while 3 in 4 TW samples showed arrythmia-like events at a lower concentration (1 sample each at 0.01, 0.1, and 1 mM, respectively).In addition, we applied the intermediate risk drug pimozide to the iCTs.While 4 control group tissues showed no arrythmia-like events at 30 nM, a concentration 70-fold higher than C max (0.4 nM), 2 out of 4 TW group tissues showed arrythmia-like events at 30 nM.These results indicated that the TW group may have higher sensitivity to the two types of high-and intermediate-risk TdP drugs.

Integrin related pathway plays a role in regulating iCT maturation
Integrins are a family of cell adhesion receptors that bind to the extracellular matrix, cell-surface, and soluble ligands. 43,44They are especially important for cell-extracellular matrix (ECM) adhesion, structural organization, and transducing mechanical signals from the ECM into cardiomyocytes. 26,45,467][48] The soft Matrigel substrate has been reported to induce the maturation of hiPSC-CM monolayers, with integrin signaling playing a vital role. 26We hypothesized that the integrin-related pathway mediated the traveling wave-induced maturation of hiPSC-CMs (Figure 8A).RNA-sequencing data indicated that most of the alpha (ITGA) and part of the beta (ITGB) subunits of integrin in the TW group showed marked upregulation compared with the control and day 0 values (Figure 8B).Similar to that in a previous report, 26 a number of downstream genes were also remarkably upregulated; these included ECM-related genes, actin (ACTB) and parvin (PARVA and PARVB), and genes related to integrin subtype-specific activation of pro-survival and pro-maturation signaling pathways, such as Ras (HRAS, NRAS), PI3K (PIK3CD), Akt (AKT1 and AKT3), and ERK (MAP2K1).These findings indicate that TWs promoted hiPSC-CM maturation through the integrin-related signaling pathway.

DISCUSSION
Rapid pacing of hiPSC-CMs is proven to be effective for their maturation. 28,29We previously developed a spontaneously originating TW for rapid pacing and maturation of hiPSC-CMs in a three-dimensional tissue ring. 30,31To date, cardiac safety paradigms such as CiPA (FDA) and JiCSA (Japan iPS Cardiac Safety Assessment) utilize 2D monolayer cardiomyocytes cultured on MEA systems for drug evaluation. 3,33In the present study, we created hiPSC-derived closed-loop cardiac tissue (iCT) on a 2D substrate.TWs can spontaneously originate within the 2D sheet and rapidly pace hiPSC-CMs.Similar to that observed in the 3D tissue ring, TWs promoted the maturation of hiPSC-CMs within the monolayer tissue, which showed improvement in multiple features such as conduction, ultrastructure, energy, and contraction.
Interestingly, the TW-trained iCT demonstrated improved oxygen consumption rate, as well as survival rate after hypoxic culture, compared with the control group.This is probably because the high expression of respiration-related proteins, myoglobin, and cytochrome c in the TW group enhanced the oxygen flux, 49 mitochondrial function, and ATP synthesis under hypoxic conditions and promoted the survival of cardiomyocytes.In future studies, it would be interesting to further analyze the survival capability of TW-trained CMs after transplantation in animal models of myocardial infarction.It has long been hypothesized that mature hiPSC-CM tissue could resemble adult myocardium and would be more suitable for the evaluation of drug response.There have been reports on the assessment of the response of cardiac tissue with improved maturation to different drugs: (1) Rapid pacing mature hiPSC-CMs showed physiological responses to isoproterenol 28 ; (2) hiPSC-CMs matured by long-term culture 34 or engineered into aligned 3D m-tissues 22 demonstrated less sensitivity to verapamil, an L-type calcium channel blocker that has a potent effect on hERG; and (3) cardiac tissue matured by a soft substrate showed modest response to the Ikr blocker E4031, which resembled the response of adult myocardium. 26In our study, TW-matured hiPSC-CMs showed similar responses to several of the previously mentioned drug types, such as isoproterenol and verapamil, but not to E4031.In addition, the TW group showed an improved response to the sodium blocker ranolazine, which is a known hERG blocker that prolongs the QT interval but has a low risk of TdP.The cessation or increased arrhythmia in the control group could be due to the immaturity of hiPSC-CMs, as discussed in previous reports. 1,41,42In addition, as indicated by the RNAseq data, the gene expression level of multiple ion channels and beta receptors is regulated toward adult levels in the TW group, compared with that of the control group.This may be related to the improved response to drugs such as beta agonists, calcium blockers, and hERG blockers.In the future, more efforts are needed to investigate the underlining mechanisms and correlation between the ion channel maturation and the drug response.
Recently, several multisite multiline studies have been performed to prove the capability of hiPSC-CMs as an in vitro proarrhythmia model. 3,33The results validated the utility of hiPSC-CMs in predicting drug-induced proarrhythmic effects as part of an evolving paradigm.To further validate the utility of TW-induced iCT as an enhanced drug assessment candidate, our future work will also include site-to-site and line-to-line variation evaluation.Since the TW is a spontaneous activity that could not be precisely controlled and there are still significant sample-to-sample variations in data such as conduction velocity and transcriptome data.More efforts are needed to investigate the most efficient way to control the origination, maintaince, and observation of TW: For example, the TW tissue could be prepared with the optogenetic cell line that could be paced by light stimulation 50 for controllable origination of TW.Moreover, the cell line with Ca 2+ indicator 51 may be used for improved observation of TW activity such as propagation velocity and the beating frequency.In addition, TW samples maybe categorized by the frequency range, which allows further reducing the sample-to-sample variations.
Instead of pacing and maturing hiPSC-CMs before drug assessment, there are also reports on using electrical 52 or optical stimulation 53,54 to pace the hiPSC-CMs during drug assessment.The pacing could control the spontaneous beating of the hiPSC-CMs and thus reduce the variation caused by the beat rate.The electrical pacing sample showed reduced response to multiple cardiac ion channel blockers and enabled more accurate rate-dependent drug evaluation. 52However, the optically stimulated group was not found to be substantially improved with respect to proarrhythmic risk predictions compared with the non-paced group. 53In the present study, the TW was stopped before the hiPSC-CMs were used for drug assessment because the high frequency of the TW (3-4 Hz) may interfere with the CM response.Following this, the beat of iCT would dramatically decrease and stabilize at approximately 1 Hz, relative to the human heart rate, allowing for a more accurate drug response.The results on stimulation during drug test, together with those reports on stimulation before a drug test, indicate that the stimulation time and stimulation delivery method may play an important role in the use of hiPSC-CM in drug assessment.
In our previous experience, it was difficult to culture 2D cardiac tissue on MEA chips for longer terms (>14 days), because the CMs would peel off from the substrate, especially when the MEA chips were reused several times.Culturing the cells on a fiber layer would resolve the problem of attachment, while allowing the recording of the signal using electrodes below the fiber sheet.This would allow long-term ($months) evaluation of the chronic response of cardiac tissue to drug candidates. 23In addition, because the cells do not need to attach to the surface of MEA chips, the chips could be reused in the same batch of tests by simply replacing the tissue with a new one.This can significantly reduce the use of expensive MEA chips.To improve throughput and reduce the cost of drug assessment, hiPSC-CM tissues have been made using as few as thousands of cells in 3D cluster microtissues. 7,55Although the cell number required for creating closed-loop cardiac tissue has been reduced to 2310 5 from 4 3 10e 5 for the 3D rings, 31 it is still not suitable for high-throughput assays.Further optimization is required to further scale down the closed-loop device and reduce the required cell number.In previous 3D ring formation, the sarcomere length of TW group has been improved, however, there is no similar improvement in 2D tissue, this may be caused by the different culture system between 2D and 3D system, as significant difference has been found between CM cultured and matured under 2D 56,57 and 3D 58 condition, in the future, we will further improve the TW device to adapted the 3D tissue while allowing the recording by MEA system.For example, the soft Matrigel could be used to coat the surface of scaffold to allow the 3D cells culture and sarcomere length improvement. 59Furthermore, it is important to note that the RNA-seq analysis in the present study is based on the assessment of average gene expression within the samples.Nonetheless, the differentiated CMs derived from hiPSCs frequently comprise a heterogeneous mixture of cell types, including atrial, ventricular cells, and various other subpopulations.In the future, the adoption of Single-cell RNA sequencing (scRNA-seq) holds the promise of providing more precise genetic information and enhanced resolution of maturation states for each of these subpopulations.''2][63] In the future, it may be more appropriate to perform the comparison within the same batch to minimize the variation.
In the present design, the PDMS is used to fabricate the device, which may have the issue of compounds sticking, 64,65 in future, the other material such as polystyrene will be used to replace the PDMS.Moreover, because cardiovascular liability of drugs commonly occurs via the altered function of the contractile myocardium, the contractility or contractile force has also been suggested as an evaluation factor for cardiac safety paradigms, in addition to the electrophysiology. 66,67Our future optimization might include integrating the evaluation of contractile force into the present device.

Limitation of the study
The study is constrained by several limitations, including: a.The inability of the TW-paced hiPSC-CM to achieve a maximum frequency of 6 Hz, as reported in previous electrical stimulation studies. 68Furthermore, restarting TW after interruption is time-consuming and demands specialized skills.b.The digestion process applied to cardiac tissue has the potential to disrupt its tissue-specific properties.c.Utilizing multiple recording systems, such as calcium and/or voltage-based imaging, is necessary for obtaining a more comprehensive dataset than that of the MEA system.This enhancement not only facilitates the recording of tissue activity in the initial days (0-4 days) but also enables the mapping of activation in a larger area than the MEA system allows.

Conclusion
We observed the TW phenomenon in a closed-loop monolayer cardiac tissue.TWs could pace the hiPSC-CMs to beat at a high frequency and maintain it for more than two weeks.The TW-paced hiPSC-CMs showed improved sarcomeric and functional maturation.According to the drug response data, the mature hiPSC-CMs demonstrated improved drug responses resembling those of the adult myocardium, holding potential as a candidate for drug discovery with improved robustness and accuracy.done gently to avoid disturbances to the traveling wave.All data in this study were collected after the tissues were cultured for 14 days.The TW was stopped by replacing the medium with 4 C medium; this will cause the CMs to stop beating for a while, and spontaneous beats would resume within one or two days.

Electrophysiological characterization
The field potentials of the CMs were recorded at 37 C using the MEA data system (USB-ME64-System, Multi Channel Systems, Germany).The data were collected and processed using MC_Rack (Multi Channel Systems).To obtain the activation map, the local activation time from a single electrode was obtained by calculating the minimum of the first-derivative plot of the raw curve.Linear interpolation between the electrodes was applied to calculate the isochronal map 23,78 using the MATLAB function (MATLAB, MathWorks, USA).The field potential included the presence of a peak corresponding to the Na + influx and depolarization, followed by the T wave determined as the repolarization phase corresponding to K + efflux. 79The QT interval was obtained and corrected using the Fridericia's correction formula: cQT interval = QT interval/ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi RR interval 3 p .To assess the effects of different drugs, 1 mL of fresh medium was added to the wells and maintained for 5 min before the baseline was recorded for 1 min.Then, another 1 mL of medium was added to twice the test concentration of the drug and gently pipetted into the well.Data were collected after 5 min.This process was repeated for all test concentrations.EC 50 /IC 50 values were calculated using a web calculator (www.aatbio.com/tools/ic50-calculator).

Transmission electron microscopy (TEM)
The iCTs were observed using a transmission electron microscope (H-7650; Hitachi Co., Tokyo, Japan).Specimens for TEM were prepared according to the following procedure: tissues were fixed with 2.5% glutaraldehyde for 120 min.The samples were then post-fixed with 1% osmium tetraoxide for 90 min and dehydrated through a graded series of ethanol (50%-100%) and propylene oxide.The tissues were then embedded in epoxy resin, sliced using an ultramicrotome (Ultracut E; Reichert-Jung, Vienna, Austria), and stained with uranyl acetate and lead citrate.

Mitochondrial respiration assay
Mitochondrial function was analyzed using a Seahorse XF96 extracellular flux analyzer (Agilent Technologies, Carlsbad, CA, USA).After culturing for 14 days, the hiPSC-CM tissue was dissociated into a single-cell suspension using 0.25% trypsin/EDTA (Thermo Fisher) and then seeded in a microplate (XF96, Agilent Technologies) at a density of 2310 4 cells/well.After 3 days of culture, the culture medium was replaced with base medium (Seahorse XF assay media; Agilent Technologies, Carlsbad, CA, USA) supplemented with 1 mM sodium pyruvate.Substrates and inhibitors were added during measurements to attain a final concentration of 3.5 mM 4-(trifluoromethoxy) phenylhydrazone (FCCP; Seahorse Bioscience, Billerica, MA, USA), 1 mM oligomycin, 0.5 mM antimycin, and 0.5 mM rotenone for the MitoStress assay.

Western blot
The iCT was washed with PBS and lysed with lysis buffer [1% CHAPS, 25 mmol/L Tris-HCl (pH 7.4), 137 mmol/L NaCl, 2.68 mmol/L KCl, and 5 mmol/L EDTA].Protein concentration was determined using a BCA Protein Assay Kit (Thermo).Lysate samples were mixed with 4 3 loading sample buffer (Bio-Rad) and mercaptoethanol (2.5%).Proteins were separated by SDS-PAGE and transferred onto polyvinylidene fluoride (PVDF) membranes.After blocking in 3% non-fat milk for 1 h, the transferred membrane was incubated with primary antibodies at 4 C overnight and thereafter with secondary antibodies at room temperature for 0.5 h.The membrane signals were recorded using ECL prime reagent (GE).Protein expression data were quantified using an ImageQuant LAS 4000 (GE) system.

Figure 1 .
Figure 1.Traveling Waves (TWs) rapidly paced the cardiomyocytes within the closed-loop route (A) Schematic and image describing the cell plating and TW origination in the device.(B) MEA field potential signal of the iCTs with or without TW on day 6 and day 14.(C) Beat rates of iCTs at different culture times (Mean G SEM; Control: n = 15; TW: n = 18 biologically independent samples from two differentiations).***p < 0.001 (Student's t test).

Figure 2 .
Figure 2. The transcriptome data indicated that TWs enhanced the expression of cardiac-related markers (A) Heatmap showing the hierarchical clustering of the correlation matrix resulting from comparison of expression values for each group.(B) Principal components analysis (PCA) of iCTs with or without TW based on RNA-sequencing data.(C) The 943 upregulated genes out of 26255 genes (adjusted p < 0.05; fold change >1.5) from the TW group (compared with the control group) were used for GO category analysis.The enriched terms are listed.(D) Heatmaps showing the expression of cardiac maturation-specific genes.To compare the gene expression in the TW and control group with fetal and adult heart data from a previous report, the FPKM data were converted into Transcripts Per Million (TPM).

Figure 3 .
Figure 3. Traveling waves (TW) enhanced sarcomere maturation (A) Representative confocal images of iCTs with or without TW on day 14.Cardiomyocytes were stained with anti-a-actinin, anti-TnT2, anti-Cx43, anti-b-MHC, and DAPI.(B) Whole-cell lysates were extracted from TW and control groups and analyzed by western blotting using the indicated antibodies.(C) TEM analysis of iCTs from both the TW and control groups on day 14.Z: Z lines; I: I band; T: T-tubules.(D) The sarcomere length.Sarcomere lengths of CMs in both groups on day 14.(Mean G SEM; Control: n = 4; TW: n = 3 biologically independent samples from two differentiations).

Figure 4 .
Figure 4. Traveling waves (TWs) improved the electrical conduction and the contractility of hiPSC-derived close-loop cardiac tissue (iCT) (A) Activation maps showing the propagation of contractility on day 14, the TWs have been removed before recording to allow spontaneous beating.Each of the points on the axis represent an electrode (8 3 8, spacing 200 mm) on MEA.(B-D) The conduction velocity (B), Max delay (C) and QTinterval (D) of the contraction of both groups (Mean G SEM; Conduction velocity: Control: n = 5; TW: n = 4 biologically independent samples from three differentiations; Max delay: Control: n = 4; TW: n = 3 biologically independent samples from two differentiations; QTinterval: Control: n = 10; TW: n = 8 independent biologically samples from four differentiations.**p < 0.01, ***p < 0.001 (Student's t test).

Figure 4 .
Figure 4. Continued (E) Representative velocity image of the control and TW groups using a motion analysis system.Red and blue represent high and low velocities, respectively.(F) Plot of a motion waveform showing contraction and relaxation velocity peaks.The green line marks the time point of the data in (e).(G-I) Contractile properties of the Control group and TW group.(g) Contraction velocity, (h) relaxation velocity, and (i) acceleration (Mean G SEM; Control: n = 10; TW: n = 13 biologically independent samples from three differentiations).*p < 0.05, **p < 0.01, ***p < 0.001 (Student's t test).

Figure 8 .
Figure 8. Proposed model for the traveling wave (TW)-induced hiPSC-derived closed-loop cardiac tissue (iCT) maturation (A) Schematic representation of an integrin pathway that could lead to the survival, adhesion/migration and growth/maturation of the cells.(B) Heatmaps showing expression of ECM-related genes.The data were collected from three biologically independent samples from two differentiations (batches).

Table 1 .
The drugs being assessed and the arrhythmia related events occurred post drug administration