Generation of high-performance human cardiomyocytes and engineered heart tissues from extended pluripotent stem cells

The availability of functional human cardiomyocytes is essential for cardiac disease modeling, drug screening, and cell therapy, whereas donor human cardiomyocytes are incredibly scarce. Human embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) are widely used to provide an unlimited supply of cardiomyocytes through differentiation schemes, such as modulating Wnt/ β -catenin signaling 1,2 . However, the use of ESCs/ iPSCs face some challenges, such as heterogeneity, the poor survival rate, and differentiation bias, limiting their application 3,4 . To solve these problems, massive efforts have been made, including optimizing culture conditions and deriving more powerful new pluripotent cell types 3 . Mouse and human extended pluripotent stem cells (EPSC) established by Dr. Deng ’ s group 5 in 2017 have the bidirectional chimeric ability that contributes to both embryonic and extraembryonic lineages. Furthermore, human EPSCs have been shown superior chimeric ability in monkey embryos very recently 6 , which further demonstrates their outstanding developmental potential. In addition, EPSC-derived hepatocytes showed improved function and a more similar transcriptome to human primary hepatocytes than ESC/iPSC-derived hepatocytes 7 . However, whether EPSCs can ef ﬁ ciently generate other lineages such as cardiomyocytes and how the EPSC-derived cardiomyocytes (EPSC-CMs) function compared with ESC/iPSC-CMs


Supplementary Video S4. iPSC-CM-and EPSC-CM-derived EHTs on day 14. CMs 144
were constructed to EHT bundles with hydrogel and cultured for 2 weeks before video 145 recording. 146 cells were converted into EPSC in feeder-free conditions as previously described 1 . In brief, 151 Cells were dissociated to single cells with Versene (Thermo Fisher, 15040066), and

Construction and culture of three-dimensional CM-derived EHTs 195
To generate 3D human cardiac tissue patches, 9×9 mm 2 polydimethylsiloxane (PDMS,196 Dow Corning) was microfabricated as previously described 2 . Hydrogel solution (24 μL 197 fibrinogen (10 mg/mL), 12 μL Matrigel, 24 μL 2× cardiac media) was mixed with 1. CMs were incubated with 2.5μM Fluo-4, AM (ThermoFisher Scientific, F14201) for 30 minutes. After incubating, CMs were washed in indicator-free medium to remove the excessive dye for 3 times, and then treated with the original medium containing 2.5 μM (-)-Blebbistatin (Sigma, B0560) to inhibit contractions for 5 min. CMs were electrically stimulated at 1 Hz to produce steady-state conditions. ZEISS LSM 980 inverted laser scanning confocal microscope with a 63X-oil objective was used for confocal fluorescence imaging by line scan. To evaluate isoproterenol response, ISO was loaded for a final concentration of 1×10 -6 M. A customized force measurement apparatus was used to measure contractile force during field stimulation. Briefly, to assess the force-frequency and force-length relationship, bundles were immersed in 37 °C Tyrode's solution with 1.8 mM CaCl2 and stimulated by the electrical pulse of 10 V and different frequencies (0, 0.5Hz, 1Hz, 1.5Hz, 2Hz, 2.5Hz, 3Hz, 3.5Hz, 4Hz). Then the bundle was stretched by linear actuator from 0% to 20% above the resting culture length in 2% increments. Inotropic responsiveness of the bundles was tested in the presence of β-adrenergic agonist isoproterenol ranging from 1×10 -11 to 1×10 -4 M in 1.8 mM Ca 2+ Tyrode's solution during 2Hz electrical stimulation at 4% stretch.
Contractile force data were analyzed using a custom MATLAB program. 255

Force assessment of EHT-bundle by video analysis 256 257 258 259
Spontaneous beating of EHT-bundles was recorded in the microscope (4×) for 10 seconds. Then 1×10 -6 M ISO was loading for further data recording. The results were then analyzed by MYOCYTER plugin, to assess the amplitude, peak time threshold (90%, 50%, 20%), and contraction frequency of the bundle. 260

Immunofluorescence staining 261
For immunofluorescence staining experiments, cells seeded on confocal dishes and 262 EHTs were cultured for 7 and 15 days, respectively. The medium was removed and washed 263 once or twice with DPBS. Cells or EHTs were fixed with 4% paraformaldehyde (Biosharp, 264 BL539A) and incubated at room temperature for 15 minutes. After washing with DPBS for 265 three times for 5 minutes, the cells or EHTs were then blocked with blocking buffer (DPBS 266 containing 5% donkey serum and 0.2% Triton X-100) at 4°C overnight. Cells were 267 incubated with of blocking buffer containing primary antibody at 4°C overnight. After 268 washing twice with DPBS, cells were incubated at room temperature for 2 hours (in the 269 shade) with a blocking buffer containing the secondary antibody. After washing 3 times with 270 using our established methods 3 . Cardiac patches were incubated with Rhod-2 AM (Thermo 277 Fisher, R1245MP, 2.5 μM) at 37°C for 25 minutes. After washing twice with DPBS, patches 278 were treated with the original medium containing 2.5 μM Blebbistatin (Sigma, B0560) to 279 inhibit cardiomyocyte contraction and eliminate motion artifacts during recording. 10-280 second episodes of electrical activity induced by stimulation with point electrode were 281 recorded in microscopic (16×) mode using a 594-channel photodiode array. Conduction 282 velocity (CV) data analysis was performed using BV Ana software and APD data analysis 283 was performed using OMapScope software. 284

Transmission electron microscopy 285
The EHT-bundle cultured for 20 days was directly fixed with a sufficient amount of pre-286 cooling 2.5% glutaraldehyde for 1 hour (cells were completely immersed in glutaraldehyde). 287 Following complete natural sedimentation for 1 hour, samples were incubated in 1mL new 288 pre-cooled 2.5% glutaraldehyde after supernatant was absorbed and stored in the 289 refrigerator at 4°C. This process should be gentle and quick, to make sure not to damage 290

cells. Then samples were sent to the Electron Microscopy and Histology Facility of 291
Shanghai DanDa company for subsequent preparation and imaging. 292

RNA-seq and gene expression analysis 293
RNA-seq was conducted by using Npvaseq PE150 (Jike, Beijing, China) with Paired-294 End reads based on Illumina Novaseq. In brief, total RNA was extracted from 1) 295 EPSC/iPSC-derived cardiomyocytes on day15; 2) EPSC, EPMC, and iPSC before initiation 296 of cardiomyocyte differentiation, using TRIzol (ThermoFisher Scientific, 15596026) 297 according to manufacturer's instruction and RNA sequencing was conducted. Each group 298 had 3 biological replicates. Raw RNA-seq data will be publicly available on 299 https://ngdc.cncb.ac.cn/ (Project number: HRA001708). Cutoffs of P-value < 0.05 and an 300 absolute value of log2(fold-change) ≥ 1 were used for differentially upregulated or 301 downregulated genes for gene ontology analysis by subjecting the gene list to DAVID 302 bioinformatics tool. For GSEA, the normalized FPKM lists of our and public gene sets 303 (Molecular Signatures Database, MSigDB) were selected for downstream analysis. 304

Cell transplantation in the infarcted rat heart 305
Following protocols approved by Hualianke Biotechnology Animal Care and Use Care 306 Committee (HLK-20201210-01), animal experiments were performed. All procedures 307 involving animals were performed according to the NIH guidelines. The nude rats (NIH-308 Foxn1rnu) aged 6-8 weeks were purchased from Beijing Vital River Laboratory Animal 309 Technology Co. (China). Three batches of animal models were assigned. The first batch of 310 rats was anesthetized and thoracotomy injected EPSC-CM directly into the left ventricle, and engraftment was assessed 4 weeks after cell transplantation (n = 6). Rats were