Transcriptomics reveal stretched human pluripotent stem cell-derived cardiomyocytes as an advantageous hypertrophy model

Left ventricular hypertrophy, characterized by hypertrophy of individual cardiomyocytes, is an adaptive response to an increased cardiac workload that eventually leads to heart failure. Previous studies using neonatal rat ventricular myocytes (NRVMs) and animal models have revealed several genes and signaling pathways associated with hypertrophy and mechanical load. However, these models are not directly applicable to humans. Here, we studied the effect of cyclic mechanical stretch on gene expression of human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) using RNA sequencing. hiPSC-CMs showed distinct hypertrophic changes in gene expression at the level of individual genes and in biological processes. We also identified several differentially expressed genes that have not been previously associated with cardiomyocyte hypertrophy and thus serve as attractive targets for future studies. When compared to previously published data attained from stretched NRVMs and human embryonic stem cell-derived cardiomyocytes, hiPSC-CMs displayed a smaller number of changes in gene expression, but the differentially expressed genes revealed more pronounced enrichment of hypertrophy-related biological processes and pathways. Overall, these results establish hiPSC-CMs as a valuable in vitro model for studying human cardiomyocyte hypertrophy. Non-standard Abbreviations and Acronyms ET-1, endothelin-1; GO, gene ontology; hESC-CM, human embryonic stem cell-derived cardiomyocyte; hiPSC, human induced pluripotent stem cell; hiPSC-CM, human induced pluripotent stem cell-derived cardiomyocyte; MAPK, mitogen-activated protein kinase; MEK1/2, mitogen-activated protein kinase kinase 1/2; NRVM, neonatal rat ventricular myocyte; PKC, protein kinase C; PBS, phosphate-buffered saline; RB+, RPMI 1640 supplemented with B-27; RB-, RPMI 1640 medium supplemented with B-27 without insulin; RT, room temperature; TF, transcription factor


39
The prevalence of cardiovascular diseases, including coronary artery disease and hypertension, is 40 increasing rapidly, from approximately 271 million in 1990 to 523 million in 2019. 1 However, 41 treatment strategies have not evolved correspondingly; hence, cardiovascular diseases are the leading 42 cause of death. 2 Hypertension and myocardial infarction increase cardiac workload, causing structural 43 and functional changes in the myocardium. 3 These changes include left ventricular hypertrophy, which 44 is characterized by cardiomyocyte enlargement. Although it is initially an adaptive response to 45 physiological and pathological stimuli, such as mechanical stretch or neurohumoral activation, 46 prolonged hypertrophy leads to contractile dysfunction and heart failure. 47 representing the differentially expressed genes. 23 Encyclopedia of RNA Interactomes (ENCORI; 158 available at http://starbase.sysu.edu.cn/) was used to predict the putative interaction partners of 159 differentially expressed lncRNAs. 24 160 Dataset comparison 161 Our dataset of stretched hiPSC-CMs was compared to datasets of the two other studies. We used data 162 from differentially expressed genes of stretched NRVMs by Rysä et al., who isolated NRVMs from 2-163 to 4-day-old Sprague-Dawley rats and stretched cells with the FlexCell vacuum system, similar to the 164 present study (0.5 Hz, 10-25% elongation). 12 We also compared our data to the data of stretched 165 hESC-CMs obtained by Ovchinnikova et al., who had used a slightly different stretching protocol: 166 cyclic stretch with elongation from 0% to 15% was applied at a frequency of 1 Hz with the FlexCell 167 system. 13 168 Immunofluorescence staining, imaging and image analysis 169 Brefeldin A (Invitrogen, Carlsbad, CA, USA) was added to the cells 3 h before fixation to block 170 exocytosis of proBNP-containing vesicles. At the end of stretching, the cells were washed twice with 171 phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde at room temperature (RT) for 172 15 min followed by 3x5 min washes with PBS. A square piece (approximately 10 mm x 10 mm) was 173 cut from the center of each flexible bottomed well of Bioflex® plate for staining. Subsequently, the 174 cells were permeabilized with 0.1% Triton X-100 in PBS at RT for 10 min. Non-specific binding sites 175 were blocked with 4% FBS in PBS for 45 min followed by addition of primary antibodies diluted in 7 4% FBS in PBS. Cardiac troponin T (cTnT) antibody (ab45932, Abcam) was used at 1:800 and 177 proBNP antibody (ab13115, Abcam) at 1:200. For the secondary antibody staining control, primary 178 antibodies were omitted. After a 60-min incubation at RT, the cells were washed 3x5 min with PBS 179 and incubated with Alexa Fluor® -conjugated secondary antibodies (Life Technologies,Eugene,180 Oregon) at 1:200, and 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) at 1 µg/ml at RT for 45 181 min. The cells were then washed 3x5 min with PBS and mounted between a microscope slide and a 182 coverslip using ProLong™ Gold Antifade Mountant (Invitrogen). The samples were cured at RT 183 overnight and then stored at 4°C until imaged. The cells were imaged with ImageXpress Micro 184 Confocal imaging system (Molecular Devices) using Nikon 10x Plan Apo 0.5 NA air objective and 185 40x C CFI APO LWD Lambda S 1.15 NA water immersion objective. MetaXpress software 186 (Molecular Devices) was used to analyze proBNP and cTnT intensity in cardiomyocyte nuclei and 187 perinuclear area. First, background was subtracted from DAPI and proBNP images using the top hat 188 filter. The nuclei were then identified based on DAPI staining and the cardiomyocyte cytoplasm was 189 identified based on cTnT staining. The average proBNP staining intensity was quantified from the 190 perinuclear region of cardiomyocytes defined as a 10-pixel ring around each cTnT positive nucleus. 191 The average cTnT staining intensity was measured from combined nuclear and perinuclear area. 192

206
The mechanical stretch model of hiPSC-CMs (Figure 1) was first validated by measuring the mRNA 207 expression of NPPA and NPPB, hallmark genes of cardiomyocyte hypertrophy. 25 After 24 h of cyclic 208 mechanical stretch, hiPSC-CMs showed increased expression of both NPPA and NPPB (Figures 2A  209   8 and 2B). At 48 h and 72 h, the upregulation of the NPPA and NPPB mRNA levels was not statistically 210 significant, although increased gene expression was observed in each independent experiment. To 211 confirm the increased BNP expression in protein level, we stained proBNP (the N-terminal fragment 212 of the BNP prohormone) in stretched hiPSC-CMs. An evident increase in perinuclear proBNP 213 expression was observed after a 24-h stretch ( Figure 2C and 2E). In agreement with the time-214 dependent gene expression, no apparent change in proBNP expression was observed after a 48-h 215 stretch. However, there was a trend towards increased cTnT intensity with increasing duration of 216 stretch ( Figure 2D and 2E). In addition, high purity of cardiomyocytes was confirmed by staining for 217 cTnT. High-resolution images acquired with 40x objective are available in Supplementary  genes encode enzymes (6) and transcription factors (6) ( Table S2). 245 We selected eleven genes for validation by qRT-PCR. The selection was first based on differential 246 expression of both up-and downregulated genes. Second, both protein-coding and noncoding 247 (LINC00648, PTPRG-AS1) genes were selected. In addition, different protein-coding genes were 248 selected, including hypertrophy-associated secreted peptides (NPPA, NPPB), cytoskeletal proteins 249 (ACTA1, ACTC1, ACTN1, TNNI3), a transcription factor (CSRP3) and a transporter protein 250 (SLC16A9). Overall, similar results were obtained with both qRT-PCR and RNAseq (Figure 4), except 251 for ACTN1, which showed downregulation in qRT-PCR and slight upregulation in RNAseq after 72 h 252 of stretch ( Figure 4A). 253

254
We compared our data with the NRVM data published by Rysä et al. to elucidate similarities and 255 differences between these two in vitro cardiomyocyte models from different species. 12 Both we and 256 Rysä et al. used time points of 24 h and 48 h; hence, these were selected for comparison, although the 257 equivalency of these time points between the species has not been proven. Overall, the number of 258 differentially expressed genes was drastically different; for example, after a 48-h stretch, over 600 259 genes were upregulated in NRVMs, while only 28 genes were upregulated in hiPSC-CMs ( Figure 5). 260 Interestingly, 21 differentially expressed genes showed similar changes in both cell models. In fact, 3 261 genes were upregulated in both cardiomyocyte types at both time points: CASQ1, TIMP1 and 262 TUBB2B. We did not identify genes that were consistently downregulated in both CM types at 24h 263 and 48 h. After 24 h of stretch, 14 genes were upregulated in both cell types, while no commonly 264 downregulated genes were identified ( Figure 5C). Additionally, we identified 8 genes that were 265 upregulated and 2 genes that were downregulated in both cell types after a 48-h stretch ( Figure 5D). 266 Cross comparison of different time points revealed 5 upregulated genes and 2 downregulated genes in 267 hiPSC-CMs after 48 h of stretch with similar expression changes in NRVMs after 24 h of stretch 268 (Table S3). In addition, 20 upregulated genes and one downregulated gene in hiPSC-CMs after a 24-h 269 stretch were similarly differentially expressed in NRVMs after a 48-h stretch. Hence, the differentially 270 expressed genes showed the most similarity between upregulated genes in hiPSC-CM at 24 h and in 271 NRVMs at 48 h. Only three genes showed opposing expression in hiPSC-CMs and NRVMs: MASP1, 272 ENO3, and CES1 were upregulated in hiPSC-CMs but downregulated in NRVMs. 273 We also compared our differential gene expression data at 48 h with that from 48 h stretched hESC-274 CMs reported by Ovchinnikova et al. 13 Of 936 differentially expressed genes in stretched hESC-CMs, 275 only 13 genes were similarly expressed in hiPSC-CMs: four genes were upregulated, and nine genes 276 were downregulated. The upregulated genes included TUBB2B, which was upregulated in all cell 277 types (hiPSC-CMs, NRVMs and hESC-CMs) after 24 h (not studied in hESC-CMs) and 48 h of 278 stretching. Other upregulated genes were DUSP13, ACAT2 and ENO3. In addition, one of the genes 279 downregulated in both hiPSC-CMs and NRVMs, ZNF519, was also downregulated in hESC-CMs. In 280 hiPSC-CMs and hESC-CMs, no changes in opposite directions were observed in any differentially 281 expressed gene. All common differentially expressed genes and their fold changes are presented in 282 Table S3. 283 We also searched for enriched GO terms for molecular function and cellular components. Again, 296 enriched GO terms were only found for upregulated genes. Three molecular functions were 297 significantly overrepresented: structural molecule activity, structural constituent of the cytoskeleton, 298 and calcium-dependent protein binding ( Figure S5). Among the cellular components, 14 GO terms, 299 especially terms related to extracellular vesicles, supramolecular complexes and cytoskeleton, were 300 enriched ( Figure S5). These analyses confirm that structural and cytoskeletal protein-coding genes are 301 among the most upregulated genes in stretched hiPSC-CMs. The complete data of the GO analyses are 302 available in Dataset S2. 303

305
To compare stretch-induced enriched biological processes in hiPSC-CMs and NRVMs, we performed 306 a similar GO analysis for upregulated genes in NRVMs reported by Rysä et al. 12 In line with a higher 307 number of upregulated genes in NRVMs compared to hiPSC-CMs, more enriched biological processes 308 were found (71 GO terms). Hence, while a very limited number of specific processes were enriched in 309 stretched hiPSC-CMs, a broad range of biological processes were detected in NRVMs. The most 310 evidently enriched biological processes associated with upregulated genes in NRVMs were RNA 311 metabolic processes, response to stimulus, biosynthetic processes, cellular component biogenesis, 312 developmental processes, and regulation of cell death ( Figure 6B). Upregulated genes from both 313 hiPSC-CMs and NRVMs thus share GO terms associated with the regulation of apoptosis and steroid 314 biosynthesis. 315 To further discover the functionality of the differentially expressed genes, KEGG and Reactome 316 pathway analyses were performed. KEGG pathway analysis revealed 11 and 10 enriched pathways in 317 upregulated genes in hiPSC-CMs and NRVMs, respectively ( Figures 7A and 7B). However, none of 318 the pathways was common to both cell types. In hiPSC-CM, the pathways included cardiac-and 319 cardiomyocyte-associated pathways, while the enriched terms in NRVMs were heterogeneous, and 320 half of them were cancer-associated. In turn, Reactome pathway analysis resulted in 31 and 17 321 enriched pathways for hiPSC-CMs and NRVMs, respectively ( Figures 7C and 7D). Three pathways 322 were enriched in both cell types: striated muscle contraction, HSP90 chaperone cycle for steroid 323 hormone receptors (SHRs), and the role of GTSE1 in G2/M progression after the G2 checkpoint. 324 were dysregulated only modestly in response to 24-h stretch in RNAseq analysis, were not 362 significantly affected by stretch in this experiment ( Figure S6). In contrast, genes showing more 363 pronounced changes in RNAseq analysis, were significantly dysregulated. Increased NPPB (5.6-fold; 364 p=0.008), ACTA1 (8.0-fold, p=0.008) and GAL (10-fold, p=0.008) expression and decreased 365 LINC00648 expression (60%, p=0.008) was detected in stretched control cells compared to 366 unstretched control cells (Figure 9). 367 Two of the inhibitors had significant effects on NPPB expression ( Figure 9A). The MEK1/2 inhibitor 368 U0126 at a concentration of 10 µM decreased the NPPB expression both in unstretched and stretched 369 hiPSC-CMs compared to control (DMSO). However, it was not able to block the stretch response 370 completely. In contrast, the inhibitor of classical PKC isoforms Gö6976 at 1 µM increased basal 371 NPPB expression, but had no effect on stretch-induced NPPB expression. The p38 MAPK inhibitor 372 SB203580 at 10 µM and the pan-PKC inhibitor Gö6983 at 1 µM had no effect on NPPB expression. 373

Enrichment of transcription factor targets sites in hiPSC-CMs and NRVMs
The expression of ACTA1 was only affected by the classical PKC inhibitor Gö6976, which slightly 374 increased the basal expression ( Figure 9B). None of the inhibitors affected the stretch-induced 375 13 upregulation of ACTA1. However, U0126 showed tendency towards stretch-reversing effect on 376 ACTA1 expression. 377 Both U0126 and Gö6976 decreased the baseline GAL expression ( Figure 9C). However, only U0126 378 was able to inhibit the stretch-induced GAL expression. Pan-PKC inhibitor Gö6983 showed no effect 379 on GAL expression, while p38 MAPK inhibitor SB203580 showed high variability but no clear 380 influence on the GAL expression. 381 The MEK1/2 inhibitor and both PKC inhibitors increased the LINC00648 expression at baseline 382 ( Figure 9D). However, none of the inhibitors affected the stretch-induced decrease in LINC00648 383 expression, although U0126 and Gö6983 showed a tendency towards stretch-reversing effect. 384 Overall, since none of the inhibitors was able to block the stretch response completely, none of the 385 pathways investigated alone is fully responsible for the stretch response of these four genes. The 386 results however suggest involvement of MEK1/2 and PKC in mediating the hiPSC-CM stretch 387 response. 388

389
Prolonged mechanical load leads to maladaptive changes in the heart, including cardiomyocyte 390 hypertrophy and left ventricular hypertrophy, which are major causes of heart failure. 5 Understanding 391 the molecular mechanisms that underlie the development of left ventricular hypertrophy is essential 392 for finding new treatments for heart failure. Identification of genes and pathways involved mechanical 393 stretch response of cardiomyocytes is therefore of great interest. 394 The optimal in vitro model of cardiomyocyte hypertrophy would utilize adult human cardiomyocytes. 395 However, they are difficult to obtain and culture long-term, therefore the most commonly used in vitro 396 hypertrophy models employ NRVMs. To study this phenomenon in human cells and reduce the use of 397 experimental animals, we used hiPSC-CMs, which are increasingly used as in vitro models in disease 398 modeling and drug development. [35][36][37] To our knowledge, the transcriptional responses of hiPSC-CMs 399 to cyclic mechanical stretch have not been characterized before. We validated the model by measuring 400 the expression of well-established mechanical stress-responsive genes, NPPA and NPPB, 25  cardiomyocytes. In the hESC-CM study, 13 >98% pure cardiomyocytes were used, but the magnitude 417 and frequency of the stretch applied were different. Moreover, the age and maturation level of the 418 cardiomyocytes were not described in the hESC-CM study, and these differences may influence the 419 comparison of the two human cardiomyocyte models. Although transcriptionally hESC-CMs and 420 hiPSC-CMs are very similar, 39 their responses to stretch were different. As NRVMs, hESC-CMs also 421 seem to respond to stretching by inducing a broad range of gene expression changes, while in hiPSC-422 CMs, differentially expressed genes are more defined. Only one gene, TUBB2B, coding for tubulin 423 beta-2A chain, a constituent of microtubules, was upregulated in all cardiomyocyte types. In hiPSC-424 CMs, other forms of alpha-and beta-tubulins were also upregulated. The increase in the expression of 425 microtubules is strongly associated with cardiac hypertrophy; thus, this change was anticipated. 40 In 426 contrast, one gene, ZNF519, coding for zinc finger protein 519, was downregulated in all cell models. 427 ZNF519 has not been characterized in cardiomyocytes, and its potential role in the development of 428 cardiomyocyte hypertrophy remains to be established. 429 In response to stretch, two central changes occur in cardiomyocytes: (1) several genes normally 430 expressed only in embryonic or fetal hearts are reactivated, and (2) the expression of sarcomeric and 431 other constitutive proteins is increased. 41,42 Here, we showed that these changes occur also in hiPSC-432 CMs in response to stretching. The upregulation of contractile proteins was also reflected in the GO 433 enrichment analysis, where most enriched processes were associated with actin-myosin filament 434 sliding and muscle contraction. 435 Although the differentially expressed genes and their numbers varied in the compared cardiomyocyte 436 models, some similarities in the enriched processes and pathways were discovered. Regulation of cell 437 death and sterol biosynthesis were enriched in the upregulated genes in all cell types. Apoptosis has 438 previously been linked to hypertrophy in multiple studies both in rodents and in humans. [43][44][45][46] Although 439 the upregulation of genes associated with steroid biosynthesis has been reported in previous 440 studies, 12,13 its role in cardiomyocyte hypertrophy has not been characterized. Increased steroid 441 synthesis might be needed for the growth of cardiomyocytes or may be associated with changes in 442 15 energy metabolism. Steroid biosynthesis is downregulated in the neonatal mouse heart within the first 443 nine days of postnatal life, during which the heart loses its regenerative capacity. 47 Hence, it can be 444 speculated that increased steroid synthesis is a part of the fetal program that is reactivated in response 445 to stress. 446 Several genes associated with both apoptosis and cardiomyocyte hypertrophy, such as CRYAB, ENO1 447 and GSTO1, were among the upregulated genes in hiPSC-CMs. 48-55 ENO1, which codes for the 448 glycolytic enzyme α -enolase and is normally highly expressed in embryonic and fetal heart but only 449 weakly in adult heart, has shown to increase during hypertrophy in animal models. 56-58 This is in line 450 with previous evidence of a metabolic switch from fatty acid to glycolysis during pathological 451 hypertrophy. 59 Furthermore, one study has shown compensatory increase in α -enolase expression to 452 protect cardiomyocytes from hypertrophy. 56 Interestingly, after a 48-h stretch, the most upregulated 453 genes were galanin and GMAP prepropeptide coding gene GAL. Galanin is expressed principally in 454 the nervous system and in some peripheral organs, but no expression in cardiomyocytes has been 455 reported. 60 However, its receptors are expressed in various cell types, including cardiomyocytes, and it 456 has been suggested to be cardioprotective. 60-64 457 All cardiomyocyte types included in the present comparisons, hiPSC-CMs, hESC-CMs and NRVMs, 458 are considered relatively immature and do not fully correspond to adult cardiomyocytes in terms of 459 their sarcomere structure, metabolism, or electrophysiological properties. 8 However, based on our 460 comparison, stretched hiPSC-CMs were the only cell model in which biological processes of muscle 461 contraction and actin-myosin filament sliding were enriched among the upregulated genes. In view of 462 in vivo cardiac overload, these are the most important processes to enhance in order to preserve 463 cardiac pump function. However, these changes could also imply maturation of hiPSC-CMs, but this 464 is unlikely because they were accompanied by upregulation of the fetal gene program and apoptosis-465 associated genes. Moreover, hiPSC-CMs were the only cells in which upregulated genes had 466 enrichment of pathways for hypertrophic cardiomyopathy. On the contrary, these pathways were 467 enriched among downregulated genes of stretched hESC-CMs. 13 Taken together, hiPSC-CMs show 468 distinct hypertrophic changes in gene expression at the levels of individual genes and biological 469 processes, indicating that cyclic stretching of hiPSC-CMs is an advantageous in vitro model for 470 studying mechanically induced cardiomyocyte hypertrophy. 471 When comparing the gene expression changes of stretched hiPSC-CMs to another widely utilized 472 hypertrophic stimulation, ET-1, we identify only few similarities.  In that study, 696, 152 and 163 genes were 474 dysregulated after 24-h, 48-h and 72-h ET-1 treatment, respectively. The 23 upregulated genes, which 475 were common to our stretch model, included for instance fetal genes (ACTA1, NPPB, TAGLN) and 476 structural protein coding genes (ACTC1, KRT8, KRT18, TUBB2A, TUBB2B, TUBB6), while the three 477 common downregulated genes were DLG2, PLCG2 and SLC35E2A. In addition, Aggarwal et al. 478 performed mRNA profiling for 18-h ET-1-treated hiPSC-CMs. 66 In that study, 235 and 290 genes 479 were up-and downregulated, respectively, in response to ET-1. The 15 upregulated genes, which were 480 common to our stretch model, included again fetal genes (ACTA1, NPPB, TAGLN) and These results suggest that p38 MAPK and MEK1/2-ERK1/2 pathways are related to the hypertrophic 507 response also in human cardiomyocytes. To our knowledge, the role of PKC in stretch-induced 508 hypertrophy has not been studied human cardiomyocytes before. However, we have previously shown 509 that activation of all PKC isoforms or inhibition of classical PKC isoforms induce pro-hypertrophic 510 changes in hiPSC-CMs. 10 Therefore, we were interested to study the genetic response of hiPSC-CMs 511 to inhibition of these kinases in combination with cyclic mechanical stretch. Interestingly, MEK1/2 512 inhibition blocked GAL expression almost completely in both unstretched and stretched hiPSC-CMs. 513 The inhibitory effect of MEK inhibition on GAL expression has previously been shown in neuronal 514 cells, but not in cardiomyocytes. 75 Hence, it seems that MEK1/2-ERK1/2 is a common pathway for 515 GAL expression regardless of the cell type. In addition, MEK1/2 inhibition markedly decreased 516 stretch-induced NPPB expression, but could not fully block it, unlike in ET-1-induced hiPSC-CM 517 hypertrophy. 10 In hiPSC-CMs, ET-1 and mechanical stimulation thus seem to induce BNP expression 518 through different signaling pathways, which is in line with the differences in gene expression profiling 519 data from stretched and ET-1-treated hiPSC-CMs. 66 Overall, these results suggest that the stretch-520 induced gene expression changes are mediated through different signaling pathways, and that at least 521 for some of the genes studied here the stretch-response is regulated via multiple intracellular signaling 522 pathways. 523 We recognize that the hypertrophy model of stretched hiPSC-CMs has its limitations. First, as 524 mentioned before, the hiPSC-CMs are not perfectly resembling adult cardiomyocytes and they are 525 relatively immature in many regards. 8 However, our results suggest that the gene expression response 526 of these cells resembles adult human cardiomyocytes better than that of NRVMs. Second, hiPSC-CMs 527 are cultured in isolation of other cardiac cell types. Hence, they lack the heterotypic cell-cell 528 interactions that are present in normal cardiac environment. Single-cell sequencing of co-cultured cells 529 could be performed to study the hypertrophic responses of cardiomyocytes in heterocellular 530 environment. Third, our model reflects the onset of cardiomyocyte hypertrophy, and is thus not 531 comparable to cardiomyocyte hypertrophy of late-stage hypertrophic heart. The model could however 532 be further utilized to study long-term effects. Finally, the study was conducted using only one healthy 533 hiPSC line. However, only minimal differences have been demonstrated in cardiomyocytes produced 534 by the differentiation protocol used here. 14 Therefore, similar stretch responses could also be predicted 535 in other hiPSC lines with no known genetic mutations affecting cardiomyocyte function. 536 In conclusion, in the present study, we showed that mechanical stretching of hiPSC-CMs is a relevant 537 in vitro model for studying human cardiomyocyte hypertrophy. We elucidated stretch-induced 538 transcriptional changes and identified biological processes and pathways associated with these gene 539 expression changes. The changes, including activation of the fetal gene program and upregulation of 540 constitutive protein coding genes, were characteristic of cardiomyocyte hypertrophy. Comparison to 541 previous data of stretched NRVMs and hESC-CMs demonstrated that hiPSC-CMs revealed more 542 defined changes in gene expression and that differentially expressed genes were restricted to cardiac 543 and hypertrophy-related genes. In addition, we identified several differentially expressed genes with 544 no or weak previous association with cardiomyocyte hypertrophy. These results can be utilized to 545 further elucidate hypertrophic signaling pathways and to discover potential pharmacological targets 546 and biomarkers of cardiomyocyte hypertrophy. 547