VENTRICULAR TISSUE CULTURE MODEL FOR

CARDIOVASCULAR RESEARCH

KRISTY PURNAMAWATI (B. Eng., NUS)

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

NUS GRADUATE SCHOOL

FOR INTEGRATIVE SCIENCES AND ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2015

DECLARATION

I hereby declare that this thesis is my original work and it has been written by me in its entirety. I have duly acknowledged all the sources of information which have been used in the thesis.

This thesis has also not been previously submitted for any degree in any university.

______

Kristy Purnamawati

06 May 2015

ii

ACKNOWLEDGEMENTS

The completion of this thesis will not be possible without the kind and generous help

from so many people both in Singapore and Germany. I wish to express my most sincere

gratitude to my supervisors and mentors Dr. William Sun, Prof. E. Birgitte Lane, A/Prof. Cao Tong,

Prof. Dr. Wolfram H. Zimmermann, Dr. Poh Loong Soong and Dr. Malte Tiburcy for their

invaluable guidance, intellectual input and constant support throughout the course of my study.

My special thanks go out to the Zimmermann lab in Göttingen for welcoming me to be part their lab for a good portion of the project.

I would also like to thank A*STAR Graduate Academy (AGA), Deutsches Zentrum für

Herz-Kreislauf-Forschung E.V. (DZHK) and Deutsche Forschungsgemeinschaft (DFG) for their financial support, as well as NGS, NUS for facilitating the multitudes of change that occurred

throughout this PhD journey.

This work was completed in 27 months, from conception to last day of data collection. I

am very grateful to my family and friends for their love and encouragement. It kept me going.

YHJ, this work is for you. RIP.

iii

TABLE OF CONTENT

TITLE PAGE ...... i

DECLARATION ...... ii

ACKNOWLEDGEMENTS ...... iii

TABLE OF CONTENT ...... iv

SUMMARY ...... vi

LIST OF FIGURES ...... ix

LIST OF TABLES ...... xiii

LIST OF APPENDICES ...... xv

LIST OF ABBREVIATIONS ...... xvi

PUBLICATIONS, PRESENTATIONS AND AWARDS ...... xvii

CHAPTER 1 ...... 18 General Introduction 1.1. Introduction ...... 19 1.2. References ...... 23

CHAPTER 2 ...... 25 Differentiation of hPSCs to Cardiomyocytes and Generation of Engineered Heart Muscles (EHMs) 2.1. Abstract ...... 28 2.2. Introduction and Literature review ...... 29 2.3. Materials and Methods ...... 43 2.4. Results ...... 54 2.5. Discussion ...... 67 2.6. References ...... 78

CHAPTER 3 ...... 85 Engineering and Functional Validation of MYL2- Knock-In Reporter Line 3.1. Abstract ...... 88 3.2. Introduction and Literature review ...... 90 3.3. Materials and Methods ...... 96 3.4. Results ...... 116 3.5. Discussion ...... 144 3.6. References ...... 150

iv

CHAPTER 4 ...... 153 In vitro Maturation of HESCs-derived Cardiomyocytes 4.1. Abstract ...... 156 4.2. Introduction and literature review ...... 158 4.3. Materials and Methods ...... 172 4.4. Results ...... 175 4.5. Discussion ...... 219 4.6. References ...... 225

CHAPTER 5 ...... 231 Utility of in vitro Ventricular Cardiomyocytes Model 5.1. Abstract ...... 233 5.2. Introduction and literature review ...... 235 5.3. Materials and Methods ...... 239 5.4. Results ...... 241 5.5. Discussion ...... 269 5.6. References ...... 278

CHAPTER 6 ...... 281 General Conclusion and Future Directions

APPENDICES ...... 286

v

SUMMARY

The ventricles of the heart are the chambers responsible for pumping blood into the lungs as well as the entire body. In these chambers of the heart, the working myocytes are termed ventricular cardiomyocytes (vCMs). Diseases associated with the ventricular chambers of the heart such as ventricular fibrillation, myocardial infarction and cardiomyopathy would ultimately affect the properties and functions of these vCMs. Limited supply (Peeters et al., 1995)

and the lack of regenerative capacity of adult human vCMs limit their utility for in vitro studies

(Mummery, 2005). Therefore, a vCM-specific in vitro model would be a useful tool for the

development of therapies for ventricular chambers-associated diseases. In addition, the vCM in

vitro model can also be used in cardiotoxicity drug screening, to screen newly developed drugs

for ventricular pro-arrhythmic liability.

In this study, human embryonic stem cells (hESCs) are used as the source of cardiomyocytes (CMs). However, cardiac differentiation of hESCs gives rise to all three subtypes of CMs. In addition, these hESCs-derived CMs are phenotypically immature, closest to fetal or at best, neonatal CMs in characteristic. To facilitate the fractionation of vCMs from a pool of mixed hESCs-derived CMs, an MYL2 specific knock-in reporter line was developed. MYL2 has been reported to be specifically expressed in vCMs and not atrial cardiomyocytes (Bird, 2003; Zhang et al., 2011). The knock-in line was generated with homologous-recombination technique with the aid of a novel genome editing tool: transcription activator-like effector nuclease (TALEN).

Two reporters (mCherry and Gaussia Luciferase) and one selectable marker (neomycin resistance) were knocked into the last exon of MYL2 , behind the coding sequence of the gene. PCR and Southern blotting confirmed that the targeting construct was inserted at the correct site. Immunohistochemistry and FACS analysis, together with karyotype data, also

vi

confirmed that the knock-in lines are pluripotent and do not show any gross chromosomal

aberrations.

Functional validation of this novel reporter line necessitates a robust hESCs cardiac

differentiation protocol. Therefore another major part of the present study was to optimize a

previously used in-house CM differentiation protocol to one that is more robust and

reproducible. Over nine iterations, the KP protocol was optimized. The final version of the

protocol consistently generated beating embryoid bodies and monolayer of cardiomyocytes by

day 9 of differentiation.

Using this optimized protocol, two newly generated MYL2 knock-in reporter lines

(H9V15 and H9VC16) were differentiated into CMs to validate their functionality. Quantitative

PCR data from mRNAs extracted from these knock-in CMs showed strong inter-reporter and

reporter–MYL2 gene expression correlation. Fluorescence microscopy analysis indicated that

mCherry was expressed in knock-in CMs starting from day 15 of differentiation. In addition, CMs

positive for mCherry and resistant to neomycin expressed significant levels of MYL2 and

negligible expression of MYL7 (a gene expressed in immature vCMs and atrial CMs). Gaussia

luciferase was also stably expressed in MYL2-expressing cells. Taken together, the novel MYL2 knock-in reporter lines are fully functional as designed.

The third part of this study makes use of the knock-in CMs to generate biomimetic cardiac tissue constructs: engineered heart muscles (EHMs), to study CM maturation in a three- dimensional tissue construct. A number of factors were tested for their ability to enhance EHMs maturation in vitro. These factors included auxotonic stretching, pre-load variation, extracellular calcium variation, electrical stimulation and in vitro aging. The in vitro aging experiment revealed a number of up-regulated and down-regulated as well as microRNAs as EHMs vii

developed in vitro; these are elaborated in Chapter 5. Auxotonic stretching, pre-load variation

and extracellular calcium variation were not found to improve EHM maturation in vitro, in that

all EHMs in all experimental groups showed negative force-frequency response, similar to fetal

CM behavior. Electrical stimulation was found in this study to enhance EHM maturation in that calcium handling, force frequency and Isoprenaline/carbachol response of these EHMs were improved in the experimental group versus the control group. The mCherry-positive CMs also appeared to be more aligned in the electrically stimulated group than in the control group.

Finally, the cell line was used in two further applications: (1) the development of a vCM- specific differentiation protocol, and (2) the investigation for the effects of cardio-active drugs in

EHMs. With this new protocol, more than 45% mCherry positive population was routinely obtained with each round of differentiation. The pilot drug testing experiment also revealed quantitative luciferase data to be a useful predictor for EHM contractile forces under normal conditions.

In summary, novel MYL2 knock-in reporter lines were developed successfully in this study. These lines would be useful tools for identifying and enriching vCMs, allowing for in vitro

studies of this CM subtype. In addition, electrical stimulation was identified as the modality of

choice for improving CM maturation in vitro. These new tools and insight will guide future

research of vCM development and maturation in vitro.

viii

LIST OF FIGURES

Figure 1.1- 1 Diagram of internal anatomy of the heart wall…………………………………………………….……………. .20 Figure 2.2.1-1 Schematic of current knowledge of factors involved in hPSCs to CMs differentiation...... 33 Figure 2.3.1-1 Burridge 2011 protocol ...... 43 Figure 2.3.1-2 KP 3D protocol (optimized) ...... 44 Figure 2.3.2-1 Lian 2012 protocol timeline ...... 45 Figure 2.3.2-2 Hudson 2014 protocol ...... 45 Figure 2.3.3-1 Technical drawing of EHM molds for casting ...... 47 Figure 2.3.3-2 Auxotonic stretchers A – Teflon® casting molds for auxotonic stretcher ...... 47 Figure 2.3.3-3 EHMs formation ...... 50 Figure 2.3.3-4 EHM ring on auxotonic stretcher...... 50 Figure 2.3.4-1 EHM ring stretched in organ bath of contraction setup ...... 51 Figure 2.3.6-1 Technical drawing of printed EHM casting plate ...... 54 Figure 2.4.1-1 EBs generated by in house protocol have variable sizes ...... 55 Figure 2.4.1-2 The Burridge protocol allows the generation of uniformly-sized EBs, depending on the number of cells seeded per well...... 56 Figure 2.4.3-1 CM viability post trypsinization was improved over 4 iterations ...... 65 Figure 2.4.3-2 Forces comparison between 30%HFF-EHMs and 50%HFF-EHMs ...... 66 Figure 2.3.3-3 Forces comparison between 30%HFF-EHMs and 50%HFF-EHMs ...... 66 Figure 2.4.4-1 Forces of EHMs casted on round and oval shaped printed wells...... 67 Figure 2.3.4-2 Forces of EHMs casted on round and oval shaped printed wells...... 67 Figure 3.4.1-1 Addition of DMSO improved MYL2 PCR amplification ...... 117 Figure 3.4.1-2 MYL2 PCR was more efficient with Long Template PCR DNA Polymerase (Roche) ...... 117 Figure 3.4.1-3 LoxP-PGK-Hyg-LoxP PCR product ...... 117 Figure 3.4.1-4 pCR2.1-MYL2 after digestion with EcoRI ...... 118 Figure 3.4.1-5 Restriction digest profile of SAC5 ...... 119 Figure 3.4.1-6 Plasmid map of SAC5 ...... 119 Figure 3.4.1-7 TALEN plasmid map ...... 120 Figure 3.4.1-8 HindIII and NotI double digest showed TALEN insert at 3.057 kb ...... 120 Figure 3.4.1-9 pEF6-TALEN plasmid map...... 121 Figure 3.4.1-10 TALEN restriction digest profile showing right orientation in expression vector ...... 121 Figure 3.4.2-1 Manually cut hPSC colonies on inactivated MEFs, ready for passaging ...... 122 Figure 3.4.2-2 Manually cut hPSC colonies on inactivated HFFs, ready for passaging ...... 122 Figure 3.4.2-3 Trypsinized hPSCs on inactivated MEFs, ready for passaging ...... 122 Figure 3.4.3-1 Restriction digest profile of TALEN F7, TALEN R2 and SAC5 before electroporation showed expected bands ...... 123 Figure 3.4.3-2 BglII completely linearized SAC5 ...... 123 Figure 3.4.3-3 Electroporated hPSCs just before Hygromycin B selection ...... 124 Figure 3.4.3-4 Maintenance of knock-in hESCs and selection of positive clones ...... 124 Figure 3.4.3-5 Electroporated hPSCs 12 days after Hygromycin B selection, ready for manual passaging 124 Figure 3.4.3-6 Electroporated hPSCs 5 days after Hygromycin B selection ...... 124 Figure 3.4.3-7 Schematic diagram showing hybridization site of Nde500L and Nde508L to gDNA of knock in line ...... 125

ix

Figure 3.4.3-8 Southern blot of MBJ5 knock in clones V43 and V74 showed that they contain both knock-in and wild type allele (heterozygous knock-in) ...... 125 Figure 3.4.3-9 Southern blot of H9 knock-in clones. All screened clones have both knock-in and wild type alleles (heterozygous knock-in) ...... 125 Figure 3.4.3-10 Southern blot of MBJ5 V167 and hes3 knock-in clones ...... 126 Figure 3.4.4-1 Result of PCR1 from all clones with SAC5 as negative control ...... 127 Figure 3.4.4-2 Result of PCR2 from all clones with SAC5 as negative control. Faint bands are labeled with asterisks ...... 128 Figure 3.4.4-3 Result of PCR3 from selected clones with SAC5 as negative control ...... 128 Figure 3.4.4-4 Result of PCR1 and PCR2 from H9VC16 indicated that PGK-HygroR cassette has been successfully floxed out ...... 128 Figure 3.4.4-5 Result of PCR3 from H9VC16 indicated that there is no genomic integration of Cre-IRES- Puro ...... 128 Figure 3.4.5-1 H3V9 karyotype ...... 129 Figure 3.4.5-2 Knock-in hESCs are positive for pluripotency markers: OCT3/4, SOX2, NANOG, TRA-1-60 and SSEA4 ...... 132 Figure 3.4.6-1 Fused EHMs, showing positive mCherry signal...... 133 Figure 3.4.6-2 EHMs under 561 nm excitation (left) and bright field (right) at different time points. MCherry expression increased with time in culture...... 134 Figure 3.4.6-3 H9V15 EHM contractile forces increased with time in culture ...... 135 Figure 3.4.6-4 Maximum H9V15 EHM forces at different time points...... 135 Figure 3.4.6-5 Lambda scan of H9V15 CM indicated that emission signal is specific for mCherry ...... 136 Figure 3.4.6-6 mCherry expression in H9 CMs and H9V15 CMs reseeded back at ...... 137 Figure 3.4.6-7 Mean fluorescence intensity of H9VC16-CMs is higher than that from H9V150CMs both on day 45 and day 60 of differentiation ...... 137 Figure 3.4.7-1 H9VC16 CMs after G418 selection. Scale bar = 50µm...... 139 Figure 3.4.8-1 Gaussia Luciferase expression in EHMs at different time points ...... 140 Figure 3.4.9-1 Relative expression of genes in H9V15 EHMs at different time points...... 141 Figure 3.4.10-1 Relative mRNA expression levels in mChr(+) and mChr(-) CMs sorted from H9V15 CMs. 143 Figure 3.4.10-2 Relative mRNA expression levels in mChr (+) and mChr (-) CMs sorted from H9VC16 CMs...... 143 Figure 3.4.10-3 Relative mRNA expression levels in selected H9VC16 CMs...... 144 Figure 4.2.2-1 Schematic representation of ionic currents measured in hESCs, early and late CMs...... 166 Figure 4.4.1-1 Action potential #1 from fused EHM...... 176 Figure 4.4.1-2 Action potential #2 of fused EHM ...... 176 Figure 4.4.2-1 Heat map of gene expression in EHMs at different time points ...... 178 Figure 4.4.2-2 Patterns of gene expression profiles in EHMs over time in culture...... 179 Figure 4.4.3-1 Heat map of miRNA expression in EHMs at different time points ...... 189 Figure 4.4.3-2 miRNA expression profiles of EHMs at different time points...... 190 Figure 4.4.3-3 Dot plot of normalized read counts of hsa-miR-146b-5p in EHMs at different time points...... 191 Figure 4.4.3-4 Dot plot of normalized read counts of hsa-let-7f-5p in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range . 194 Figure 4.4.3-5 Dot plot of normalized read counts of hsa-miR-425-5p in EHMs at different time points. . 195 Figure 4.4.3-6 Dot plot of normalized read counts of hsa-miR-4461 in EHMs at different time points. .... 199 Figure 4.4.3-7 Dot plot of normalized read counts of hsa-miR-6087 in EHMs at different time points. .... 199 x

Figure 4.4.4-1 Electrically stimulated EHM rings showed rougher surface and areas of thinning (arrows) ...... 200 Figure 4.4.4-2 Electrically stimulated EHMs contracted with lower forces than non-electrically stimulated EHMs ...... 203 Figure 4.4.4-3 Electrical stimulation did not affect sensitivity of EHMs to extracellular calcium concentration ...... 204 Figure 4.4.4-4 Electrically stimulated EHMs showed more positive force-frequency response than non- electrically stimulated EHMs ...... 204 Figure 4.4.4-5 C16 EHMs response to isoprenaline and carbachol ...... 206 Figure 4.4.5-1 Gaussia Luciferase expression of stimulated H9 and V15 EHMs ...... 209 Figure 4.4.5-2 Electrically stimulated EHMs contracted with lower forces than non-electrically stimulated EHMs. 20% increase in preload was sufficient to abolish contractions in EHMs ...... 210 Figure 4.4.5-3 Electrical stimulation lessened calcium hypersensitivity in V15 EHMs at 60mm ...... 210 Figure 4.4.5-4 Electrical stimulation induced a more positive force-frequency response in V15 EHM at 60mm ...... 211 Figure 4.4.5-5 Responses to isoprenaline and carbachol in electrically stimulated vs non-stimulated EHMs ...... 212 Figure 4.4.5-6 Electrical stimulation improved response to cardioactive drugs ...... 212 Figure 4.4.5-7 Gaussia Luciferase expression of stimulated H9 and C16 EHMs ...... 213 Figure 4.4.5-8 Electrical stimulation and 10% increase in preload lowered contraction force in C16 EHMs ...... 215 Figure 4.4.5-9 10% increase in preload further worsened calcium hypersensitivity; 5V stimulation was not sufficient to lessen calcium hypersensitivity...... 216 Figure 4.4.5-10 Force frequency response of electrically stimulated vs non-stimulated EHMs ...... 217 Figure 4.4.5-11 Responses to isoprenaline and carbachol in electrically stimulated vs non-stimulated EHMs ...... 219 Figure 4.4.5-12 %change in forces after isoprenaline and carbachol treatment ...... 219 Figure 5.4.1-1 More luciferase was expressed by EHMs cultured in higher concentrations of calcium (up to 1.6mM) ...... 243 Figure 5.4.1-2 Forces of EHMs cultured in different Ca2+ concentrations ...... 244 Figure 5.4.1-3 Forces of EHMs treated with different calcium concentrations as percentage of forces at 2.8mM Ca2+ ...... 245 Figure 5.4.1-4 Force frequency response of EHMs cultured in different concentrations of calcium ...... 246 Figure 5.4.1-5 EHMs response to Isoprenaline and carbachol stimulation ...... 247 Figure 5.4.1-6 Carbachol response of EHMs cultured in different calcium concentrations ...... 247 Figure 5.4.2-1 Different Wnt inhibitors did not significantly affect luciferase expression per cell ...... 248 Figure 5.4.2-2 Different Wnt inhibitors did not significantly affect mCherry(+) fraction ...... 249 Figure 5.4.2-3 Percentage of MYL2 positive cardiomyocytes in flasks differentiated with different Wnt inhibitors ...... 249 Figure 5.4.3-1 Luciferase expression of EHMs treated with different concentrations of Endothelin-1 ...... 252 Figure 5.4.3-2 Forces of EHMs treated with different Endothelin-1 concentrations ...... 252 Figure 5.4.3-3 Force frequency response of EHMs treated by different concentrations of Endothelin-1 .. 253 Figure 5.4.3-4 Isoprenaline-carbachol response of EHMs treated with different concentrations of Endothelin-1 ...... 254 Figure 5.4.3-5 Increase in force after Isoprenaline treatment ...... 255 Figure 5.4.3-6 Reduction in force after carbachol treatment ...... 255

xi

Figure 5.4.3-7 Luciferase expression of EHMs treated with different concentrations of Noradrenaline ... 256 Figure 5.4.3-8 Forces of EHMs treated with different concentrations of Noradrenaline ...... 257 Figure 5.4.3-9 Force frequency response of EHMs treated with different concentrations of Noradrenaline ...... 257 Figure 5.4.3-10 Noradrenaline-treated EHMs response to Isoprenaline and Carbachol ...... 258 Figure 5.4.3-11 Reduction in forces after Carbachol stimulation...... 259 Figure 5.4.3-12 Response to Isoprenaline stimulation in Noradrenaline-treated EHMs ...... 259 Figure 5.4.3-13 Luciferase expression of EHMs treated with different concentrations of IBMX ...... 259 Figure 5.4.3-14 Forces of EHMs treated with different concentrations of IBMX ...... 260 Figure 5.4.3-15 Force frequency response of EHMs treated with different concentrations of IBMX ...... 261 Figure 5.4.3-16 Isoprenaline and Carbachol response from EHMs treated with different concentrations of IBMX ...... 262 Figure 5.4.3-17 Reduction in contratile forces in IBMX-treated EHMs after Carbachol treatment ...... 262 Figure 5.4.3-18 Response to Isoprenaline stimulation in IBMX-treated EHMs ...... 262 Figure 5.4.3-19 Luciferase expression of EHMs treated with different concentrations of Metformin ...... 263 Figure 5.4.3-20 Forces of EHMs treated with different concentrations of Metformin ...... 264 Figure 5.4.3-21 Force frequency response of EHMs treated with different concentrations of Metformin 264 Figure 5.4.3-22 Metformin-treated EHMs response to Isoprenaline and Carbachol ...... 265 Figure 5.4.3-23 Reduction in forces in Metformin-treated EHMs stimulated with Carbachol ...... 266 Figure 5.4.3-24 Response to Isoprenaline stimulation in Metformin-treated EHMs ...... 266 Figure 5.4.3-25 Gaussia Luciferase expression of EHMs treated with different concentrations of Gemfibrozil ...... 267 Figure 5.4.3-26 Contractile forces o EHMs treated with different concentrations of Gemfibrozil ...... 268 Figure 5.4.3-27 Force frequency response of Gemfibrozil-treated EHMs ...... 268 Figure 5.4.3-28 Responses of Gemfibrozil-treated EHMs to Isoprenaline and Carbachol stimulation ...... 269 Figure 5.4.3-29 Reduction in Gemfibrozil-treated EHM force after carbachol stimulation ...... 269 Figure 5.4.3-30 Increase in Gemfibrozil-treated EHM forces after isoprenaline stimulation ...... 269 Figure 5.5.1-1 Combined luciferase expression of EHMs treated with different hypertrophic-inducing drugs ...... 273 Figure 5.5.1-2 Combined EHM contractile forces treated with different hypertrophic-inducing drugs ..... 273 Figure 5.5.1-3 Summary of isoprenaline response ...... 274 Figure 5.5.1-4 Summary of carbachol response ...... 274 Figure 5.5.1-5 Summary of luciferase expression from EHMs treated with Metformin and Gemfibrozil .. 275 Figure 5.5.1-6 Summary of forces for EHMs treated with Metformin and Gemfibrozil ...... 276 Figure 5.5.1-7 Summary of isoprenaline response from EHMs treated with Metformin and Gemfibrozil . 277 Figure 5.5.1-8 Summary of carbachol response from EHMs treated with Metformin and Gemfibrozil ..... 277

xii

LIST OF TABLES

Table 2.2.2-1 Summary of published 3D based cardiac differentiation protocols ...... 38 Table 2.2.2-2 Summary of published monolayer based cardiac differentiation protocol...... 39 Table 2.3.1-1 Composition of CARM medium ...... 43 Table 2.3.1-2 Components of media used in Burridge protocol ...... 44 Table 2.3.1-3 Media composition used in KP 3D protocol ...... 44 Table 2.3.2-1 Media composition used in Hudson protocol ...... 46 Table 2.3.3-1 Composition of EHM master mix ...... 49 Table 2.4.1-1 Comparison of EB sizes from the Burridge and KP protocol ...... 57 Table 2.4.1-2 Different versions of KP 3D protocol...... 59 Table 2.4.2-1 The Lian protocol generates >80% α-actinin (+) cardiomyocytes ...... 61 Table 2.4.2-2 The Hudson protocol generates >75% α-actinin (+) cardiomyocytes ...... 63 Table 2.4.2-3 The KP 2D protocol generates >90% α-actinin (+) cardiomyocytes ...... 63 Table 3.3.1-1 PCR amplification protocol for MYL2 homology arms ...... 97 Table 3.3.1-2 PCR amplification protocol for adding LoxP to PGK-Hygro ...... 98 Table 3.3.1-3 Protocol of Chemical and Electrical transformations ...... 98 Table 3.3.1-4 Details of primers and templates used for PCR amplification of inserts for SAC ...... 99 Table 3.3.1-5 PCR amplification parameters of different inserts for SAC ...... 100 Table 3.3.1-6 PCR amplification parameters to generate F- and R- TALEN PCR products ...... 101 Table 3.3.2-1 Compositions of feeder cell culture media ...... 102 Table 3.3.3-1 Compositions of hESCs cell culture media ...... 105 Table 3.3.4-1 PCR cycling parameters for screening positive knock-in clones ...... 109 Table 3.3.5-1 Cycling parameters for PCR1 and PCR2 to screen for successfully excised clones ...... 112 Table 3.3.5-2 Cycling parameters for PCR3 to check for integration of Cre-IRES-Puro into knock-in line genome...... 112 Table 3.3.6-1 PCR cycling parameters for gDNA sequencing ...... 113 Table 3.3.6-2 List of primary antibodies used for pluripotency immunofluorescence staining ...... 114 Table 3.4.3-1 Experimental values of electroporation done on different hPSCs lines ...... 123 Table 3.4.3-2 Summary of homologous recombination efficiency ...... 124 Table 3.4.5-1 Percentage match of knock-in gDNA to expected sequence ...... 129 Table 3.4.5-2 List of mutations from direct sequencing of gDNA. Highlighted in red are the mutations that are common between the three different reporter cells lines ...... 130 Table 3.4.6-1 mCherry live FACS of CMs at different time points. MCherry expression increased with time in culture ...... 138 Table 3.4.9-1 Correlation of reporter genes to MYL2 and to each other are significant ...... 141 Table 4.2.1-1 Summary of differences between immature and adult cardiomyocytes...... 164 Table 4.3.2-1 Passive mechanical stretch (increasing CMs preload) procedures...... 174 Table 4.4.2-1 of genes which belong to profile 3 ...... 180 Table 4.4.2-2 Gene Ontology of gene which belong to profile 4 ...... 181 Table 4.4.2-3 Gene ontology of genes which belong to profile 5 ...... 181 Table 4.4.2-4 Gene Ontology of genes which belong to profile 9 ...... 182 Table 4.4.2-5 Gene ontology of genes which belong to profile 10 ...... 182 Table 4.4.2-6 Gene Ontology of genes which belong to profile 16 ...... 183

xiii

Table 4.4.2-7 Gene ontology of genes which belong to profile 17 ...... 184 Table 4.4.2-8 Gene Ontology of genes which belong to profile 18 ...... 185 Table 4.4.2-9 Functions of genes which belong to profile 22...... 185 Table 4.4.2-10 Gene Ontology of genes which belong to profile 23 ...... 186 Table 4.4.2-11 Functions of highly-upregulated genes in profile 24 ...... 187 Table 4.4.3-1 List of up-regulated miRNAs ...... 188 Table 4.4.3-2 Gene Ontology of genes targeted by hsa-miR-214-5p ...... 192 Table 4.4.3-3 Gene Ontology of genes targeted by hsa-miR-98-5p ...... 193 Table 4.4.3-4 Gene Ontology of genes targeted by hsa-miR-199b-5p ...... 195 Table 4.4.3-5 Gene ontology of genes targeted by hsa-miR-425 ...... 196 Table 4.4.3-6 List of down-regulated miRNAs ...... 196 Table 4.4.3-7 Gene ontology for genes targeted by hsa-miR-302d ...... 198 Table 4.4.3-8 Gene ontologies of genes targeted by hsa-miR-4461 ...... 199 Table 4.4.4-1 Stimulation parameters for first electrical stimulation experiment ...... 200 Table 4.4.4-2 Alpha-actinin staining result of cells extracted from electrically stimulated EHMs ...... 201 Table 4.4.4-3 Electrical stimulation parameters for experiment 3 ...... 201 Table 4.4.4-4 Gaussia Luciferase expression of control EHMs versus electrically stimulated EHMs from 2 different cell lines ...... 202 Table 4.4.4-5 Electrically stimulated V15 and C16 EHMs showed higher mCherry expression than non- stimulated EHMs ...... 202 Table 4.4.5-1 Electrical stimulation parameters for experiment 1 ...... 206 Table 4.4.5-2 mCherry signal from H9 and C16 EHMs (mechanical only vs mechanical + electrical stimulation) ...... 207 Table 4.4.5-3 Electrical stimulation parameters for experiment 2 ...... 208 Table 4.4.5-4 mCherry signal from V15 EHMs (non-stimulated vs stimulated) ...... 208 Table 4.4.5-5 Electrical stimulation parameters for experiment 3 ...... 213 Table 4.4.5-6 Electrical stimulation lowered mCherry signal in C16 EHMs. (Dotted white circles indicate position of printed poles) ...... 214 Table 5.4.1-1 mCherry expression increased in EHMs cultured in increasing calcium concentrations (up to 1.6mM) ...... 242 Table 5.4.1-2 EC50 values of EHMs cultured in different calcium concentrations ...... 245 Table 5.4.3-1 mCherry expression from EHMs cultured in different drugs for 10 days ...... 251 Table 5.5.1-1 Summary of EC50 values ...... 274 Table 5.5.1-2 Summary of EC 50 values of EHMs treated with Metformin and Gemfibrozil ...... 277

xiv

LIST OF APPENDICES

Appendix 4 - 1 G.Luc expression of 2D CMs ...... 288 Appendix 4 - 2 C16 time point qPCR data relative to GAPDH and d18 ...... 288 Appendix 5 - 1 Gene expression profile 3 (ratio is ratio of log2expression at different time points) ...... 289 Appendix 5 - 2 Gene expression profile 18 (ratio is ratio of log2expression at different time points) ...... 293 Appendix 5 - 3 Gene expression profile 23 (ratio is ratio of log2expression at different time points) ...... 295 Appendix 5 - 4 Gene expression profile 24 (ratio is ratio of log2expression at different time points) ...... 299 Appendix 5 - 5 Gene expression profile 4 (ratio is ratio of log2expression at different time points) ...... 305 Appendix 5 - 6 Genes targeted by miRNA hsa-miR-302d which enrich for “ion binding” ...... 311 Appendix 5 - 7 Genes targeted by miRNA hsa-miR-302d which enrich for “Regulation of transcription, DNA-templated” ...... 312 Appendix 5 - 8 List of miRNA that are differentially expressed in d90 and d30 EHMs. Down-regulations are shown in red ...... 313 Appendix 5 - 9 Dot plots of different miRNAs that are expressed differentially in d90 and d30 EHMs...... 315 Miscellaneous Appendix - 1 List of primers ...... 320 Miscellaneous Appendix - 2 Materials ...... 322

xv

LIST OF ABBREVIATIONS

α/βMHC Alpha/beta myosin heavy chain aCMs Atrial cardiomyocytes BMP Bone morphogenic CaMKs Calcium/calmodulin-dependent protein kinases CASQ2 Calsequestrin 2 (cardiac muscle) CMs Cardiomyocytes CPCs Cardiac progenitor cells DDR2 Discoidin domain receptor tyrosine kinase 2 EB Embryoid body EHM Engineered heart muscle EHT Engineered heart tissue ERK Extracellular signal-regulated kinases FGF Fibroblast growth factor G.Luc Gaussia Luciferase GAPDH Glyceraldehyde-3-phosphate dehydrogenase GSK Glycogen synthase kinase hERG Human ether-à-go-go related gene hESC Human embryonic stem cell HFF Human foreskin fibroblast hiPSC Human induced pluripotent stem cell hPSC Human pluripotent stem cell IRES Internal ribosomal entry site MAPK Mitogen-activated protein kinases MChr MCherry MEF Mouse embryonic fibroblast MEF2C Myocyte enhancer factor 2C mESCs Mouse embryonic stem cells MG Matrigel MI Myocardial infarction MYL7 Myosin light chain 2a MYL2 Myosin light chain regulatory, cardiac, slow NeoR Neomycin resistance RLX Relaxin ROCK Rho-associated kinase SFMM Serum free maturation media vCMs Ventricular cardiomyocytes VEGF Vascular endothelial growth factor Wnt Wingless-type MMTV integration site

xvi

PUBLICATIONS, PRESENTATIONS AND AWARDS

CONFERENCE PRESENTATIONS

Purnamawati, K., Soong, P.L., Tiburcy, M., Zimmermann, W.H., Sun, W. (2014) TALEN mediated engineering of a cardiac reporter line. Keystone Abstracts. Poster presentation delivered at the Keystone Symposia on Molecular and Cellular Biology, Olympic Valley, CA., April 2014.

PEER-REVIEWED PUBLICATIONS

Hu, M., Deng, R., Schumacher, K.M., Kurisawa, M., Ye, H., Purnamawati, K., Ying, J.Y. Hydrodynamic spinning of hydrogel fibers. Biomaterials. 2010 Feb; 31(5):863-9. DOI: 10.1016/j.biomaterials.2009.10.002. Epub 2009 Oct 29.

REVIEW ARTICLES

Verma, V., Purnamawati, K., Manasi, Shim, W. Steering signal transduction pathway towards cardiac lineage from human pluripotent stem cells: a review. Cell Signaling. 2013 May; 25(5):1096-107. doi:10.1016/j.cellsig.2013.01.027. Epub 2013 Feb 13.

AWARDS

A*STAR Graduate Scholarship August 2010 – July 2014 SembCorp Undergraduate Scholarship August 2005 – July 2009 ASEAN Pre-University Scholarship January 2003 – December 2004 ASEAN Scholarship January 2001 – December 2002

xvii

CHAPTER 1

General Introduction

18

1.1. Introduction

The heart consists of two upper atrial chambers and two lower ventricular chambers. Atria

serve as blood collection points whereas ventricles serve as pumps. The walls of the heart

comprise three main layers: the epicardium, myocardium and endocardium (figure 1.1-1). It is the myocardium layer that contracts and makes up the atria and ventricles. The myocardium, in turn, consists of a number of components: cardiomyocytes (CMs), extracellular matrices, blood vessels, lymphatic vessels, nerves and interstitial cells (of Cajal) (Filipoiu, 2014). CMs in the atria are accordingly termed atrial cardiomyocytes (aCMs), whereas those in the ventricles are termed ventricular cardiomyocytes (vCMs). The third CM subtype: the nodal CMs are located at the sinoatrial (SA) and atrioventricular (AV) nodes at the right atrium.

Figure 1.1- 1 Diagram of internal anatomy of the heart wall

19

Atrial and ventricular cardiomyocytes differ in their electrophysiology and molecular

biology (Ng et al., 2010). Ventricular CMs (vCMs) have action potentials with lower resting

potential, higher amplitude and longer phase 2 (figure 1.1-2). This is due to the differential

expressions of ion channels on the cell surface membranes of the two CM subtypes (Ng et al.,

2010). Details on each of the ion channels responsible for each phase of the action potentials

are available elsewhere (Ng et al., 2010). The main differences in molecular biology of the two different subtypes are summarized in table 1.1-1 (Ng et al., 2010).

-80mV -85mV

Figure 1.1- 2 Action potentials of (A) Atrial CMs versus (B) Ventricular CMs

Parameter Atrial cardiomyocytes Ventricular cardiomyocytes Transcription factor HRT1 HRT2 IRX4 Contractile protein MYL7 MYL2 Gap junctions CX40 CX43 Membrane protein Sarcolipin - Secretory protein/peptides Atrial natriuretic factor (ANF)

Table 1.1- 1 Differences in molecular biology of atrial CMs versus ventricular CMs

The study described in this thesis aims to develop an in vitro model for the vCMs of the heart. Such a model will be useful as there is a limited supply of adult human vCMs (Peeters et

al., 1995) for use in research since adult human vCMs are difficult to obtain and they also have

very little regenerative capacity (Mummery, 2005). The availability of vCM in vitro model would

facilitate more preclinical studies aimed at developing new therapies such as cell-based therapy

for heart failure patients, in whom the pumping action of the heart has been severely

20

compromised. In addition, such model will also be useful as cardiotoxicity screening platforms

for new drugs, especially in detecting their ventricular pro-arrhythmic liability (Cavero and

Holzgrefe, 2014; Vunjak Novakovic et al., 2014). Current cardiotoxicity screenings employ animal models which are suboptimal in predicting human responses.

Human embryonic stem cells (hESCs) are used in the present study as the source of CMs.

Previous reports have described successful attempts of differentiating these cells into contractile CMs in vitro (Wei et al., 2012; Yang et al., 2008; Zhang et al., 2011; Burridge et al.,

2011; Elliott et al., 2011; Laflamme et al., 2007; Paige et al., 2010; Lian et al., 2012; Zhang et al.,

2012). Yet despite the encouraging data reported in these publications, reproducibility remains

to be an issue in this laboratory and others. Until today, researchers in the field continue to

publish their own methods of generating CMs, indicative of the difficulties encountered in

adopting protocols already published by their contemporaries.

The first part of the thesis describes efforts to develop a more robust and reproducible

cardiac differentiation protocol in order to improve CM yield. An optimized protocol was

obtained after nine rounds of iterations; the final version of this protocol consistently generated

contractile embryoid bodies and CM monolayer by day 9 of differentiation.

The second part of the thesis describes the development of a reporter line which would

facilitate the identification of and enrichment for vCMs from a mixture of hESCs-derived CM

subtypes. From table 1.1-1, there are four possible genes to target for this purpose. However,

out of the four, only MYL2 gene has been described to be expressed specifically in vCMs (Bird,

2003; Zhang et al., 2011). HRT2 is also expressed in endothelial and vascular smooth muscle cells

(Xin et al., 2007). IRX4, although important for ventricular genes regulation, has not been found

to be sufficient to confer ventricular identity (Bruneau et al., 2001). Therefore, in this study, 21

MYL2 gene is used as a marker to identify vCMs. The MYL2 knock-in line was designed to have two reporters (mCherry, Gaussia luciferase) and one selectable marker (neomycin resistance) inserted into the last exon and behind the coding sequences of the MYL2 gene. Endogenous

expression of MYL2 gene in knock-in hESCs will lead to stoichiometric expression of the inserted foreign genes, thus enabling enrichment for vCMs through neomycin selection or cell sorting for mCherry positive CM population by fluorescent activated cell sorting (FACS). In addition, Gaussia luciferase expression would also enable real-time monitoring of vCMs development in vitro.

The subsequent parts of the thesis describe attempts at improving the phenotypic

maturation of hESCs-derived CMs as these cells are relatively immature – they are closest in

phenotype to human fetal CMs (Cao et al., 2008; Snir et al., 2003). Immaturity of these hESCs- derived CMs limits their potential in applications such as cell therapy and drug screening. In the present study, different in vitro modalities were attempted to improve the phenotypic maturation of hESCs-derived CMs: (1) three-dimensional versus monolayer culture; (2) continuous culture for 90 days in vitro (aging); (3) electrical and/or mechanical stimulation and

(4) treatment with different extracellular calcium concentrations. CM maturation was determined by assessing the expression of reporter genes, contractile forces, calcium handling capacity, force-frequency response as well as responses to isoprenaline and carbachol stimulations. The results from these investigations are described in detail in Chapters 4-5 of this thesis.

The final chapter of the thesis provides a general discussion to unify the different research findings from this study.

22

1.2. References

Bird, S. (2003). The human adult cardiomyocyte phenotype. Cardiovascular Research 58, 423-434. Bruneau, B.G., Bao, Z.Z., Fatkin, D., Xavier-Neto, J., Georgakopoulos, D., Maguire, C.T., Berul, C.I., Kass, D.A., Kuroski-de Bold, M.L., de Bold, A.J., et al. (2001). Cardiomyopathy in Irx4-deficient mice is preceded by abnormal ventricular gene expression. Molecular and cellular biology 21, 1730-1736. Burridge, P.W., Keller, G., Gold, J.D., and Wu, J.C. (2012). Production of de novo cardiomyocytes: human pluripotent stem cell differentiation and direct reprogramming. Cell Stem Cell 10, 16 - 28. Burridge, P.W., Matsa, E., Shukla, P., Lin, Z.C., Churko, J.M., Ebert, A.D., Lan, F., Diecke, S., Huber, B., Mordwinkin, N.M., et al. (2014). Chemically defined generation of human cardiomyocytes. Nat Methods 11, 855-860. Cavero, I., and Holzgrefe, H. (2014). Comprehensive in vitro Proarrhythmia Assay, a novel in vitro/in silico paradigm to detect ventricular proarrhythmic liability: a visionary 21st century initiative. Expert Opinion on Drug Safety 13, 745-758. Dubois, N.C., Craft, A.M., Sharma, P., Elliott, D.A., Stanley, E.G., Elefanty, A.G., Gramolini, A., and Keller, G. (2011). SIRPA is a specific cell-surface marker for isolating cardiomyocytes derived from human pluripotent stem cells. . Nat Biotechnol 29, 1011-1018. Dunwoodie, S.L. (2007). Combinatorial signaling in the heart orchestrates cardiac induction, lineage specification and chamber formation. Seminars in Cell & Developmental Biology 18, 54-66. Elliott, D.A., Braam, S.R., Koutsis, K., Ng, E.S., Jenny, R., Lagerqvist, E.L., Biben, C., Hatzistavrou, T., Hirst, C.E., Yu, Q.C., et al. (2011). NKX2-5eGFP/w hESCs for isolation of human cardiac progenitors and cardiomyocytes. Nature Methods 8, 1037-1040. Filipoiu, F.M. (2014). Atlas of Heart Anatomy and Development (Springer London). Goyal, S.K. (2014). Ventricular Fibrillation J.N. Rottman, ed. (Medscape). Guzzo, R.M., Foley, A.C., Ibarra, Y.M., and Mercola, M. (2007). Signaling Pathways in Embryonic Heart Induction. 18, 117-151. Laflamme, M.A., Chen, K.Y., Naumova, A.V., Muskheli, V., Fugate, J.A., Dupras, S.K., Reinecke, H., Xu, C., Hassanipour, M., Police, S., et al. (2007). Cardiomyocytes derived from human embryonic stem cells in pro-survival factors enhance function of infarcted rat hearts. Nat Biotechnol 25, 1015-1024. Lian, X., Hsiao, C., Wilson, G., Zhu, K., Hazeltine, L.B., Azarin, S.M., Zhang, J., Kamp, T.J., Palecek, S.P., and Raval, K.K. (2012 ). Robust cardiomyocyte differentiation from human pluripotent stem cells via temporal modulation of canonical Wnt signaling. PNAS 109, E1848 - 1857. Michael, W., and M. A. Q., S. (2007). Signal Transduction in Early Heart Development (I): Cardiogenic Induction and Heart Tube Formation. Exp Biol Med 232, 852–865. Mummery, C.L. (2005). Cardiology: solace for the broken-hearted? Nature 433, 585-587. Ng, S.Y., Wong, C.K., and Tsang, S.Y. (2010). Differential gene expressions in atrial and ventricular myocytes: insights into the road of applying embryonic stem cell-derived cardiomyocytes for future therapies. American journal of physiology Cell physiology 299, C1234-1249. Paige, S.L., Osugi, T., Afanasiev, O.K., Pabon, L., Reinecke, H., and Murry, C.E. (2010). Endogenous Wnt/beta-catenin signaling is required for cardiac differentiation in human embryonic stem cells. PLoS One 5, e11134.

23

Peeters, G.A., Sanguinetti, M.C., Eki, Y., Konarzewska, H., Renlund, D.G., Karwande, S.V., and Barry, W.H. (1995). Method for isolation of human ventricular myocytes from single endocardial and epicardial biopsies. American Journal of Physiology - Heart and Circulatory Physiology 268, H1757-H1764. Sumi, T., Tsuneyoshi, N., Nakatsuji, N., and Suemori, H. (2008). Defining early lineage specification of human embryonic stem cells by the orchestrated balance of canonical Wnt/ -catenin, Activin/Nodal and BMP signaling. Development 135, 2969-2979. Tohyama, S., Hattori, F., Sano, M., Hishiki, T., Nagahata, Y., Matsuura, T., Hashimoto, H., Suzuki, T., Yamashita, H., Satoh, Y., et al. (2013). Distinct Metabolic Flow Enables Large-Scale Purification of Mouse and Human Pluripotent Stem Cell-Derived Cardiomyocytes. Cell Stem Cell 12, 127 - 137. Verma, V., Purnamawati, K., Manasi, and Shim, W. (2013). Steering signal transduction pathway towards cardiac lineage from human pluripotent stem cells: a review. Cellular signalling 25, 1096-1107. Vunjak Novakovic, G., Eschenhagen, T., and Mummery, C. (2014). Myocardial tissue engineering: in vitro models. Cold Spring Harbor perspectives in medicine 4. Wei, H., Tan, G., Manasi, Qiu, S., Kong, G., Yong, P., Koh, C., Ooi, T.H., Sze, Y.L., Wong, P., et al. (2012). One-step derivation of cardiomyocytes and mesenchymal stem cells from human White, R.B., and Ziman, M.R. (2008). Genome-wide discovery of Pax7 target genes during development. Physiol Genomics 33, 41-49. Xin, M., Small, E.M., van Rooij, E., Qi, X., Richardson, J.A., Srivastava, D., Nakagawa, O., and Olson, E.N. (2007). Essential roles of the bHLH transcription factor Hrt2 in repression of atrial gene expression and maintenance of postnatal cardiac function. Proceedings of the National Academy of Sciences of the United States of America 104, 7975-7980. Yang, L., Soonpaa, M.H., Adler, E.D., Roepke, T.K., Kattman, S.J., Kennedy, M., Henckaerts, E., Bonham, K., Abbott, G.W., Linden, R.M., et al. (2008). Human cardiovascular progenitor cells develop from a KDR+ embryonic-stem-cell-derived population. Nature 453, 524-528. Zhang, Q., Jiang, J., Han, P., Yuan, Q., Zhang, J., Zhang, X., Xu, Y., Cao, H., Meng, Q., Chen, L., et al. (2011). Direct differentiation of atrial and ventricular myocytes from human embryonic stem cells by alternating retinoid signals. Cell Res 21, 579-587.

24

CHAPTER 2

Differentiation of hPSCs to

Cardiomyocytes and Generation of

Engineered Heart Muscles (EHMs)

25

TABLE OF CONTENT

2.1. Abstract ...... 28

2.2. Introduction and Literature review ...... 29

2.2.1. Embryonic heart development ...... 30 Mesoderm formation ...... 30 Mesoderm Patterning and Cardiac Mesoderm Formation ...... 31 Cardiomyocytes terminal differentiation ...... 31

2.2.2. Differentiation of hPSCs to CMs ...... 34 Co-culture with endoderm ...... 34 2D (monolayer) and 3D (EB)-based Differentiation Techniques ...... 35

2.2.3. Phenotypic Status of hPSC-Derived CMs ...... 40 2.2.4. Cardiomyocyte Maturation in vitro ...... 42

2.3. Materials and Methods ...... 43

2.3.1. Differentiation of hESCs to immature CMs - Embryoid body (3D) approach...... 43 In-house protocol ...... 43 Burridge 2011 protocol ...... 43 KP 3D protocol (unpublished) ...... 44

2.3.2. Differentiation of hESCs to immature CMs - Monolayer (2D) approach ...... 45 Lian 2012 protocol ...... 45 KP 2D protocol (unpublished) ...... 45 Hudson (in press, 2014) protocol ...... 45

2.3.3. Generation of EHMs ...... 46 Generation of EHMs molds ...... 46 Generation of auxotonic stretchers ...... 46 Cardiomyocytes dissociation to single cell suspension ...... 48 Generation of engineered heart muscles (EHMs) ...... 49

2.3.4. Measurements of EHM contractile forces ...... 50 EHM rings force measurement ...... 50

2.3.5. Automation of casting and transfer ...... 52 Printing casting plates ...... 52

2.3.6. Statistical analysis ...... 52

2.4. Results ...... 54

26

2.4.1. Optimization of EB-based cardiac differentiation protocol ...... 54 In house protocol ...... 54 Burridge 2011 protocol ...... 54 KP 3D protocol ...... 56

2.4.2. Optimization of 2D differentiation protocol ...... 59 Lian 2012 protocol ...... 59 Hudson protocol ...... 61 KP 2D protocol ...... 62

2.4.3. Engineered Heart Muscles ...... 63 Trypsinization of cardiomyocytes ...... 63 Effect of composition of EHMs ...... 64

2.4.4. Printed plate parameters and handling affects functionality of EHMs ...... 66

2.5. Discussion ...... 67 3-dimensional approach to cardiac differentiation ...... 67 2-dimensional approach to cardiac differentiation ...... 72 Cardiomyocyte trypsinization protocol ...... 74 In vitro cardiac tissue mimetic - Engineered Heart Muscles (EHMs) ...... 75 3D-printing technology for the automated formation of EHMs ...... 76 Concluding remarks ...... 77

2.6. References ...... 78

27

2.1. Abstract

A number of bioactive molecules were investigated for their ability to direct human

embryonic stem cells (hESCs) differentiation towards cardiomyocytes (CMs). Over 9 iterations, it was determined that the synergistic effect of BMP4, Activin A and BIO-acetoxime (a GSK-3α/β inhibitor) for phase 1; Wnt inhibitor and p38/MAPK inhibitor for phase 2 efficiently drove cardiac differentiation, generating beating embryoid bodies (EBs) or CM monolayer by day 9 of differentiation.

Once contractile CMs were obtained, they were dissociated into single cell suspension

(or “trypsinization”) and combined with human foreskin fibroblasts (HFFs) at a preset cell ratio.

This mixed cell suspension was then resuspended in a collagen gel for casting in molds to make engineered heart muscle (EHM) rings. Enzymatic composition as well as trypsinization

techniques affected CM yield as they varied CM viability from 25.7% to 70.13% post-digestion.

The cell ratio of CMs to HFFs was also found to affect the functionality of the EHMs. EHMs made

from 70% CMs and 30% HFFs generated significantly higher forces than those made from 50%

CMs and 50% HFFs (0.226 ± 0.045mN vs 0.093± 0.018 mN, respectively; p<0.05, n=4, data is

presented as mean ± SEM).

In sum, a robust differentiation protocol was successfully developed over nine

iterations to efficiently generate CMs from hESCs in nine days. Enzymatic composition used as

well as trypsinization techniques are critical for maintaining viability of trypsinized CMs and 30%

HFFs are optimal for the generation of contractile EHMs.

28

2.2. Introduction and Literature review

It is widely accepted that the adult heart has limited regenerative capacity (Mummery,

2005). Events leading to cell loss or dysfunction, such as myocardial infarction (MI), are mostly

irreversible and may result in progressive heart failure and mortality. Progressive heart failure is

a burgeoning public health problem and despite advances in the medical field, prognosis for

patients with heart failures remains poor. Heart transplantation still is the gold standard of

treatment for these patients. However, lack of donors presents a major challenge to this

approach.

Cell therapy and tissue engineering present an alternative approach to the problem.

However, to be clinically relevant, large amounts of cardiomyocytes, CMs (in the order of 109) will be needed. Human pluripotent stem cells (hPSCs), given their pluripotency and inherent capacity for indefinite cell renewal, present a potential solution. Unfortunately as will be evident in the following sections, state-of-the-art methods of differentiating hPSCs to CMs are still rather inefficient and are hard to reproduce. This part of the thesis therefore describes work that aims to generate a robust cardiac differentiation protocol to enable downstream investigations utilizing CMs.

The development of such a differentiation protocol hinges on the understanding of cardiac embryonic development as well as its molecular events in vivo. Therefore, much of the following literature review will focus on the different steps of cardiac embryogenesis. Once a robust differentiation protocol was developed, the next step was to optimize the protocol for generating single cell suspension of CMs (or “trypsinization”). An inefficient trypsinization protocol could lead to massive cell death and reduce CM yield for downstream experiments.

29

2.2.1. Embryonic heart development

There are four contiguous stages of cardiogenesis in mammals: (1) mesoderm formation,

(2) mesoderm patterning, (3) cardiac mesoderm formation and (4) cardiomyocyte maturation

(Willems et al., 2009). Cardiogenesis is a highly dynamic process that is tightly regulated by the

interactions of multiple signaling pathways, some of which will be briefly discussed below. Good

in-depth reviews on cardiac development have been published elsewhere (Abu-Issa and Kirby,

2007; Olson, 2006; Olson and Schneider, 2003)

Mesoderm formation

Cardiogenesis in mammals starts prior to gastrulation, whereby a number of cells destined to be heart cells are scattered throughout the pre-gastrula epiblast (Guzzo et al., 2007).

Gastrulation begins with the signaling of the transforming growth factor β (TGF-β) family member NODAL signaling in the proximal epiblast, which keeps BMP4 expression in the adjacent extra-embryonic ectoderm, which then in turn induces canonical Wnt signaling in the proximal epiblast (Burridge et al., 2012). Soon after, Wnt expression gets restricted to the posterior epiblast only, due to the expression of Wnt antagonist Dkk1 and the NODAL antagonists LEFTY1 and CER1 (CER1 is also Wnt and BMP signaling inhibitor) in the anterior visceral endoderm. Cells destined to be heart cells then ingress and migrate through the Primitive Streak (PS) antero- laterally, going through epithelial-mesenchymal transition (EMT) to form mesodermal cells expressing markers such as T (Brachyury) and Mixl1. Early mesodermal precursor cells migrate through PS from posterior region of low activin and high BMPs towards anterior region of low

BMPs and high activin (Sumi et al., 2008). Just before migration however, these cells transiently activate the basic helix-loop-helix transcription factor mesoderm posterior 1 (Mesp1) to enter cardiac mesodermal lineage (Bondue et al., 2011; Wu, 2008). Mesp1 has been termed the

30

“master regulator” of cardiac lineage specification, more specifically in cardiac progenitor

specification or cardiovascular lineage commitment (Bondue and al, 2008; Bondue and Blanpain,

2010).

Mesoderm Patterning and Cardiac Mesoderm Formation

Once the cells reach the anterior position, they have become fully committed to the cardiac lineage (Sumi et al., 2008). Mesp1 triggers the expression of cardiac mesodermal markers which include Nkx2-5, Tbx5, Hand1/2, Gata4, Mef2c, Myocd, Foxh1 and Isl1 (Bondue and Blanpain, 2010; Dierickx et al., 2012). In mice, Nkx2-5 is important for the interpretation of patterning signals within the primitive heart tube and is likely to act in tandem with Tbx5 during formation of both the atrial and left ventricular compartments (Mummery et al., 2012). The cardiac mesoderm gives rise to the first heart field (FHF), second heart field (SHF), endocardium, and the pro-epicardial mesenchyme. The differentiation of FHF proceeds ahead of the differentiation of SHF, a process regulated by Bmp4. Meanwhile, SHF is still proliferative and

regulated by hedgehog and canonical Wnt signaling (down-regulation) (Dierickx et al., 2012).

FGF and Notch signaling come into play once heart fields have been established, controlling

proliferation and fate determination of cardiac progenitor cells (CPCs) (Dierickx et al., 2012).

Specifically, Notch induces the expression of Wnt5a, Bmp6, and Sfrp1, which altogether serve to

increase the proliferation of CPCs (Chen et al., 2008).

Cardiomyocytes terminal differentiation

A number of pathways have been reported to regulate CM terminal differentiation during cardiogenesis. Firstly, Hedgehog signaling and canonical Wnt signaling are down- regulated, thus stopping proliferation of CPCs and initiating their differentiation (Dierickx et al.,

2012). In addition, down-regulation of canonical Wnt signaling induces Wnt11 expression, thus

31

switching on non-canonical Wnt signaling (Dierickx et al., 2012). BMP signaling is also needed for

terminal differentiation by maintaining Tbx20 expression (Shi et al., 2000; Walters et al., 2001).

Nkx2-5 and Tbx5 associate with members of GATA family of zinc finger transcription factors

(GATA4/5/6) and with serum response factor (SRF) to activate cardiac structural genes, such as actin, myosin light chain, myosin heavy chain (MHC), troponins and desmins (Mummery et al.,

2012). Tbx5 also cooperates with Nkx2-5 to activate expression of atrial natriuretic factor and

the junctional protein connexin 40 (Mummery et al., 2012). In addition, SRF regulates the

expression of microRNAs essential in heart development such as miR-1 and miR-133 (Chen et al.,

2006; Zhao et al., 2005). The P38-MAPK pathway regulates the activation of myocyte enhancer factor 2c (Mef2c), which in turn regulates cardiac muscle structural genes (Mummery et al.,

2012). Altogether, these different regulatory factors and pathways direct the differentiation of

CPCs to neonatal stage CMs. These cells are still somewhat proliferative and the process is modulated by factors such as IGF1, activating the PI3K/AKT signaling pathway (Evans-Anderson et al., 2008). The maturation of CMs beyond this point, however, has not been fully elucidated.

In conclusion, cardiogenesis in the embryo is governed by precise spatial and temporal regulations, as well as interactions of four main signaling pathways: nodal/activin/TGFβ, Wnt,

BMP, and FGF. In vitro differentiation from hPSCs to CMs is thought to follow similar trends as in vivo. Figure 2.2.1-1 below summarizes the current knowledge of in vitro cardiac differentiation.

32

.

adapted with permissions (Burridge et al., 2012) Image Image Schematic of current knowledge of factors involved in hPSCs to CMs to hPSCs involveddifferentiation factors of current of knowledge in Schematic

1 - 2.2.1 Figure

33

2.2.2. Differentiation of hPSCs to CMs

There are many published protocols describing methods to differentiate hPSC to CMs.

Some methods are using two-dimensional (monolayer) technique while some others are using three-dimensional (EB) technique. Regardless of the different technical details, all protocols have four similar stages: (1) maintenance of PSCs; (2) differentiation of PSCs to mesodermal cells;

(3) cardiac lineage specification; and (4) cardiomyocytes proliferation and maturation.

Differences lie in the techniques in generating the EBs, growth factors/inhibitor cocktail and basal media used as well as their time of application and/or removal. Outlined below are some of the more common methods used in literature.

Co-culture with endoderm

This method attempts to recapitulate the in vivo differentiation environment and was

first described by Mummery et al. (2003) with subsequent improvement by Passier et al. (2005).

HESCs and/or hiPSCs were co-cultured with mitomycin-C inactivated visceral endoderm-like cell

line (END-2) and contractile foci of beating cells were observed within 12 days of culture, with about 85% of CMs being ventricular-like in nature as determined through electrophysiological measurements. Although this method is relatively inefficient, it consistently induced cardiomyocyte (CM) differentiation, indicating the importance of factors secreted/depleted by adjacent anterior endoderm in cardiac lineage induction. This protocol served as a model to optimize later differentiation protocols. Some enhancements to the protocol involved removing serum from the differentiation medium, adding ascorbic acid (Passier et al., 2005), removing insulin (Freund et al., 2008), adding low concentration of p38 MAPK signaling inhibitor (Gaur et al., 2010; Graichen et al., 2008) and adding prostaglandin I (Xu et al., 2008). These modifications have increased the CM differentiation efficiency up to 25 %.

34

2D (monolayer) and 3D (EB)-based Differentiation Techniques

EB-based differentiation is the most widely-used method to differentiate hPSCs to CMs.

EBs are three-dimensional spheroids with an inner layer of ectoderm-like cells and a single outer layer of endoderm. They are named so to reflect their similarity to early post-implantation embryos. Cells in EBs differentiate into derivatives of the three primary germ layers: endoderm, mesoderm and ectoderm. Earliest generation of CMs from EBs was seen in mouse EBs generated from mouse embryonic stem cells (mESCs) and mouse embryonal carcinoma cells in

1985. Typically, these cells were dissociated into single cells and aggregated in hanging drops on the lids of petri dish to form EBs. Over time, a fraction of the EBs would develop spontaneously beating foci containing CMs (Doetschman et al., 1985).

First few efforts to replicate this method in the hESCs, however, showed that unlike mESCs, hESCs tend to die upon single-cell isolation, thus making human EB formation less straightforward. The earlier attempt of successful CMs differentiation were from EBs derived by collagenase IV treatment: hESC colonies were detached from feeder layer into clumps of 3 to 20 cells and grown in suspension on non-adherent petri dish (Kehat et al., 2001). In these EBs, cell- cell contacts induce mesodermal and early cardiac markers (Tran, 2009). Typically after a few days, the EBs were transferred to 0.1% coated culture dish for attachment and beating foci were observed 4 days post plating. Similar technique (spontaneously differentiating EBs) was also utilized to generate CMs from human induced pluripotent stem cells (hPSCs) (Gai, 2009; Nelson,

2009; Zhang et al., 2009). Despite the low efficiency of this protocol (lower than 1% CMs formation), electrophysiological analysis showed that all three subtypes of CMs were generated: atrial, nodal and ventricular CMs (He et al., 2003; Kehat et al., 2001; Xu et al., 2002). Other than the low differentiation efficiency, this technique is hampered by significant line-to-line

35

variability and low rate of reproducibility. This was caused by the use of undefined media

components, heterogeneous EB sizes as well as different culture conditions in different labs.

Improvements to this early protocol require better-defined culture conditions and generation of

EBs with uniform sizes.

Some groups tried to overcome the problem of EB size heterogeneity by forced aggregation methods in which defined numbers of enzymatically adapted hPSCs were seeded into low-attachment U- or V-well bottomed multi-well plates and aggregated with or without the aid of centrifugation (Burridge et al., 2007; Ng et al., 2005; Ungrin et al., 2008). In this method, EB sizes can be controlled by varying the number of cells seeded into each well. Stem

Cell Technologies combined this idea with silicon wafer based micro-fabrication technology to commercially produce AggreWells. This method, though attractive, is unfortunately still not

suitable for large scale production as it is rather expensive and time consuming. Other methods

to keep the EB sizes uniform include micro-well EBs (Khademhosseini et al., 2006; Mohr et al.,

2006; Mohr et al., 2010) and micro-patterned EBs (Bauwens et al., 2008; Peerani et al., 2007).

These methods focus on limiting the size of clonal growth of each hPSC colony such that upon

manual scraping of hPSC colony from the substrate, uniform size EBs would be generated.

However, these approaches necessitate the use of micro-engineering technologies that are not

publically accessible (Mummery et al., 2012).

Serum-free media have recently been preferred over serum-containing media since there is lot-to-lot variability in serum that might affect the reproducibility of a protocol. To circumvent this problem, some groups opt to use serum-free media such as APEL (Stem Cell

Technologies, Vancouver, BC, Canada) (Elliott et al., 2011), StemPro-34 (Invitrogen) (Yang et al.,

2008), homemade media such as the selenium-transferrin based media (Wei et al., 2012) or the

36

Burridge RPMI+PVA media (Burridge et al., 2011) (similar to APEL). Growth factors and/or

pathway inhibitors can be added at specific time-points into these serum free media to direct

the stochastic differentiation process towards the cardiac lineage. Some factors that have been

shown to improve the efficiency of differentiation include: (1) BMP4 addition from day 0-4

(Takei et al., 2009); (2) WNT3A or GSK3β-inhibitor addition for the first 24 (Lian et al., 2012) or

48 hours (Tran et al., 2009); (3) addition of inhibitors of Wnt signaling such as IWR1 from day 4-6

(Ren et al., 2011) or IWP2/4 from day 3-5 (Lian et al., 2012) or Dkk1 from day 3-5 (Paige et al.,

2010) and (4) addition of p38/MAPK inhibitor SB203580 from day 4-6 (Gaur et al., 2010).

Furthermore, factors such as G-CSF (Shimoji et al., 2010) and L-ascorbic acid (Cao et al., 2011) have been demonstrated to improve cardiac differentiation by enhancing CPCs proliferation whereas IGF and IGF2 were shown to enhance the hPSC-derived CM proliferation (Burridge et al., 2012; McDevitt et al., 2005). Different groups have used different combination of factors at different concentrations and time points in their protocol. Some protocols are discussed in greater detail below.

37

Group Primitive streak Mesoderm induction Cardiac specification and Efficiency formation terminal differentiation Wei, H. et Day 1-5: Day 5-13: ~22% al 1% Selenium Transferrin medium + 1% ST medium + P38-MAPK inhibitor β-MHC (+) (2012) P38-MAPK inhibitor SB 203580 (5 µM) (3D) SB 203580 (5 µM) Day 13 onwards: Plated on 0.1% gelatin coated plates in DMEM+2% FBS Yang et al Day 0-1: Days 1–4: Days 4–8: 40-50% (2008) StemPro 34 + StemPro34 + StemPro34+ CTNT+ cells (3D) BMP4 (10 ng/ml) BMP4 VEGF (10 ng/ml) + 5% CO2/5% O2 (10 ng/ml) + human DKK1 (150 ng/ml) /90% N2 bFGF (5 ng/ml) + activin A Days 8 onwards: (3 ng/ml) VEGF+ DKK1+ bFGF. 5% CO2/5% O2 5% CO2/5% O2/90% N2 (until day 12) /90% N2 Zhang, Q et Day 0-1: Day 1-2: Day 3-5: RPMI+B27+ about 73% al RPMI+B27+ RPMI+B27+ Noggin (250 ng/ml) CTNT+ cells (2011) BMP4 (25 ng/ml)+ Activin A (3D) bFGF (6 ng/ml) (100 ng/ml) Day 5-8: RPMI+B27+ DKK1 (200 ng/ml) + Retinoic acid inhibitor (1 µM)

Day 8 onwards:RPMI+B27+ DKK1 (200 ng/ml) Burridge, P. Day 0-2: Day 2-4: Day 4 onwards: ~ 91.2% et al RPMI + PVA ( 4 mg/ml)+ RPMI + RPMI + contracting (2011) 1-thioglycerol (400µM) + 1-thioglycerol 1-thioglycerol (400µM ) + EBs from all (3D) hr-insulin (10µg/ml) + (400µM)+ hr-insulin (10µg/ml) + EBs BMP4 (25ng/ml) + 20%FBS 1x chemically defined lipids generated FGF2 (5ng/ml) + 1x chemically defined lipids + 1µM Y27632 Elliott. D.A Day 0-2: Low Insulin-APEL + Day 3-onwards: ~96.4% et al Insulin (1µg/ml) + BMP4 (20-40ng/ml ) + LI-AEL without growth factors contracting (2011) Activin A (20ng/ml) + VEGF (30ng/ml )+ EBs from all (2D&3D) SCF (40ng/ml ) + WNT3a (50-80ng/ml) EBs generated

Table 2.2.2-1 Summary of published 3D based cardiac differentiation protocols

38

Group Primitive streak Mesoderm induction Cardiac specification and Efficiency formation terminal differentiation Laflamme, M.A. et al Day 0-1: RPMI+B27+ Day 1-5: RPMI+B27+ Day 6 onwards: ~30% (2007) activin A (100ng/ml) BMP4 (10ng/ml) RPMI + B27 (2D) Paige, S.L. et al Day 0-1: RPMI+B27+ Day1-5: Day 5-11: RPMI+B27+ ~48% (2010) Activin A (100 RPMI+B27+ DKK1 (200 ng/ml) sarcomeric MHC (+) (2D) ng/ml)+ BMP4 (10 ng/ml) Wnt3a (100 ng/ml) Lian, X. et al Day 0-1: Day 1-3: Day 3-5: ~85% (2012) RPMI+ RPMI+B27 (-insulin) RPMI+B27 (-insulin) + cTnT (+) (2D) B27 (-insulin) + IWP2 or IWP4 (5 µM) CHIR99021 (12µM) Day 5-7:

RPMI+B27 (-insulin)

Day 7 onwards: RPMI+B27 (+insulin) Zhang, J. et al Day (-2) - 0: Day1-5: Day 5 onwards: ~98% (2012) mTeSR1 + RPMI + B27(-insulin)+ RPMI+B27 (+insulin) cTnT (+) (2D) Matrigel overlay BMP4 (10ng/ml)+ bFGF (5 ng/ml) Day 0 - 1: RPMI+ B27 (-insulin) Matrigel overlay + Activin A (100 ng/ml)

Table 2.2.2-2 Summary of published monolayer based cardiac differentiation protocol

Although each study reported moderate to high differentiation efficiencies, reproducibility remains an issue in the field. Sometimes, a given protocol and cell line can yield significant differences from one experiment to the next in the same laboratory. This inconsistency is one of the major setbacks for advancement in the field. A robust and reproducible protocol that is applicable to the majority of cell lines is very much needed to fully realize hPSCs potential in this area of research.

39

2.2.3. Phenotypic Status of hPSC-Derived CMs

Immunostaining and action potential analysis reveal that hPSC-derived CMs are mostly

heterogeneous in nature: nodal-like, atrial-like and ventricular-like action potentials can all be

observed from these CMs (Mummery et al., 2003). The phenotypes of all three CM subtypes

correspond to their fetal or at most neonatal counterparts. It is interesting to see that different

protocols generate different proportions of CM subtypes. EB-based protocols generate almost

equal proportions of the three subtypes while co-culture protocols seem to generate a majority

of ventricular-like cells based on morphological and electrophysiological parameters (Mummery

et al., 2003; Rajala et al., 2011).

HPSC-derived CMs show automaticity and have fetal type: (1) ion channel expression

(Beqqali et al., 2006); (2) electrophysiological signals with slower action-potential upstroke and a relatively depolarized maximum diastolic potential (Brito-Martins et al., 2008; Davis et al., 2011;

He et al., 2003); (3) gene expression patterns and (4) physical phenotypes (Cao et al., 2008).

They express expected ionic currents, such as fast sodium current, L-type calcium current,

pacemaker currents, as well as transient outward and inward rectifier potassium currents (Brito-

Martins et al., 2008; Mummery et al., 2003; Rajala et al., 2011). Among the cardiac-specific

transcription factors expressed are GATA4, NKX2-5, ISL1, TBX5, TBX20, and MEF2C (Kehat et al.,

2001; Norström et al., 2006; Xu et al., 2002; Xu et al., 2009; Yoon et al., 2006). Structural

such as sarcomeric proteins α-actinin, cardiac troponins T and I, atrial- and ventricular

myosin light chains (MYL7 and MYL2); desmin, tropomyosin and gap junction proteins are also

expressed (He et al., 2003; Mummery et al., 2003; Xu et al., 2002; Xu et al., 2009). Similar to

adult CMs, contraction in hPSC-derived CMs is a caused by a rise in intracellular calcium

concentration (Mummery et al., 2003; Sartiani et al., 2007). However, due to the immature

40

sarcoplasmic reticulum in hPSC-derived CMs, intracellular calcium handling in these cells differs

from their adult counterparts. It has been postulated that hPSC-derived CMs lack the expression

of phospholamban and calsequestrin, which are two of the main intracellular calcium handling

proteins (Binah et al., 2007; Dolnikov et al., 2005; Sartiani et al., 2007). Also, unlike the

characteristic response of a typical mature cardiomyocyte whereby contraction amplitude

increases in response to drug-increased stimulation (i.e., the force-frequency relationship), hPSC-derived CMs have been reported to respond in the opposite manner and display negative force-frequency behavior (Kehat et al., 2002).

Morphologically, hPSC-derived CMs have less defined rod shape and display very low

frequency of multi-nucleation (<1%) compared to adult human CMs (20%) (Snir et al., 2003).

Ultra-structurally, hPSC-derived CMs show clearly identifiable sarcomeres with A, I, and Z bands

and intercalated discs with gap junctions and desmosomes, similar to the adult CMs, although

the myofibrillar and sarcomeric organization indicate an immature phenotype in the stem cell-

derived population (Kehat et al., 2001; Liu et al., 2007; Snir et al., 2003; Yoon et al., 2006).

Mitochondria are also present in these cells in the vicinity of the sarcomeres (Rajala et al., 2011).

The immaturity of hPSC-derived CMs provided a good platform to study CM maturation in vitro. Furthermore, since MYL2 has been reported to be a marker of mature ventricular CMs

(Franco et al., 1999; Kubalak et al., 1994; Segev et al., 2005), the MYL2 knock-in reporter line could be used to monitor in real time the maturation process of these ventricular CMs in vitro.

41

2.2.4. Cardiomyocyte Maturation in vitro

Culture environment plays an important role in CM maturation. Studies have shown 3-

dimensional (3D) culture in proteinaceous matrix to be superior to monolayer approach for

maturation of CMs (Tiburcy et al., 2011; Zhang et al., 2013). In comparison to monolayer culture,

the CMs in 3D culture: (1) appeared to be more rod-shaped (Tiburcy et al., 2011) and had longer

sarcomeres (Zhang et al., 2013); (2) showed significantly higher conduction velocities (Zhang et

al., 2013); (3) had higher frequency of bi-nucleation (Tiburcy et al., 2011); (4) expressed higher levels of genes expressed in adult CMs such as CASQ2, SERCA2 and the contractile genes cTnT,

αMHC, and MYL2 (Zhang et al., 2013); (5) expressed higher levels of adult CMs proteome, tropomyosin isoforms, desmin and M-type creatine kinase (Tiburcy et al., 2011); as well as (6) exhibited significant positive inotropy with isoproterenol treatment (Schaaf et al., 2011; Tiburcy et al., 2011). In addition, CMs cultured in 3D exhibited decreased expressions in ANP and CARP proteins and had higher responsiveness to triiodothyronine (T3) treatment, indicative of tissue maturation (Akins et al., 2010). As such, a three-dimensional culture system in the form of previously reported engineered heart muscles (EHMs) (Soong et al., 2012; Tiburcy et al., 2014) will be used for downstream study of maturation.

42

2.3. Materials and Methods

2.3.1. Differentiation of hESCs to immature CMs - Embryoid body (3D) approach

In-house protocol

CARM medium HESCs were scaled up with manual passaging onto Matrigel-coated DMEM (4.5g/L glucose) 1x Selenium Transferrin mix 2mM L-Glutamine 10cm dishes, making sure that each passaged fragment had similar 100µM NEAA 100µM 2-mercaptoethanol sizes. 5 days later, hESCs were treated with 1mg/ml Dispase at 370C Table 2.3.1-1 Composition of CARM medium for 6 minutes and washed 3 times in DMEM/F12 basal media. HESC colonies were dislodged carefully from culture surface using cell scraper and transferred to low attachment 6 well plates to form embryoid bodies (EBs) overnight in hESC medium without bFGF. On the next day (d1 of differentiation), the culture medium was changed to CARM + 5µM

SB203580. On day 5 of differentiation, the culture medium was changed to fresh CARM + 5µM

SB203580 and kept until day 9 when the culture medium was changed to CARM only.

Contraction of EBs was monitored from day 10 of differentiation onwards.

Burridge 2011 protocol

The published protocol is summarized in the following diagram:

Change medium every 3 days. Day Day 48 hours Day Day Day Once contracting, use RPMI + PVA (-1) (0) in 5% O2 (2) (4) (7) or RPMI + Insulin

Seed Dissociate hESCs Change Change Change Observe for 6 2x10 into single cells medium to medium to medium to beating from hESCs on with TrypLE RPMI + FBS RPMI + Insulin RPMI + day 9 T25 pre- Express and Transfer EBs to Insulin seed into V-well 96-well U coated 3 with 96 WP at 5x10 bottom tissue Matrigel hESCs/well in culture plate 100µl RPMI+PVA Figure 2.3.1-1 Burridge 2011 protocol 43

RPMI + PVA Medium RPMI + FBS RPMI 1640 RPMI 1640 400µM 1-thiglycerol 400µM 1-thioglycerol 4 mg/ml PVA 20% FBS 10µg/ml hr-insulin RPMI + Insulin 25ng/ml BMP4 RPMI 1640 5ng/ml FGF2 400µM 1-thiglycerol 1% chemically defined lipids 10µg/ml rh-insulin 1µM Y27632 1% chemically defined lipids

Table 2.3.1-2 Components of media used in Burridge protocol

KP 3D protocol (unpublished)

This protocol was first developed to improve differentiation efficiency of the then in house protocol. KP 3D protocol was optimized by experimenting on the differentiation media components, their concentrations and time of applications during differentiation. The final version of the protocol is shown below:

Day Day Day Day Day Change medium every 2 days. (-1) (0) (3) (6) (8) Once contracting, B27 (+) medium

Seed Dissociate hESCs Change Change Change 6 4x10 into single cells medium to medium to medium to hESCs on with 0.05% Phase 2 Phase 3 Phase 4 T25 pre- Trypsin/EDTA and medium medium medium, seed into V-well 96 Observe coated 4 with WP at 5x10 for beating GFR- hESCs/well in 100µl Phase1 medium Figure 2.3.1-2 KP 3D protocol (optimized)

Phase 1 Medium Phase2 Medium Phase4 Medium RPMI 1640 RPMI 1640 RPMI 1640 B27(-) 4mg/ml PVA 1x B27 minus insulin 1x B27 medium 200µM Ascorbic acid 1x Penicillin/Streptomycin 1x Penicillin/Streptomycin 1mM Sodium Pyruvate 200µM Ascorbic acid 5µM IWR1 400µM 1-thiglycerol 5µM SB203580 5µM SB203580 1µg/ml rh-insulin 5µM IWR1 2µM BIO 20ng/ml BMP4 Phase3 Medium B27(+) medium 50ng/ml Activin A Phase 2 medium RPMI 1640 20ng/ml bFGF 1µM BMS453 1x B27 1% chemically defined lipids 1x Penicillin/Streptomycin 2µM Y27632

Table 2.3.1-3 Media composition used in KP 3D protocol 44

2.3.2. Differentiation of hESCs to immature CMs - Monolayer (2D) approach

Lian 2012 protocol

The published protocol is summarized in the timeline below:

Figure 2.3.2-1 Lian 2012 protocol timeline

KP 2D protocol (unpublished)

This protocol is a derivation of KP 3D protocol. Each version is exactly the same as the

3D counterpart, in terms of differentiation media compositions and time-points of media change. The only difference in the 3D and 2D protocols was that the phase 1 to 4 differentiation media used in 3D protocol had to be diluted 1: 4 to be used in 2D differentiation protocol.

Hudson (in press, 2014) protocol

This protocol is outlined below:

Figure 2.3.2-2 Hudson 2014 protocol

45

Base Medium Mesodermal Medium Cardiac differentiation RPMI 1640 with Glutamax Base medium Medium 200µM Ascorbic acid 1µM CHIR Base medium 1mM Sodium Pyruvate 5 ng/ml BMP4 5µM IWP4 1x Penicillin/Streptomycin 9 ng/ml Activin A 1x B27 5 ng/ml bFGF

Table 2.3.2-1 Media composition used in Hudson protocol

2.3.3. Generation of EHMs

Generation of EHMs molds

Shown on figure 2.3.3-1 is the technical drawing of small circular molds used to cast

EHMs. Briefly, polydimethylsiloxane (PDMS, or silicone), was poured into a 6cm diameter glass petri dish to 0.5cm depth and allowed to crosslink at room temperature overnight. On the next

day, 4-polytetrafluoraethylene (Teflon®) molds in the form of donuts were arranged on top of

the semi cross-linked PDMS. Each Teflon® donut was put together with a straw insert and PDMS stick inside the donut and more PDMS was poured onto the glass dish to reach the upper surface of the Teflon® donut. The PDMS was let to crosslink at room temperature to get rid of small bubbles overnight before being baked at 600C oven for 3 hours.

Generation of auxotonic stretchers

Auxotonic stretchers were made from PDMS by injecting into pre-made Teflon® molds as shown on figure 2.3.3-2. The PDMS was allowed to crosslink overnight at room temperature before being baked at 600C oven for 3 hours the next day. After the PDMS was fully cross-linked, the resulting stretcher can be extracted from the mold.

46

Figure 2.3.3-1 Technical drawing of EHM molds for casting (Tiburcy et al., 2014)

Figure 2.3.3-2 Auxotonic stretchers A – Teflon® casting molds for auxotonic stretcher B – Technical drawing of auxotonic stretcher

47

Cardiomyocytes dissociation to single cell suspension

A few methods were employed: (1) collagenase I (with or without Y27632) followed by

“Accutase mix”; (2) collagenase I followed by trypsin in differing concentration; (3) “Accutase mix” with prior PBS wash – PBS composition was varied in this case. Collagenase I (Sigma) was reconstituted to a final concentration of 2mg/ml in 20% FBS in PBS with calcium and magnesium.

“Accutase mix” was really Accutase mixed with trypsin/EDTA diluted to a final concentration of

0.125% and DNAse I diluted to a final concentration of 20µg/ml.

Method 1 and 2: collagenase I (with or without Y27632) was added to cardiomyocytes

and incubated at 370C for 30 minutes. Y27632 is a cell-permeable small molecule Rho-associated

kinase (ROCK) inhibitor (Ishizaki et al., 2000). Cardiomyocytes (CMs) were washed with PBS (+) once, triturated to break down to small clumps and transferred to a 50ml Falcon tube. Culture flask was washed again with PBS (+). CMs were then centrifuged at 1000rpm for 4 minutes and the pellet resuspended and incubated for 5-10 minutes at 370C in pre-warmed “Accutase mix”

or differing concentration of trypsin to produce single cell suspension. Trypsin was deactivated

by addition of FBS-containing media.

Method 3: CMs were incubated in either PBS (-) or PBS (+) for 5 minutes at room

temperature before incubating with “Accutase mix” for 5-10 minutes at room temperature. PBS

(+) contains 2mg/L CaCl2, 2mg/L MgCl2-6H2O, 4mg/L KCl, 4mg/L KH2PO4, 160 mg/L NaCl,

43.2mg/L Na2HPO4-7H2O, 20mg/L D-Glucose and 0.72mg/L sodium pyruvate. Enzymes were deactivated either by FBS-containing media or by diluting in PBS (+).

48

Generation of engineered heart muscles (EHMs)

Master mix components for Volume 4x small EHMS (µl) SFBM medium SFMM medium Bovine Collagen I (5.83 mg/ml) 322 485ml Iscove 48ml SFBM 2x DMEM 322 45mg L-Ascorbic acid 2-phosphate 2ml B27 0.1 M NaOH 52 sesquimagnesium salt hydrate 100ng/ml IGF-1 HFF = 1.753x106 cells 1408 1x NEAA 5ng/ml VEGF Cardiomyocytes = 5.26x106 1x Penicillin/Streptomycin 10ng/ml FGF Total volume of master mix 2104 2mM L-Glutamine 5ng/ml TGF-β1** Table 2.3.3-1 Composition of EHM master mix Table 2.3.3-2 Media composition in EHM casting and culture ** only added when EHM rings are still in molds

EHMs are cardiac muscle tissue-mimetic constructs generated by growing CMs and human foreskin fibroblasts (HFFs) in a collagen matrix. The cell-in-matrix suspension was cast in a ring format using the above mold to make EHM rings. Table 2.3.3-1 lists the composition of master mix for every 4 molds in 6 cm glass beaker. Typically, each EHM would contain 0.4 mg of collagen and a total of 1.5x106 cells (CMs + HFFs mixture). All components for master mix

(except CMs and HFFs) were kept on ice to facilitate uniform ring formation. 450 µl of master

0 mix was pipetted into each mold and allowed to condense in CO2 incubator at 37 C for 1 hour.

Subsequently, 6 ml of pre-warmed SFMM medium was added into the dish. Media was changed the next day and every other day subsequently. Figure 2.3.3-3 shows the process of EHMs condensation inside the mold for the first 3 days post-casting. Once fully condensed, EHM rings were transferred manually using pipet tips to pre-sterilized auxotonic stretchers. Figure 2.3.3-4 shows an EHM ring cultured on such stretcher.

49

Figure 2.3.3-3 EHMs formation (Tiburcy et al., 2014)

Figure 2.3.3-4 EHM ring on auxotonic stretcher. Scale bar = 5mm

2.3.4. Measurements of EHM contractile forces

EHM rings force measurement

Figure 2.3.4-1 EHM ring stretched in organ bath of contraction setup 50

Isometric forces were measured inside an organ bath supplied by Föhr Medical

Instruments GmbH (FMI) under a fixed frequency (1.5-2.5Hz) and 200mA field stimulation in modified tyrode solution. Tyrode solution contains: 120mM NaCl, 5.36mM KCl, 0.2mM CaCl2,

1.05mM MgCl2, 22.62mM NaHCO3, 0.42mM NaH2PO4, 1 g/L glucose and 0.1g/L ascorbic acid.

Briefly, EHM rings were stretched gradually to increase their preload according to Frank-Starling

law in presence of physiological 1.8mM Ca2+. The Frank starling law describes the positive

relationship of cardiac muscle stretch and its contractility, up to an optimum length (Iaizzo,

2009). Thus stretching was stopped once the max-min plot showed plateau phase, which is

indicative of the sarcomeric optimal length being reached. The EHMs were washed with fresh

tyrode solution to bring down the calcium concentration back to 0.2mM.

The calcium response experiment was carried out at Frank Starling length to assess

cardiomyocytes calcium handling capacity. EHMs were stimulated at 200 mA and 1.5 Hz and

their forces at increasing calcium concentration (0.2 mM – 4.0 mM) were recorded. Stimulation

was stopped after the contractile force at 4.0mM calcium was recorded and the EHMs were

washed with fresh tyrode. To measure the force frequency behavior, EHMs were stimulated at

constant current of 200 mA at their Frank-Starling length and in presence of 1.2 mM Ca2+. The

stimulation frequency was changed from 1 Hz to 3 Hz at interval of 5 minutes per frequency. The forces generated by EHMs were recorded.

To assess the responses to β-adrenergic stimulation, EHMs were stimulated at 200 mA

current, 1.5 Hz frequency in presence of 0.8 mM Ca2+. Isoprenaline was then added to the organ

bath at a final concentration of 1 µM. Following washing step with fresh tyrode, a muscarinic

receptor agonist, carbachol, was added to a final concentration of 10 µM.

51

2.3.5. Automation of casting and transfer

The labor intensive nature of EHM casting and transfer prevented the system to be used

for high throughput analysis. Since high throughput analysis was necessary for maturation studies of the project, we turned to 3D-printing technology to automate molds and stretchers production.

Printing casting plates

Casting plate was printed with Objet 3D printer from Stratasys, using tango black material for the poles and veroclear for the rest. The plate was washed for a week in water and then sterilized by autoclaving before use. The V-bottom design around the poles allowed for automatic EHM ring transfer onto the poles for auxotonic stretching.

2.3.6. Statistical analysis

Statistical significance was calculated using Prism 5.0 (Graphpad) with the student t-test or

ANOVA when applicable. Bonferroni's Multiple Comparison Test comparing all groups was used as post-hoc test when ANOVA was employed.

52

casting plate plate casting

(Courtesy of Dr. Tim Meyer) Technical drawing of printed EHM

1 - 2.3.6 Figure

53

2.4. Results

2.4.1. Optimization of EB-based cardiac differentiation protocol

In house protocol

The in-house SB protocol forms EBs at high frequency of

success. However, there were a number of challenges: (1)

the EBs formed were of different sizes (figure 2.4.1-1), (2)

the EBs tend to coalesce to form bigger EBs and (3) the

500µM Figure 2.4.1-1 EBs generated by in house differentiation efficiency was low (about 10 - 20 % protocol have variable sizes beating EBs by day 21 of differentiation).

Burridge 2011 protocol

Reproducing Burridge 2011 protocol was found empirically to be challenging due to a number of factors:

1. The protocol required enzymatic cell passaging to single cells on a strict timescale of every 3 days using feeder cells-conditioned media. All hPSCs lines used in this study (MBJ5, Hes3, and H9) differentiated spontaneously when cultured in conditioned media. 2. The protocol required day (0) to day (2) of differentiation to be carried out in a hypoxic chamber, which is not available in this laboratory. 3. Extensive cell death was observed during EB formation step, resulting in non-uniform EB sizes 4. EB formation step was not always successful for yet unknown reasons 5. Strictly following the published protocol resulted in cystic EBs. Unfortunately, these EBs were not beating even with extended period of culture (≥ 20 days).

Shown in figure 2.4.1-2 are images of EBs derived with Burridge protocol, with different

starting number of hPSCs.

54

No of cells/ Day 4 Day 10 Day 13 well

5000

10000

25000

50000

Figure 2.4.1-2 The Burridge protocol allows the generation of uniformly-sized EBs, depending on the number of cells seeded per well. Scale = 500µm

55

KP 3D protocol

A new protocol, (KP protocol) was developed to improve the then in house differentiation protocol efficiency. The protocol was first conceptualized by combining the

Burridge and Lian protocols and adapting it to a 3D format. To form EBs with similar sizes,

Burridge’s forced aggregation method was employed. Table 2.4.1-1 shows the diameters of EBs formed from different numbers of cells seeded per well using the Burridge and KP protocols on day 5 of differentiation. Similar to the Burridge protocol, the first iteration of the KP protocol resulted in morphologically cystic EBs, yet no beating was observed. Version 9 is the most robust version of the protocol in generating beating EBs as well as bioengineered heart muscles (BHMs).

BHMs are the resulting tissues when hESCs were directly differentiated in collagen matrix.

Number of cells Burridge’s protocol/ µm KP’s protocol /µm per well (Mean ± SD) (Mean ± SD) 5000 283.14 ± 36.86 87.20 ± 8.92 10000 232.26 ± 37.04 118.29 ± 11.85 25000 450.69 ± 46.56 370.18 ± 45.29 50000 661.38 ± 35.81 540.23 ± 22.31 Table 2.4.1-1 Comparison of EB sizes from the Burridge and KP protocol

Table 2.4.1-2 summarizes the different versions of the KP 3D protocol. Over 9 iterations, the protocol improved cardiac differentiation efficiencies from 0% - 100%. Differentiation efficiency was measured with the following formula

= 100 Efficiency (%) Number of beating EBs Total number of EBs ∗

56

Days 0 1 2 3 4 5 6 7 8 9 10 >10

2 µM CHIR + 400µM 1- 400µM 1-thiglycerol + 10µg/ml rh-insulin thiglycerol + 4 mg/ml PVA + Day 0 400µM 1-thiglycerol + 400µM 1-thiglycerol + 1% chemically defined lipids V1 10µg/ml hr-insulin + 25 ng/ml medium + 10µg/ml hr-insulin + 1% 10µg/ml rh-insulin

BMP4 + 5ng/ml FGF2 + 1% 5 µM chemically defined 1% chemically defined lipids Efficiency: chemically defined lipids + SB203580 lipids + 5 µM IWP2 0% 2µM Y27632

5 µM BIO + 400µM 1- 400µM 1-thiglycerol 400µM 1-thiglycerol + 10µg/ml rh-insulin thiglycerol V2 10µg/ml rh-insulin 1% chemically defined lipids + 150 ng/ml DKK1 4 mg/ml PVA + RPMI only 1% chemically defined lipids + 1% chemically defined lipids + 150 ng/ml DKK1 + 5 µM SB203580 Efficiency: No EBs 2µM Y27632

D7-10 medium 2 µM BIO + 400µM 1-thiglycerol 400µM 1-thiglycerol without V3 4 mg/ml PVA + 10µg/ml rh-insulin 2% B27 (-) + 150 ng/ml DKK1 10µg/ml rh-insulin DKK1 40 ng/ml BMP4 + 5ng/ml FGF2 + 40 ng/ml Activin A + 5µM SB 203580 1% chemically defined lipids +

1% chemically defined lipids + 2µM Y27632 150 ng/ml DKK1 Efficiency: 6.25%

D(-1): matrigel overlay + 2µM 2% B27 (-) + 10 ng/ml BMP4 + V4 2% B27 (+) BIO 5 ng/ml bFGF + 25 - 100 ng/ml 2% B27 (-) + 150 ng/ml DKK1 + 5 µM SB203580 Efficiency: ≤1.7% Scrape colonies to make EBs Activin A

2 µM BIO + 400µM 1-thiglycerol 2% B27 (-) V5 4 mg/ml PVA + 1 µg/ml rh-insulin 2% B27 (-) + 2 – 5 µM IWP4 + 5 µM SB Efficiency: 20 ng/ml BMP4 + 5ng/ml FGF2 + 40 ng/ml Activin A 203580 H9: ≤46.7% 1% chemically defined lipids + 2µM Y27632 HES3: ≤100%

57

Days 0 1 2 3 4 5 6 7 8 9 10 >10 Efficiency

2% B27 (-)

Efficiency: 2µM IWP4 50% 2 – 5 µM BIO + 400µM 1-thiglycerol 5µM BIO 25nM C59 33% 2% B27 (-) + 4 mg/ml PVA + 1 µg/ml rh-insulin 50nM C59 16.7% 6 2 µM IWP4 or 25 – 50 nM C59 + 5 µM SB 20 ng/ml BMP4 + 5ng/ml FGF2 + 40 ng/ml Activin A 50nM C59 62.5% 203580 1% chemically defined lipids + 1µM Y27632 3µM BIO 2µM IWP4 5µM SB 80% 25nM C59 100% 2µM IWP4 100% 2 µM BIO 25nM C59 100% 50nM C59 85.7%

2% B27 (-)

Efficiency: 2 µM BIO + 400µM 1-thiglycerol 2% B27 (-) + 2µM IWP4 86.7% 4 mg/ml PVA + 1 µg/ml rh-insulin 2 – 5 µM IWP4 or 5µM SB V7 5µM IWP4 76.7% 20 ng/ml BMP4 + 5ng/ml FGF2 + 50 ng/ml Activin A 25 – 50 nM C59 + - 66.7% 2 µM BIO 25nM C59 1% chemically defined lipids + 1µM Y27632 with or without 5 µM SB 203580 5µM SB 96% - 10.34% 50nM C59 5µM SB 96.7%

2 µM BIO + 400µM 1-thiglycerol 2% B27 (+) 4 mg/ml PVA + 1 µg/ml rh-insulin 2% B27 (-) + V8 20 ng/ml BMP4 + 5ng/ml FGF2 + 50 ng/ml Activin A 2% B27 (-) 25 nM C59 + 5 µM SB 203580 Efficiency: 1% chemically defined lipids + 1µM Y27632 + 98.6% 1mM Sodium pyruvate

2 µM BIO + 400µM 1-thiglycerol 2% B27 (-) + 4 mg/ml PVA + 1 µg/ml rh-insulin 2% B27 (+) + 2% B27 (+) 5 µM IWR1 + 20 ng/ml BMP4 + 20 ng/ml FGF2 + 2% B27 (-) + 5 µM IWR1 + 5 µM SB 203580 5 µM IWR1 + V9 5 µM SB 203580 50 ng/ml Activin A + 200 µM Ascorbic acid 5 µM SB Efficiency: + 200 µM Ascorbic acid 1% chemically defined lipids + 2 µM Y27632 + 203580 100% + 1 µM BMS453 1mM Sodium pyruvate + 200 µM Ascorbic acid

Table 2.4.1-2 Different versions of KP 3D protocol. Version 8-9 was validated for more than 5 times in independent experiments per version.

58

2.4.2. Optimization of 2D differentiation protocol

In the initial phases of the study reported in this thesis (before the optimization of CMs trypsinization protocol), quantification of differentiation efficiency in 2D format was challenging in this laboratory due to a number of reasons: FACS analysis of alpha-actinin staining required cardiomyocytes monolayer to be trpysinized prior to fixing and staining. This trypsinization step was found to be detrimental to cells, killing almost 50%-70% of cardiomyocytes. This high amount of cell death subsequently affected the alpha-actinin staining result. Therefore, analysis of beating efficiency was done by qualitative observation. Video clips of contracting CM monolayer are included in a CD accompanying this thesis.

Lian 2012 protocol

In the Lian protocol, extensive cell death occurred after the addition of GSK3 inhibitor (BIO or CHIR). Strictly following the published protocol did not generate contractile CMs. Thus the

Lian protocol was optimized in this laboratory; the optimization steps taken were:

1. Varying hPSCs culture density 4 days before the start of differentiation (d(-4)) 2. Varying length of CHIR treatment on d0 - d1 of differentiation (from 20 – 24 hours) 3. Varying concentration of CHIR and BIO 4. Varying concentration of IWP4 on d3 of differentiation

The following parameters were determined empirically to give the most efficient differentiation using Lian protocol, leading to the generation of beating cells in all wells of NUNC

12 well plates:

1. Cell density of 7.5x105 H9 hESCs per well of 12 well plate on d(-4) 2. Treatment length of strictly 22 hours with 12µM CHIR 3. Adding a 1:1 mixture of spent and fresh B27(-) medium, with final concentration of 5µM IWP4

59

Although the Lian protocol was optimized, having to differentiate in 12 well plates was time consuming, especially when differentiating large numbers of CMs (in the order of 106) at a time. There was a need to scale up the system. Adapting the Lian protocol to T25 and T75 tissue culture flasks required further rounds of optimization with regards to cell seeding density. It was noted that different hESCs lines (H9, H9V15 and H9VC16) required different cell density on day 4 before the start of differentiation (d(-4)). H9 and H9V15 required 1.6x105 cells/cm2; while

H9VC16 required 1.6x105 –2.2x105 cells/cm2.

Control Ms IgG Alpha-actinin 0.7% 87.4%

Table 2.4.2-1 The Lian protocol generates >80% α-actinin (+) cardiomyocytes

The Lian protocol was used as the main 2D differentiation protocol as it routinely generated >80% alpha-actinin positive CMs (table 2.4.2-1). However, for unknown reasons, the differentiation efficiency was suddenly lost after a few good rounds of differentiation. It was hypothesized that trypsin-passaging of hESCs might have changed the characteristics of hESCs over time thus reducing the efficiency of the Lian differentiation protocol. This prompted the switch to the Hudson protocol, as well as parallel optimization of KP 2D protocol.

60

Hudson protocol

The Hudson protocol (Hudson et al, in press) uses trypsin-passaged hESCs cultured on

HFFs feeder cells for maintenance. This protocol was the main protocol used in the Zimmermann lab in Göttingen for CMs differentiation from hESCs. H9 and H9V15 were more adaptable to trypsin passaging and culture on HFFs. H9VC16, however, was less adaptable and adaptation process to HFFs feeder frequently resulted in spontaneous differentiation. This hampered immediate utilization of the Hudson protocol. When healthy looking colonies were finally obtained on HFFs, strictly following the protocol did not generate beating cells. The protocol had to be optimized.

The first modification of the Hudson protocol involved using hESCs cultured on MEFs instead of on HFFs but this modification did not generate any contractile CMs. A different culture medium was then used to culture hESCs from the 4th day prior to differentiation to the day just before differentiation (d(-4) to d(-1)) and cell density on d(-1) was optimized further. The cell density for H9 and H9V15 was most optimal at 4x106 cells/T25 flask and H9VC16 at 4.5x106 cells/T25 flask. Furthermore, hESCs were best cultured on MEFs in normal hESCs medium instead of Matrigel in mTeSR1 medium prior to differentiation. In other words, the d(-4) to d(-1) step of the Hudson protocol had to be omitted.

Following these modifications, the Hudson protocol proved to be robust, consistently generating CMs. However, differentiation efficiencies also dropped as the hESCs passage number increased, similar to the observation with the Lian protocol. It was found that over time, later passages of hESCs required higher cell density for previously observed differentiation efficiencies to be achieved. H9VC16 of passage 13 and above required 1.6x107 cells/T75 flask in comparison with earlier passages (1.35 - 1.4x107/T75).

61

KP 2D protocol

This protocol was optimized in parallel to hESCs trypsin adaptation to HFFs feeder for the Hudson protocol. Extensive cell death was observed with the use of the KP 3D protocol

Phase 1 medium (table 2.4.1-3) for monolayer-based differentiation. Thus different dilutions of phase 1 medium were tested. Empirically, 4x dilutions of version 8 and 9 of KP 3D phase 1 medium was found to be most optimal for monolayer-based differentiation on T25 tissue culture flasks (>80% success of all differentiation efforts). Phase 1 medium from version 8 of KP 3D protocol was diluted in B27 (-) medium + 1mM sodium pyruvate without ascorbic acid whereas

Phase 1 medium from version 9 was diluted in B27 (-) + 1mM sodium pyruvate + 200µM ascorbic acid. Beating was routinely observed from day 8-10 of differentiation.

Control Ms IgG Alpha-actinin

0.2% 76.0%

Table 2.4.2-2 The Hudson protocol generates >75% α-actinin (+) cardiomyocytes

Control Ms IgG Alpha-actinin

6.3% 91.8%

Table 2.4.2-3 The KP 2D protocol generates >90% α-actinin (+) cardiomyocytes

62

2.4.3. Engineered Heart Muscles

Trypsinization of cardiomyocytes

CM trypsinization to generate single cell suspension is important for downstream experiments. Suboptimal trypsinization technique could result in loss of up to 70% CMs. First attempts to digest CMs involved the use of collagenase I for 30 minutes at 370C followed by trypsinization with 0.25% trypsin/EDTA for 10 minutes at 370C. This resulted in only 25.7% viable

CMs.

10µM Y27632, a rho-associated protein kinase (ROCK) signaling pathway inhibitor, was then added into collagenase I; and the effective trypsin concentration in “Accutase mix” (AM) was reduced from 0.25% to 0.125%. These changes almost doubled the viability of trpysinized

CMs to 46.99% ± 5.2% (mean ± SD), as can be seen from bar A in figure 2.4.3-1. Further decreasing effective trypsin (T) concentration from 0.125% to 0.05% further improved the viability of trpysinized CMs to 65.09% ± 16.1% (mean ± SD).

It was observed that collagenase I digestion step could be eliminated thus reducing trypsinization process from 45 to 20 minutes. Bar C and D in figure 2.4.3-1 show the viability of trpysinized CMs when only AM containing 0.05% Trypsin/EDTA + 20µg/ml DNAse I was used.

Methods C and D differed in the PBS used for washing and the medium used to deactivate the digestion enzymes. Method C used PBS (-) to wash the cells whereas D used PBS (+); and C used

FBS-containing media to deactivate the enzymes while D used serum-free PBS (+) solution.

Method D was the most optimal, generating 70.13% ± 11.9% (mean ± SD) live CMs post- digestion. However, the difference was not statistically significant to the results obtained from method C.

63

Figure 2.4.3-1 CM viability post trypsinization was improved over 4 iterations

Effect of composition of EHMs

The cardiomyocytes (CMs) were then tested for their ability to self-assemble in a collagen matrix to form contractile tissue constructs: the engineered heart muscles (EHMs).

Different ratios of CMs to HFFs (100%: 0% to 50%: 50%) were tested to investigate the optimal ratio for highest force generation. EHMs containing 90% and 100% CMs did not condense properly to form compact tissues; these EHMs did not contract and were excluded from contraction experiment.

EHMs condensation, i.e. aggregation of the cells and establishment of cell-cell- connections within the collagen matrix, is a prerequisite for EHM contractile force generation. A minimum of 25% HFFs was empirically determined to be critical for EHMs condensation.

Comparing EHMs containing 30% HFF and 50% HFF, 30%-HFF EHMs significantly (p<0.05, n=4)

64

beat with higher contractile forces: 0.226 ± 0.045mN vs 0.093± 0.018 mN (mean ± SEM). Thus

70-75% CMs: 30-25% HFF ratio was employed throughout the project.

* * * * * * * 0.1 50% HFF 30% HFF n = 4 per group 0.01 Contraction force (mN) 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 Ca2+ (mM)

Figure 2.4.3-2 Forces comparison between 30%HFF-EHMs and 50%HFF-EHMs

65

2.4.4. Printed plate parameters and handling affects functionality of EHMs

A number of parameters were optimized for printed plate: (1) shape of wells (round and oval); (2) materials for poles; (3) diameter of poles; (4) cleaning protocol and; (5) volume of master mix. Beating EHMs were obtained successfully when plate was made from veroclear and poles from tango black. Poles with 1.3mm diameter allowed better visualization of EHM contraction than 1.5mm (1.5 mm is the diameter of auxotonic poles). Cleaning procedure prior to plate sterilization was very critical to the success of experiment. Immersing in water for at least two weeks, with agitation and daily water change, followed by ethanol and UV sterilization was optimal in generating healthy EHMs post casting. Finally, only a max of 200µl of master mix should be cast per well to prevent the two arms of the EHMs rings from fusing.

n = 4 per group

Figure 2.4.4-1 Forces of EHMs casted on round and oval shaped printed wells

As evident from figure 2.4.4-1, EHMs casted in oval shaped wells (n=4) gave EHMs which beat with higher force than those casted in round shaped wells. The significance of this difference, however, could not be calculated as EHMs casted on round shaped wells snapped during force measurement. 66

2.5. Discussion

3-dimensional approach to cardiac differentiation

HESCs can be differentiated towards CMs using 3D (EB) or 2D (monolayer) approach. The

EB approach necessitates the formation of uniformly sized EBs. The conventional in house EB formation protocol generated non-uniform EB sizes which tend to coalesce to form bigger EBs.

Furthermore, this protocol resulted in low cardiac differentiation efficiency: only 10% – 20% of all differentiated EBs were beating by day 21 of differentiation.

The Burridge 2011 protocol appeared to offer a simple and efficient approach to make uniform-sized EBs via the forced aggregation method. However, the protocol did not generate beating EBs. In addition, extensive cell death was observed during EB formation step, resulting ultimately in non-uniform EBs. Initially, it was thought that the high concentration of insulin at

10 µg/ml in the RPMI+PVA medium might prevent efficient cardiac differentiation. It was observed previously from the in-house protocol that EBs grown in medium containing 20%

KnockOut™ serum replacement (KOSR) failed to differentiate towards CMs. KOSR has a high level of insulin at approximately 100 mg/l, resulting in an approximately 20 µg/ml in 20% solution (Odorico et al., 2004). Thus insulin was removed from RPMI+PVA medium and the

Burridge protocol was repeated. EBs failed to form in absence of insulin (as reported in KP 3D version 2). Removing PVA from the mix also resulted in no EB formation. Therefore, although the

Burridge 2011 protocol did not result in successful cardiac differentiation, it introduced an efficient technique for EB formation, as well as the insight that PVA and insulin are both critical for human EB formation. In addition, from experimentation on initial cell seeding number, an optimal hPSCs seeding number was determined for making EBs for cardiac differentiation, i.e.

5x104cells/well.

67

This initial understanding of EB formation led to further experimentation leading to the development of final KP 3D protocol. It was observed that adding a GSK3 inhibitor, 2 µM BIO, and 40 ng/ml activin A to the RPMI + PVA mix (as in KP 3D version 3, table 2.4.1-2) resulted in cell death during EB formation, which explains the smaller diameter of EBs generated with KP 3D v3 protocol (table 2.4.1-1). However, despite cell death, some EBs were observed to be beating, and the method improved the cardiac differentiation efficiency from 0% to 6.25%. This initial finding highlighted the importance of canonical Wnt signaling, as well as the importance of having both Activin A and BMP4 to synergistically activate TGFβ signaling for cardiac development. The improvement in cardiac differentiation efficiency might also be attributed to the use of a Wnt signaling inhibitor Dkk1 together with p38/MAPK inhibitor SB203580 from day

4-8 of the differentiation (instead of RPMI+FBS medium in Burridge 2011 protocol). Despite this initial success, the low differentiation efficiency indicated that there was still something in the medium which might inhibit cardiogenesis.

Armed with this knowledge, the next iteration involved experimentation with the concentrations of some of the active components. In comparison to version three of the KP 3D protocol, BIO concentration was kept the same at 2 µM, BMP4 concentration was reduced from

40 ng/ml to 20 ng/ml; activin A concentration was kept constant at 40 ng/ml and the rh-insulin concentration was reduced from 10 µg/ml to 1 µg/ml. Also, since Dkk1 had to be used at very high concentrations of 150 ng/ml in version three to four, the cost became prohibitively expensive. Therefore, instead of Dkk1, KP 3D version five onwards made use of small molecule

Wnt signaling inhibitors IWP4, C59 and IWR1 (Chen et al., 2009). In version five of KP 3D protocol, two different concentrations were tested: 2 µM and 5 µM IWP4. The p38/MAPK inhibitor

SB203580 was kept at the same concentration as in version three to four at 5 µM. With these

68

modifications in place, cardiac differentiation efficiency was increased significantly in comparison to version three. 46.7% and 35.7% of all differentiated H9 EBs, were beating with

2µM IWP4 and 5µM IWP4 treatment, respectively. HES3 EBs showed higher differentiation efficiencies: 75% and 100% of all differentiated EBs were beating with 2 µM IWP4 and 5 µM

IWP4 treatment, respectively.

A couple of reasons could be behind this observed improvement: (1) insulin concentration was too high in version 3 at 10 µg/ml and this somehow inhibited phase 1 of cardiac differentiation; (2) the ratio of 20 ng/ml of BMP to 40 ng/ml of activin A was more optimal than 40 ng/ml of BMP4 to 40 ng/ml of activin A and (3) IWP4 at the 2 different concentrations tested (2µM and 5µM) was more effective than 150 ng/ml Dkk1. It has been understood that insulin inhibits cardiac differentiation in the early stages of END-2 cell co-culture cardiac differentiation system (Xu et al., 2008). A recent publication by Lian et al (2013) also support our observation that 10 μg/ml insulin at day 0 fully blocked the synergistic effects of activin A and BMP4 in cardiac differentiation as it inhibits cardiac mesoderm differentiation of hPSCs.

Version six of the KP 3D protocol kept concentrations of insulin, BMP4 and activin A at the same level as version five. The ROCK inhibitor (Y27632) concentration was reduced from

2µM to 1µM and BIO was varied from 2 – 5 µM. For phase 2 of the differentiation, on top of

2µM and 5µM IWP4, a new Wnt inhibitor “C59” was tested at 25nM and 50nM. Generally, lower concentrations of BIO (2µM – 3µM) in phase 1 media gave better differentiation efficiencies, generating 62.5% – 100% beating EBs from all differentiated EBs. 5µM BIO gave 16.7% – 50% beating EBs. When 2µM BIO was used in phase 1 of differentiation, EBs treated with 50nM C59 in phase 2 of the differentiation showed lower differentiation efficiencies than EBs treated with

69

25nM C59 and 2µM IWP4. EBs treated with 2µM IWP4 and 25nM C59 were all beating while only 85.7% of all EBs treated with 50nM C59 were beating. When 3 µM BIO was used in phase 1,

80%, 100% and 62.5% of all EBs were beating when treated with 2 µM IWP4, 25nM C59 and

50nM C59, respectively. When 5 µM BIO was used in phase 1, 50%, 33% and 16.7% of all EBs were beating when treated with 2 µM IWP4, 25nM C59 and 50 nM C59, respectively. Altogether, the result indicated that insulin at 1 µg/ml, BMP4 at 20 ng/ml, Activin A at 40 ng/ml, Y27632 at

1µM and BIO at 2µM were optimal for phase 1 of cardiac differentiation. For phase 2 of differentiation, 2µM IWP4 or 25nM C59 was more optimal than 50nM C59.

Version seven investigated the importance of p38/MAPK inhibition in phase 2 of cardiac differentiation. Activin A concentration was now increased from 40 ng/ml to 50 ng/ml in phase 1 media and for phase 2 of differentiation, 25nM and 50nM C59 was tested with or without 5µM

SB203580. For comparison, 2µM IWP4 + 5µM SB203580 and 5µM IWP4 + 5µM SB203580 groups were also tested. The result indicated that the p38/MAPK inhibitor SB203580 was needed for a more efficient cardiac differentiation. 66.7% and 96% of all differentiated EBs were beating after

25nM C59 and 25nM C59 + 5µM SB203580 treatment, respectively. Treatment with 50 nM C59 and 50 nM C59 + 5 µM SB203580 generated 10.34% and 96.7% beating EBs, respectively from all differentiated EBs. Finally, from all differentiated EBs, 86.7% and 76.7% of EBs were beating after treatment with 2µM IWP4 + 5µM SB203580 and 5µM IWP4 + 5µM SB203580, respectively. C59 +

SB203580 treated EBs gave higher percentages of beating EBs than those treated with IWP4 +

SB203580. Thus, it can be inferred that C59 could be a more effective Wnt inhibitor than IWP4.

This prompted the use of 25 nM C59 + 5 µM SB203580 for phase 2 of cardiac differentiation from version eight onwards.

70

Version eight of the protocol consolidated knowledge gained to date. Multiple rounds of differentiation showed consistently high EBs differentiation efficiencies (>90%) with version 8 of the protocol. Also, it was empirically observed that addition of sodium pyruvate to a final concentration of 1mM reduced the extent of cell death during EB formation step. It has been known that GSK3β phosphorylates pyruvate dehydrogenase thus inactivating it; this results in mitochondrial dysfunction and possibly cell death (Hoshi et al., 1996). GSK3β inhibition, conversely, will lead to glycogen synthase and pyruvate dehydrogenase (PDH) activation, resulting in increased mitochondrial respiration. In addition, 8 – 24 hours of GSK3 inhibition has been reported to result in an approximately twofold increase in glucose uptake due to a similar increase in protein expression of the facilitative glucose transporter 1 (GLUT1) (Buller et al.,

2008). Taken together, GSK3 inhibition seems to increase cellular glucose uptake and metabolism. This theory could explain the observed cell death when BIO-acetoxime (GSK3β inhibitor) was used in phase 1 media. Cell apoptosis was subsequently ameliorated when 1mM sodium pyruvate was added into phase 1 media. The cells could have been metabolizing glucose at such a fast rate upon GSK3β inhibition by BIO that glucose was rapidly depleted in the medium; the cells could be dying due to starvation in absence of pyruvate supplement.

Version nine (final version) of the protocol serves to further fine-tune the KP 3D protocol by increasing bFGF concentration to 20 ng/ml in order to further enhance cell proliferation in the early phase of differentiation. To reduce oxidative stress in the culture system during enhanced cell proliferation, 200µM of ascorbic acid was added into phase 1 media. Addition of ascorbic acid has been shown to improve cardiac differentiation in a number of studies (Burridge et al.,

2012; Burridge et al., 2011; Cao et al., 2012; Kattman et al., 2011) possibly due to increasing collagen synthesis that might enhance proliferation of cardiovascular progenitor cells (Burridge

71

et al., 2014; Cao et al., 2012). A recent publication by Burridge et al (2014) also highlighted the use of ascorbic acid in E8 medium for the culture of hPSCs, thus postulating an additional role of ascorbic acid in cardiac differentiation. Furthermore, another report pinpointed the role of ascorbic acid as a regulator of DNA methylation in embryonic stem cells by Tet enzymes

(Blaschke et al., 2013). Thus ascorbic acid might in itself serve to regulate a number of genes which altogether serve to enhance cardiac differentiation.

In addition, the Wnt inhibitor used in phase 2 of differentiation was changed from C59 to IWR1. IWR1 has been reported to block Wnt signaling pathway more efficiently as it inhibits downstream Wnt responses whereas IWP2 and 4 inhibit novel Wnt production; there could still be some residual Wnt responses observed with the use of IWPs. Detailed mechanisms of these two different classes of Wnt inhibitors have been well reviewed elsewhere (Chen et al., 2009).

Recently, IWR1 has been demonstrated to play an important role in ventricular cardiomyocyte

(vCMs) subtype specification (Karakikes et al., 2014; Weng et al., 2014) Thus the Wnt inhibition step was prolonged from 3 days (day 3-6 of differentiation) to 7 days (day 3-10 of differentiation).

Furthermore, 1µM BMS453 was added from day 6-8 of differentiation in the hope of promoting differentiation towards vCM subtype (Zhang et al., 2011), together with the use of IWR1. This part will be further elaborated in Chapter 5. In addition, insulin was added back into the medium starting from day 8-10, to promote cardiac progenitor proliferation. Using this latest version of the protocol, 100% all formed EBs routinely started to beat by day 9 of differentiation.

2-dimensional approach to cardiac differentiation

If the success of 3D differentiation protocols hinged on the successful formation of uniformly-sized EBs, the success of 2D differentiation protocols was found to depend strongly on the cell seeding density on day 0 of differentiation. Cell density had to be just right: sub-

72

confluent culture did not differentiate efficiently towards CMs and similarly, over-confluent culture also often failed to be differentiated to CMs. Cell-to-cell signaling could be at the root of this observation. Bauwens et al (2008) explained that the maintenance of hESCs pluripotency is enhanced in larger colonies due to an antagonistic interaction between secreted factors from pluripotent hESCs and the extra-embryonic endoderm cells, which regulates Smad1 activation

(Peerani et al., 2007). Increased hESCs density leads to higher levels of secreted BMP antagonists

(such as GDF3) by these cells, thus promoting self-renewal, as well as the expression of neuro- ectodermal markers Pax6 and nestin via the reduction in Smad1 activation (Bauwens et al., 2008;

Pera et al., 2004). However, lower hESCs densities render the hESCs more susceptible to BMP-2 secreted by extra-embryonic endoderm. BMP2 will activate Smad1 and self-renewal of hESCs will be suppressed instead. As such, hESCs cultured in low density tend to differentiate towards endodermal lineages (higher Gata6/Pax6 ratio), whereas hESCs cultured in high density tend to differentiate towards ectodermal lineages (Bauwens et al., 2008).

In the effort to reproduce the Lian protocol, a number of challenges were encountered: there was (1) a drift in optimal cell seeding density; (2) failure to replicate the protocol when fresh RPMI + B27 (-) + 5 µM IWP4 was added on the third day of differentiation (mixture of spent and fresh media was required instead). As described above, it was also observed experimentally that as hESCs grew older, a more dense seeding density was required on day 0 of differentiation to achieve previously observed differentiation efficiencies. Enzymatic passaging methods of hESCs (collagenase or trypsin-based) have been reported to promote chromosomal aberrations in hESCs lines that had originally been propagated with manual methods over as few as 23 passages (Mitalipova et al., 2005). In addition, the same authors also reported that hESCs passaged with bulk passaging methods (cell-dissociation reagent based and/or enzymatic based)

73

demonstrated an increase in the expression of early differentiated genes, indicating that these methods had compromised the undifferentiated hESCs normal gene expression profile

(Mitalipova et al., 2005). Therefore, the observed drift in optimal seeding density might be caused by bulk passaging and high density culture conditions; which might have affected the gene expression profiles of hESCs used in this study such that it affected the “balance point” in which the cells define their optimal cell density on day 0 of differentiation.

It is still not understood how mixing spent media with fresh RPMI + B27 (-) + 10µM IWP4 in 1:1 ratio promotes the differentiation process. Presumably, the differentiating cells secreted bio-active factors into the media, which, when in combination with 5µM IWP4, effectively tuned the signaling pathways efficiently to enhance cardiac differentiation in the differentiating cell population.

Cardiomyocyte trypsinization protocol

The first attempt of CMs isolation with collagenase I and 0.25% trypsin/EDTA resulted in

25.7% viable CMs. CM viability was observed to increase as trypsin concentration was decreased, reaching about 47% ± 5.2% when the trypsin concentration was halved in conjunction with the use of Accutase mix. CM viability was further enhanced to 65.1% ± 16.1% as trypsin concentration was further lowered to 0.05%. Thus it appears that trypsin concentration plays a major role in CM viability post digestion.

However, experimental observations indicated that trypsin concentration was not the only parameter determining CM viability post-digestion. It was observed that the state of cardiomyocytes just before trypsinization was more important – the CMs had to be quiescent prior to trypsinization. In house protocol for isolating rat CMs from rat EHMs involve the use of

0.035 mg/ml Liberase Blendzyme III (Roche) and 30 mmol/L BDM (2,3-butanedione monoxime) 74

at 37 °C for 60 min. The use of BDM improved isolated cell number, viability and morphology and is effective in that respect. However, BDM use is also associated with a number of disadvantages: BDM acts as a non-specific phosphatase in the isolation solutions and especially in culture media. Prolonged perfusion with BDM may result in cells with better morphology, but cells either enter a state of terminal contracture and die, are not excitable, or have significantly altered cellular electrical properties once BDM was removed (Verrecchia and Herve, 1997;

Watanabe et al., 2001; Wolska and Solaro, 1996). Thus different CM trypsinization methods were explored. However, the most effective protocol was the simplest, necessitating only 5 minutes of CM incubation in PBS (+) at room temperature prior to trypsinization. There is very low level of calcium in PBS (-) or PBS (+), thus contractile forces of the CMs are reduced following

PBS wash. At this point, trypsinization into single cell suspension with Accutase will result in minimal CM death. This current protocol led to improved CM viability post digestion to around

70% ± 11.9%.

In vitro cardiac tissue mimetic - Engineered Heart Muscles (EHMs)

The heart is composed of different cell types: cardiac fibroblasts and cardiomyocytes

(majority); endothelial cells; vascular smooth muscle cells (VSMCs); macrophages and neurons

(Banerjee et al., 2007; Baudino et al., 2006; Camelliti et al., 2005; Harvey and Rosenthal, 1999;

Zimmermann, 2011). All cellular components interact collectively and dynamically with the extracellular matrix to regulate cardiac mechanical, biochemical and electrophysiological functions (Baudino et al., 2006). In this light, in order to construct a functional in vitro model of heart tissue, one needs to incorporate these different components accordingly. Published reports have demonstrated the indispensable role of non-myocytes fraction for the generation of functional cardiac tissue patch (Kim et al., 2010; Naito et al., 2006; Zimmermann et al., 2006;

75

Zimmermann and Eschenhagen, 2003). Some groups used fibroblasts (Soong et al., 2012; Tiburcy et al., 2014), while some others used endothelial cells and/or smooth muscle cells (Liau et al.,

2011; Naito et al., 2006; Narmoneva et al., 2004).

In this study, fibroblasts were used to represent the non-myocyte fraction due to two main reasons (1) they form the majority of the non-myocyte fraction in native human hearts and

(2) they produce extracellular matrix for tissue growth and maintenance (Zimmermann, 2011).

During EHM casting, collagen type-I was used as the matrix to hold the different cells together, as the native human heart contains mostly collagen type I and III (Zimmermann, 2011).

Interestingly, the observed experimental results seem to recapitulate the in vivo conditions, with

30% fibroblast-containing EHMs contracting with higher forces than 50% fibroblasts-containing

EHMs (Banerjee et al., 2007). EHM contractile forces also increased with increasing extracellular calcium concentrations, and they responded well to isoprenaline stimulation. As such, EHMs could be used as in vitro model of the heart tissue for downstream studies of hPSC-derived CM maturation.

3D-printing technology for the automated formation of EHMs

3D-printing technology allows immediate prototyping of ideas. In this study, 3D-printing was used to print a semi-automated plate system which enabled EHMs casting and transfer to auxotonic stretchers to be done in automatic and seamless manner. Unfortunately, although veroclear and tango black were biocompatible materials, the support and binding materials were not. It was observed that leeching of residual support material to culture medium was cytotoxic.

In addition, the tango black material did not provide the same mechanical properties to the conventionally used PDMS, thus altering the properties of 3D-printed stretchers. The utility of

76

this technology depends on future optimizations on the plate design as well as printing materials and sterilization techniques. Due to the time limitation, this effort was discontinued.

Concluding remarks

Cardiogenesis is a tightly regulated process governed by precise spatial and temporal regulation, as well as interactions of four main signaling pathways: nodal/activin/TGFβ, Wnt,

BMP, and FGF. However, in vitro recapitulation of these events are as not as straightforward.

Successful in vitro cardiac differentiation protocols do not stop at simply incorporating these pathways at specific timings in the regime. Special attention to: (1) techniques used in maintenance of pluripotent hPSC culture (Burridge et al., 2014), (2) EB sizes for 3D-based differentiation, (3) hPSC density on day 0 of differentiation for 2D-based differentiation, as well as (4) the compositions of differentiation media is of utmost importance and can make or break the differentiation protocol at hand.

In this part of the study, both 3D and 2D differentiation protocols for CM generation from hESCs have been successfully optimized. With the new protocol described here, it is possible to generate 100% beating EBs and monolayers by day 9 of differentiation.

Similar attention to detail is also needed for the generation of a functional in vitro heart tissue model: the EHM tissue constructs. Successful in vitro engineering of functional heart muscles pivots on the understanding of native heart components and their respective functions.

Additionally, efficient trypsinization protocol of CMs can improve CM yield, which is the bottleneck in making large numbers of EHMs for downstream studies. In this study, a simple yet efficient CM trypsinization protocol has been successfully developed. In addition, early experiments with EHMs composition indicated 30% HFFs: 70% CMs ratio to be the most optimal for EHM formation. 77

2.6. References

Abu-Issa, R., and Kirby, M.L. (2007). Heart field - from mesoderm to heart tube. Annu Rev Cell Dev Biol 23, 45-68. Akins, R.E., Jr., Rockwood, D., Robinson, K.G., Sandusky, D., Rabolt, J., and Pizarro, C. (2010). Three- dimensional culture alters primary cardiac cell phenotype. Tissue Eng Part A 16, 629-641. Banerjee, I., Fuseler, J.W., Price, R.L., Borg, T.K., and Baudino, T.A. (2007). Determination of cell types and numbers during cardiac development in the neonatal and adult rat and mouse. American Journal of Physiology - Heart and Circulatory Physiology 293, H1883-H1891. Baudino, T.A., Carver, W., Giles, W., and Borg, T.K. (2006). Cardiac fibroblasts: friend or foe? American Journal of Physiology - Heart and Circulatory Physiology 291, H1015-H1026. Bauwens, C.L., Peerani, R., Niebruegge, S., Woodhouse, K.A., Kumacheva, E., Husain, M., and Zandstra, P.W. (2008). Control of Human Embryonic Stem Cell Colony and Aggregate Size Heterogeneity Influences Differentiation Trajectories. STEM CELLS 26, 2300 –2310. Beqqali, A., Kloots, J., Ward-van Oostwaard, D., Mummery, C., and Passier, R. (2006). Genome-wide transcriptional profiling of human embryonic stem cells differentiating to cardiomyocytes. Stem Cells 24, 1956-1967. Binah, O., Dolnikov, K., Sadan, O., Shilkrut, M., Zeevi-Levin, N., Amit, M., Danon, A., and Itskovitz-Eldor, J. (2007). Functional and developmental properties of human embryonic stem cells-derived cardiomyocytes. J Electrocardiol 40, S192-196. Blaschke, K., Ebata, K.T., Karimi, M.M., Zepeda-Martinez, J.A., Goyal, P., Mahapatra, S., Tam, A., Laird, D.J., Hirst, M., Rao, A., et al. (2013). Vitamin[thinsp]C induces Tet-dependent DNA demethylation and a blastocyst-like state in ES cells. Nature 500, 222-226. Bondue, A., and al, e. (2008). Mesp1 acts as a master regulator of multipotent cardiovascular progenitor specification. Cell stem cell 3, 69-84. Bondue, A., and Blanpain, C. (2010). Mesp1: a key regulator of cardiovascular lineage commitment. Circulation research 107, 1414-1427. Bondue, A., Tannler, S., Chiapparo, G., Chabab, S., Ramialison, M., Paulissen, C., Beck, B., Harvey, R., and Blanpain, C. (2011). Defining the earliest step of cardiovascular progenitor specification during embryonic stem cell differentiation. J Cell Biol 192, 751-765. Brito-Martins, M., Harding, S.E., and Ali, N.N. (2008). β1- and β2-adrenoceptor responses in cardiomyocytes derived from human embryonic stem cells: comparison with failing and non-failing adult human heart. British journal of pharmacology 153, 751-759. Buller, C.L., Loberg, R.D., Fan, M.H., Zhu, Q., Park, J.L., Vesely, E., Inoki, K., Guan, K.L., and Brosius, F.C., 3rd (2008). A GSK-3/TSC2/mTOR pathway regulates glucose uptake and GLUT1 glucose transporter expression. Am J Physiol Cell Physiol 295, C836-843. Burridge, P.W., Anderson, D., Priddle, H., Barbadillo, M.M.D., Chamberlain, S., Allegrucci, C., Young, L.E., and Denning, C. (2007). Improved human embryonic stem cell embryoid body homogeneity and cardiomyocyte differentiation from a novel v-96 plate aggregation system highlights interline variability. Stem Cells 25, 929 -938. Burridge, P.W., Keller, G., Gold, J.D., and Wu, J.C. (2012). Production of de novo cardiomyocytes: human pluripotent stem cell differentiation and direct reprogramming. Cell Stem Cell 10, 16 - 28.

78

Burridge, P.W., Matsa, E., Shukla, P., Lin, Z.C., Churko, J.M., Ebert, A.D., Lan, F., Diecke, S., Huber, B., Mordwinkin, N.M., et al. (2014). Chemically defined generation of human cardiomyocytes. Nat Methods 11, 855-860. Burridge, P.W., Thompson, S., Millrod, M.A., Weinberg, S., Yuan, X., Peters, A., Mahairaki, V., Koliatsos, V.E., Tung, L., and Zambidis, E.T. (2011). A universal system for highly efficient cardiac differentiation of human induced pluripotent stem cells that eliminates interline variability. PLoS One 6, e18293. Camelliti, P., Borg, T.K., and Kohl, P. (2005). Structural and functional characterisation of cardiac fibroblasts. Cardiovascular Research 65, 40-51. Cao, F., Wagner, R.A., Wilson, K.D., Xie, X., Fu, J.D., Drukker, M., Lee, A., Li, R.A., Gambhir, S.S., Weissman, I.L., et al. (2008). Transcriptional and functional profiling of human embryonic stem cell-derived cardiomyocytes. PloS one 3, e3474. Cao, N., Liu, Z., Chen, Z., Wang, J., Chen, T., Zhao, X., Ma, Y., Qin, L., Kang, J., Wei, B., et al. (2012). Ascorbic acid enhances the cardiac differentiation of induced pluripotent stem cells through promoting the proliferation of cardiac progenitor cells. Cell Res 22, 219-236. Chen, B., Dodge, M.E., Tang, W., Lu, J., Ma, Z., Fan, C.W., Wei, S., Hao, W., Kilgore, J., Williams, N.S., et al. (2009). Small molecule-mediated disruption of Wnt-dependent signaling in tissue regeneration and cancer. Nat Chem Biol 5, 100-107. Chen, J.-F., Mandel, E.M., Thomson, J.M., Wu, Q., Callis, T.E., Hammond, S.M., Conlon, F.L., and Wang, D.-Z. (2006). The role of microRNA1 and microRNA133 in skeletal muscle proliferation and differentiation Nature Genetics 38, 228-233. Chen, V.C., Stull, R., Joo, D., Cheng, X., and Keller, G. (2008). Notch signaling respecifies the hemangioblast to a cardiac fate. Nature biotechnology 26, 1169-1178. Davis, R.P., van den Berg, C.W., Casini, S., Braam, S.R., and Mummery, C.L. (2011). Pluripotent stem cell models of cardiac disease and their implication for drug discovery and development. Trends Mol Med 17, 475-484. Dierickx, P., Doevendans, P.A., Geijsen, N., and van Laake, L.W. (2012). Embryonic Template-Based Generation and Purification of Pluripotent Stem Cell-Derived Cardiomyocytes for Heart Repair. Journal of cardiovascular translational research. Doetschman, T.C., Eistetter, H., Katz, M., Schmidt, W., and Kemler, R. (1985). The in vitro development of blastocyst-derived embryonic stem cell lines - formation of visceral yolk sac, blood islands and myocardium. J Embryol Exp Morphol 87, 27-45. Dolnikov, K., Shilkrut, M., Zeevi-Levin, N., Danon, A., Gerecht-Nir, S., Itskovitz-Eldor, J., and Binah, O. (2005). Functional properties of human embryonic stem cell-derived cardiomyocytes. Annals of the New York Academy of Sciences 1047, 66-75. Elliott, D.A., Braam, S.R., Koutsis, K., Ng, E.S., Jenny, R., Lagerqvist, E.L., Biben, C., Hatzistavrou, T., Hirst, C.E., Yu, Q.C., et al. (2011). NKX2-5(eGFP/w) hESCs for isolation of human cardiac progenitors and cardiomyocytes. Nat Methods 8, 1037-1040. Evans-Anderson, H.J., Alfieri, C.M., and Yutzey, K.E. (20008). Regulation of Cardiomyocyte proliferation and mycardial growth during development by FOXO transcription factors. Circulation research 102, 686- 694. Franco, D., Markman, M.M., Wagenaar, G.T., Ya, J., Lamers, W.H., and Moorman, A.F. (1999). Myosin light chain 2a and 2v identifies the embryonic outflow tract myocardium in the developing rodent heart. Anat Rec 254, 135-146.

79

Freund, C., Ward-van Oostwaard, D., Monshouwer-Kloots, J., van den Brink, S., van Rooijen, M., Xu, X., Zweigerdt, R., Mummery, C., and Passier, R. (2008). Insulin redirects differentiation from cardiogenic mesoderm and endoderm to neuroectoderm in differentiating human embryonic stem cells. Stem Cells 26, 724-733. Gai, H.e.a. (2009). Generation and characterization of functional cardiomyocytes using induced pluripotent stem cells derived from human fibroblasts. Cell Biology International 33, 1184-1193. Gaur, M., Ritner, C., Sievers, R., Pedersen, A., Prasad, M., Bernstein, H.S., and Yeghiazarians, Y. (2010). Timed inhibition of p38MAPK directs accelerated differentiation of human embryonic stem cells into cardiomyocytes. Cytotherapy 12, 807-817. Graichen, R., Xu, X., Braam, S.R., Balakrishnan, T., Norfiza, S., Sieh, S., Soo, S.Y., Tham, S.C., Mummery, C., Colman, A., et al. (2008). Enhanced cardiomyogenesis of human embryonic stem cells by a small molecular inhibitor of p38 MAPK. Differentiation 76, 357-370. Guzzo, R.M., Foley, A.C., Ibarra, Y.M., and Mercola, M. (2007). Signaling Pathways in Embryonic Heart Induction. 18, 117-151. Harvey, P.R., and Rosenthal, N. (1999). Heart Development, Vol 1 (New York: New York: Academic). He, J.Q., Ma, Y., Lee, Y., Thomson, J.A., and Kamp, T.J. (2003). Human embryonic stem cells develop into multiple types of cardiac myocytes: action potential characterization. Circulation research 93, 32-39. Hoshi, M., Takashima, A., Noguchi, K., Murayama, M., Sato, M., Kondo, S., Saitoh, Y., Ishiguro, K., Hoshino, T., and Imahori, K. (1996). Regulation of mitochondrial pyruvate dehydrogenase activity by tau protein kinase I/glycogen synthase kinase 3beta in brain. Proceedings of the National Academy of Sciences of the United States of America 93, 2719-2723. Ishizaki, T., Uehata, M., Tamechika, I., Keel, J., Nonomura, K., Maekawa, M., and Narumiya, S. (2000) Pharmacological properties of Y-27632, a specific inhibitor of rho-associated kinases. Mol Pharmacol 57: 976-983. Karakikes, I., Senyei, G.D., Hansen, J., Kong, C.W., Azeloglu, E.U., Stillitano, F., Lieu, D.K., Wang, J., Ren, L., Hulot, J.S., et al. (2014). Small molecule-mediated directed differentiation of human embryonic stem cells toward ventricular cardiomyocytes. Stem Cells Transl Med 3, 18-31. Kattman, S.J., Witty, A.D., Gagliardi, M., Dubois, N.C., Niapour, M., Hotta, A., Ellis, J., and Keller, G. (2011). Stage-specific optimization of activin/nodal and BMP signaling promotes cardiac differentiation of mouse and human pluripotent stem cell lines. Cell Stem Cell 8, 228-240. Kehat, I., Gepstein, A., Spira, A., Itskovitz-Eldor, J., and Gepstein, L. (2002). High-resolution electrophysiological assessment of human embryonic stem cell-derived cardiomyocytes: a novel in vitro model for the study of conduction. Circulation research 91, 659-661. Kehat, I., Kenyagin-Karsenti, D., Snir, M., Segev, H., Amit, M., Gepstein, A., Livne, E., Binah, O., Itskovitz- Eldor, J., and Gepstein, L. (2001). Human Embryonic Stem Cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J Clin Invest 108, 407-414. Khademhosseini, A., Ferreira, L., Blumling, J.r., Yeh, J., Karp, J.M., Fukuda, J., and Langer, R. (2006). Co- culture of human embryonic stem cells with murine embryonic fibroblasts on microwell-patterned substrates. Biomaterials 27, 5968 -5977. Kim, C., Majdi, M., Xia, P., Wei, K.A., Talantova, M., Spiering, S., Nelson, B., Mercola, M., and Chen, H.S. (2010). Non-cardiomyocytes influence the electrophysiological maturation of human embryonic stem cell- derived cardiomyocytes during differentiation. Stem cells and development 19, 783-795.

80

Kubalak, S.W., Miller-Hance, W.C., O’Brien, T.X., Dyson, E., and Chien, K.R. (1994). Chamber specification of atrial myosin light chain-2 expression precedes septation during murine cardiogenesis. The Journal of biological chemistry 269, 16961-16970. Laflamme, M.A., Chen, K.Y., Naumova, A.V., Muskheli, V., Fugate, J.A., Dupras, S.K., Reinecke, H., Xu, C., Hassanipour, M., Police, S., et al. (2007). Cardiomyocytes derived from human embryonic stem cells in pro-survival factors enhance function of infarcted rat hearts. Nat Biotechnol 25, 1015-1024. Lian, X., Hsiao, C., Wilson, G., Zhu, K., Hazeltine, L.B., Azarin, S.M., Zhang, J., Kamp, T.J., Palecek, S.P., and Raval, K.K. (2012 ). Robust cardiomyocyte differentiation from human pluripotent stem cells via temporal modulation of canonical Wnt signaling. PNAS 109, E1848 - 1857. Lian, X., Zhang, J., Zhu, K., Kamp, T.J., and Palecek, S.P. (2013). Insulin inhibits cardiac mesoderm, not mesendoderm, formation during cardiac differentiation of human pluripotent stem cells and modulation of canonical Wnt signaling can rescue this inhibition. Stem cells 31, 447-457. Liau, B., Christoforou, N., Leong, K., and Bursac, N. (2011). Pluripotent Stem Cell-derived Cardiac Tissue Patch with Advanced Structure and Function. Biomaterials 32, 9180-9187. Liu, J., Fu, J.D., Siu, C.W., and Li, R.A. (2007). Functional sarcoplasmic reticulum for calcium handling of human embryonic stem cell-derived cardiomyocytes: insights for driven maturation. Stem Cells 25, 3038- 3044. McDevitt, T.C., Laflamme, M.A., and Murry, C.E. (2005). Proliferation of cardiomyocytes derived from human embryonic stem cells is mediated via the IGF/PI 3-kinase/Akt signaling pathway. J Mol Cell Cardiol 39, 865-873. Mitalipova, M.M., Rao, R.R., Hoyer, D.M., Johnson, J.A., Meisner, L.F., Jones, K.L., Dalton, S., and Stice, S.L. (2005). Preserving the genetic integrity of human embryonic stem cells. Nat Biotech 23, 19-20. Mohr, J.C., de Pablo, J.J., and Palecek, S.P. (2006). 3-D microwell culture of human embryonic stem cells. Biomaterials 27, 6032- 6042. Mohr, J.C., Zhang, J., Azarin, S.M., Soerens, A.G., de Pablo, J.J., Thomson, J.A., Lyons, G.E., Palecek, S.P., and Kamp, T.J. (2010). The microwell control of embryoid body size in order to regulate cardiac differentiation of human embryonic stem cells. Biomaterials 31, 1885-1893. Mummery, C.L. (2005). Cardiology: solace for the broken-hearted? Nature 433, 585-587. Mummery, C., Ward-van Oostwaard, D., Doevendans, P.A., Spijker, R., van den, B.S., Hassink, R., van der, H.M., Opthof, T., Pera, M., de la Riviere, A.B., et al. (2003). Differentiation of human embryonic stem cells to cardiomyocytes: role of coculture with visceral endoderm-like cells. Circulation Journal 107, 2733-2740. Mummery, C.L., Zhang, J., Ng, E.S., Elliott, D.A., Elefanty, A.G., and Kamp, T.J. (2012). Differentiation of human embryonic stem cells and induced pluripotent stem cells to cardiomyocytes: a methods overview. Circulation research 111, 344-358. Naito, H., Melnychenko, I., Didie, M., Schneiderbanger, K., Schubert, P., Rosenkranz, S., Eschenhagen, T., and Zimmermann, W.H. (2006). Optimizing engineered heart tissue for therapeutic applications as surrogate heart muscle. Circulation 114, I72-78. Narmoneva, D.A., Vukmirovic, R., Davis, M.E., Kamm, R.D., and Lee, R.T. (2004). Endothelial cells promote cardiac myocyte survival and spatial reorganization: implications for cardiac regeneration. Circulation 110, 962-968. Nelson, T.J.e.a. (2009). Repair of acute myocardial infarction by human stemness factors induced pluripotent stem cells. Circulation 120, 408-416.

81

Ng, E.S., Davis, R.P., Azzola, L., Stanley, E.G., and Elefanty, A.G. (2005). Forced aggregation of defined numbers of human embryonic stem cells into embryoid bodies fosters robust, reproducible hematopoietic differentiation. . Blood 106, 1601-1603. Norström, A., Åkesson, K., Hardarson, T., Hamberger, L., Björquist, P., and Sartipy, P. (2006). Molecular and Pharmacological properties of human embryonic stem cell-derived cardiomyocytes. Experimental Biology and Medicine 231, 1753-1762. Odorico, J., Pedersen, R., and Zhang, S.C. (2004). Human Embryonic Stem Cells. (Taylor & Francis). Olson, E.N. (2006). Gene Regulatory Networks in the evolution and development of the heart. Science 313, 1922-1927. Olson, E.N., and Schneider, M.D. (2003). Sizing up the heart - development redux in disease. Genes Dev 2003, 1937-1956. Paige, S.L., Osugi, T., Afanasiev, O.K., Pabon, L., Reinecke, H., and Murry, C.E. (2010). Endogenous Wnt/beta-catenin signaling is required for cardiac differentiation in human embryonic stem cells. PLoS One 5, e11134. Passier, R., Oostwaard, D.W., Snapper, J., Kloots, J., Hassink, R.J., Kuijk, E., Roelen, B., de la Riviere, A.B., and Mummery, C. (2005). Increased cardiomyocyte differentiation from human embryonic stem cells in serum-free cultures. Stem Cells 23, 772-780. Peerani, R., Rao, B.M., Bauwens, C., Yin, T., Wood, G.A., Nagy, A., Kumacheva, E., and Zandstra, P.W. (2007). Niche-mediated control of human embryonic stem cell self-renewal and differentiation. . EMBO J 26, 4744-4755. Pera, M.F., Andrade, J., Houssami, S., Reubinoff, B., Trounson, A., Stanley, E.G., Ward-van Oostwaard, D., and Mummery, C. (2004). Regulation of human embryonic stem cell differentiation by BMP-2 and its antagonist noggin. J Cell Sci 117, 1269-1280. Rajala, K., Pekkanen-Mattila, M., and Aalto-Setala, K. (2011). Cardiac differentiation of pluripotent stem cells. Stem Cells Int 2011, 383709. Ren, Y., Lee, M.Y., Schliffke, S., Paavola, J., Amos, P.J., Ge, X., Ye, M., Zhu, S., Senyei, G., and Lum, L. (2011). Small molecule Wnt inhibitors enhance the efficiency of BMP-4-directed cardiac differentiation of human pluripotent stem cells. J Mol Cell Cardiol 51. Sartiani, L., Bettiol, E., Stillitano, F., Mugelli, A., Cerbai, E., and Jaconi, M.E. (2007). Developmental changes in cardiomyocytes diffrentiated from human embryonic stem cells: a molecular and electrophysiological approach. Stem Cells 25, 1136-1144. Schaaf, S., Shibamiya, A., Mewe, M., Eder, A., Stohr, A., Hirt, M.N., Rau, T., Zimmermann, W.H., Conradi, L., Eschenhagen, T., et al. (2011). Human engineered heart tissue as a versatile tool in basic research and preclinical toxicology. PLoS One 6, e26397. Segev, H., Kenyagin-Karsenti, D., Fishman, B., Gerecht-Nir, S., Ziskind, A., Amit, M., Coleman, R., and Itskovitz-Eldor, J. (2005). Molecular analysis of cardiomyocytes derived from human embryonic stem cells. Development, Growth & Differentiation 47, 295-306. Shi, Y., Katsev, S., Cai, C., and Evans, S. (2000). BMP signaling is required for heart formation in vertebrates. Dev Biol 224, 226-237. Shimoji, K., Yuasa, S., Onizuka, T., Hattori, F., Tanaka, T., Hara, M., Ohno, Y., Chen, H., Egasgira, T., and Seki, T. (2010). G-CSF promotes the proliferation of developing cardiomyocytes in vivo and in derivation from ESCs and iPSCs. Cell stem cell 6, 227-237.

82

Snir, M., Kehat, I., Gepstein, A., Coleman, R., Itskovitz-Eldor, J., Livne, E., and Gepstein, L. (2003). Assessment of the ultrastructural and proliferative properties of human embryonic stem cell-derived cardiomyocytes. Am J Physiol Heart Circ Physiol 285, H2355-H2363. Soong, P.L., Tiburcy, M., and Zimmermann, W.H. (2012). Cardiac differentiation of human embryonic stem cells and their assembly into engineered heart muscle. Curr Protoc Cell Biol Chapter 23, Unit23 28. Sumi, T., Tsuneyoshi, N., Nakatsuji, N., and Suemori, H. (2008). Defining early lineage specification of human embryonic stem cells by the orchestrated balance of canonical Wnt/beta-catenin, Activin/Nodal and BMP signaling. Development 135, 2969-2979. Takei, S., Ichikawa, H., Johkura, K., Mogi, A., No, H., Yoshie, S., Tomotsune, D., and Sasaki, K. (2009). Bone morphogenetic protein-4 promotes induction of cardiomyocytes from human embryonic stem cells in serum-based embryoid body development. Am J Physiol Heart Circ Physiol 296, H1793-H1803. Tiburcy, M., Didie, M., Boy, O., Christalla, P., Doker, S., Naito, H., Karikkineth, B.C., El-Armouche, A., Grimm, M., Nose, M., et al. (2011). Terminal differentiation, advanced organotypic maturation, and modeling of hypertrophic growth in engineered heart tissue. Circulation research 109, 1105-1114. Tiburcy, M., Meyer, T., Soong, P.L., and Zimmermann, W.H. (2014). Collagen-based engineered heart muscle. Methods in molecular biology 1181, 167-176. Tran, T.H., Wang, X., Browne, C., Zhang, Y., Schinke, M., Izumo, S., and and Burcin, M. (2009). Wnt3a- induced mesoderm formation and cardiomyogenesis in human embryonic stem cells. Stem Cells 27, 1869- 1878. Tran, T.H.e.a. (2009). Wnt-3a induced mesoderm formation and cardiomyogenesis in human embryonic stem cells. Stem Cells 27, 1869-1878. Ungrin, M.D., Joshi, C., Nica, A., Bauwens, C., and Zandstra, P.W. (2008). Reproducible, ultra highthroughput formation of multicellular organization from single cell suspension-derived human embryonic stem cell aggregates. PLoSOne 3, e1565. Verrecchia, F., and Herve, J.C. (1997). Reversible blockade of gap junctional communication by 2,3- butanedione monoxime in rat cardiac myocytes. Am J Physiol 272, C875-885. Walters, M.J., Wayman, G.A., and Christian, J.L. (2001). Bone morphogenetic protein function is required for terminal differentiation of the heart but not for early expression of cardiac marker genes. Mech Dev 100, 263-273. Watanabe, K., Ueno, M., Kamiya, D., Nishiyama, A., Matsumura, M., Wataya, T., Takahashi, J.B., and Nishikawa, S.e.a. (2007). A ROCK inhibitor permits survival of dissociated human embryonic stem cells Nature biotechnology 25, 681-686. Watanabe, Y., Iwamoto, T., Matsuoka, I., Ohkubo, S., Ono, T., Watano, T., Shigekawa, M., and Kimura, J. (2001). Inhibitory effect of 2,3-butanedione monoxime (BDM) on Na+/Ca2+ exchange current in guinea- pig cardiac ventricular myocytes. British journal of pharmacology 132, 1317-1325. Wei, H., Tan, G., Manasi, Qiu, S., Kong, G., Yong, P., Koh, C., Ooi, T.H., Sze, Y.L., Wong, P., et al. (2012). One-step derivation of cardiomyocytes and mesenchymal stem cells from human pluripotent stem cells. Stem cell research 9, 87-100. Weng, Z., Kong, C.W., Ren, L., Karakikes, I., Geng, L., He, J., Chow, M.Z., Mok, C.F., Keung, W., Chow, H., et al. (2014). A simple, cost-effective but highly efficient system for deriving ventricular cardiomyocytes from human pluripotent stem cells. Stem cells and development 23, 1704-1716. Willems, E., Bushway, P.J., and Mercola, M. (2009). Natural and synthetic regulators of embryonic stem cell cardiogenesis. Pediatric cardiology 30, 635-642.

83

Wolska, B.M., and Solaro, R.J. (1996). Method for isolation of adult mouse cardiac myocytes for studies of contraction and microfluorimetry. Am J Physiol 271, H1250-1255. Wu, S.M. (2008). Mesp1 at the heart of mesoderm lineage specification. Cell stem cell 3, 1-2. Xu, C., Police, S., Rao, N., and Carpenter, M.K. (2002). Characterization and enrichment forcardiomyocytes derived from human embryonic stem cells. Circ Res 91, 501-508. Xu, X.Q., Graichen, R., Soo, S.Y., Balakrishnan, T., Rahmat, S.N., Sieh, S., Tham, S.C., Freund, C., Moore, J., Mummery, C., et al. (2008). Chemically defined medium supporting cardiomyocyte differentiation of human embryonic stem cells. Differentiation 76, 958-970. Xu, X.Q., Soo, S.Y., Sun, W., and Zweigerdt, R. (2009). Global expression profile of highly enriched cardiomyocytes derived from human embryonic stem cells. Stem Cells 27, 2163-2174. Xu, Y.Z., X.; Hahm, H. S.; Wei, W.; Hao, E.; Hayek, A.; Ding, S. (2010). Revealing a core signaling regulatory mechanism for pluripotent stem cell survival and self-renewal by small molecules. Proceedings of the National Academy of Sciences 107, 8129-8134. Yang, L., Soonpaa, M.H., Adler, E.D., Roepke, T.K., Kattman, S.J., Kennedy, M., Henckaerts, E., Bonham, K., Abbott, G.W., Linden, R.M., et al. (2008). Human cardiovascular progenitor cells develop from a KDR+ embryonic-stem-cell-derived population. Nature 453, 524-528. Yoon, B.S., Yoo, S.J., Lee, J.E., You, S.H., Lee, T., and Yoon, H.S. (2006). Enhanced dfferentiation of human embryonic stem cells into cardiomyocytes by combining hanging drop culture and 5-azacytidine treatment. Differentiation; research in biological diversity 74, 149-159. Zhang, D., Shadrin, I.Y., Lam, J., Xian, H.Q., Snodgrass, H.R., and Bursac, N. (2013). Tissue-engineered cardiac patch for advanced functional maturation of human ESC-derived cardiomyocytes. Biomaterials 34, 5813-5820. Zhang, J., Klos, M., Wilson, G.F., Herman, A.M., Lian, X., Raval, K.K., Barron, M.R., Hou, L., Soerens, A.G., Yu, J., et al. (2012). Extracellular Matrix Promotes Highly Efficient Cardiac Differentiation of Human Pluripotent Stem Cells: The Matrix Sandwich Method. Circulation research. Zhang, J., Wilson, G.F., Soerens, A.G., Koonce, C.H., Yu, J., Palecek, S.P., Thomson, J.A., and Kamp, T.J. (2009). Functional cardiomyocytes derived from human induced pluripotent stem cells. Circulation research 104, e30 -e41. Zhang, Q., Jiang, J., Han, P., Yuan, Q., Zhang, J., Zhang, X., Xu, Y., Cao, H., Meng, Q., Chen, L., et al. (2011). Direct differentiation of atrial and ventricular myocytes from human embryonic stem cells by alternating retinoid signals. Cell Res 21, 579-587. Zhao, Y., Samal, E., and Srivastava, D. (2005). Serum response factor regulates a muscle-specific microRNA that targets Hand2 during cardiogenesis. Nature 436, 214-220. Zimmermann, W.H. (2011). Myocardial Tissue Engineering. In, A.R. Boccaccini, and S. Harding, eds. (Springer). Zimmermann, W.H., Didie, M., Doker, S., Melnychenko, I., Naito, H., Rogge, C., Tiburcy, M., and Eschenhagen, T. (2006). Heart muscle engineering: an update on cardiac muscle replacement therapy. Cardiovasc Res 71, 419-429. Zimmermann, W.H., and Eschenhagen, T. (2003). Cardiac tissue engineering for replacement therapy. Heart Fail Rev 8, 259-269.

84

CHAPTER 3

Engineering and Functional Validation

of MYL2- Knock-In Reporter Line

85

TABLE OF CONTENT

3.1. Abstract ...... 88

3.2. Introduction and Literature review ...... 90

3.3. Materials and Methods ...... 96

3.3.1. Cloning of targeting plasmid ...... 96 Preparing 5’ and 3’ homology arms to MYL2 gene ...... 96 Adding 5’-LoxP and 3’ LoxP sites to PGK-Hygro selection cassette ...... 96 Cloning MYL2 PCR product into TOPO vector ...... 97 Generating targeting vector – Self Assembly Cloning ...... 98 Generation of TALENs expression vectors ...... 99

3.3.2. Feeder cell culture ...... 101 MEF culture – thawing frozen stock ...... 101 Propagation of MEFs ...... 102 Inactivation of MEFs and HFFs ...... 102

3.3.3. Human Pluripotent Stem Cells (hPSCs) culture...... 104 Cell culture media composition ...... 104 Thawing and freezing of hPSCs ...... 104 Feeder-dependent propagation and freezing of hPSCs ...... 104

3.3.4. Homologous recombination and screening for positive clones ...... 106 Preparation of F- and R- TALEN plasmids and targeting vector ...... 106 HPSCs electroporation for generation of knock-in line ...... 106 Selection and screening for positive clones ...... 107

3.3.5. Removal of Floxed PGK-Hygromycin cassette ...... 109 Preparation of Cre-recombinase plasmid ...... 109 H9V20 transfection with Cre-IRES-PuroR ...... 109 Selection for successfully transfected clones ...... 110

3.3.6. In vitro characterization of knock in lines ...... 111 Karyotyping and sequencing ...... 111 Pluripotency markers – immunofluorescence staining and FACS ...... 112 Pluripotency markers – FACS analysis ...... 112

3.3.7. Validation of gene expression and transcription of MYL2 knock-in reporter line ...... 113 Microscopy analysis of knock-in CM in 2D and 3D format ...... 113 Live cell FACS of knock-in CMs ...... 114 Investigation for effective Neomycin concentration ...... 114 Luciferase assay ...... 114

86

3.3.8. Statistical analysis ...... 115

3.4. Results ...... 116

3.4.1. Cloning of Targeting Plasmid ...... 116 MYL2 and Phlox PCR products ...... 116 Targeting construct generation by self-assembly cloning (SAC) ...... 117 Generation of TALEN expression vector ...... 118

3.4.2. Human pluripotent stem cells (hPSCs) culture ...... 120 HPSCs colony morphology ...... 120

3.4.3. Generation of knock-in line ...... 121 Validation of TALEN expression constructs and the targeting vector ...... 121 HPSCs electroporation ...... 122 Efficiency of homologous recombination (HR) ...... 123

3.4.4. PGK-Hygromycin cassette was removed using Cre-recombinase ...... 126 3.4.5. Knock-in lines are pluripotent and have no chromosomal aberrations ...... 128 Knock in lines have normal karyotypes ...... 128 gDNA sequencing gave inconclusive result ...... 128 Knock-in lines are pluripotent ...... 129

3.4.6. MCherry (mChr) was expressed in knock-in cardiomyocytes ...... 132 3.4.7. Neomycin resistance gene was expressed in knock-in cardiomyocytes ...... 138 3.4.8. Gaussia Luciferase was expressed in knock-in cardiomyocytes...... 139 3.4.9. Expression of reporter genes was correlated to the endogenous expression level of MYL2 gene ...... 140 3.4.10. MYL2 positive cells (ventricular cardiomyocytes - vCMs) can be enriched by live FACS sorting as well as Neomycin selection ...... 141

3.5. Discussion ...... 144

3.6. References ...... 150

87

3.1. Abstract

State-of-the-art cardiomyocyte differentiation protocols result in a heterogeneous population of cardiomyocytes (CMs) and other cell types (Burridge et al., 2014; Mummery et al.,

2012; Rajala et al., 2011). This heterogeneity is suboptimal for downstream experiments as it can affect the accuracy of results from studies aimed at investigating just a specific subtype of CMs.

To enable enrichment for ventricular cardiomyocytes (vCMs), an MYL2 knock-in line was generated by homologous recombination mediated by a novel genome editing technology, TAL- effector Nucleases (TALENs). A targeting vector was cloned and electroporated along with forward and reverse TALENs to generate the recombinant knock-in line.

Efficiency of homologous recombination was determined by PCR and ranged from 5.33%

- 27.2%, depending on the wild type pluripotent stem cell line used. A few clones were chosen for Southern blotting to validate successful homologous recombination event and Southern blot showed positive heterozygous knock-in at the gene locus of interest. FACS and immunohistochemistry data showed that knock-in lines are positive for pluripotency markers

SOX2, OCT3/4, NANOG, TRA-1-60 and SSEA4. Karyotyping data also showed no gross chromosomal aberrations.

Initial validation experiments revealed that MYL2 was expressed later than day 14 of differentiation. RNA from knock-in CMs (day 18 – 90 of differentiation) was extracted for quantitative PCR (qPCR) experiment to assess correlation between the expression level of MYL2 and inserted reporters. QPCR data showed that reporter genes were expressed in high correlation to MYL2 gene. Each reporter gene’s expression was also highly correlated with the expression of other reporter genes.

88

MCherry expression was observable from day 18 of differentiation onwards in CM monolayer culture. In the engineered heart muscles (EHMs), mCherry expression was observable from day 24 of differentiation onwards. MCherry expression was observed to increase qualitatively with time, with day 90 of differentiation EHMs showing significantly brighter signal than day 24 of differentiation EHMs. MCherry-positive CMs were sorted out with FACS and their

RNA extracted for qPCR experiment. QPCR data showed that mCherry-positive CMs expressed significantly higher level of MYL2 gene than mCherry-negative CMs (p ≤ 0.001). Treating knock-in

CM monolayer with G418 for one week resulted in enrichment for mCherry-positive CMs, indicating that knock-in CMs expressed neomycin resistance gene. Finally, Gaussia luciferase (G.

Luc) was also expressed as early as day 21 of differentiation in CM monolayer as well as in EHMs.

G. Luc expression increased with time, with a maximum level of expression observed on day 45 of differentiation in EHMs.

It was observed that CMs derived from H9VC16 (a knock-in line in which the PGK-HygroR cassette was removed) expressed significantly higher level of mCherry than H9V15-derived CMs

(p ≤ 0.001). However, G. Luc and neomycin resistance (NeoR) reporter genes expression did not differ significantly between H9VC16 and H9V15 (parental line of H9VC16 with PGK-HygroR cassette still intact).

In conclusion, the engineered MYL2 knock-in reporter lines are functional as designed with the three knocked-in foreign genes being expressed in significant correlation with the MYL2 gene.

These novel reporter lines can function as an enrichment tool for vCMs, thus allowing for downstream studies utilizing this specific subtype of CMs.

89

3.2. Introduction and Literature review

Human embryonic stem cells (hESCs)-derived ventricular cardiomyocytes (vCMs) are of interest in the field of cardiovascular research. Diseases involving the ventricular chambers of the heart will ultimately affect the vCMs. A pure population of vCMs would enable in vitro studies of these diseases, as well as studies aimed at developing potential therapies for them.

Yet, supply of vCMs is limited and when they become available, the yield is often insufficient as these cells are terminally differentiated and have lost most of their proliferative capacity. With unlimited proliferative potential, hESCs becomes an attractive source of CMs. Yet, hESCs give rise to immature CMs from all different subtypes (more on this in chapter 4 of this thesis). Therefore, an enrichment tool for hESCs-derived vCMs will be useful and relevant for this field of research. A homogenous population of vCMs would also provide in vitro toxicology data with higher accuracy and predictive power, especially for assessing drug-induced ventricular pro- arrhythmic events.

There have been a few reports describing attempts at deriving a relatively homogenous population of vCMs. Zhang et al. (2011) reported the role of retinoic acid (RA) in CMs subtype specification, with RA agonist increasing the proportion of atrial-like CMs to 94% and RA inhibition (by using RA inhibitor BMS 189453) increasing the proportion of ventricular-like cells to 83%. Müller et al. (2010) transfected mouse embryonic stem cell (mESC) line with eGFP which expression was under the control of MYL2 promoter and enhancer element of CMV. As mentioned previously, MYL2 gene is expressed in vCMs but not atrial CMs (Bird, 2003; Zhang et al., 2011), thus prompting its use as vCM marker in this laboratory and others. MYL2 promoter sequence -497 to +70 relative to transcription start site has also been used to drive the expression of GFP reporter using recombinant adenoviral constructs (Bizy et al., 2013; Huber et 90

al., 2007). These approaches employed standard transfection technique, which although is simpler than homologous recombination strategy, carries with it inherent disadvantages.

Compared with standard transfection strategy, homologous recombination enables a better control of the targeting construct integration site and it reduces the likelihood of the targeting construct integrating into multiple locations in the genome. Site-specific knock-in will also result in a more consistent level of expression of the transgenes from generation to generation (Koch Institute for integrative cancer research at MIT, http://ki.mit.edu/sbc/escell/models/knock).

To generate a multifunctional reporter line, a multicistronic reporter construct has been designed in the present study. This construct contains two reporters: a fluorescent protein mCherry (mChr) and bioluminescence enzyme Gaussia Luciferase (G.Luc); and a selectable marker: Neomycin resistance (NeoR). For simplification purpose, these three inserts will henceforth be referred to as “reporters.”

This reporter construct was designed such that all reporters are transcribed along with the MYL2 gene. As such, the levels of their expressions should be in stoichiometric ratio with the endogenous level of MYL2 gene expression, thus enabling real time monitoring of MYL2 gene expression.

Having a multicistronic targeting construct with three reporters side-by-side, a cleaving strategy during translation was needed. Traditionally, internal ribosomal entry site (IRES) has been used for multicistronic vectors. However, IRES is relatively large at around 600 base pairs.

In addition, the expression of the downstream cistron strongly depends on its specific location relative to the IRES and hence this system might not be dependable as the downstream coding

91

sequence often suffers from much lower levels of expression than the upstream sequence

(Trichas et al., 2008). As such, IRES usually does not provide equivalent levels of expression of the genes separated by its elements (Trichas et al., 2008). There was a need for an alternative.

The use of 2A peptides has recently been favored over IRES. The 2A sequences are relatively short peptides (about 20 amino-acids long, depending on the virus of origin) containing the consensus motif Asp-Val/Ile-Glu-X-Asn-Pro-Gly-Pro. They cause “ribosomal skip” during translation as they prevent the formation of a peptide bond between the penultimate glycine and the last proline residue (Donnelly et al., 2001). The ribosome would skip to the next codon during translation, with the nascent peptide cleaving between the glycine and proline residue of the 2A peptide. After cleavage, the short 2A peptide remains fused to the C-terminus of the

'upstream' protein, while the proline is added to the N-terminus of the 'downstream' protein.

There are four types of 2A peptides, each derived from different virus: foot-and-mouth disease virus, equine rhinitis A virus, Thosea Asigna virus and porcine teschovirus-1. Of the four aforementioned types, the one derived from porcine teschovirus-1 (P2A) has the highest cleavage efficiency as tested in three commonly used human lines, zebrafish embryos and adult mice (Kim et al., 2011). Advantages of using 2A peptide have been well reviewed by de Felipe and colleagues (de Felipe et al., 2006). This superiority over IRES has prompted the use of P2A sequence in our targeting vector.

Having the targeting vector at hand, the appropriate method of getting this vector into the hESCs genome had to be identified. In hESCs, chemical methods of transfection generated homologous clones 10–5 of the time (Eiges et al., 2001). However, attempts using ExGen500 and

FuGene-6 for homologous recombination into the human POU5F1 gene resulted in no resistant colonies to selection drugs (Zwaka and Thomson, 2003). When these authors employed 92

electroporation method however, they obtained 103 G418-resistant clones, 28 of which (27%) were demonstrated to be positive for homologous recombination through PCR and Southern

Blotting. This recombination efficiency was further improved to almost 40% when they used a second targeting vector with a longer 3' homologous arm. The same authors further reported a

26:1 ratio of stable transfected clones to homologous recombination events for the POU5F1 construct when hESCs were electroporated with a DNA construct containing a neo cassette under the control of the tk promoter. A lower ratio of 50:1 was reported for transfection of the HPRT1 vector (Zwaka and Thomson, 2003).

In recent years, the advancement in genome engineering technology has made available more tools for improving homologous recombination efficiency. Zinc Finger Nucleases (ZFNs) technology gained popularity as a mediator of homologous recombination process in the mid-

21st century. This technology is a development of initial work by Chandrasegaran and colleagues

(Kim et al., 1996). ZFNs can be engineered to introduce a double stranded break into the desired target gene locus, thus stimulating gene targeting 102 - 104 fold by activating cellular DNA repair pathways (Cathomen and Joung, 2008). Although this improved efficiency is highly desirable, the technology also carries with it some drawbacks. One of the most concerning drawbacks is the cytotoxic effect of ZFNs due to off-target effects; this has been reported in several studies (Alwin et al., 2005; Beumer et al., 2006; Bibikova et al., 2002; Miller et al., 2007; Porteus, 2006;

Szczepek et al., 2007). In addition, designing ZFNs to bind to specific sequences of the DNA can be rather challenging (Ramirez et al., 2008) as the binding specificity of ZFNs array assembly does not always correlate with the specificity of the ZFs building blocks (Porteus and Baltimore,

2003).

93

A more recent advancement in genome editing technology came in the form of transcription activator-like effector endonucleases (TALENs). TALENs, like ZFNs, consist of DNA binding modules coupled to FokI endonuclease (Boch and Bonas, 2010; Boch et al., 2009;

Christian et al., 2010; Moscou and Bogdanove, 2009; Li et al., 2011; Mahfouz et al., 2011; Miller et al., 2011). The DNA-binding modules originate from plant pathogens Xanthomonas. Each module consists of a tandem repeat of 34 amino acids and is nearly identical to the others except for the residues at position 12 and 13, which are termed repeat variable diresidues

(RVDs). These RVDs are the ones conferring nucleotide binding specificity of the module (Boch et al., 2009; Moscou and Bogdanove, 2009). “NI” binds to A; “HD” to C; “NN” or “NK” to G and “NG” to T (Boch et al., 2009). In this way, DNA binding region can be engineered to target a specific site of known sequences in the genome.

Some studies have reported the efficiencies of non-homologous end joining (NHEJ)- induced mutagenesis mediated by TALENs of up to 45% of transfected cells (Miller et al., 2011;

Mussolino et al., 2011). In addition, TALENs have been shown to edit the genome with low toxicity (Mussolino et al., 2011). The most attractive advantage of using TALENs over ZFNs is the apparent lack of context dependence and 1:1 correspondence of RVDs with a specific nucleotide pair (Carlson et al., 2012). This is in contrast to ZF modules, where there is redundancy of ZFs for a given triplet of base pairs. This redundancy is further complicated with contextual interaction within the ZFNs array assembly (Carlson et al., 2012). Thus engineering DNA binding specificity is more straightforward in TAL effector than ZFs. Very recently, another genome editing tool becomes available: CRISPR/Cas technology (Jinek et al., 2012). However, the use of this technology as genome editing tool was only gaining popularity midway through this project.

94

Therefore, in this study, TALENs were used to enhance homologous recombination efficiency to generate the knock-in cell line described herein.

Once a stable MYL2 knock-in reporter line was obtained, this line would be differentiated to CMs and the reporter genes expression would be analyzed. In particular, the faithfulness of the reporter genes expression, i.e. the degree of correlation between their expressions and the expression of MYL2 gene as well as to one another, would be examined in this part of the study. Finally, the effects of removing PGK-Hygromycin cassette from the knock- in line were also explored.

95

3.3. Materials and Methods

3.3.1. Cloning of targeting plasmid

Preparing 5’ and 3’ homology arms to MYL2 gene

MBJ6, HES3, and H9 hESCs genomic DNA (gDNA) were extracted using Qiagen DNeasy®

Blood and Tissue kit. The sequences flanking about 1kb 5’ and 3’ from the stop codon of MYL2 gene were PCR amplified using Expand Long Template PCR System from Roche. The cycling protocol for the amplification is as follows:

Step Temperature (0C) Time (s) Cycle Initial denaturation 94 120 1x Denaturation 94 10 Annealing 62 30 10x Elongation 68 150 Denaturation 94 15 Annealing 62 30 20x Elongation 68 150 + 20 per cycle Final elongation 68 480 1x Cooling 4 ∞ 1x

Table 3.3.1-1 PCR amplification protocol for MYL2 homology arms

Adding 5’-LoxP and 3’ LoxP sites to PGK-Hygro selection cassette

MHC-Luc plasmid contains PGK-Hygro cassette and was obtained from Dr. William Sun.

The PGK-Hygro cassette was amplified using long primers named Phlox (forward and reverse), which contain LoxP sequence as well as complimentary bases to the template of interest. For this PCR amplification, Expand High Fidelity enzyme by Roche as well as 5% DMSO additive was used. Product obtained was 1.83kb in length. PCR product of LoxP-PGK-Hygro-LoxP (herein,

“Phlox”) was cut from the gel and extracted using Qiaquick gel extraction kit from Qiagen. The cycling parameters are as follows:

96

Step Temperature (0C) Time (s) Cycle Initial denaturation 94 120 1x Denaturation 94 10 Annealing 55 30 3x Elongation 72 120 Denaturation 94 10 Annealing 65 30 7x Elongation 72 120 Denaturation 94 30 Annealing 65 30 15x Elongation 72 120 + 10 per cycle Final elongation 72 480 1x Cooling 4 ∞ 1x

Table 3.3.1-2 PCR amplification protocol for adding LoxP to PGK-Hygro

Cloning MYL2 PCR product into TOPO vector

Once MYL2 PCR product was obtained at suitable amount, they were cloned into pCR™

2.1-TOPO® vector which was part of Invitrogen’s TOPO® TA Cloning® Kit. Different molar ratio of

PCR inserts: vector ration was tested for efficient transformation into TOP10 chemically competent of TG1 electrocompetent cells. The general methodology is outlined below:

Chemical transformation Electrical transformation 1. Prepare PCR products and mix with commercial 1. Prepare PCR products and mix with commercial vector as per manual vector as per manual 2. Take out 4ul of this mix and put into TOP10 2. Take out 2ul of this mix and put into 50µl TG1 chemically competent cells electro-competent cells 3. Incubate on ice for 30 minutes 3. Transfer mixture to 1mm cuvette 4. Heat shock cells at 420C for 45s without shaking 4. Transform with 10 µF capacitance, 600Ω resistance and 1800V voltage 5. Put back on ice immediately and recover with 5. Within 10s, add 950µl of recovery medium into the 250µl SOC medium cuvette and resuspend well 6. Incubate at 370C for 1hour with 200rpm shaking 6. Transfer to culture tube and incubate tube at 370C for 1hour with 250rpm shaking 7. Spread 100µl of cells per ampicillin selection 7. Spread 100µl of cells per ampicillin selection plate plates and incubate overnight and incubate overnight

Table 3.3.1-3 Protocol of Chemical and Electrical transformations

White colonies were picked from TOPO-MYL2 selection plates and grown overnight at 370C with

200rpm shaking frequency in LB medium. Plasmids were then extracted from the bacterial cells using Qiaprep spin miniprep kit from Qiagen. To confirm successful cloning of MYL2 into pCR™

2.1-TOPO® vector, restriction digest with EcoRI-HF was carried out. True transformant clones will

97

generate 3 bands: 6.278kb, 2.365kb and 3.913kb. Once positive clones are identified, their plasmids were sent for sequencing to check for mutations at MYL2 insert.

Generating targeting vector – Self Assembly Cloning

The targeting vector consists of three parts: (1) homology arms to endogenous MYL2 gene, (2) reporters cassette, and (3) loxP flanked-selection cassette PGK-HygroR. To combine all three parts of the targeting vector, self-assembly cloning (SAC) method as previously reported was employed (Matsumoto and Itoh, 2011). In short, each insert of interest is to be PCR amplified two times using two pairs of primers that would generate “sticky-ends” either only on

5’ end or 3’ end of the insert. PCR amplicons from the same template but different sets of primers would be termed sister amplicons (A and B; C and D; E and F). Each sister amplicon together with its respective primers are outlined in table 3.3.1-4; and the cycling parameters of

PCR reactions used to amplify amplicons are outlined in table 3.3.1-5. All amplifications were done using High Fidelity Phusion Hot Start II enzyme in presence of 5% DMSO additive.

Reporters insert Phlox MYL2 homology arms Primers used Primer 1-2 and 2-1 (A) Primer 3-2 and 4-1 (C) Primer 5-2 and Primer 6-1 (E) Primer 1-1 and 2-2 (B) Primer 3-1 and 4-2 (D) Primer 5-1 and Primer 6-2 (F) Template DNA pEF6-MNCL2A Phlox pCR2.1-MYL2-C1

Table 3.3.1-4 Details of primers and templates used for PCR amplification of inserts for SAC

Sister amplicons would be denatured and reannealed back together such that each insert would have sticky ends on both 5’ and 3’ ends. The sticky ends of first insert would complement the sticky ends of the second insert, those of the second will complement those of the third; and the third will complement the first thus sticking all inserts together to form a nicked vector. Upon successful transformation into bacterial hosts, these nicks would be sealed off and functional nick-free plasmids are generated.

98

Reporters insert Phlox MYL2 homology arms Cycling Initial denaturation: Initial denaturation: Initial denaturation: para- 980C, 240s, x1 980C, 90s, x1 980C, 180s, x1 meters Denaturation: 980C, 30s Denaturation: 980C, 30s Denaturation: 980C, 30s Annealing: 620C, 90s x25 Annealing: 580C, 30s x25 Annealing: 620C, 80s x25 Extension: 720C, 85s Extension: 720C, 60s Extension: 720C, 225s Final extension: 720C, 300s, 1x Final extension: 720C, 300s, 1x Final extension: 720C, 600s, 1x Cooling: 40C, ∞ Cooling: 40C, ∞ Cooling: 40C, ∞

Table 3.3.1-5 PCR amplification parameters of different inserts for SAC

Nicked SAC plasmids were transformed into TOP10 chemically competent cells. For each transformation experiment, 20 colonies or more were picked and screened using restriction digest and DNA sequencing. All picked colonies were first digested with NheI for 2 hours (to show presence of MYL2 homology arms), followed by SalI-HF for 2 hours (to show presence of reporters insert) and finally with NdeI for another 2 hours (to show presence of Phlox). At each digestion step, an aliquot was taken out of the reaction and analyzed by gel electrophoresis. A positive clone would show bands at the following lengths:

1. After NheI: 6.071kb and 4.629kb 2. After SalI-HF: 4.629kb, 3.216kb, 2.855kb 3. After NdeI: 4.629kb, 2.855kb, 1.616kb, 1.600kb

Generation of TALENs expression vectors

TALENs were commercially generated by Life Technologies. They were designed to target the last exon of MYL2 gene just before the stop codon. In this laboratory, each of the forward (F-) and reverse (R-) TALEN construct was cloned into pEF6-expression vectors. To amplify F- and R- TALEN plasmids, 1.5µl of each F- and R- TALEN plasmids was mixed with 25µl

E. Cloni cells. This mixture was transferred immediately into 1cm cuvette and electroporated with the same parameter as previously described (table 3.3.1-3) for electroporating TG1 electro- competent cells. After 1 hour incubation, cells were spread onto Kanamycin selection plates. 10

99

clones from F- and R- TALEN plates were picked for scaling up. The plasmids from these clones were then extracted from the cells using Qiagen miniprep kit, and digested with HindIII and NotI to check for the presence of TALEN inserts. Presence of insert will be indicated by a band at

3.075kb. 100ng of plasmids from positive clones F2 and R2 were used as templates in PCR amplification of the F- and R- TALEN sequences with high fidelity hot start II Phusion polymerase using TAL-5 and TAL-3 primers. The cycling parameter is outlined below:

Step Temperature (0C) Time (s) Cycle Initial denaturation 98 120 1x Denaturation 98 30 25x Annealing 62.5 45 Elongation 72 110 Final elongation 72 300 1x Cooling 4 ∞ 1x

Table 3.3.1-6 PCR amplification parameters to generate F- and R- TALEN PCR products

The resulting F- and R-TALEN PCR products were loaded onto agarose gel for gel electrophoresis and the appropriate band at 3.075kb was cut and gel purified with Promega

Wizard® SV Gel and PCR Clean-Up System. To clone F- and R- TALENs into expression vector, a single “A” overhang was first added onto extracted PCR products using Dream Taq polymerase from Thermo Scientific. This fresh PCR product was mixed with pEF6-TOPO vector from

Invitrogen as per supplier’s manual and transformed into TOP10 chemically competent cells. 10 clones each were picked from F- and R- TALEN selection plates and grown overnight in

2xTY+ampicillin medium at 370C with 200rpm shaking. Plasmids from cultured clones were extracted using Qiagen miniprep kit and further digested with SpeI and NheI to determine if F- and R- TALEN inserts had been cloned in the right orientation into the expression vector.

100

3.3.2. Feeder cell culture

Throughout this project, a number of feeder cells were used: (1) CF1 mouse embryonic fibroblasts (MEFs) (Millipore), (2) hygromycin resistant MEFs (Millipore), (3) DR4 MEFs (ATCC), (4)

Puromycin-resistant MEFs (Stem cell technologies), (5) HFFs (ATCC) and (6) normal human dermal fibroblasts (NHDF) (Lonza). All mouse fibroblasts and human fibroblasts were cultured in

MEF medium and HFF medium, respectively. The compositions of each medium are outlined in table 3.3.2-1.

MEF medium 440ml DMEM with 4.5g/L glucose 5ml 100x Penicillin-Streptomycin-Glutamine 5ml 100x NEAA (100x) 50ml FBS HFF medium 415ml DMEM with 4.5g/L glucose 5ml 100x Penicillin-Streptomycin-Glutamine 5ml 100x NEAA (100x) 75ml FBS Add bFGF to a final concentration of 5ng/ml when above passage 28 Table 3.3.2-1 Compositions of feeder cell culture media

MEF culture – thawing frozen stock

While in Singapore, CF1 MEFs (normal and hygromycin resistant) at p3 was obtained commercially from Millipore, while HFFs were obtained commercially from ATCC. While in

Göttingen, Germany, MEFs were derived from mouse embryos by the animal facility staff and immediately passed to the laboratory in a test tube for propagation.

The protocols for MEFs and HFFs culture are very similar; they only differ in the culture medium compositions. In short, frozen vial of cell was thawed quickly inside a 370C water bath.

Partially thawed vial was then sprayed with 70% ethanol and brought inside sterile biosafety hood. Once the vial was opened, 1ml of pre-warmed culture medium was added drop-wise into the vial, with shaking after addition of each drop. The cell mixture was then aspirated into a 15

101

ml falcon tube. Fresh 8 ml of pre-warmed culture medium was added into the cell suspension and the cells were pelleted down by centrifuging at 250 xg for 4 minutes at room temperature.

At the end of centrifugation, the supernatant was aspirated, and the cells were resuspended by pipetting up and down gently in 5 ml of culture medium. Cell number was counted using hemocytometer. Once the cell number was obtained, the cells were seeded onto tissue culture flask at a density of 10-15x104/cm2. For MEFs and HFFs maintenance, 20 ml and 35 ml of culture medium was added per T75 flask and T175 flask, respectively. Cells were then allowed to grow to 80-90% confluence by incubating in 370C incubator. The culture medium was changed every 3 days for cells maintenance.

Propagation of MEFs

MEFs and HFFs were routinely passaged when they reach about 80-90% confluence.

Briefly, confluent cells were first incubated in PBS (-) (PBS without Ca2+ and Mg2+) for 3 minutes at room temperature to aid cell detachment. Subsequently, cells were digested for 5 minutes at

370C with pre-warmed 0.05% trypsin/versene (5ml and 10ml per T75 and T175 flask, respectively). The trypsin was then deactivated by the addition of pre-warmed fresh MEFs/HFFs culture medium (10ml and 20ml per T75 and T175 flask, respectively), and the cell mixture triturated up and down with pipet to break the cell clumps. Afterwards, cell suspension was transferred to 50 ml Falcon tube and centrifuged at 250 xg for 4 minutes at room temperature.

MEFs and HFFs were seeded back onto tissue culture flasks pre-coated with 0.1% gelatin at a density of 10-15x104/cm2. The culture medium was changed every 3 days for maintenance.

Inactivation of MEFs and HFFs

In Singapore, MEFs and HFFs were mitotically inactivated by Mitomycin C (Kyowa). In

Göttingen, Germany, MEFs and HFFs were inactivated by gamma rays irradiation. Briefly,

102

Mitomycin C (MMC) was resuspended in water to a final concentration of 2 mg/ml and then filter-sterilized. This stock solution was then added into fresh MEFs/HFFs medium to a final concentration of 20 µg/ml.

Beginning with 90%-confluent layer of MEFs and HFFs, spent MEFs/HFFs medium was removed from the flask and replaced with fresh MEFs/HFFs medium containing MMC (10ml and

15ml per T75 and T175 flask, respectively). MEFs/HFFs were incubated in this medium for 2 hours 45 minutes in CO2 incubator for mitotic inactivation. Subsequently, inactivation medium was removed from the flask(s) and the cells were washed 3 times with PBS (-). During the last washing step, the cells were incubated in PBS (-) for 3 minutes before trypsinization with pre- warmed 0.05% trypsin/versene. After cell number was obtained, cells were resuspended at

5x105/ml of freezing medium (90% FBS + 10% DMSO) and aliquoted at 1ml per cryovial. Cryovials were then transferred to -800C freezer in a pre-cooled cell freezing containers (CoolCell) or

Nalgene Mr. Frosty (Thermo Scientific). The next day, frozen vials were transferred to liquid nitrogen tank or -1500C freezer for storage.

For gamma rays inactivation, 90%-confluent MEFs/HFFs were first washed and incubated in PBS (-) for 3 minutes, trpysinized with 0.05% trypsin/EDTA and resuspended in cold

MEFs/HFFs medium at cell density of 8x106 cells/10ml medium. Cell suspension was transferred to 15ml Falcon tubes and kept on ice prior to gamma irradiation. Cells were then irradiated with

60 Gray of gamma rays (MEFs) or 30 Gray of gamma rays (HFFs) at room temperature. Irradiated

MEFs/HFFs were then frozen in freezing medium at -800C overnight before they were transferred to -1500C freezer for storage.

103

3.3.3. Human Pluripotent Stem Cells (hPSCs) culture

Throughout this study, H9 (WiCell), HES3, MBJ5 and MBJ6 (hiPSCs) lines were used.

MBJ5 and MBJ6 were derived from BJ fibroblasts with RNA reprogramming method (Mehta et al.,

2014). HPSCs were cultured on feeder cells (MEFs or HFFs) for maintenance.

Cell culture media composition

hESCs medium 390ml DMEM/F12 5ml 100x NEAA 100ml KOSR 910µl 55mM 2-mercaptoethanol 5ml 100x Penicillin-Streptomycin-Glutamine *Add bFGF to a final concentration of 10ng/ml just before using hESCs recovery medium 25 ml FBS 25 ml hESCs medium without bFGF 200 µl Matrigel hESCs freezing medium mFreSR™ for stock hESCs 90% FBS+10% DMSO for working hESCs

Table 3.3.3-1 Compositions of hESCs cell culture media

Thawing and freezing of hPSCs

A frozen vial of hPSCs was taken out from the liquid nitrogen tank and thawed quickly in

370C water bath. Partially thawed vial was sterilized with 70% ethanol and brought into biosafety hood. 1ml of pre-warmed hESCs medium was then added drop-wise with shaking into the vial and the cell suspension was transferred into 15ml Falcon tube. 8 ml of hESCs medium was further added drop-wise into this cell suspension before centrifugation at 200xg for 4 minutes at room temperature. HPSCs were then resuspended in hESCs medium + 10µM Y27632 (ROCK inhibitor) and plated onto feeder-coated or matrigel-coated culture dishes.

Feeder-dependent propagation and freezing of hPSCs

Mitotically-inactivated feeder cells were seeded on tissue culture dishes pre-coated with

0.1% gelatin at a density of 1.6x105cells/cm2. On the next day, each undifferentiated hPSC colony

104

was cut manually into small grids or squares and each small square of hPSC fragment was transferred onto this feeder layer at a maximum of 50 pieces per 6cm dish. If hPSCs were to be frozen, these pieces were transferred into cryo-vials filled with mFreSR™ at maximum of 70 pieces/vial.

Alternatively, hPSCs were trypsin-passaged onto feeder cells. Briefly, a rho-associated protein kinase (ROCK) signaling pathway inhibitor, Y27632, was added to hPSCs on culture dishes to a final concentration of 10µM and allowed to incubate for 1 hour in the CO2 incubator before trypsinization. Subsequently, hPSCs were washed with PBS (-) for 3 minutes at room temperature before trypsinization with 0.05% Trypsin/Versene or 0.05% Trypsin/EDTA for 4 minutes at 370C. Once hPSCs were dissociated from culture dishes, hESCs recovery media was added to the culture dishes to deactivate the trypsin. Culture dishes were washed with hESCs medium and the cell suspension was passed through 40µm cell strainer before being transferred to a falcon tube for centrifugation at 200xg for 4 minutes. HPSCs pellet was resuspended in hESCs medium + 10µM Y27632 and seeded at 1.2-1.5x104 cells/cm2. The culture medium was changed the next day to hESCs medium without Y27632. The culture medium was added at 5ml per 6cm dish; 10ml per 10cm dish and 15 ml per T75 tissue culture flask. For maintenance, the culture medium was changed daily until hPSCs reached 80%-90% confluence. If hPSCs were to be frozen, a determined cell number was centrifuged and resuspended in mFreSR™ at a density of

2x106/ml. hPSCs were then frozen at 1ml per vial in -800C freezer overnight before transferring to liquid nitrogen tank for storage.

105

3.3.4. Homologous recombination and screening for positive clones

Preparation of F- and R- TALEN plasmids and targeting vector

Transformed Top10 chemically competent cells carrying plasmids with the following inserts of interest were scaled up in 1L LB + 100µg/ml Ampicillin for maxiprep: (1) SAC5 (2) F7-

TALEN and (3) R2-TALEN. Plasmids were extracted using Purelink Hi-pure plasmid filter maxiprep kit and precipitator modules from Invitrogen, Life Technologies and subsequently digested with restriction enzymes to confirm presence of inserts of interest. SAC5 was sequentially digested with NheI-HF, SalI-HF and NdeI, while F-TALEN and R-TALEN were each double-digested with

SpeI and NheI-HF. Once the expected restriction digest profiles were obtained for all three plasmids, 100 µg of SAC5 was linearized with BglII at 370C for 4 hours followed by phenol chloroform extraction. The resulting SAC5 pellet was dissolved in 40µl of sterile water overnight at room temperature. To check if BglII digestion was complete, 5µl of the digestion product was size-fractionated with gel electrophoresis.

HPSCs electroporation for generation of knock-in line

Before electroporation, hPSCs on tissue culture dishes were incubated with 10 µM

Y27632 for 1 hour at 370C. During incubation, a DNA mix was prepared in a total volume of 300

µl PBS (-); this mix consists of 40 µg linearized SAC5, 5 µg F7-TALEN and 5 µg R2-TALEN. At the end of 1 hour incubation with Y27632, hPSCs were washed with PBS (-) for 3 minutes at room temperature before trypsinization with 0.05% trypsin/versene at 370C for 4 minutes.

To inactivate trypsin, three times the volume of hESCs recovery medium was added into each dish. Cell suspension was then passed through 40µm cell strainer before cell counting.

4x107 cells were pelleted down and resuspended in 450µl hESCs medium + 10 ng/ml bFGF +

10µM Y27632. 300µl of the DNA mix was then added into this cell suspension and the resulting 106

mixture transferred to 0.4cm electroporation cuvette for electroporation with the following parameters: 320 V voltage, 200 µF capacitance, ∞ Ω resistance and 4 mm cuvette size.

The experimental time constant and voltage were noted down and within 10 seconds of electroporation, hPSCs were recovered by adding 1 ml of hESCs medium + 10ng/ml bFGF + 10µM

Y27632 (complete hESCs medium) into the cuvette. Electroporated cell suspension was immediately transferred into 7 ml of complete hESCs medium in a Falcon tube and seeded back at a density of 5x106 cells/dish. Each dish was pre-coated the day before with 1x106 hygromycin- resistant MEFs. These dishes were incubated for two days in a CO2 incubator. On day 3 onwards, the culture medium was changed daily to hESCs medium + 10ng/ml bFGF without Y27632 until the hPSCs reached 80% confluence.

Selection and screening for positive clones

Once 80% confluence was reached, the culture medium was changed daily to hESCs medium + 10ng/ml bFGF + 50 µg/ml hygromycin B. HESCs culture was maintained in this way until hygromycin-resistant colonies were big enough for manual passaging.

Each hygromycin-resistant colony was cut manually and the pieces transferred to (1) one

OCD pre-coated with mitotically inactivated MEFs (1.6x105 cells/OCD) for maintenance; and (2) one well of 24 well plate pre-coated with hESCs qualified Matrigel. The pieces on Matrigel- coated plate were allowed to grow until 90% confluence so that their genomic DNA could be extracted in sufficient amount using DNeasy Blood & Tissue Kit from Qiagen. Genomic DNA from each hygromycin-resistant colony was then PCR-amplified using MYL2KI-F and MYL2KI-R primers with Taq DNA polymerase, 5% DMSO and 2mM MgCl2 additive with the cycling parameter (table

3.3.4-1). Knock-in hPSCs should generate a band at 1.794kb (henceforth positive clones) while negative clones would not generate any bands. 107

Step Temperature (0C) Time (s) Cycle Initial denaturation 95 180 1x Denaturation 95 30 Annealing 60 30 30x Elongation 72 110 Final elongation 72 420 1x Cooling 4 ∞ 1x

Table 3.3.4-1 PCR cycling parameters for screening positive knock-in clones

Positive clones were further scaled up for Southern blotting while negative clones were disposed of. For Southern blotting, 15µg of genomic DNA (gDNA) from positive clones was first digested with NdeI at 370C for 4 hours. Subsequently, this digested gDNA was extracted using phenol-chloroform method and reconstituted in 35 µl of water. Reconstituted gDNA was loaded onto 1% agarose gel and run in TBE buffer at 110V for 2½ hours before being soaked in dH2O containing SYBR green for 45 minutes for subsequent DNA bands visualization under UV exposure.

Genomic DNA on the gel was then depurinated, denatured and neutralized as per manufacturer’s recommendations to prepare for DNA transfer. Afterwards, gel was equilibrated in 20x SSC; nylon membrane in 2x SSC and the transfer sandwich was assembled for transfer of gDNA onto nylon membrane overnight. On the next day, gDNA was cross-linked onto the nylon membrane through exposure to UV light at 120mJ before being prepared for hybridization by incubating in Easyhyb DIG buffer (Roche) for 45 minutes at 450C (for Nde500L probe) or 47.50C

(for Nde508L probe).

While the membrane was being primed for hybridization, a hybridization mix was prepared. Briefly, DIG-labeled probe (either Nde500L or Nde508L) was mixed with 50µl of H2O, denatured at 990C for 5 minutes and immediately chilled on ice before being transferred into pre-warmed Easyhyb DIG buffer. This hybridization mix was added onto the membrane at the

108

end of membrane priming step for overnight incubation. On the next day, nylon membrane was washed sequentially with low stringency buffer, high stringency buffer and washing buffer before blocking with blocking solution (Roche) and addition of primary antibody solution (ant-

DIG-AP). Finally, the membrane was washed and equilibrated with detection buffer before the addition of ready to use CDP-star to show bands of interest for imaging.

3.3.5. Removal of Floxed PGK-Hygromycin cassette

Preparation of Cre-recombinase plasmid

Cre recombinase was obtained commercially from Addgene as “Cre-IRES-Puro”

(Catalogue number 30205) and streaked onto LB + 100µg/ml Ampicillin agar plate. To obtain sufficient amount of plasmids for transfection, a colony was picked and scaled up in LB +

100µg/ml ampicillin overnight. Extracted plasmid was confirmed for presence of Cre- recombinase by restriction digest experiment. In short, the extracted plasmids were first double- digested with AscI and NdeI before subsequent digestion with XhoI. Positive clone would yield bands at 7.903kb and 1.329kb after the first double digest and 1.329kb, 1.868kb and 6.035kb after XhoI digestion.

H9V20 transfection with Cre-IRES-PuroR

5x6cm tissue culture dishes were pre-coated the day before with inactivated Puromycin- resistant MEFs at 1x106 MEFs/dish. H9V15 was trpysinized into single cells and seeded onto these culture dishes at seeding density of 3x105 cells/dish. On the next day, the culture medium was changed to hESCs medium + 10ng/ml bFGF without Y27632. Meanwhile, 517 µl of

DMEM/F12 was mixed with 33µl Fugene HD transfection reagent and incubated at room temperature for 5 minutes. 100 µl of this diluted Fugene was set aside in a new tube for control

109

dish while to the remaining, 4 µg of Cre-IRES-Puro plasmid was added. Plasmid-Fugene HD mixture was incubated at room temperature for 40 minutes. Subsequently, 100 µl of diluted

Fugene was added into a control dish while 100 µl of the plasmid-Fugene-HD mixture was added into each of the 4x 6cm dishes drop-wise, with swirling of dish after the addition of every drop.

Time of addition was recorded.

24 hours later, the culture medium was changed to hESCs medium containing 1µg/ml puromycin. This puromycin selection medium was applied again on the 48th hour post- transfection. On the 72nd hour post transfection, the culture medium was changed back to hESCs medium + 10ng/ml bFGF. For hESCs maintenance, the culture medium was subsequently changed every 2 days and culture was maintained until puromycin-resistant colonies were big enough for manual passaging (7days).

Selection for successfully transfected clones

Forty clones were picked and each clone was passaged manually onto (1) one OCD pre- coated with irradiated MEFs, and (2) one well of 24-well plate pre-coated with Matrigel. The culture medium was changed daily and seven days later, cells on Matrigel were trpysinized and their gDNA extracted using Qiagen DNeasy Blood and Tissue Kit.

Three rounds of independent PCR amplification were carried out to check for successful removal of floxed PGK-hygromycin cassette in the knock-in line. The first PCR used Luc-F and qPCR-WT-R primers; a band would show at 2.775kb and 1.021kb for non-excised clones and successfully excised clones, respectively. The second PCR made use of Luc-F and Seq-hyg-R2 primers; a band would show at 1.723kb for non-excised clones, while no bands would show for successfully excised clones. The third PCR was used to check if cre-IRES-puro was integrated into the knock-in line genome. For this third PCR, Puro-F and Puro-R primers were used and cre-IRES- 110

puro integration into the genome would be indicated by a band at 516bp. Each PCR cycling parameters are outlined below:

Step Temperature (0C) Time (s) Cycle Initial denaturation 98 180 1x Denaturation 98 30 Annealing 63 for PCR1; 64 for PCR2 35 30x Elongation 72 84 Final elongation 72 180 1x Cooling 10 ∞ 1x

Table 3.3.5-1 Cycling parameters for PCR1 and PCR2 to screen for successfully excised clones

Step Temperature (0C) Time (s) Cycle Initial denaturation 98 120 1x Denaturation 98 30 Annealing 64 20 30x Elongation 72 30 Final elongation 72 120 1x Cooling 10 ∞ 1x

Table 3.3.5-2 Cycling parameters for PCR3 to check for integration of Cre-IRES-Puro into knock-in line genome

3.3.6. In vitro characterization of knock in lines

Karyotyping and sequencing

Karyotyping was done by The DNA Diagnostic & Research Laboratory (DDRL) at KK

Women’s and Children’s Hospital (KKH). HPSCs were passaged manually onto Matrigel-coated tissue culture dish and this dish was delivered to KKH for analysis.

For sequencing, gDNA was extracted from knock-in hESCs and sent directly for sequencing to 1st BASE or Axil Scientific using primers KIgDNA-F and KIgDNA-R. Alternatively, gDNA was PCR amplified using the same primers, run on gel electrophoresis with band of interest at 2.595kb cut and gel purified before sending it to Seqlab (Germany) for sequencing.

PCR amplification was done with high fidelity hot start Phusion DNA polymerase using 5% DMSO as additive. The cycling parameters are as follows:

111

Step Temperature (0C) Time (s) Cycle Initial denaturation 98 180 1x Denaturation 98 60 Annealing 60 or 63 45 25x Elongation 72 75 Final elongation 72 360 1x Cooling 4 ∞ 1x

Table 3.3.6-1 PCR cycling parameters for gDNA sequencing

Pluripotency markers – immunofluorescence staining and FACS

Knock-in hPSCs cultured on mitotically-inactivated MEFs on OCDs were fixed in 4% paraformaldehyde (PFA) at room temperature for 15 minutes. The colonies were washed three times with PBS (-) before incubation with blocking buffer for 30 minutes at room temperature.

For Tra-1-60 and SSEA4, the blocking buffer without Triton-X100 was used while for Oct3/4,

Nanog and Sox2, the blocking buffer containing 0.1% Triton-X100 was used. The blocking buffer consisted of 5% goat serum + 1% bovine serum albumin (BSA) in PBS (-).

Primary antibodies were then diluted in blocking buffer to desired concentration (table

3.3.6-2) and incubated with the hPSCs at 40C overnight. The next day, the dishes were washed three times with PBS (-) before incubation with Hoechst 33342 and secondary antibody for 1 hour at room temperature. Hoechst counterstain and secondary antibodies (goat anti rabbit

Alexa 546, goat anti mouse Alexa 488) were all diluted 1:1000 in blocking buffer. Subsequently, dishes were washed three times with PBS (-) and mounted with coverslips for imaging.

Pluripotency markers – FACS analysis

The procedure of staining for FACS analysis is very similar to immunofluorescence staining protocol. The difference lies in the cells being stained as colonies for immunofluorescence and as single cell suspension for FACS analysis.

112

Primary Antibody Dilutions for Dilutions Host Composition of Immunofluorescence for FACS species Blocking buffer Tra-1-60 1:50 1:100 Mouse Oct3/4 1:50 1:100 Mouse PBS(-) Nanog 1:50 1:100 Rabbit 5% goat serum SSEA4 1:200 1:400 Mouse 1% BSA Sox2 1:50 1:100 Rabbit With or without IgG isotype control 1:1000 1:1000 Mouse 0.1% Triton X-100 IgG isotype control 1:2000 1:2000 Rabbit

Table 3.3.6-2 List of primary antibodies used for pluripotency immunofluorescence staining

3.3.7. Validation of gene expression and transcription of MYL2 knock-in reporter line

All reporter genes were analyzed at transcript level as well as protein level (except for

NeoR). Analysis was done either in monolayer (2D) or in EHMs (3D) format. For EHMs format, wild type H9 hESCs as well as knock-in lines H9V15 and H9VC16 were differentiated to CMs and cultured on culture dishes until day 18. CMs were digested into single cells on day 18, mixed with human foreskin fibroblasts (HFFs) in collagen I matrix to form EHMs in the molds. EHMs were transferred to auxotonic stretchers on day 21 and cultured in SFMM as described in chapter 2.

Total RNA was extracted at each different time points from day 18-90 of differentiation and expressions of MYL7, MYL2, mChr, NeoR, G.Luc and CASQ2 were analyzed with qPCR. Ct values from all samples were normalized to their respective GAPDH expression. MChr protein expression was analyzed by fluorescence microscope while G.Luc protein expression was analyzed using luciferase assay.

Microscopy analysis of knock-in CM in 2D and 3D format

Differentiated CMs from H9 and H9V15 or H9VC16 either in 2D or EHM format were analyzed by fluorescence microscope for mChr expression from day 18-90 of differentiation. 2D analysis was done with an Xcellence rt (Olympus) microscope whereas 3D analysis was done with a SteREO Lumar.V12 (Zeiss). 113

Live cell FACS of knock-in CMs

Differentiated CMs from H9V15 and H9VC16 were cultured on Matrigel until day 30 of differentiation and trypsinized into single cell suspension for live cell FACS sorting. MCherry positive population was gated and sorted for. Total RNA was extracted from mChr (+) and mChr(-)

CMs and qPCR was performed to analyze MYL7, MYL2 and CASQ2 expression levels in each sample. Ct values were normalized to the respective GAPDH expression level in each sample.

Investigation for effective Neomycin concentration

H9V15 and H9VC16 knock-in hESCs were differentiated to CMs and cultured to day 15 of differentiation. On day 15, CMs were digested into single cells and reseeded on Matrigel-coated dishes at density of 2.5x105 cells/well of a 12-well plate in B27 (+) medium. On day 27 of differentiation, selection medium containing different concentrations of G418 was applied to the

CM monolayer (0 - 50 µg/ml). Cells were cultured in this selection medium for a week; with medium change every other day. On day 35 of differentiation, selection medium was changed back to B27 (+) without G418 and CMs were cultured for additional 9 days, with media change every other day. On day 44 of differentiation, cells were imaged by fluorescence microscopy for mChr signal, before trypsinization and total RNA extraction. Finally, qPCR analysis was carried out for MYL7 and MYL2. All qPCR data was normalized to respective GAPDH expression level in all samples.

Luciferase assay

EHMs as well as CM monolayers from H9 wild type and knock-in H9V15 and H9VC16 were cultured until day 90 of differentiation. Culture media were changed every three days and media samples were collected at day 21, 24, 27, 30, 45, 60, 75 and 90 of differentiation. Assay was carried out in white 96-well plate using GloMax® 96 microplate luminometer with dual

114

injectors from Promega. BioLux® Gaussia Luciferase Flex assay kit from NEB was used and assay was carried out as per manufacturer’s recommendations. In short, 20 µl of medium sample was pipetted into each well of the 96-well plate and the injector was set with the following parameters: 50 µl of injection volume, delay between injection and measurement of 40s and 10s of integration time.

3.3.8. Statistical analysis

Data is presented as mean ± standard error mean (SEM). Statistical analyses were performed with Student’s t-test (2-tailed) or ANOVA when applicable followed by post-hoc

Bonferroni's Multiple Comparison test. Results were considered to be significant at p ≤ 0.05 (*); very significant at p ≤ 0.01(**) and extremely significant at p ≤ 0.001 (***). Correlation studies were carried out using Prism 5.0.

115

3.4. Results

3.4.1. Cloning of Targeting Plasmid

MYL2 and Phlox PCR products

PCR amplification for MYL2 homology arms was most efficient with long template PCR

DNA polymerase (Roche) and 5% DMSO additive. Product obtained was 2.347kb in length (figure

3.4.1-1 and -2). “Phlox” insert was obtained readily with High Fidelity (HF) Expand DNA

Polymerase (Roche) and gave a bright band at 1.83kb (figure 3.4.1-3).

HF Expand Roche HF Long DMSO Expand Template 3 (-) (+) (-) (+) Roche PCR Roche 2 3 2 1.5 1.2 1 1 0.5

0.5 0.4

Figure 3.4.1-1 Addition of DMSO improved Figure 3.4.1-2 MYL2 PCR was more efficient MYL2 PCR amplification with Long Template PCR DNA Polymerase (Roche)

To clone MYL2 PCR product into the expression vector pCR™ 2.1-

3 TOPO® vector (Invitrogen), four different insert: vector molar ratios 2 1.5 1 were tested (1:1, 5:1, 10:1, 50:1). Of the four ratios tested, 5:1 ratio

gave the most number of transformants on the selection plates. Ten

clones each were picked from chemically competent and electro-

competent groups and scaled up for plasmid analysis. Figure 3.4.1-3 LoxP-PGK-Hyg- LoxP PCR product

116

6 C1 C2 C3 C4 C5 6 C6 C7 C8 C9 C10 5 5 4 4 3 3

E1 E2 E3 E4 E5 E6 E7 E8 E9 E10 6 6 5 5 4 4 3 3

Figure 3.4.1-4 pCR2.1-MYL2 after digestion with EcoRI

The PCR result showed that the chemically competent group (C1 – C10) showed three positive transformants (in red) out of ten picked clones while the electrically competent group

(E1 – E10) showed one positive transformant out of ten picked clones. Clones C1, C3, C5, E1, E3 and E5 were sent for sequencing to check for mutations.

Sequencing result showed that clone C1 from chemically competent group had no mutations and therefore it was used for subsequent experiments (henceforth pCR2.1-MYL2-C1).

Targeting construct generation by self-assembly cloning (SAC)

Strictly following published SAC protocol did not yield desired construct. It was determined empirically that denaturation and annealing of sister amplicons had to be done in cycles (ten cycles), instead of one cycle as per published protocol. In addition, denaturation period had to be extended and an extra annealing step at 700C had to be added before the slow cooling step. A further incubation at 40C for 2 hours was also found to be necessary for successful SAC. 117

Of 48 transformant clones picked, only one clone showed expected restriction digest profile (SAC5). Illustrated on the next page is the diagram of SAC5 together with its restriction digest result. SAC5 was sent for DNA sequencing and the results showed no mutations on SAC5.

Thus it was used for subsequent homologous recombination to generate the knock in line.

Sequential digest should yield fractions:

1. After NheI: 6071bp and 4629bp 2.After SalI-HF: 4629bp, 3216bp, 2855bp

3. After NdeI: 4629bp, 2855bp, 1616bp, 1600bp

10000 Nhe I Sal I Nde I 8000 6000 5000 4000 3500 3000 2500

1500 1000 750

Figure 3.4.1-5 Restriction digest profile of SAC5 Figure 3.4.1-6 Plasmid map of SAC5

Generation of TALEN expression vector

Figure 3.4.1-7 shows the diagram of a TALEN plasmid (forward TALEN has the same map as reverse TALEN). After transformation into TOP10 chemically competent cells, 9 transformants each were picked for F- and R-reverse TALENs. These transformants were scaled up and their plasmids extracted for a double digest with HindIII and NotI to confirm the presence of TALEN insert. Positive transformants would yield 2 bands at 3.057kb (TALEN insert) and 2.582kb (other half of the plasmid).

118

3.5 3 5.6 kb 2.5 2

1 FokI TAL repeats Figure 3.4.1-8 HindIII and NotI double digest Figure 3.4.1-7 TALEN plasmid map showed TALEN insert at 3.057 kb

Once positive clones were identified, F- and R- TALENs were PCR amplified using the F7 R2 positive plasmids as template, the expected 3.075kb band was cut, gel extracted and purified. 3´

Adenine overhang was then added to this extracted PCR products before they were mixed with pEF6-TOPO vector. pEF6-F- and R- TALENs were transformed into TOP10 chemically competent cells. Shown in figure 3.4.1-9 is the diagram of pEF6-TALEN vector. Constructs inserted in the right orientation would generate a band at 2.288kb while those inserted in the wrong orientation would generate a band at 852bp. This is because there are 33 bases from SpeI digest site to the start of TALEN sequence and there are 819 bases from NheI digest site to the end of

TALEN sequence.

For F-TALEN construct, one out of ten (F7) picked clones had the construct inserted in the right orientation (positive clones), while for R-TALEN construct, there were six positive clones out of ten picked clones (R1, R2, R4, R5, R8, R9). Plasmids from F7 and R2 were sent for sequencing and the result showed no mutations in both sequences. Thus, pEF6-TALEN-F7 and pEF6-TALEN-R2 were the TALEN expression vectors used for homologous recombination.

119

F7 R2

00 10000 80 1 0 0 8000 0 6000 5000 S

p e

I

SpeI 4000 0 0 0 7 pEF6-TALEN 3500 2

0

0

0

pEF6-TALEN-F (expression 3000

construct)8923 bp 2500 8.923kb 2000 6 1500 0 0 0

t 1000

c u r t s 0 n o 0 c 0 750 - F 3 N E L TA 500

5 00 250 0 4000

eI Nh NheI Figure 3.4.1-10 TALEN restriction digest profile showing right orientation in Figure 3.4.1-9 pEF6-TALEN plasmid map expression vector

3.4.2. Human pluripotent stem cells (hPSCs) culture

HPSCs colony morphology

Shown in figure 3.4.2-1 is the typical morphology of manually-passaged human pluripotent stem cells (hPSCs) colony grown on mitotically inactivated mouse embryonic fibroblasts (MEF) feeder layer: the colonies look roundish and flat, with clear edges. Manually- passaged hPSCs grown on Human Foreskin Fibroblasts (HFFs) look slightly different, a little bit irregularly shaped instead of roundish, although still flat and having distinct edges.

Trypsinized hPSCs grow at a faster rate than mechanically passaged hPSCs and shown in figure 3.4.2-3 is the typical appearance of trypsinized hPSCs grown on inactivated MEFs feeder.

120

Figure 3.4.2-1 Manually cut hPSC colonies on Figure 3.4.2-2 Manually cut hPSC colonies on inactivated MEFs, ready for passaging inactivated HFFs, ready for passaging

Figure 3.4.2-3 Trypsinized hPSCs on inactivated MEFs, ready for passaging

3.4.3. Generation of knock-in line

Validation of TALEN expression constructs and the targeting vector

Before electroporating hESCs with F- and R- TALENs as well as the targeting vector, a validation round of restriction digest was carried out to make sure that the right vectors were used. Figure 3.4.3-1 below shows that the restriction digest profiles of pEF6-F7-TALEN and pEF6-

R2-TALEN as well as targeting vector SAC5 are as expected. Figure 3.4.3-2 shows that linearization of SAC5 was complete prior to electroporation.

121

Bgl-II F7 R2 NheI SalI NdeI -HF (-) (+) (+)

10 kb

Generuler 1kb DNA ladder Generuler 1kb DNA ladder Figure 3.4.3-1 Restriction digest profile of TALEN F7, TALEN R2 and SAC5 before electroporation showed Figure 3.4.3-2 BglII completely expected bands linearized SAC5

HPSCs electroporation

Different hPSCs line used gave rise to different numbers of hygromycin B-resistant colonies. Tabulated below are the experimental values of voltage of time constant of each electroporation experiment. HPSCs were viable after electroporation (figure 3.4.3-4) and reached 80-90% confluence in 4-5 days (figure 3.4.3-3). Hygromycin B selection caused extensive cell death, but resistant colonies started to be visible 5 days after hygromycin selection (figure

3.4.3-6). Good-sized hygromycin-resistant colonies, such as shown in figure 3.4.3-5 were manually passaged and propagated for maintenance and PCR screening.

hPSCs line Experimental Experimental Time Voltage (V) constant (ms) MBJ5 317 3.8 Hes3 319 3.5 H9 318 3.6

Table 3.4.3-1 Experimental values of electroporation done on different hPSCs lines

122

Figure 3.4.3-3 Electroporated hPSCs just before Figure 3.4.3-4 Maintenance of knock-in hESCs and selection of positive clones Hygromycin B selection

Figure 3.4.3-6 Electroporated hPSCs 5 days after Figure 3.4.3-5 Electroporated hPSCs 12 days after Hygromycin B selection Hygromycin B selection, ready for manual passaging

Efficiency of homologous recombination (HR)

Listed in table 3.4.3-2 is the summary of 3 independent efforts in generating knock in line from various hPSCs lines. Only very few MBJ5 clones could be scaled up as majority of MBJ5

PCR positive clones senesced after a few passages.

Colonies PCR positive Efficiency of HR Colonies scaled up for Colonies positive by hPSCs picked colonies through PCR (%) southern blot southern blot MBJ5 162 44 27.2 3 2 Hes3 11 1 9.09 1 1 H9 75 4 5.33 4 4

Table 3.4.3-2 Summary of homologous recombination efficiency

123

As can be seen from table 3.4.3-2, the HR efficiency varied widely across different attempts (5.33% - 27.2%). Nevertheless, each round successfully generated at least one positive

HR clone as screened through PCR. Shown below is the diagram of endogenous MYL2 gene in the knock-in line and wild type hPSCs. DIG labeled probe was designed to target the 3’ end of the homology arm for southern blot analysis. Therefore, NdeI-cut knock-in gDNA would give a band at 2.2kb while the wild type NdeI-cut gDNA would give a larger band at 6.3kb.

NdeI NdeI NdeI NdeI Knock in line

2.223 kb Wild type NdeI

Nde NdeI 6.352 kb NdeI DIG labeled probe

Figure 3.4.3-7 Schematic diagram showing hybridization site of Nde500L and Nde508L to gDNA of knock in line

MBJ5 MBJ5 H9 H9 WT V43 V74 WT V43 V74 6 WT V15 V20 V48 V55 WT V15 V20 V48 V55 5 6 4 3 3 2 2.5 1.6 2 1.5 1

1

Figure 3.4.3-8 Southern blot of Figure 3.4.3-9 Southern blot of H9 knock-in clones. All MBJ5 knock in clones V43 and screened clones have both knock-in and wild type alleles V74 showed that they contain (heterozygous knock-in) both knock-in and wild type allele (heterozygous knock-in)

124

MBJ5 HES3 Out of 44 MBJ5 knock-in clones (positive through PCR WT V167 WT V9 MBJ5 HES3 WT V167 WT V9 screening), 3 were scaled up for southern blot: V43, V74 and 6

3 V167. Figure 3.4.3-8 shows that V43 and V74 are both 2.5 2 heterozygous knock-in, showing bands corresponding to both 1.5

1 wild type allele (6.352kb) and knock-in allele (2.223kb). V167

did not show any bands for either wild type or knock-in allele.

This could be due to the insufficient amount of starting gDNA

Figure 3.4.3-10 Southern blot of used for Southern blotting. MBJ5 V167 and hes3 knock-in clones

V167 senesced after 10 passages and could not be further propagated. V43 and V74 also grew very slowly, taking up more than 1 month to reach confluence, whereas HES3 and H9 took

1 week. This slow growth rate of the generated knock-in cells prevented their use for downstream experiments.

Homologous recombination was repeated, this time using HES3 as the starting population. Unfortunately, HES3 generated only one knock-in clone (H3V9) and it was deemed too risky as this clone might carry mutations. While waiting for karyotyping and sequencing result of H3V9, HR was repeated using H9 hESCs. Four putative clones from H9 screened positive through PCR and the successful knock-in was confirmed by Southern blotting. All four

H9 knock-in clones were heterozygous, containing both wild type and knock in alleles (figure

3.4.3-9).

125

3.4.4. PGK-Hygromycin cassette was removed using Cre-recombinase

When knock-in CMs were first analyzed, it was noted that the reporters’ expression was

not as strong as expected. There have been reports which suggested that the presence of strong

constitutive promoter driving the expression of selection marker might interfere with the

endogenous enhancer/promoter elements and affect the physiological expression of the gene or

neighboring genes (Calegari and Waskow, 2014). This possibility prompted the use of Cre-

recombinase to remove the floxed PGK-Hygromycin cassette in the knock-in line. H9V15 was

transfected with Cre-IRES-Puro as outlined in the methods section. Puromycin-resistant colonies

were propagated for PCR to identify colonies in which the PGK-Hygromycin selection cassette

has been successfully removed.

40 colonies were screened with PCR and out of these 40, 38 showed a band at 1.021kb

for PCR1, indicating that the PGK-Hygromycin cassette had been floxed out in these clones. For

PCR2, gel appeared mostly empty, very faint bands were seen at 1.723kb for 18 clones,

indicating that in these clones, the PGK-Hygromycin cassette had not been removed. Combining

this result with result from PCR1, this means that 18 of these clones were not pure, that they

contain some non-floxed cells. First round of PCR3 showed that all clones had Cre-IRES-Puro

plasmid integrated into the genome.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 3 33 34 35 36 37 38 39 40 SAC5

6 3 2 1.5 1

Figure 3.4.4-1 Result of PCR1 from all clones with SAC5 as negative control

126

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 SAC5

6 3 * * * * * * * * * * * * * * * * * * 2 1.5 1

Figure 3.4.4-2 Result of PCR2 from all clones with SAC5 as negative control. Faint bands are labeled with asterisks

2 8 9 11 12 13 14 15 17 19 21 22 4 5 6 7 23 24 27 28 29 30 31 33 34 35 40 SAC5 2 25

1 0.8

0.5

0.3 Figure 3.4.4-3 Result of PCR3 from selected clones with SAC5 as negative control

Colony 16, henceforth H9VC16, was chosen for downstream experiments. Shown below are gel pictures showing results of PCR1-3 from H9VC16, indicating that its PGK-Hygromycin cassette had been floxed out, with minimal (or possibly, negative) integration of Cre-IRES-Puro in the genome (since H9V15 also showed a very faint band at 500bp, this might be an unspecific band).

C16 H9 Cre- C16 SAC5 C16 SAC5 V15 IRES- Puro 6 2 1.5 3 1 2 0.8 1.5 1 0.5 0.4 0.3 PCR1 PCR2 PCR3 Figure 3.4.4-4 Result of PCR1 and PCR2 from H9VC16 indicated that Figure 3.4.4-5 Result of PCR3 from PGK-HygroR cassette has been H9VC16 indicated that there is no successfully floxed out genomic integration of Cre-IRES-Puro

127

3.4.5. Knock-in lines are pluripotent and have no chromosomal aberrations

Knock in lines have normal karyotypes

Shown on the left is the karyotype of the

knock-in line derived from HES3: H3V9.

H9VC16 was also sent for karyotype analysis.

Analysis undertaken by the Cytogenetics

Laboratory in KKH indicated that there are

no major chromosomal aberrations both in

Figure 3.4.5-1 H3V9 karyotype H3V9 and H9VC16. gDNA sequencing gave inconclusive result

Genomic DNA from H9V15 and H9VC16 were sent for sequencing. The resulting sequences, however, did not align 100% with the expected sequence. The results are tabulated below:

Cell line Match to expected sequence H3V9 98% using KIgDNA-F and 97% using KIgDNA-R H9V15 98% using KIgDNA-F and 97% using KIgDNA-R H9V20 99% using KIgDNA-F and 97% using KIgDNA-R

Table 3.4.5-1 Percentage match of knock-in gDNA to expected sequence

H3V9 H9V15 H9V20 Position Mutation Position Mutation Position Mutation 1228 GAC 1172 TG 1203 CT 1246 Insertion of T 1228 Insertion C or G 1226 GA 1252 Insertion of G 1230 GA 1233 Insertion A 1265 Insertion of C 1258 Insertion A or G 1241 CG 1281 C  T 1261 Insertion C or G 1243 Insertion T 1314 T  G 1264 GC 1252 Insertion G 1331 C  G 1281 CT 1256 GA 1353 Insertion of G 1282 TC 1259 Insertion G 1371 AC 1287 Insertion T 1267 Insertion A 1385 CT 1310 AC 1275 Insertion C 1312 Insertion T or C 1279 AT

128

1334 CT 1281 CT H3V9 continued H9V15 continued H9V20 continued Position Mutation Position Mutation Position Mutation 1366 CG 1282 TC 1371 AC 1294 Insertion C 1377 CG 1298 GA 1385 CT 1301 Insertion A 1394 CG 1317 CG 1321 TA

Table 3.4.5-2 List of mutations from direct sequencing of gDNA. Highlighted in red are the mutations that are common between the three different reporter cells lines

This experiment was not repeated due to time constraints, as at the time, the author had to relocate to Germany for the second half of the study.

Knock-in lines are pluripotent

Shown below are the representative results of knock-in lines stained with pluripotency markers.

The results indicated that despite having gone through a number of manipulations, the knock-in lines still maintained their pluripotency.

129

Anti- 100x total magnification 200x total magnification FACS data body

3.9%

Ms IgG control

3.5%

Rb IgG control

94.1%

Oct3/4

98%

Sox2

130

86.8%

Nanog

91.9%

Tra-1-60

99.6%

SSEA4

Figure 3.4.5-2 Knock-in hESCs are positive for pluripotency markers: OCT3/4, SOX2, NANOG, TRA-1-60 and SSEA4

131

3.4.6. MCherry (mChr) was expressed in knock-in cardiomyocytes

First attempts at investigating mChr expression from day 10 – 14 of differentiation had failed to detect its expression and it was thought that the knock-in line was not functional.

However, when 43 days old EHMs (figure 3.4.6-1) were observed by chance under a Lumar microscope, mChr expression was observed, leading to a new understanding that the failure of previous attempts stemmed from the fact that MYL2 expression appeared later in human CMs.

Thus a time series experiment was carried out in which mChr expression was analyzed every three days starting from day 18 – 60 of differentiation. Day 90 EHM videos are provided in the supplementary CD with this thesis.

Figure 3.4.6-1 Fused EHMs, showing positive mCherry signal. Scale = 1mm

132

A D

B E

C F

Figure 3.4.6-2 EHMs under 561 nm excitation (left) and bright field (right) at different time points. MCherry expression increased with time in culture. Scale = 1 mm. A – H9 EHM on day60 D – V15 EHM on d30 B – V15 EHM on day24 E – V15 EHM on d45 C – V15 EHM on day27 F – V15 EHM on d60

133

MChr expression was observable from day 24 of differentiation in EHMs, or 6 days after

CMs digestion into single cells and EHMs casting (figure 3.4.6-2). Interestingly, mChr expression level seemed to indicate forces generated by the corresponding EHMs. Older EHMs expressed brighter mChr signal qualitatively and they also beat with significantly higher forces (figure 3.4.6-

4). Day 90 EHMs beat with the highest forces at maximum of 1.96 ± 0.156 mN, followed by day

60 EHMs at 1.38 ± 0.127 mN, day 45 EHMs at 1.06 ± 0.087 mN and lastly day 30 EHMs at 0.271 ±

0.019 mN (figure 3.4.6-3). Day 90 EHMs corresponded to 72 days post EHM casting, since

EHMs were casted on day 18 of differentiation.

2.5

d90 2.0 d60 d45 1.5 d30

1.0 Force (mN)

0.5

0.0 0 1 2 3 4 [Ca2+] in mM

Figure 3.4.6-3 H9V15 EHM contractile forces increased with time in culture

Figure 3.4.6-4 Maximum H9V15 EHM forces at different time points. P-value was calculated using d30 as standard 134

mCherry Excitation = 587 nm Emmision =610 nm

Figure 3.4.6-5 Lambda scan of H9V15 CM indicated that emission signal is specific for mCherry

To assess whether the mChr signal was specific, lambda scan was performed with a confocal microscope. Emission was observed at the expected wavelength around 610 nm (figure

3.4.6-5).

CMs trypsinized on day 18 of differentiation were also seeded back onto Matrigel at high and low density for prolonged in vitro monolayer culture until day 90 of differentiation. CMs cultured in different cell densities showed different morphologies, as well as mChr expression pattern (Figure 3.4.6-6).

On day 45 and day 60 of differentiation, CMs were digested into single cells and subjected to live-cell sorting with FACS for mChr expression analysis. At both time points, the mean fluorescence intensity of CMs derived from H9VC16 was significantly higher than those derived from H9V15 (figure 3.4.6-7), indicating that removal of PGK-hygromycin cassette resulted in higher expression of reporter gene, in this case mChr. Strong CM contractions 135

resulted in the CM monolayer detaching from the culture surface around day 80 of differentiation, thus there was no live-cell FACS data for day 90 monolayer CMs.

561 nm excitation brightfield overlap

A

B

Figure 3.4.6-6 mCherry expression in H9 CMs and H9V15 CMs reseeded back at A – High density and B – Low density. Scale = 50 µm

Figure 3.4.6-7 Mean fluorescence intensity of H9VC16-CMs is higher than that from H9V150CMs both on day 45 and day 60 of differentiation 136

Days of H9 wild type CMs H9V15 knock-in CMs H9VC16 knock-in CMs differentiation

7.7% 40.3% 45.6%

Day 45

1.2% 56.5% 70.6%

Day 60

Table 3.4.6-1 mCherry live FACS of CMs at different time points. MCherry expression increased with time in culture

137

3.4.7. Neomycin resistance gene was expressed in knock-in cardiomyocytes

Knock-in CMs from H9V15 and H9VC16 showed G418 resistance up to 50 µg/ml (figure

3.4.7-1). Extensive cell death was observed in wells with higher concentration of G418, with only very bright mChr (+) cells surviving ≥30 µg/ml G418. Weak mChr (+) cells could only tolerate a maximum of 20 µg/ml G418.

Figure 3.4.7-1 H9VC16 CMs after G418 selection. Scale bar = 50µm. CMs which survived G418 selection ≥30 µg/ml expressed bright mCherry signal 138

3.4.8. Gaussia Luciferase was expressed in knock-in cardiomyocytes

Days of differentiation

Figure 3.4.8-1 Gaussia Luciferase expression in EHMs at different time points

Knock-in CMs secreted G.Luc into the culture medium, thus generating luminescence when appropriate substrate was added. G.Luc expression increased markedly in the first 9 days post EHMs casting and peaked at day 45 of differentiation (day 24 of differentiation) at 3.3x106

RFU for H9VC16 CMs and 2.25x106 RFU for H9V15 CMs. Knock-in CMs cultured in monolayer format also show similar trend for G.Luc expression, although the sharpest increase in expression was delayed from day 30 to day 45 of differentiation (appendix 4.1).

139

3.4.9. Expression of reporter genes was correlated to the endogenous expression level

of MYL2 gene

10

1

0.1 MYL7MLC2a

MYL2MLC2v mCherry 0.01 NeoR

Relative expression to GAPDH GLuc

0.001 18 30 45 60 90 Days post differentiation

Figure 3.4.9-1 Relative expression of genes in H9V15 EHMs at different time points. Reporter genes expression are significantly correlated with MYL2 gene expression

Correlation coefficient Pair R2 P value Significance ( Pearson r) MLC2v and mCherry 0.9335 0.8714 0.0021 ** MLC2v and NeoR 0.9882 0.9765 <0.0001 *** MLC2v and G.Luciferase 0.7812 0.6102 0.0381 * MLC2v and MYL7 0.8759 0.7672 0.0097 ** mCherry and NeoR 0.9750 0.9506 0.0002 *** mCherry and G.Luciferase 0.9460 0.8948 0.0013 ** NeoR and G.Luciferase 0.8563 0.7332 0.0139 *

Table 3.4.9-1 Correlation of reporter genes to MYL2 and to each other are significant

Figure 3.4.9-1 shows the qPCR result of MYL2, MYL7, together with the three reporter genes from EHMs harvested at different time points. MYL2 started to be expressed on day 18, at about 0.04x the expression level of GAPDH. This expression level was about 50x lower than the expression of MYL7. There was a dip of expression from all genes assessed on day 21. From day

21 – 27, MYL2 expression increased at a sharper rate than MYL7; from day 27 – 30, MYL2

140

expression remained stable while there was a dip in MYL7 expression. Expression of all genes increased steadily from day 30 – 45 of differentiation, with MYL2 expression during this period being 2x higher than the MYL7 expression. There was a decrease in all genes expression from day 30 – 45 of differentiation, with MYL7 decreasing at a higher rate than MYL2. Finally, all genes expression increased steadily to reach a level of expression similar to that observed between day

45 and day 60 of differentiation. In addition, the graph also showed that the expressions of MYL2 and reporter genes are very closely correlated. This correlation was found to be significant with p-values ranging from 0.0381 to <0.0001 (table 3.4.9-1). The expression of each reporter gene was also significantly correlated to each other with p-values ranging from 0.0139 to <0.0001

(table 3.4.9-1). Similar trend was obtained when all gene expression data was normalized to both GAPDH as well as the gene’s respective fold change on day 18 (appendix 4.2).

3.4.10. MYL2 positive cells (ventricular cardiomyocytes - vCMs) can be enriched by live

FACS sorting as well as Neomycin selection

To determine if FACS sorting for mCherry signal is efficient for vCMs enrichment, mChr(+) and mChr(-) fractions from H9V15 and H9VC16 CMs were sorted and their total RNA was extracted. Figure 3.4.10-1 showed relative expression levels of a number of marker genes from

H9V15 CMs, normalized to GAPDH. The data implied that vCMs sorted out were still immature, as they still express a high level of MYL7. MCherry FACS sorting was also specific for vCMs as the

MYL2 expression in mChr(+) cells was significantly higher than in mChr(-) cells. The sorting process was very stringent as cells expressing low levels of mChr were gated out. This was evident from the fact that the mChr(-) cells still showed detectable expression of MYL2. Finally,

141

reporter gene expressions were expressed at a similar level to MYL2. Similar level of marker genes expression was observed with H9VC16 CMs (figure 3.4.10-2).

*** 10 * *** *** *** mChr (+) mChr (-) 1 *

0.1 Relative gene expressiongene Relative 0.01 MLC2a MYL2 mChr NeoR G.Luc CASQ2

Figure 3.4.10-1 Relative mRNA expression levels in mChr(+) and mChr(-) CMs sorted from H9V15 CMs. MChr(+) CMs expressed significantly higher level of MYL2 than mChr(-) CMs. Data is normalized to GAPDH

* *** * 10 mChr (+) mChr (-) 1 * 0.1 Relative gene expressiongene Relative 0.01 MLC2a MYL2 mChr NeoR G.Luc CASQ2

Figure 3.4.10-2 Relative mRNA expression levels in mChr (+) and mChr (-) CMs sorted from H9VC16 CMs. MChr(+) CMs expressed significantly higher level of MYL2 than MYL2(-) CMs. Data is normalized to GAPDH

142

Total RNA from surviving H9VC16 CMs post G418 selection was extracted and converted to cDNA for qPCR of MYL7 and MYL2. The result indicated that there was significant enrichment for vCMs with increasing G418 concentration from 10 µg/ml to 50 µg/ml. Averaged MYL2 expression was significantly higher in the CMs monolayer treated with 30µg/ml and 50µg/ml

G418 than with 10µg/ml. The analysis also indicated that CMs were still relatively immature, as they still expressed significant levels of MYL7.

50 * * * * 10 10 µg/ml 20 µg/ml 5 30 µg/ml 40 µg/ml 50 µg/ml 1 Relative expression gene to GAPDH MLC2a MLC2v MYL7 MYL2

Figure 3.4.10-3 Relative mRNA expression levels in selected H9VC16 CMs. Significant enrichment for vCMs was obtained with 30 µg/ml and 50 µg/ml G418. Data was normalized to GAPDH.

143

3.5. Discussion

In the present study, MYL2 knock-in reporter lines were generated from 3 different hPSC lines using homologous recombination (HR) technique with the aid of a genome editing tool,

TALENs. The cloning of targeting vector was accomplished successfully using self-assembly cloning method (Matsumoto and Itoh, 2011), with slight modification of the denaturing and re- annealing step of sister amplicons.

HR efficiencies differed depending on the stem cell line used, with H9 showing the lowest HR efficiency at 5.33% from 75 clones screened. HES3 HR efficiency was higher at 9.09% from 11 hygromycin resistant clones. MBJ5 showed the highest HR efficiency at 27.2% from 162 clones screened. It must be pointed out that there were more than 75 hygromycin resistant clones for H9 HR experiment. Thus the low efficiency of 5.33% might be attributed to the higher frequency of negative clones being picked from the parent dishes for propagation and/or not enough clones being screened. The same cannot be said for hES3 HR experiment, however, as there were only a total of 11 hygromycin resistant clones from the HES3 HR experiment. This low number might be attributed to the fact that only 2.5x107 cells were used, in comparison to

4.0x107 cells for the other two hPSC lines.

Initially, the MYL2 reporter knock-in line was to be derived from MBJ6, the human induced PSCs (hiPSCs). However, for yet unknown reasons, the knock-in lines derived from MBJ6 stopped proliferating after passage 10, preventing their use for downstream experiments. It is postulated that knocking-in a rather large of DNA piece (around 3kb) into a specific gene locus might have interfered with global physiological gene expression, leading to cell cycle arrest.

However, this phenomenon was puzzling as it was specific to MBJ6, a hiPSC line; no such effect was observed with knock-in lines derived from HES3 or H9. It was observed, nonetheless, that 144

the knock-in lines derived from hESCs lines were more susceptible to spontaneous differentiation than their corresponding wild-type hESCs. As such, it was harder to maintain the knock-in lines in the pluripotent state. Outside this susceptibility, knock-in lines behaved very similarly to their wild-type counterparts, staining positive for pluripotency markers, showing no chromosomal aberrations, as well as maintaining the ability to be differentiated towards CMs.

Thus the phenomenon observed in MBJ6-derived knock-in lines might be due to an incomplete reprogramming process of the parental BJ cells to the pluripotent state.

It was also observed that the genomic sequencing data of H9V15, an H9-derived knock-in line, revealed many mutations in the middle part of the sequence. The sequencing primers

KIgDNA-F and KIgDNA-R would give a product of 2.595kb in length. The earliest mutation was observed at position 1172 from 5‘- end of KIgDNA-F, which is beyond the accuracy range for the primer (a primer has accuracy range of about 850bp). In addition, PCR reaction could also introduce mutations in the PCR products to be sequenced. Adding another pair of sequencing primers somewhere in the mid-region of the sequence of interest should improve the accuracy of the sequencing data. Nevertheless, since the cDNA sequencing results showed no mutations in the cDNA of the knock-in lines, the reporters should be expressed as designed, if there were no other confounding factors in the system. Indeed as has been described in this chapter, this was the case.

In conclusion, with the aid of TALENs, a number of MYL2 knock-in reporter lines have been successfully generated. These lines should allow for enrichment as well as real time monitoring of ventricular CMs development, amongst other potential applications. The use of

TALENs reduced the need to screen hundreds of colonies before positive transgenic clones could be identified. In comparison to Thomson and Zwaka’s report in 2003, the approach taken here

145

was most similar to the POU5F1 knock-in experiment whereby a construct containing a neo cassette under the control of the tk promoter was introduced to hESCs. The authors reported a 26:1 ratio of stable transfected clones to HR events (Zwaka and Thomson, 2003).

Using TALENs in this study, at the lowest HR efficiency of 5.33%, we obtained four homologous recombinant clones from screening 75 clones (similar to a ratio of 18.75:1). At the highest HR efficiency, 44 homologous recombinant clones were obtained from screening 162 colonies

(similar to a ratio of 3.68:1). Thus TALENs improved homologous recombination events in hESCs.

Furthermore, TALENs also did not contribute to cytotoxicity, or gross chromosomal aberrations in the knock-in clones.

Through in vitro aging experiment, it was observed that the three knocked-in reporters were expressed in close correlation with MYL2 gene (figure 3.4.9-1). The earliest mChr signal was observable on day 24 of differentiation in the EHMs; this corresponded to MYL2 gene expression at 10x lower than that of GAPDH. On day 24 of differentiation, the G. Luc signal was around

1x105 RFU, around 5x higher than the background signal from the wild-type CMs. Knock-in CMs were also resistant to G418 treatment up to 50 µg/ml.

The correlation of MYL2 gene expression with each of the reporter was shown to be significant (p ≤ 0.05 for G. Luc; p ≤ 0.01 for mCherry and p ≤ 0.001 for NeoR). Lambda scan also showed that observed red fluorescence was specific for mChr with emission signal detected exclusively around 610nm, i.e. the theoretical emission wavelength of mCherry. While mChr reporter provides a qualitative real time signal for vCMs, G. Luc reporter provides a quantitative one. By day 30, knock-in CMs generated G. Luc signal 3 orders of magnitude higher than the wild type CMs (figure 3.4.8-1). Taken together, the data demonstrated reporter genes were well expressed in the knock-in CMs.

146

G418 selection of up to 50 µg/ml enabled enrichment for very bright mChr (+) CMs.

However, considerable cell death was observed in cultures treated with higher concentrations

(30 – 50 µg/ml) of G418. In fact, some mChr (+) CMs have been observed to float (and selected against G418) after a few days of antibiotic treatment. This was puzzling at first since NeoR gene was expressed as shown through qPCR data (figure 3.4.9-1). NeoR expression correlation to

MYL2 gene was also the strongest (p ≤ 0.001).

Presumably, G418 killed the MYL2 (-) cell population, which included not only the matrix producing non-myocytes, but also the other CMs subtypes (atrial and nodal). These MYL2 (-) cells, however, might have been tightly coupled to MYL2 (+) cells. Thus killing these cells might have resulted in detachment of the closely coupled MYL2 (+) adjoining cells from the cell culture surface.

General protocols for G418 killing curve experiment require sub-confluence of cells before antibiotic selection. However, directly differentiated CMs in tissue culture flasks tend to be tightly coupled to one another. The need for trypsinization and subsequent CM replating at

20-25% confluence conflicted with the original purpose of including this reporter in the knock-in line, i.e. to provide a one-step enrichment method for MYL2 positive cells from a mixed cell populations following a directed cardiac differentiation from hPSCs. Due to time limitation, G418 selection on CMs derived from 3D-differentiation method was not performed, thus the effect of selection on suspension culture of CMs could not be determined.

When comparing the reporter genes signals from H9V15 CMs versus H9VC16 CMs, it was observed that H9VC16 CMs expressed mChr at higher intensities than the H9V15 CMs (figure

3.4.6-7). This data is in agreement with literature which suggested that the strong expression of constitutive PGK promoter in PGK-Hygromycin selection cassette might interfere with the 147

endogenous enhancer/promoter elements and affect the physiological expression of the gene or neighboring genes (Calegari and Waskow, 2014). Although the G. Luc expression was higher in

H9VC16 CMs (figure 3.4.8-1), this difference was not shown to be statistically significant.

However, considering that on day 21, H9VC16 EHMs expressed lower level of G. Luc, it might be the case that H9VC16 EHMs contained fewer MYL2 (+) CMs, but that each of these CM produced more G. Luc than each of the MYL2 (+) H9V15 CM. This resulted in overall higher level of G.Luc expression (albeit not statistically significant) in H9VC16 EHMs from day 27 onwards, in comparison with H9V15 EHMs. Only if this presupposition holds, can we make the conclusion that the removal of the PGK-Hygromycin selection cassette resulted in a higher level of reporter gene expression.

Although qPCR data indicated a slight dip in MYL2 and all reporter genes expression from day 45 – 60, this dip in expression at protein level was only observed with G.Luc signal (figure

3.4.8-1). MChr signal was observed to increase in intensity through time in culture (figure 3.4.6-

2). Due to the qualitative nature of mCherry signal, however, no definitive conclusion could be made yet on the correlation of mChr signal with EHM contractile forces.

The G. Luc expression level on day 90 of differentiation was relatively similar to day 60.

However, the contractile forces of day 90 EHMs were on average 42% higher (1.96 ± 0.16 mN vs

1.38 ± 0.13 mN) than the forces of day 60 EHMs. Considering G. Luc expression to be proxy of

MYL2 gene expression, this increase in contractile forces must be attributed to other sources besides MYL2 contractile protein density in the CMs. Nevertheless, the in vitro aging experiment gave us insights on the expression pattern of MYL2. The gene starts to be expressed strongly in

CMs cultured as EHMs from day 24 of differentiation. This expression increased steadily until its peak expression on day 45 of differentiation. Beyond day 45, MYL2 expression level reached a

148

plateau. As such, it can be inferred that MYL2 is probably a good proxy for vCM early development up until day 45 of differentiation. Beyond day 45, MYL2 expression level does not seem to be a good proxy for vCM maturation.

149

3.6. References

Alwin, S., Gere, M.B., Guhl, E., Effertz, K., Barbas, C.F.r., Segal, D.J., and al, e. (2005). Custom zinc-finger nucleases for use in human cells. Mol Ther 12, 610–617. Beumer, K., Bhattacharyya, G., Bibikova, M., Trautman, J.K., and Carroll, D. (2006). Efficient gene targeting in Drosophila with zinc-finger nucleases. . Genetics 172, 2391–2403. Bibikova, M., Golic, M., Golic, K.G., and Carroll, D. (2002). Targeted chromosomal cleavage and mutagenesis in Drosophila using zinc-finger nucleases. Genetics 161, 1169–1175. Bird, S. (2003). The human adult cardiomyocyte phenotype. Cardiovascular Research 58, 423-434. Bizy, A., Guerrero-Serna, G., Hu, B., Ponce-Balbuena, D., Willis, B.C., Zarzoso, M., Ramirez, R.J., Sener, M.F., Mundada, L.V., Klos, M., et al. (2013). Myosin light chain 2-based selection of human iPSC-derived early ventricular cardiac myocytes. Stem Cell Res 11, 1335-1347. Boch, J., and Bonas, U. (2010). Xanthomonas AvrBs3 family-type III effectors: discovery and function. Annu Rev Phytopathol 48, 419-436. Boch, J., Scholze, H., Schornack, S., Landgraf, A., Hahn, S., Kay, S., and al., e. (2009). Breaking the code of DNA binding specificity of TAL-type III effectors. Science 326, 1509–1512. Bondue, A., and Blanpain, C. (2010). Mesp1: a key regulator of cardiovascular lineage commitment. Circulation research 107, 1414-1427. Bondue, A., Lapouge, G., Paulissen, C., Semeraro, C., Iacovino, M., Kyba, M., and Blanpain, C. (2008). Mesp1 acts as a master regulator of multipotent cardiovascular progenitor specification. Cell Stem Cell 3, 69-84. Burridge, P.W., Matsa, E., Shukla, P., Lin, Z.C., Churko, J.M., Ebert, A.D., Lan, F., Diecke, S., Huber, B., Mordwinkin, N.M., et al. (2014). Chemically defined generation of human cardiomyocytes. Nat Methods 11, 855-860. Calegari, F., and Waskow, C. (2014). Stem Cells: From Basic Research to Therapy: Basic Stem Cell Biology, Tissue Formation during Development, and Model Organisms (Taylor & Francis). Carlson, D.F., Fahrenkrug, S.C., and Hackett, P.B. (2012). Targeting DNA With Fingers and TALENs. Molecular Therapy Nucleic Acids 1. Cathomen, T., and Joung, J.K. (2008). Zinc-finger Nucleases: The Next Generation Emerges. Molecular Therapy 16, 1200-1207. Koch Institute for integrative cancer research at MIT. ES Cell and Transgenics: Mouse Models 101: Knockins and Knockouts. Christian, M., Cermak, T., Doyle, E.L., Schmidt, C., Zhang, F., Hummel, A., Bogdanove, A.J., and Voytas, D.F. (2010). Targeting DNA double-strand breaks with TAL effector nucleases. Genetics 186, 757-761. de Felipe, P., Luke, G.A., Hughes, L.E., Gani, D., Halpin, C., and Ryan, M.D. (2006). E unum pluribus: multiple proteins from a self-processing olyprotein. Trends in Biotechnology 24. Dierickx, P., Doevendans, P.A., Geijsen, N., and van Laake, L.W. (2012). Embryonic template-based generation and purification of pluripotent stem cell-derived cardiomyocytes for heart repair. J Cardiovasc Transl Res 5, 566-580. Donnelly, M.L.L., Luke, G.M., Amit , Li, X., Hughes, L.E., Gani, D., and Ryan, M.D. (2001). Analysis of the aphthovirus 2A/2B polyprotein ‘cleavage’ mechanism indicates not a proteolytic reaction, but a novel translational effect: a putative ribosomal ‘skip’. J Gen Virol 82, 1013-1025.

150

Eiges, R., Schuldiner, M., Drukker, M., Yanuka, O., Itskovitz-Eldor, J., and Benvenisty, N. (2001). Establishment of human embryonic stem cell-transfected clones carrying a marker for undifferentiated cells. Curr Biol 11, 514-518. Huber, I., Itzhaki, I., Caspi, O., Arbel, G., Tzukerman, M., Gepstein, A., Habib, M., Yankelson, L., Kehat, I., and Gepstein, L. (2007). Identification and selection of cardiomyocytes during human embryonic stem cell differentiation. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 21, 2551-2563. Jinek, M., Chylinski, K., Fonfara, I., Hauer, M., Doudna, J.A., and Charpentier, E. (2012). A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science 337, 816-821. Kim, J.H., Lee, S.-R., Li, L.-H., Park, H.-J., Park, J.-H., Lee, K.Y., Kim, M.-K., Shin, B.A., and Choi, S.-Y. (2011). High Cleavage Efficiency of a 2A Peptide Derived from Porcine Teschovirus-1 in Human Cell Lines, Zebrafish and Mice. PLoS One 6, e18556. Kim, Y.G., Cha, J., and Chandrasegaran, S. (1996). Hybrid restriction enzymes: zinc finger fusions to FokI cleavage domain. . Proc Natl Acad Sci USA 93, 1156-1160. Li, T., Huang, S., Jiang, W.Z., Wright, D., Spalding, M.H., Weeks, D.P., and Yang, B. (2011). TAL nucleases (TALNs): hybrid proteins composed of TAL effectors and FokI DNA-cleavage domain. Nucleic Acids Res 39, 359-372. Mahfouz, M.M., Li, L., Shamimuzzaman, M., Wibowo, A., Fang, X., and Zhu, J.K. (2011). De novo- engineered transcription activator-like effector (TALE) hybrid nuclease with novel DNA binding specificity creates double-strand breaks. Proc Natl Acad Sci USA 108, 2623-2628. Matsumoto, A., and Itoh, T.Q. (2011). Self-assembly cloning: a rapid construction method for recombinant molecules from multiple fragments. Biotechniques 51, 55-56. Mehta, A., Verma, V., Nandihalli, M., Ramachandra, C.J.A., Sequiera, G.L., Sudibyo, Y., Chung, Y., Sun, W., and Shim, W. (2014). A Systemic Evaluation of Cardiac Differentiation from mRNA Reprogrammed Human Induced Pluripotent Stem Cells. PloS one 9, e103485. Miller, J.C., Holmes, M.C., Wang, J., Guschin, D.Y., Lee, Y.L., Rupniewski, I., and al., e. (2007). An improved zinc-finger nuclease architecture for highly specific genome editing. . Nat Biotechnol 25, 778-785. Miller, J.C., Tan, S., Qiao, G., Barlow, K.A., Wang, J., Xia, D.F., Meng, X., Paschon, D.E., Leung, E., Hinkley, S.J., et al. (2011). A TALE nuclease architecture for efficient genome editing. Nat Biotechnol 29, 143-148. Moscou, M.J., and Bogdanove, A.J. (2009). A simple cipher governs DNA recognition by TAL effectors. Science 326, 1501. Müller, M., Fleischmann, B.K., Selbert, S., JI, G.J., Middeler, G., Müller, O.J., Schlenke, P., Frese, S., Wobus, A.M., Hescheler, J., et al. (2010). Selection of ventricular-like cardiomyocytes from ES-cells in vitro. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 14, 2540 - 2548. Mummery, C.L., Zhang, J., Ng, E.S., Elliott, D.A., Elefanty, A.G., and Kamp, T.J. (2012). Differentiation of human embryonic stem cells and induced pluripotent stem cells to cardiomyocytes: a methods overview. Circulation research 111, 344-358. Mussolino, C., Morbitzer, R., Lütge, F., Dannemann, N., Lahaye, T., and Cathomen, T. (2011). A novel TALE nuclease scaffold enables high genome editing activity in combination with low toxicity. . Nucleic Acids Res 39, 9283-9293. Porteus, M.H. (2006). Mammalian gene targeting with designed zinc finger nucleases. Mol Ther 13, 438– 446.

151

Porteus, M.H., and Baltimore, D. (2003). Chimeric nucleases stimulate gene targeting in human cells. . Science 300, 763. Rajala, K., Pekkanen-Mattila, M., and Aalto-Setala, K. (2011). Cardiac differentiation of pluripotent stem cells. Stem Cells Int 2011, 383709. Ramirez, C.L., Foley, J.E., Wright, D.A., Müller-Lerch, F., Rahman, S.H., Cornu, T.I., and al., e. (2008). Unexpected failure rates for modular assembly of engineered zinc fingers. Nat Methods 5, 374-375. Szczepek, M., Brondani, V., Buchel, J., Serrano, L., Segal, D.J., and Cathomen, T. (2007). Structure-based redesign of the dimerization interface reduces the toxicity of zinc-finger nucleases. . Nat Biotechnol 25, 786-793. Trichas, G., Begbie, J., and Srinivas, S. (2008). Use of the viral 2A peptide for bicistronic expression in transgenic mice. BMC Biology 6, 40. Zhang, Q., Jiang, J., Han, P., Yuan, Q., Zhang, J., Zhang, X., Xu, Y., Cao, H., Meng, Q., Chen, L., et al. (2011). Direct differentiation of atrial and ventricular myocytes from human embryonic stem cells by alternating retinoid signals. Cell Res 21, 579-587. Zwaka, T.P., and Thomson, J.A. (2003). Homologous recombination in human embryonic stem cells. Nature Biotechnology 21, 319 - 321.

152

CHAPTER 4

In vitro Maturation of HESCs-derived

Cardiomyocytes

153

TABLE OF CONTENT

4.1. Abstract ...... 156

4.2. Introduction and literature review ...... 158

4.2.1. When are CMs truly “mature”? ...... 158 Cell behavior, structure and morphology...... 159 Electrophysiology and calcium handling ...... 160 Contractile forces, cytosolic proteins and extracellular matrix ...... 161

4.2.2. Old is better than young ...... 164 In vitro aging ...... 164 Physical manipulation ...... 166 Electrical stimulation ...... 168 Biochemical stimulation ...... 169 Co-culture with other cell types ...... 171

4.3. Materials and Methods ...... 172

4.3.1. Temporal in vitro maturation in maturation medium...... 172 Total RNA and microRNA sequencing ...... 172

4.3.2. Increasing pre-load resistance for EHMs ...... 172 4.3.3. Electrical stimulation of EHMs ...... 173 4.3.4. Increasing preload and electrical stimulation ...... 173 4.3.5. Statistical analysis ...... 174

4.4. Results ...... 175

4.4.1. EHMs action potentials ...... 175 4.4.2. RNA profiling of in vitro aged CMs in EHM tissue constructs ...... 176 Down-regulated mRNAs ...... 179 Up-regulated mRNAs ...... 182

4.4.3. miRNA up- and down-regulation as EHMs age in vitro ...... 187 Upregulated miRNAs ...... 190 Down-regulated miRNAs ...... 195

4.4.4. Electrical stimulation improved force frequency and isoprenaline/carbachol response of cardiomyocytes in EHMs ...... 199 Experiment 1 (06.01.2014) ...... 199 Experiment 2 (23.04.2014) ...... 200

4.4.5. Cardiomyocytes in EHMs can tolerate a maximum of 10% increase in strain ...... 205 Experiment 1 (07.07.2014) - gradual 20% stretch per week ...... 205

154

Experiment 2 (22.07.2014) concurrent with experiment 3 – total of 20% stretch ...... 206 Experiment 3 (22.07.2014) – 10% stretch ...... 212

4.5. Discussion ...... 219

4.6. References ...... 225

155

4.1. Abstract

It is widely accepted that hPSC-derived cardiomyocytes (CMs) are phenotypically immature, akin to fetal CMs (Cao et al., 2008; Snir et al., 2003). This part of the thesis summarizes efforts to mature hESCs-derived CMs in vitro. The first attempt used a combination of growth factors, auxotonic mechanical conditioning and extended culture period of 3 months. The contractile forces as well as reporter genes expression generally increased over time of culture; this was described in the previous chapter. Day 30, 45, 60 and 90 EHMs beat with an average of 0.229 ±

0.019 mN, 0.869 ±0.087mN, 1.09 ±0.13mN and 1.62 ±0.16mN, respectively.

This chapter focuses on the dynamic regulation of EHM transcriptomes as they aged in vitro.

A multitude of genes were up-regulated and down-regulated over time in culture. Up-regulated genes included many associated with the extracellular matrix development and remodeling, regulation of developmental processes and cardiac morphogenesis, potassium transport, lipid and ions binding and transport, and cell adhesion. Down-regulated genes were functionally clustered around cell proliferation, some cell signaling cascade and carbohydrate metabolism.

The second approach involved electrical stimulation in conjunction with auxotonic stretching at basal (60 mm inter-pole distance) or at greater extension lengths (66 mm, 72 mm and 84 mm).

66 mm was found to be the maximum tolerable distance; anything farther was detrimental to the function of CMs inside the EHMs. Of all different electrical stimulation protocols tested, the following protocol was found to enhance CM maturation the most in EHM format: 8V basic square wave with 2 days each of the following frequencies: 1Hz, 1.5Hz, 2Hz and 2.5Hz. EHMs stimulated with the latter protocol showed (1) higher EC50 values than the control group of unstimulated EHMs (0.8 mM vs 0.6 mM Ca2+); (2) more positive force-frequency response from

1Hz to 2Hz and less negative response from 2Hz to 3Hz (figure 4.4.5-4); and (3) better response 156

to cardioactive drugs: isoprenaline (89.2% vs 5.5% in unstimulated EHMs) and carbachol (36.5% vs 4.4% in unstimulated EHMs).

These experiments demonstrated that in vitro manipulations could enhance the maturation level of cardiac tissue mimetic constructs (EHMs), but current in vitro protocols still fell short in generating CMs with the with the degree of maturity expected of CMs in vivo.

157

4.2. Introduction and literature review

At the phenotypical and transcriptional level, human pluripotent stem cell (hPSC)-derived cardiomyocytes (CMs) are most similar to fetal CMs (Cao et al., 2008; Snir et al., 2003); more specifically, the ones from primary heart tube (Fijnvandraat, 2003). This fact makes the use of hPSC-derived CMs suboptimal for number of applications such as drug screening and cell therapy due to their immature electrophysiology and mechanical properties (Chen et al., 2009). There is a need to improve the maturation level of hPSC-derived CMs in vitro, and this chapter describes different attempts to reach this goal.

3-dimensional CMs culture as engineered heart muscle constructs (EHMs), coupled with auxotonic stretching, has been shown to improve terminal differentiation and maturation of

CMs in vitro in comparison with the monolayer (2D) culture approach (Tiburcy et al., 2011; Zhang et al., 2013). However, there are still reported signs of CM immaturity in EHMs (Schaaf et al.,

2011). Thus, there is a need for further development for in vitro maturation techniques.

A number of maturation methods have been reported in literature and they will be explored in greater depth in the literature review section below. Before delving into these protocols, however, it is useful to understand what the term “mature” entails.

4.2.1. When are CMs truly “mature”?

Excellent reviews on CM maturation are available (Robertson et al., 2013; Yang et al.,

2014). In this thesis, some of these parameters will be summarized; some other important factors not yet covered in the reviews will also be outlined. Readers are referred to Chapter 2 section 2.2.3 for the phenotypes of hPSC-derived CMs.

158

Cell behavior, structure and morphology

Adult CMs are generally rod-shaped/cylindrical, with length-to-width ratio of 7 to 9.5

(Gerdes et al., 1992). The relatively large CM size plays an important role in action potential (AP) propagation and maximal rate of depolarization (Vmax = dV/dtmax) (Spach et al., 2004). Adult CMs have well organized and long sarcomeres, with relaxed state length of around 2.2 µm (Bird,

2003). They show the typical periodic Z-lines, M lines and the transverse-tubule network opposed tightly to terminal cisternae forming dyads (Lieu et al., 2009; Ziman et al., 2010). T- tubules in adult vCMs are closely apposed to the sarcoplasmic reticulum (SR) to facilitate calcium

Induced calcium release (CICR) events, thus promoting coordinated contraction throughout the entire cytoplasm. In addition, adherens junctions marked by N-cadherin are localized to intercalated disks of adult CMs to promote cell-to-cell attachment; and gap junctions in the same location containing connexin 43 serve to enhance conduction velocity (Angst et al., 1997).

In human neonates, CMs continue to proliferate in the first few months after birth: only around 25% neonatal CMs are binucleate (Laflamme and Murry, 2011). Adult human CMs however, are mostly multinucleate as they replicate their DNA without cytokinesis. The great majority of adult CMs have exited the cell cycle, with only 1% annual turn-over at the age of 25 to 0.45% at the age of 75 (Bergmann et al., 2009). In these cells, hypertrophic growth predominates; with circa 30-40 fold size increase in both rodents and humans (Laflamme and

Murry, 2011).

The cytoplasm of adult CMs is filled with contractile proteins and mitochondria

(Robertson et al., 2013); myofibrils make up 40% - 52% (Barth et al., 1992; Tashiro et al., 1990) and mitochondrial volume makes up 15% - 40% of the total CM volume (Barth et al., 1992;

159

Hattori et al., 2010; Schaper et al., 1985; Sulkin et al., 2014; Tashiro et al., 1990). In addition, mitochondria of the adult CMs are more elongated with higher membrane potential and well- structured lamellar cristae arranged in crystal like lattice pattern (Schaper et al., 1985). This improved structural organization leads to higher oxidative capacity and lower levels of reactive oxygen species (ROS) production in adult CM mitochondria (Yang et al., 2014). Notably, the mitochondria permeability transition pores (MPTPs) are closed in adult CMs (Yang et al., 2014).

Metabolically, fatty acid oxidation accounts for 90% of acetyl-CoA production and 60% –

80% of ATP production (Huss and Kelly, 2005; Yang et al., 2014) in adult CMs, indicating their preference for oxidative metabolism and their dependence on oxygen (Giordano, 2005; Harris and Das, 1991).

Electrophysiology and calcium handling

During the first few years of life, CMs increase 2.4-fold in size, yet AP duration (APD) only increases by around 20% (Lakatta and Sollott, 2002; Roberts and Wearn; Zak, 1974). The different AP characteristics of all hPSC-derived CMs have been excellently reviewed elsewhere

(Chen et al., 2009a). Adult CMs generally have hyperpolarized maximal diastolic potential (MDP) around -85mV (Drouin et al., 1998), relatively long APD (Schram et al., 2002), and extremely fast maximum rate of depolarization dv/dt (max) or Vmax around 300 V/s (Robertson et al., 2013). In addition, adult vCMs are generally quiescent and do not exhibit spontaneous phase 4 depolarization (Schram et al., 2002). The left ventricle of myocardium in humans has been shown to have a conduction velocity of 0.3 to 1.0 m/s (Yang et al., 2014).

A recent study in zebrafish by Hou, et al. (2014) demonstrated that early in development, action potentials were initiated via a calcium current through L-type calcium channels (LTCC). As the heart develops, there was a switch in the ventricular AP to a sodium-driven upstroke, while 160

the atrial AP remained LTCC-driven. In the adult state, APs are initiated by sodium current, both in the atria or ventricles (Hou et al., 2014). Thus, there is an increased dependence in sodium channels for AP initiation in adult hearts, at least in the zebrafish.

Donald M. Bers (2002) has described cardiac excitation-contraction coupling in great detail. A number of membrane proteins are important for excitation-contraction coupling in adult CMs. The intracellular Ca2+ concentration is reduced by calcium clearance from the cytosol mediated by SERCA (sarcoplasmic reticulum Ca2+ ATPase pump), the sarcolemmal sodium- calcium exchange pump, the sarcolemmal Ca2+-ATPases and/or mitochondrial Ca2+ uniport (Bers,

2002). Calcium import into the cytosol, on the other hand, is mostly driven by calcium induced calcium release (CICR) events from the sarcoplasmic reticulum (SR). In fact, the SR contributes around 70% of the total calcium release from adult CMs (Bers, 2002). CICR mostly acts through L- type calcium channels (LTCC), transmembrane calcium channels located in T-tubules, and ryanodine receptors (RYRs) (Bers, 2002). These proteins lie in close proximity to each other in the region termed the “calcium cleft” (Bers, 2002). RYRs are coupled to triadin, junctin and calsequestrin in the luminal side of SR, thus they also act as intra-SR calcium sensor (Bers, 2002).

Contractile forces, cytosolic proteins and extracellular matrix

The beating force of adult human and rat myocardium has been reported to be around

44 ± 11.7 mN/mm2 and 56.4 ± 4.4 mN/mm2 respectively (Hasenfuss et al., 1991). Adult CMs beat along a single axis with a positive force-frequency relationship, i.e. faster pacing frequency will result in increased beating force (Bers, 2002; Dolnikov et al., 2006). In addition, they also exhibit post-rest potentiation: calcium uptake is increased during resting phase after rapid pacing.

Treatment of adult CMs with isoprenaline (beta receptor agonist) would generally lead to

161

chronotropic and inotropic responses, while treatment with carbacholine (muscarinic receptor) would lead to the opposite: reduced heart rate and contractile forces.

Adult CMs show isoform specificity in their expression of titin, troponin and MHC. Unlike their younger counterparts, adult CMs express a shorter and stiffer form of titin of 3000 kDa

(N2B), which aids in the regulation of passive tension (Lahmers et al., 2004; Opitz et al., 2004;

Warren et al., 2004). Adult CMs also contain the cardiac troponin I (cTnI) instead of the slow skeletal troponin I (ssTnI) isoform (Bhavsar et al., 1991; Hunkeler et al., 1991; Sabry and Dhoot,

1989; Saggin et al., 1989). CTnI contributes to decreased calcium sensitivity for contraction in adult CMs, in comparison with complexes containing ssTnI (Metzger et al., 2003; Reiser et al.,

1994; Siedner et al., 2003). Therefore, a decrease in extracellular calcium sensitivity is predictive for improved CM maturation. Very recently, the cTnI: ssTnI isoform ratio has been proposed as a

CM-maturation signature (Bedada et al., 2014). In adult CMs there is also more β-MHC (myosin heavy chain) isoform than α-MHC, the latter being more prevalent in fetal hearts (Xu et al.,

2009). The increased abundance of the slower β-isoform might contribute to the observed postnatal heart rate decrease.

With regard to cell-extracellular matrix attachment proteins, adult CMs were found to attach efficiently to laminin and type IV collagen, but weakly to fibronectin (Borg et al., 1984). In contrast, neonatal CMs were demonstrated to attach to collagen I, III and IV, fibronectin and laminin with similar efficiency (Borg et al., 1984).

162

Parameters Immature cardiomyocytes Mature cardiomyocytes Cell shape Circular Rod-shaped Morphology Membrane 17.5±7.6 pF ≈150 pF capacitance Disarrayed; Highly organized; Structure Sarcomere Non-periodic Z lines Periodic Z-lines Length ≈1.6 µm ≈2.2 µm Titin N2BA N2B Myofibrillar Troponin I ssTnI cTnI isoform MHC β > α β >> α T-tubules No Yes Irregular reticular network in Regularly distributed; Mitochondria cytoplasm; occupies a small occupies ≈ 20% - 40% of fraction of cell volume cell volume Metabolic substrate Glucose Fatty acid Multinucleation Mononucleated ≈25% multinucleated Proliferative capacity Proliferative (limited) Non-proliferative Upstroke ≈50 V/s ≈250 V/s velocity Electrophysiological Resting properties membrane ≈−60 mV ≈−90 mV potential Automaticity Spontaneous contraction Quiescent Mature; Partially developed; Depends on Calcium- E-C coupling Contraction depends on trans- induced-calcium release sarcolemmal calcium influx (CICR) Force frequency response Negative Positive Contractile force ≈nN range/cell ≈µN range/cell Polarized to intercalated Gap junction distribution Circumferential disks Conduction velocity ≈0.1 m/s 0.3 – 1.0 m/s Responses to β-adrenergic Chronotropic response; Chronotropic response; stimulation Lack of inotropic reaction Inotropic reaction

Table 4.2.1-1 Summary of differences between immature and adult cardiomyocytes. Adapted from Yang, et al. (2014)

163

4.2.2. Old is better than young

CMs are regulated by differing environmental signals during embryonic development in vivo. Amongst these factors are mechanical cues, electrical fields, soluble factors as well as dynamic composition and characteristics of extracellular matrices. These factors in turn may determine spatial organization and induce tissue morphogenesis (Henderson and Chaudhry,

2011; Nuccitelli, 1992). Therefore, most successful in vitro maturation approaches attempt to recapitulate the in vivo environment in order to mature CMs in culture. Methods to enhance CM maturation can use either monolayer or three-dimensional culture approach. Once a culture system is determined, state-of-the-art maturation techniques can be broadly grouped into 5 different approaches: (1) in vitro aging, (2) physical manipulation, (3) electrical pacing, (4) biochemical treatment and (5) co-culture with other cell types (often in combination with other techniques).

In vitro aging

In vivo, human neonatal CMs developed over 6 – 10 years to the reach their adult phenotype

(Peters et al., 1994). Practical constraints limit in vitro aging experiment for similar duration.

Nevertheless, long term culture has been used as the most instinctive way to enhance CM maturation in vitro. HPSCs-derived CMs have been reported to exhibit morphological and ultrastructural maturation as well as withdrawal from cell cycle within 60 days of in vitro culture

(Snir et al., 2003). CMs were shown to be more elongated and bigger in size, myofibril morphology, organization and density were also improved, with A, I and Z bands being observed in older CMs. The gap junction proteins connexin 43 and 45 were also expressed (Kehat et al.,

2002). T tubules, however, were still absent in these CMs. More recent publications (Kamakura et al., 2013; Lundy et al., 2013) using 180 day old and 80 – 120 day old hPSC-derived CMs,

164

respectively, have also supported these observations. Kamakura et al. (2013) continued to culture their CMs for 1 year and noted that M bands “were finally detected in 360-day old EBs,” although the level of expression was still lower than that in the adult heart. Lundy et al. (2013) also reported a 10-fold increase in the fraction of multinucleated CMs. In a separate study, electrophysiology of hESCs-derived CMs was shown to mature over 3-month culture period

(Sartiani et al., 2007). Figure 4.2.2-1 captures the essence of quantitative and qualitative changes in the different currents during maturation. Lundy et al (2013) also reported similar electrophysiological behavior with their older hPSC-derived CMs: late-stage CMs were shown to have hyperpolarized maximum diastolic potentials (MDP), increased action potential amplitudes, and faster upstroke velocities.

Developmental maturation is illustrated by gray triangles, indicating an increase in the current density and by a dashed stripe, which represents the slowing activation rate for If. Presence of INa has been described (Satin et al., 2004). Abbreviations: ICa,L, L-type calcium current; If, pacemaker current; IK1, inward rectifier potassium current; IKr, rapid delayed rectifier potassium current; INa, sodium current; Ito, transient outward potassium current

Figure 4.2.2-1 Schematic representation of ionic currents measured in hESCs, early and late CMs. Image is reproduced from (Sartiani et al., 2007) with permission.

Additionally, 80 – 120 day old CMs were shown to beat with shortening magnitude double that of the younger (20 – 40 day old) CMs. Their contraction kinetics was also slower in comparison to the early stage cells (Lundy et al., 2013). Furthermore, these authors reported an increase in calcium release and reuptake rates with no change in the maximum contraction amplitude. Despite the encouraging results from in vitro aging experiments, the nature of the

165

experiments prevents time-effective in vitro CM maturation. In addition, despite 1 year of culture, hPSC-derived CMs still did not reach the maturity level of adult CMs in vivo (Kamakura et al., 2013).

Physical manipulation

A number of approaches fall into this category:

1. Varying the mechanical properties of the attachment substrate

Hazeltine and colleagues (2012) seeded hPSC-derived CMs on polyacrylamide hydrogels of varying stiffness from 4.4 – 99.7 kPa. Contraction stress was found to increase with substrate stiffness. Strain, beating rate and morphology, however were not significantly affected. CMs seeded on all 4 different substrate stiffness nevertheless responded to isoprenaline stimulation, indicating presence of beta adrenergic receptors. A similar result was reported by Rodriguez et al.

(2011) who cultured neonatal rat ventricular CMs on micro-fabricated arrays of flexible posts of different stiffness. It was found that CMs cultured on stiffer substrates developed 6-fold greater twitch forces but slower twitch velocity. In addition, CMs on stiffer arrays were found to exhibit more mature myofibril structure, as they had longer sarcomeres and wider Z-bands.

Furthermore, the CMs cultured on stiffer arrays also showed improved intracellular calcium levels during a twitch, indicating better intracellular calcium storage capacity.

2. Attachment surface engineering for anisotropic cell attachment

Some groups have taken the approach of using biomimetic topographical cues. CM shape has been shown to regulate sarcomere alignment (Bray et al., 2008), and cell alignment has been shown to improve when cells are cultured on patterned substrates (Au et al., 2009;

Wang et al., 2011). Kim et al. (2010b) constructed an anisotropically nanofabricated substrate of

166

polyethylene glycol patterned with ridges and grooves ranging from 150 to 800 nm wide, designed to closely reproduce a nanoscale structure of the myocardial extracellular matrix composed of aligned fibrils around 100 nm in diameter. Each CM spanned >10 nano-ridges and was shown to align along the direction of the topographical cue. Aligned CMs exhibited cell geometry, conduction velocity, and Cx43 expression more similar to that of the native heart. In a more recent attempt, surface topography and substrate stiffness approaches were combined

(Wang et al., 2011). Morphology and orientation of neonatal rat ventricular CMs were mainly influenced by nano-grooved structures, whereas the contractile function of the CMs was regulated by both surface topography and substrate stiffness, with better contraction on the soft substrate with deep grooves (Yang et al., 2014).

3. Mechanical conditioning In another set of experiments, neonatal rat CMs were seeded in neutralized collagen I into 3-dimensional engineered heart tissues (EHTs) which were then subjected to phasic unidirectional stretch for 6 days (Fink et al., 2000). The authors showed that atrial natriuretic factor (ANF) was up-regulated by 98% and alpha-sarcomeric actin by 40%. In addition, stretched

EHTs developed hypertrophic CMs with increased cell length and width, longer myofilaments, and increased mitochondrial density. The CMs in stretched EHTs also showed marked improvement in structural organization, being arranged into parallel arrays of rod-shaped cells.

Finally, stretched EHTs demonstrated significant improvements in contractile forces. Stretched

EHTs beat with 2 – 4 fold higher forces both under basal conditions and after stimulation with calcium or isoprenaline.

More recently, hESCs-derived CMs were exposed to cyclic equiaxial mechanical stretch at 0.5 Hz with pulsation of 10–25% elongation of cells for 24 h (Foldes et al., 2011). CMs

167

stretched for 24 hours showed increase in cell size (1.6-fold, P < 0.001) and percentage of CMs showing organized sarcomere structure (P < 0.001). Mechanically stretched CMs also showed significantly higher mRNA expression levels of αMHC + βMHC and ANF (P < 0.05 and 0.001 vs. unstretched control, respectively, triplicate determinations). The ratio of α/β-MHC, however, was not found to change significantly with stretching. These results were in accordance with those of another study (Tulloch et al., 2011), which demonstrated that uniaxial mechanical stress conditioning resulted in 2-fold increase in alignment of CMs and matrix fibers as well as enhanced myofibrillogenesis and sarcomeric banding. The cyclic stress conditioning also induced

CMs hypertrophy (2.2-fold) and proliferation rates (21%).

Electrical stimulation

In hESCs-derived CMs, electrical stimulation led to longer action potential duration and higher caffeine-induced calcium flux (Martherus et al., 2010). Also, it has been reported that electrically stimulated hESCs-derived CMs had higher level of Kir2.1 expression (Deng et al.,

2000). In another study, hPSC-derived CMs were cultured in 3-dimensional format and electrically paced with either progressively increasing stimulation frequency from 1 Hz to 3 Hz or from 1 Hz to 6 Hz over 1 week of culture at 3-4V/cm (Nunes et al., 2013). The authors demonstrated that electrical stimulation enhanced CMs sizes and phenotype, with the ones paced up to 6 Hz showing better results than the ones paced up to 3 Hz. In particular, electrically-stimulated CMs showed significantly more rod-like cells, enhanced CMs self- organization with presence of Z disks, I bands, H zones, and desmosomes. The number of H zones per sarcomere, the ratio of I bands to Z discs and the desmosome density were shown to increase significantly with increasing frequency of stimulation (Nunes et al., 2013). The same authors also described down-regulation of NPPA, NPPB, MYH6 and up-regulation of KCNJ2 in

168

their electrically stimulated CMs, implying maturation of these CMs. Electrically stimulated CMs also exhibited lower proliferation rates, improvements in electrical properties, electrophysiology and calcium handling properties. A more recent study employed a more modest electrical stimulation frequency (Hirt et al., 2014). The authors mixed hiPSCs-derived CMs with fibroblasts to form engineered heart tissues (EHTs) and subsequently paced the EHTs with biphasic pulses of 4 ms and field strength of 2V/cm from day 4 post-casting onwards, at 2 Hz for the first week and 1.5Hz thereafter (Hirt et al., 2014). In comparison with the non-paced group, the paced EHTs were found to beat with 1.5x higher forces, and contained more anisotropically aligned CMs with higher cytoplasm-to-nucleus ratio.

Biochemical stimulation

1. Hormones

To test the effects of hormones on CM differentiation, hESCs-derived CMS were subjected to 100 μM angiotensin II or 10 μM α-adrenergic phenylephrine (PE) treatment for 48 hours (Foldes et al., 2011). The authors found that CMs treated with PE had a significant increase in cell area (1.8-fold, P < 0.0001), cell volume (2-fold, P < 0.001) and fraction of CMs with organized sarcomeres (3.8-fold, P < 0.0001). Messenger RNA and protein expression of ANF (2.1- fold, P < 0.001) were also increased in PE-treated CMs, in comparison with untreated CMs. CMs treated with angiotensin II also showed increase in cell size (1.2-fold, P < 0.05), fraction of CMs with organized sarcomere structure and αMHC, βMHC, and ANF mRNA levels). The level of increase for each parameter, however, was less than that observed with PE. The ratio of α/β-

MHC was also not found to change significantly with PE or angiotensin treatment.

169

Thyroid hormone has an indispensable role in maintaining normal heart function (Klein and Ojamaa, 2001). T3 has also been reported to modulate the developmental titin isoform switch in cultured fetal rat CMs from fetal isoform N2BA (>3200kDa) to adult isoform N2B

(3000kDa) (Kruger et al., 2008), indicating that the hormone promotes CM maturation. In ovine background, T3 was demonstrated as a prime driver of CM maturation (Chattergoon et al., 2012).

T3 treatment increased the expression levels of phospho-mTOR, atrial natriuretic peptide (ANP) and SERCA2a. In additional, T3 treatment also increased CM width by 14%, bi-nucleation rate by

31% and p21 protein expression by 200%. T3-treated CMs also showed a decrease in proliferation by 39% and a decrease of cyclin D1 protein expression by 50% (Chattergoon et al.,

2012). The role of T3 is not limited to CM maturation, however. The hormone has also been shown to promote CM differentiation from mouse embryonic stem cells (Lee et al., 2010). T3 treatment improved CM yield as well as up-regulation of a number of cardiac markers such as

Nkx2.5, MYL2, αMhc and βMhc (Lee et al., 2010). Furthermore, T3-treated CMs displayed more adult-like electrophysiology and calcium homeostasis: they had more negative resting membrane potential (−74.72 ± 4.12 mV; P < 0.05) and higher expression of Serca2a and RyR2

(Lee et al., 2010). T3-treated CMs also had larger peak amplitude of calcium transients, F/F0

(2098 ± 208.30); higher maximal upstroke velocity, F/F0/sec (3.38 ± 0.32 vs. 1.17 ± 0.10) and maximal decay velocity, F/F0/sec (−1.34 ± 0.08 vs. −0.34 ± 0.09) (Lee et al., 2010).

The effect of another hormone, relaxin (RLX), was tested in mouse neonatal CMs in primary culture (Nistri et al., 2012). The authors found that RLX affected cell proliferation, enhancing it at 2 and 12 hours yet inhibiting it at 24 hrs. RLX also induced the expression of

Gata4, Nkx2-5, cx43, troponin T and Hcn4 ion channel at both the mRNA and protein level.

Furthermore, RLX-treated CMs showed ultrastructural and electrophysiological maturity.

170

In another paper, transgenic mice overexpressing insulin growth factor receptor (Igf1R) in the heart were reported to exhibit cardiac hypertrophy (McMullen et al., 2004). The transgenic mice also had enhanced systolic function at 3 months of age, which was maintained at 12-16 months of age. Thus it seems that IGF1 might enhance CM maturation in vitro.

2. miRNA

MiR-1 has been identified as a potential modulator of CM maturation (Fu et al., 2011).

Indeed, miR-1 overexpression decreased action potential duration (APD) and hyperpolarized both the resting membrane potential and the maximum diastolic potential in hESCs-CMs, due to increased Ito, IKs, and IKr, and decreased If. Furthermore, miR-1 was demonstrated to improve calcium transient amplitude and kinetics. MiR208 is yet another miRNA identified as a CM maturation modulator. This has been shown to be involved in thyroid hormone responsiveness and myosin isoform switching (Callis et al., 2009; van Rooij et al., 2007).

Co-culture with other cell types

Using genetically purified early hESCs-CMs, Kim et al (2010a) showed that non-CMs are required for the development of expression of some intracellular calcium handling proteins.

Presence of non-CMs was demonstrated to enhance ion channel development (HCN4) and CMs electrophysiological maturation. Non-myocyte fraction has been shown to be critical for functionality of 3-dimensional culture of hPSC-derived CMs (Tiburcy et al., 2014). Defined non- myocyte fraction (fibroblasts) was demonstrated to enhance the maturation of CMs inside the

EHTs, which were measured through higher contractile force of 9.8±0.9 nN/CM, higher EC50 for calcium, and enhanced inotropic response to isoprenaline. In addition, there was evidence for cell cycle withdrawal, increased actinin protein expression per CM; and reduction in MYL2/MYL7

171

double positive CMs fraction. Taken together, non-myocyte fraction seems to improve the maturation of the CMs in EHTs by improving their functionality.

4.3. Materials and Methods

4.3.1. Temporal in vitro maturation in maturation medium

Total RNA and microRNA sequencing

At day 18 of differentiation, H9V15 and H9VC16 CMs were digested into single cells and mixed with HFFs in collagen I matrix to make EHMs. These EHMs were cultured in vitro for 90 days in presence of VEGF, IGF, FGF and auxotonic stretching but no electrical stimulation. RNA samples were extracted from these EHMs on day 30, 45, 60 and 90 of differentiation and then sent for RNA and small RNA sequencing experiment. The raw data was provided by the DNA

Microarray and Deep-Sequencing Facility, Göttingen.

The raw data was analyzed by the author of this study in Singapore using Partek

Genomics Suite® on trial version. Firstly, an ANOVA table was created to determine the probability values which would indicate that the differential gene expression observed was not a stochastic effect. Genes with false discovery rate (FDR) ≤ 0.001 and absolute log2fold change value ≥ 3.000 were shortlisted for further analysis.

4.3.2. Increasing pre-load resistance for EHMs

H9VC16 EHMs were transferred to printed auxotonic stretchers of varying inter-pole distances: 60mm (normal), 66mm, 72mm, and 84mm. Three different sets of experiments were

172

carried out (table 4.3.2-1). CMs level of maturation was assessed through EHM force measurements, force frequency response and calcium handling capacity (EC50).

Experiment number Treatment 1 60mm for 2 weeks; 72mm for 1 week; 84mm for 1 week 2 60mm for 8 days; 66mm for 9days; 72mm for 8 days 3 60mm for 5 days; 66mm for 2 weeks

Table 4.3.2-1 Passive mechanical stretch (increasing CMs preload) procedures

4.3.3. Electrical stimulation of EHMs

H9VC16 EHMs were transferred to 60mm printed auxotonic stretchers and cultured for a minimum of 2 weeks. They were then paced with ION Optix C-Pace EP Culture Pacer with square bipolar pulses at 5 – 8V with 5ms duration in between pulses. Frequency of pulse was adjusted manually to 1Hz, 1.5Hz, 20Hz and 2.5Hz. Each frequency was applied to EHMs for one or two days per frequency. The level of CM maturation was assessed through the expression level of reporter genes, EHM contraction force measurement, EHM force-frequency response as well as calcium handling capacity (EC50).

4.3.4. Increasing preload and electrical stimulation

In a few experiments, the preload of H9VC16 EHMs were gradually increased from 60mm to

84mm across three weeks of culture in vitro. EHMs were first cultured on 60 mm printed auxotonic stretchers for a week, before being transferred to 72 mm stretchers. A week later, these EHMs were transferred again to 84 mm stretchers and kept at this distance for another week of culture. EHMs were electrically stimulated as per section 4.3.3 whilst being stretched at

84 mm. In another set of experiments, EHMs were electrically stimulated whilst being stretched at 72 mm or 66 mm.

173

4.3.5. Statistical analysis

Data is presented as means ± standard error mean (SEM). Statistical analyses were performed with Student’s t-test (2-tailed) or ANOVA when applicable, with post-hoc Bonferroni's

Multiple Comparison test. Results were considered to be significant at p ≤ 0.05 (*); very significant at p ≤ 0.01(**) and extremely significant at p ≤ 0.001 (***).

174

4.4. Results

4.4.1. EHMs action potentials

Fused EHMs as on figure 3.4.6-1 were set on electrophysiology viewing chamber and impaled with glass electrodes. The following figures show the typical action potentials obtained.

50 Average resting -70.3 ± 1.50 mV membrane potential 0 (RMP)

Average action potential 106.5 ± 0.70 mV amplitude (APA)

-50 Average maximum speed 169.66 ± 2.26 V/s of depolarization (dV/dt) Average APD40 147.05 ± 9.48 ms Average APD90 202.61 ± 10.43 ms -100 1 2 3 4 5 6 Figure 4.4.1-1 Action potential #1 from fused EHM

50

Average RMP -60.57 ± 1.77 mV

0 Average APA 99.81 ± 4.04 mV

Average dV/dt 72.41 ± 12.93 V/s Average APD40 108.4 ± 6.74 ms

-50 Average APD 90 171.75 ± 8.11 ms

-100 1:47 1:48 1:49 1:50 1:51 1:52 1:53 1:54 1:55 1:5

Figure 4.4.1-2 Action potential #2 of fused EHM

175

4.4.2. RNA profiling of in vitro aged CMs in EHM tissue constructs

Using FDR ≤ 0.001 and the absolute value of Log2Ratio (treatment/control) ≥ 3 as initial cut off criteria, 4900 genes were shortlisted for further analysis. A column statistics was then generated for these genes to facilitate hierarchical clustering and the generation of Self

Organizing Map (SOM) in the form of heat map. The profile trellis function was also used in this study to display the different patterns of gene expression from different samples. The profile trellis enables simple identification of the group of genes which displayed a certain pattern of expression. In this particular experiment, the different samples were obtained from EHMs of different age. Thus a continuously increasing pattern shown on profile trellis e.g. sub-profile 17 of figure 4.4.2-2 on the next page would indicate genes which were continuously up-regulated as

EHMs aged in vitro for a period of 90 days.

The heatmap shown in figure 4.4.2-1 of the RNA-sequencing results shows genes that were up- and down-regulated through time in culture (red shows higher expression and blue shows lower expression). From the heat-map, one can see that there are more genes that are up-regulated than down-regulated the longer EHMs were cultured. Almost all of the gene functions and site of expressions outlined below are adapted from GeneCards® database at www..org as well as gene ontology module of Genomics Suite® software, unless stated otherwise. Up-regulation and down-regulation fold changes for mRNA expressions are quoted as log2foldchange values, unless stated otherwise. Thus three times down-regulation refers to eight times absolute fold change.

176

gene expression level in this particular sample 2.27 Heat map of gene expression in EHMs different at genein of points time map Heat expression

1 - 4.4.2 Figure Log2(fold change) scale. “2.77” change) red2 scale. = Log2(fold

:

Dendogram the shows Scale hierarchy of the clusters which allows for exploratory analysis see to how the group microarrays on based together similarity of features 177

The different number boxes show one unique expression profile.

profiles in EHMs over time in culture. Grey area shows 1 standard deviation replicates at each time point time replicates each at deviation Grey 1 standard shows area Patterns of gene expression

2 - 4.4.2

Figure

expression - gene log2 Normalized

178

Figure 4.4.2-2 shows the different types of expression patterns of the 4900 shortlisted genes, across different time points (day 30, 45, 60 and 90). Genes which showed continuous differential up-regulation or down-regulation from day 30-90 of culture will be further analyzed.

Down-regulated mRNAs

Profile 3

Genes belonging to profile 3 showed a general down-regulation in gene expressions through time in culture, with steepest gradient from d60 – 90 of differentiation. Data was sorted for ratio of log2 expression on d60 vs d90 of >= 2.0, indicating greater than 4 absolute fold change in expression. Gene ontology was also filtered for p-value of enrichment < 0.05. The full list of genes together with their normalized level of expression is available in appendix 5.1.

Gene Ontology Gene list and Sites of expression/Functions Carbohydrate catabolic PGK1, OGDHL, GAPDH, ENO3 and PFKP processes Cholesterol TM7SF2, APOA1 biosynthesis and transport Transporters ATP4A, COX7A1, ATP6V1G  proton transporters CHRNA3, TMEM38, ATP1B4, KCNV2, and ABCB1  other transporters Signal transducer MSX2, expressed in atrio-ventricular canal cells (Seddon et al., 2004), heart tube and cardiomyocyte-like progenitor cells (Tate, 2010) Cell adhesion PROM1 (expressed in cardiomyocytes progenitors), molecules CDH13, SELP (expressed in atrioventricular node cells) and RND2 (expressed in cardiac crescent and cushion mesenchymal cells)

Table 4.4.2-1 Gene Ontology of genes which belong to profile 3

179

Profile 4

Genes in profile 4 showed steady down-regulation as EHMs age in vitro.

Gene Ontology Gene list and Sites of expression/Functions Cell-cell signaling EREG  positively regulates cell proliferation and mitosis, CRHBP, GRIA1, TRH, TFAP2C, STC1, HOXA11, FGFBP1, COMT, CHRNA9 and ADCY8 Extracellular matrix reorganization and Ca2+-ion binding MMP3 Transporters SLC16A6  monocarboxylate transporter and symporter, SLC2A2  facilitative glucose transporter Lipid-binding APOA2  involved also in cellular lipid catabolic process and cholesterol import, proteins SORL1 Membrane proteins CLDN6, NLGN4Y and KIRREL3

Table 4.4.2-2 Gene Ontology of gene which belong to profile 4

Profile 5

Genes in this profile group showed down-regulation of genes as EHMs aged from d30 – 60. The down-regulated genes then show steadily low expression from d60 – d90.

Gene Ontology Gene list and Sites of expression/Functions Antiporter activity SLC4A9  anion-anion antiporter activity Bicarbonate transporter SLC4A9, SLC24A3, SLC4A8, SLC26A9 Membrane-associated transporters CNGA2, SLC4A9, SYP, KCNK16, TRPC3, TRPA1, GJD2, SYT10, ATP13A4, SLC4A8, GABRB2, SLC24A3 and SLC26A9 Glucose transports regulation IL1B, OSTN Short-chain fatty acid metabolic process IRG1 and THNSL2 Striated muscle tissue development ALDH1A2, EYA1 and RXRG  also involved in fatty acid storage Lipid-binding protein CLVS2, SYP, SH3GL2, SYT14, VNN1, CADPS, YBX2 and ALDH1A2 Cell-cell adhesion regulatory genes IL1B, L1CAM and DPP4 Cell-cell signaling genes IL1B, TRHDE, SALL1, KCNK16, GJD2, ADCY2, ABAT, GABRB2 and CORT

Table 4.4.2-3 Gene ontology of genes which belong to profile 5

180

Profile 9 shows sharp down-regulation in EHMS day 30-45, steady expression in EHMs day 45-60 and further down-regulation in day 60-90.

Gene Ontology Gene list and Sites of expression/Functions Cation transporters KCNQ2, CHRNB4, CCR7 Anion transporters ABCC3, SLC6A18, CRYM, SLC51B Very-low-density lipoprotein assembly protein and SOAT2 cholesterol-binding Positive regulation of glycoprotein biosynthetic CCR7 and SLC51B process

Table 4.4.2-4 Gene Ontology of genes which belong to profile 9

Profile 10 shows sharp down-regulation from day 30-45 EHMs and steady low level of expression beyond day 45 of differentiation.

Gene Ontology Gene list and Sites of expression/Functions G-protein coupled receptor signaling pathway NPVF, GUCA1C, PDC and NPFFR1 Channel proteins CYBB, PANX3, CACNA2D2, GABRQ, and CACNA1D Signaling receptor activity proteins TAS2R42, NPFFR1, OR51Q1, PTCHD2, GABRQ, CCRL2, GPR64, HNF4G Gap junction hemi-channel protein PANX3

Table 4.4.2-5 Gene ontology of genes which belong to profile 10

181

Up-regulated mRNAs

Profile 16

Genes in this profile were up-regulated rather sharply from day 30-45 of differentiation which continued up to day 90 in culture.

Gene Ontology Gene list and Sites of expression/Functions Signal transduction APLN, GPR116, ANGPT2, NRK, RAPGEF5, CXCR4, COL15A1, CASQ2, KIT, PDE3A, GRAP, ELTD1, RHOJ, TNFSF10, ACKR3, ARHGDIB, ARHGAP18, F2RL3, RASSF9, F2R, RASL12, FZD4 HTR2B, KITLG, GNG2, RAMP2, SPARC, TLR4, TEK, AKR1C3, ENTPD1, EFEMP1, DLC1, FAT4, F2RL1, PTN, ACTG1, HEY1, CTHRC1, RAB15, PLCE1, P2RX1, DIRAS3, PLCL2, DCBLD2, NRP2, PRKD3, RALA, NEK6, RAB7L1, RAB31, HOMER3, RAB23, YES1, GFPT1, IQGAP1, MAP4K4. Negative regulation of cell OGN  regulator of left ventricular mass (Petretto et al., 2008), proliferation CDH5, NRK, F2R, DLC1, ETS1 and CD9 Heart development and KIT, RAMP2, SPARC, TEK, FBN1, PLCE1 and NRP2 cardiac progenitor differentiation calcium-mediated signaling CXCR4, CASQ2, HTR2B and PLCE1 calcium ion binding CD93, CDH5, PCDH12, CASQ2, EDIL3, ELTD1, PLSCR4, MMP16, SPARC, SYT4, EFEMP1, FAT4, FBN1, PLCE1, PLCL2, TPM4 and IQGAP1 Phospholipid translocation ATP8B1 and ATP10D A cell-surface membrane protein: CD93 Mediates cell-cell adhesion, linkages of transmembrane proteins to cytoskeleton, Ca2+ and carbohydrate binding. An endogenous ligand for G-protein coupled APJ receptor: APLN Regulates heart contractile forces, cardiac tissue remodeling, inhibits adenylylcyclase, increases intracellular calcium and activates extracellular signal-regulated kinases (ERKs), PI-3K, AKT and p70S6 kinase (S6K1). At the cellular level, apelin signaling participates in cell relaxation (smooth muscle cells), migration and proliferation, as well as contraction in cardiomyocytes (Audigier, 2006). Cell-cell junction proteins CDH5, CLDN5, CXCR4, GJA4, PCDH12, ESAM Cell adhesion, regulates transient increase in intracellular calcium MCAM concentration Basement membrane protein: stabilizes muscle cells in the heart COL15A1

Table 4.4.2-6 Gene Ontology of genes which belong to profile 16

182

Profile 17

Genes which belong to this profile showed a rate of increase less steep than genes which belong to profile 16 from day 30-45 of differentiation but the final level of expression at d90 of differentiation was higher than those genes which belong to profile 16.

Gene Ontology Gene list and Sites of expression/Functions Angiogenesis PGF, DLL4, PIK3CG, PLXND1, TIE1, PLXDC1, RASIP1, CYP1B1, ACVRL1, COL8A2, WARS, EFNA1, GPR124, HTATIP2, SRPX2, S1PR1, ITGA5, VASH1 and ERAP1 Regulation of cell motility MMP9, DLL4 and HGF and migration Regulation of localization LPL, MMP9 Lipid metabolic process ELOVL2, PTGIS, CYP1B1, PLA2G4A, C2CD4B, PIK3CG, PPARG, TRIB3, SNAI2, IGF2, NR3C1 and PDP1 Biological adhesion POSTN, ISLR, COL14A1 and PIK3CG Negative regulation of PIK3CG, POSTN cardiac contractility NOTCH signaling pathway DLK1, DLL4, SNAI2, CCND1, ADAM17 and POSTN Regulation of cardiac GATA6, FOXP1 muscle tissue growth

Table 4.4.2-7 Gene ontology of genes which belong to profile 17

Profile 18

Genes belonging to this profile showed a general up-regulation of genes, with a rather level expression from d30 – d45 post-differentiation in vitro, but rapid up-regulation from d45 to d90 post-differentiation in vitro. The full list of genes which belong to this profile is available in appendix 5.2.

183

Gene Ontology Gene list and Sites of expression/Functions Extracellular matrix TGFBR3, ADAMTS4, CILP2, COL6A6; development and SMOC and ADAMTS5 (both expressed by mature cardiac fibroblasts) remodeling Regulation of heart TGFBR3, TBX18 and SOX9 (altogether regulate TGFβ, BMP and Wnt signaling pathways) developmental processes Low-density lipoprotein COLEC12 and SCARF1; these genes mediate the uptake of modified LDL into cells for binding degradation Cell adhesion molecules B4GALT1 mediates cell-cell, cell-matrix interactions (by binding to specific oligosaccharide ligands on cells or matrix), as well as lactose biosynthesis; CBLN1 mediates cell-cell interactions and is expressed in atrio-ventricular canal cells (Seddon et al., 2004)

Table 4.4.2-8 Gene Ontology of genes which belong to profile 18

Profile 22

Genes in this profile group showed up-regulation from day 30-45, steady level of expression from day 45-60 and further up-regulation from day 60-90 in vitro.

Gene Functions SOX18 Heart tube development, endocardial cell differentiation and outflow tract morphogenesis. CHST1, BGN and CHST15 Glycosaminoglycan biosynthesis PDGFB, SPRY1, P2RY1, KDR and PDGFRA ERK1 and ERK2 cascade regulatory genes CD34, LYVE1, COL12A1, PCDH1, PODXL Cell adhesion genes RASSF2 Cell cycle arrest SPARCL1, PCDH1, FAT3and PCDHA7 calcium ion binding genes PIEZO2 and SLC40A1 Cation transport ANO1 Regulation of membrane potential XIRP2, ZFPM2, SOX18 Heart development XIRP2, PDGFB, EPS8 and FHL3 Actin and cell-cell junction cytoskeleton organization P2RY1, PDGFRA Regulation of cytosolic calcium concentration P2RY1 Positive regulation of carbohydrate metabolic process P2RY1, PDGFB, KDR, ZEB2, and PDGFRA MAPK cascade

Table 4.4.2-9 Functions of genes which belong to profile 22

184

Profile 23

Genes in this profile showed continuous and steady up-regulation of genes throughout in vitro culture, with higher rate of up-regulation in late phase of culture (d60 – 90). Complete list of genes which belong to this category is available in appendix 5.3.

Gene Ontology Gene list and Sites of expression/Functions Extracellular matrix COL1A1, COL5A1, TRAM2, CHST2, CSPG4, and PDGFRB (altogether for collagen and processes glycosaminoglycan biosynthesis) ADAMTS2, ADAMTS7, CILP, ECM2, TENC1, OLFM2A, PXDN, SPON1 and VWF (altogether to organize and up-regulate the expression of extracellular matrix proteins) PXDN also plays a role in mediating peroxidative reactions in the cardiovascular system (Shin et al., 2003). Cell adhesion SORBS3, COL5A1, ECM2, ITGA8, ITGA  expressed in endocardium, heart atrium and ventricle as well as pericardium SPON1  expressed in heart atrium STAB1  expressed in intercalated discs in heart muscle and endocardium KIAA1462, PCDHGA10, PCDHGB6, PCDHGC3, RAPH1, APBA1 and VWF. Cardiac NOTCH1 (heart valve, coronary vasculature and atrium formation) and HEG1 (atrium morphogenesis formation) Heart rate regulation AVPR1A and EPAS1 Transporters ABCC9 and KCNJ8  Potassium transport SLC43A3 (strongly expressed in endocardium, heart atrium and ventricle as well as pericardium) Cytoskeleton ARHGEF28, DTNBP1, FAM101B, FGD5, FMNL3, INF2, PDGFRB and WASF2 organization Signaling pathway Serine/Threonine (TGFβ) signaling pathway by NOTCH1, DAB2 and ITGA8 regulation Wnt signaling pathway by NOTCH1,COL1A1, DAB2 and GPRC5B MAPK signaling cascade by CSPG4, DAB2, PDGFRB, SASH1 and SORBS3

Table 4.4.2-10 Gene Ontology of genes which belong to profile 23

185

Profile 24

Profile 24 of figure 5.4.1-3 shows genes that are expressed in a relatively steady level up until day 60 of differentiation with sudden rapid up-regulation from day 60 onwards.

Gene Functions CD36 TGFβ binding, thrombospondin receptor activity, lipid binding, storage and transport, positive regulation of cellular metabolic process and protein phosphorylation, positive regulation of MAPK cascade (together with HIPK2, EGFR, PRKD2, TGFB1 and INSR), I-kappaB kinase/NF- kappaB signaling, positive regulation of reactive oxygen species metabolic process and cell adhesion (together with SFRP2, MYO10, PRKD2 and GSK3B), regulation of nucleocytoplasmic transport, as well as ions transmembrane transport. SCN7A A voltage gated sodium channel: Regulates membrane depolarization during action potential generation, maintains sodium ion homeostasis. Strongly expressed in adult human heart (http://www.nextprot.org/). GPC4 Connects cells to heparin sulphate proteoglycan in the ECM, aids in cellular lipid metabolic process HSPG2 Binds to glycosaminoglycan, involved in cardiac tissue development, cellular lipid and organic acid metabolic process TTYH3 A large-conductance Ca2+-activated chloride channel: Plays a role in Ca2+ signal transduction. KCNMB2 A negative regulator of Ca2+-activated K+-channel KCNMA1 ENG, HIPK2, ULK1, EGFR, DYRK1B, PDK4, CAMK4, Serine/threonine kinases, DYRK1B and GSK3B are involved RPS6KA2, MAST4, WNK1, PRKD2, MAPK7, LATS1, PAK2 in syncytium formation by plasma membrane fusion of and GSK3B myoblasts. PEAK1, EGFR, DYRK1B and INSR Tyrosine kinases BCL2, EGFR, TGFB1, RPS6KA2, JMY, and LATS1 Regulation of cell cycle process, the genes in bold negatively regulate mitotic cell cycle.

Table 4.4.2-11 Functions of highly-upregulated genes in profile 24

186

4.4.3. miRNA up- and down-regulation as EHMs age in vitro

Micro RNA (miRNA) expressions across different time points can be summarized by a heat map and profile trellis (figure 4.4.3-1 and -2). Based on the patterns of expression, up- regulated miRNAs belonging to profile 6 and 9 as well as down-regulated miRNAs belonging to profile 4 and 7 would be further analyzed. As each miRNA targets a large set of genes, only miRNAs that are expressed highly (>1000 copies at day 30 for down-regulated miRNAs or at day

90 for up-regulated miRNAs) and showed significant level of up- or down-regulation (> 8x) will be analyzed. These miRNAs are underlined in the following table. Unlike the mRNA expression data format above (log2fold-change data), the fold change of miRNA quoted below is the absolute fold change. Table 4.4.3-1 summarizes all miRNAs that are up-regulated as EHMs aged in vitro.

Profile 6 Profile 9 hsa-miR-582-5p hsa-let-7g-5p hsa-miR-150-5p hsa-miR-326 hsa-miR-483-3p hsa-miR-424-3p hsa-miR-16-1-3p hsa-miR-132-5p hsa-miR-10a-5p hsa-miR-214-3p hsa-miR-1247-3p hsa-miR-3129-3p hsa-miR-146b-5p hsa-miR-137 hsa-miR-1247-5p hsa-miR-450a-2-3p hsa-miR-98-5p hsa-miR-148a-5p hsa-miR-3613-5p hsa-miR-378a-5p hsa-miR-10a-3p hsa-miR-155-5p hsa-miR-216b-5p hsa-miR-574-3p hsa-miR-676-3p hsa-let-7a-5p hsa-miR-98-3p hsa-miR-92a-1-5p hsa-miR-3120-5p hsa-miR-2682-5p hsa-let-7c-3p hsa-miR-186-5p hsa-miR-214-5p hsa-miR-132-3p hsa-let-7c-5p hsa-miR-99a-5p hsa-miR-146b-3p hsa-miR-125b-1-3p hsa-let-7b-5p hsa-miR-15b-3p hsa-let-7i-3p hsa-miR-374a-3p hsa-let-7d-5p hsa-miR-629-5p hsa-miR-335-3p hsa-miR-16-2-3p hsa-let-7d-3p DQ590704 hsa-let-7f-2-3p hsa-miR-145-3p hsa-let-7i-5p hsa-miR-125b-2-3p hsa-miR-483-5p hsa-let-7e-5p hsa-let-7f-1-3p SNORD51 hsa-miR-99a-3p hsa-miR-425-3p hsa-let-7f-5p hsa-miR-22-5p hsa-miR-450b-5p hsa-miR-7706 hsa-miR-4532 SNORD12C hsa-miR-503-5p hsa-miR-199b-5p SNORA75 hsa-miR-212-3p hsa-miR-130a-5p hsa-miR-3613-3p hsa-miR-212-5p hsa-miR-6716-3p hsa-miR-887-3p hsa-miR-26a-2-3p

Table 4.4.3-1 List of up-regulated miRNAs

187

this particularthis sample gene expression level in

3.13 ” = red = 2 3.13 Heat map of miRNA expression EHMs expression different at in points miRNA time of map Heat

1 - 4.4.3 Figure Log2(fold change) scale. “

:

Dendogram the shows Scale hierarchy of the clusters which allows for exploratory analysis see to how the group microarrays on based together similarity of features 188

at time each

replicates of

expression profiles of EHMs at different time points. Grey area shows 1 standard deviation deviation Grey EHMs of expression different profiles at points. area time 1 standard shows

miRNA

2 - 4.4.3

Figure

0 at centered expression of mean with expression, Normalized

189

Upregulated miRNAs

Profile 6

Out of miRNAs in profile 6, hsa-miR-10a-5p shows the highest level of expression at day

90 at 78198 copies (12.62x higher in day 90 vs day 30 EHMs). The miRNA targets cardiac muscle cell differentiation genes (down-regulating TBX5 and up-regulating GATA6) and biosynthetic process genes (up-regulating ELOVL2, NFIX, and NR5A2). Next on the list with high level of expression at 44812 copies (11.87x higher in d90 vs d30 EHMs) is hsa-miR-146b-5p. It down- regulates cell-cell signaling gene STC1 and up-regulates biological adhesion and signaling gene

PCDH1. In addition, it up-regulates voltage-gated sodium channel gene involved in cardiac muscle cell action potential: SCN3B; actin filament bundle assembly genes ABL2 and GAS7 and mitochondrial fusion gene: VAT1.

has-miR-146b-5p

Normalized readscount) (Total Normalized

Figure 4.4.3-3 Dot plot of normalized read counts of hsa-miR-146b-5p in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range

Hsa-miR-483-3p targets a number of signal transduction genes, down-regulating RAB3B

(8.89x) and SHC3 (6.29x) and up-regulating GUCY1B3 (3.59x), HIPK2 (3.98x), DLC1 (4.78x), KITLG

(7.85x), IGF1 (10.45x) and PDGFB (13.43x). IGF1 and PDGFB also regulate polysaccharide metabolism. In addition, together with KITLG and HIPK2 (3.98x), IGF1 and PDGFB positively

190

regulates cell proliferation and MAPK cascade. Hsa-miR-483-3p down-regulates protein–binding gene involved in vesicular traffic RAB3B (8.89x). The table below summarizes the gene ontologies of all genes targeted by miR-214, while the following table summarizes the gene ontologies of all genes targeted by miR-98.

Target localization Developmental process System development Regulation of cell communication Down-regulation: Down-regulation: Up-regulation: Down-regulation: ZIC1 (21.2x) TFAP2C (16.33x) TRPS1 (4.24x) RUNX1 ZIC1 (21.2x) SORL1 (6.50) SALL1 (6.57x) (4.79x) SULF1 (4.79x) SALL1 (6.57x) MECOM (4.92x) SEMA3D -214-5p Regulation of lipid Up-regulation: Up-regulation: (5.54x) DCLK1 (5.64x)

miR transport EHD2 (4.14x) TRPS1 (4.24x) - PCDH1 (14.18x) TRPS1 (4.24x) RUNX1 (4.79x)

hsa Up-regulation: IRS2 (3.05x) RUNX1 (4.79x) SULF1 (4.79x) MECOM LIPG (3.37x) SULF1 (4.79x) (4.92x) RUNX1 (4.79x) FMNL3 (5.96x) BMP8A (7.67x) PGF (29.43x)

Table 4.4.3-2 Gene Ontology of genes targeted by hsa-miR-214-5p

Heart development and Blood vessel Extracellular matrix Positive regulation of transport growth development organization Up-regulation: Up-regulation: Up-regulation: Down-regulation: PDGFB (13.43x) PDGFB (13.43x) COL15A1 (26.02x) CACNA1D (3.39x) COL3A1 (4.76x) FZD4 (8.75x) COL14A1 (17.40x) RASL10B (3.53x) COL14A1 (17.40) PRRX1 (5.56x) COL1A1 (16.63x) Up-regulation: GJC1 (2.81x) TGFBR1 (2.21x) PDGFB (13.43x) PDGFB (13.43x) EDN1 (3.34x) COL1A1 (16.63x) SULF2 (9.72x) IGF1 (10.45x) MEF2D (2.57x) COL1A2 (8.94x) COL1A2 (8.94x) STXBP5 (3.76x) TGFBR1 (2.21x) TGFBR3 (6.44x) ADAMTS5 (7.15x) EDN1 (3.34x) TGFBR3 (6.44x) COL3A1 (4.76x) PXDN (4.84x) IRS2 (3.05x) DLC1 (4.78x) ITGB8 (4.42x) SULF1 (4.79x) EDA (3.05x) COL1A1 (16.63x) COL3A1(4.76x) TGFBR3 (6.44x) Positive COL11A1(4.63x) Positive regulation of MAPK cascade -98-5p regulation of ADAMTS1 (2.15x) COL9A3 (4.53x) Down-regulation: miR protein kinase B - FOXP1 (2.73x) ITGB8 (4.41x) EPHA4 (3.03x) signaling hsa COL5A2 (3.74x) Down-regulation: Down-regulation: CCR7 (6.11x) NID2 (2.67x) TBX5 (2.98x) CCR7 (6.11x) TGFBR1 (2.21x) Up-regulation: Innervation, also TSPYL5 (8.61x) HAS2 (2.12x) PDGFB (13.43x) negative regulation of Up-regulation: CRTAP (2.02x) IGF1 (10.45x) FGF receptor signaling IGF1 (10.45x) FZD4 (8.75x) pathway Down-regulation: HIPK2 (3.98x) SULF2 (9.72x) COL4A6 (4.44x) EDN1 (3.34x) SULF1 (4.79x) Positive regulation of Plasma membrane part Receptor signaling protein serine/ Wnt Signaling pathway threonine kinase activity Up-regulation: Up-regulation: MSR1 (2.80x) Up-regulation: COL1A1 (16.63x) PDGFB (13.43x) IGF1R (2.59x) NRL (41.65x)

191

SULF2 (9.72) PLXND1 (9.89x) SLC20A1 (2.55x) SULF1 (4.79x) FZD4 (8.75x) SLC16A10 (2.40x) EDA (3.5x) CD200 (7.33x) INSR (2.38x) Positive regulation of TGFBR3 (6.44x) TGFBR1 (2.21x) Transmembrane signaling receptor cell cycle EPHA3 (5.64x) HAS2 (2.12x) activity Up-regulation: DLC1 (4.78x) Up-regulation: PDGFB (13.43x) OSMR (4.75x) Down-regulation: PLXND1 (9.89x) IGF1 (10.45x) ITGB8 (4.42x) EPHA4 (3.03x) FZD4 (8.75x) CCND2 (4.61x) STXBP5 (3.76x) CACNA1D (3.39x) TGFBR3 (6.44x) CCND1 (2.69x) SLC4A7 (3.66x) TSPEAR (4.28x) EPHA3 (5.64x) EDN1 (3.34x) EDA (3.05x) SYT11 (4.34x) OSMR (4.75x) CBL (2.98x) IGDCC3 (5.10x) Down-regulation: CYTH3 (2.97x) CCR7 (6.11x) EPHA4 (3.03x) P2RX1 (2.93x) TRHDE (8.71x) CCR7 (6.11x) GJC1 (2.81x)

Table 4.4.3-3 Gene Ontology of genes targeted by hsa-miR-98-5p

Hsa-miR-582-5p targets genes that are involved in heart development: SFRP2, SALL1,

FOXF1, MEF2D, FOXC1, FBN1, HHEX and NOTCH1. It also up-regulates cell-cell junction proteins

DLC1, EPS8 and FZD4; up-regulates TGFβ signaling pathway, transmembrane receptor protein tyrosine kinase signaling pathway ADAM12, Wnt signaling pathway (TBX18, SFRP2, SOX9 and

NOTCH1) and down-regulates BMP signaling pathway (SFRP2 and NOTCH1).

Profile 9

Hsa-miR-3613-5p shows highest level of fold change at 15.46x higher in d90 vs d30 EHMs.

It targets metabolic regulating genes: such as the lipid metabolic process regulators, up- regulating IGF1R (2.59x), RORA (2.48x), SAMD8 (2.03x) and down regulating PDK1 (2.59x). It also regulates glucose metabolic process by up-regulating RORA (2.48x) and down-regulating PDK1

(2.59x). In addition, it also regulates lipid transport by up-regulating RUNX1 (4.79x). Finally, it up- regulates vasculature development genes RUNX1 (4.79x), VASH2 (3.13x) and RAP1B (2.39x) and cell surface receptor signaling pathway transducers TRPS1 (4.24x) and IGF1R (2.59x).

192

Second highest fold-change, at 8.23x higher in d90 vs d30 EHMs was shown by hsa-let-

7d. Interestingly, hsa-let-7d-5p belongs to the same family as hsa-miR-98-5p (a highly expressed miRNA from profile 6): the let-7 family. Other miRNAs in profile 9 that belong to this family include: hsa-let-7i-5p, hsa-let-7c-5p and hsa-let-7b-5p. Incidentally, these miRNAs belong to top

6 of most highly expressed miRNAs in profile 9. Thus, the gene ontology for let-7 family is the same as those of hsa-miR-98-5p and can be found on table 4.4.3-3.

An exception in the cut off process was made for miRNA hsa-miR-199b-5p, which shows a high level of expression in day 90 EHMs at 40695 copies and 7.11x fold change with respect to day 30 EHMs. The gene ontologies for genes targeted by this miRNA are listed on table 4.4.3-4.

hsa-let-7f-5p

Normalized readscount) (Total Normalized Figure 4.4.3-4 Dot plot of normalized read counts of hsa-let-7f-5p in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range

Heart Heart morpho- Syncytium Angiogenesis Positive regulation of development genesis formation Wnt-signaling pathway Up-regulation: Up-regulation: Up-regulation: Up-regulation: Up-regulation: ITGA4 (3.99x) DLC1 (4.78x) MYH9 (2.10x) PLXND1 (9.89x) SULF1 (4.79x) SOX4 (2.67x) MKL2 (2.35x) GSK3B (2.02x) ETS1(4.25x) CAV1 (2.93x)

-199b-5p MEF2D (2.57x) GPR124 (3.39x) SOX4 (2.67x) Down-regulation: CAV1 (2.93x) miR - Down-regulation: TGFB2 (2.16x) MYH9 (1010x) Extra-cellular matrix

hsa TGFB2 (2.16x) TGFB2 (2.16x) organization Trans-membrane Membrane raft Integral component of plasma Up-regulation: signaling receptor membrane CSGALNACT (14.38x) activity SULF1 (4.79x)

193

Up-regulation: Up-regulation: Up-regulation: Down- ITGA4 (3.99x) KIT (18.42x) PODXL (7.02x) PLXND1 (9.89x) regulation: COL5A3 (3.34x) PLXND1 (9.89x) SULF1 (4.79x) FZD4 (8.75x) PDPN (3.02x) LAMC1 (2.22x) FZD4 (8.75x) DLC1(4.78x) CD200 (7.33x) EPHA7(3.23x) GPRC5A (5.23x) RGMB (3.31x) PODXL (7.02x) FLRT3 (4.53x) Down-regulation: GPR124 (3.39x) CAV1 (2.93x) GPRC5A (5.23x) TGFB2 (2.16x) IGDCC3 (5.10x) PLXNA2 (2.94x) GSK3B (2.02x) CAV1 (2.93x) PTPRE (2.66x)

Down-regulation: Integrin complex EPHA7 (3.23x) Up-regulation: ITGA4 (9.99x) with KIT and PDPN also mediate MYH9 (2.10x) cell-cell adhesion

Table 4.4.3-4 Gene Ontology of genes targeted by hsa-miR-199b-5p

One miRNA that showed up with separate Partek flow® analysis was hsa-miR-425-5p with expression profile shown below. The genes targeted by this miRNA enrich to a number of ontologies. These are outlined in table 4.4.3-5

hsa-miR-425-5p Normalized readscount) (Total Normalized

Figure 4.4.3-5 Dot plot of normalized read counts of hsa-miR-425-5p in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range

194

Cell adhesion Regulation of Carbohydrate derivative Extracellular matrix transcription metabolic process organization Up-regulation: Up-regulation: Up-regulation: Up-regulation: FZD4 (+8.75x) IGF1 (+10.45x) IGF1 (+10.45x) ACAN (+4.54x) ACAN (+4.54x) FZD4(+8.75x) ACAN (+4.54x) FBN1 (+3.68x) RGMB (+3.31x) BNC2 (+4.26x) TIMP2 (+3.08x)

-425-5p CDON (+3.27x) FZD1 (+3.78x) Transmembrane-signaling Positive regulation of RGMB (+3.31x) receptor activity muscle cell miR - ZNF423 (+3.27x) differentiation

hsa CDON (+3.26x) Up-regulation: Up-regulation: CBL (+2.98x) FZD4 (+8.75x) CDON (+3.27x) ZFHX3 (+2.89x) FZD1 (+3.78x) ZFHX3 (+2.89x) CCND1(+2.69x) TGFBR2 (+3.63x) MAML1 (+2.33x) HNRNPD (+2.46x) EFNB2 (+2.20x)

Table 4.4.3-5 Gene ontology of genes targeted by hsa-miR-425

Down-regulated miRNAs

Profile 4 Profile 7 hsa-miR-154-5p hsa-miR-381-3p hsa-miR-485-3p MT-RNR1 hsa-miR-411-3p hsa-miR-376c-3p hsa-miR-615-3p hsa-miR-302a-5p hsa-miR-323b-3p hsa-miR-379-5p hsa-miR-377-5p hsa-miR-302c-3p hsa-miR-1185-2-3p SNORD114-14 SNORD114-22 hsa-miR-302b-3p SNORD114-9 hsa-miR-369-5p SNORD114-17 hsa-miR-494-3p hsa-miR-337-3p DQ590021 hsa-miR-412-5p hsa-miR-485-5p SNORD113-3 SNORD113-8 SNORD113 hsa-miR-490-5p hsa-miR-431-5p SNORD113-9 hsa-miR-363-3p hsa-miR-302a-3p SNORD114-23 SNORD113-7 hsa-miR-382-5p hsa-miR-302d-3p hsa-miR-196a-5p SNORD114-3 hsa-miR-370-3p hsa-miR-184 hsa-miR-431-3p SNORD113-6 hsa-miR-409-3p SNORD115-23 hsa-miR-323a-3p hsa-miR-411-5p SNORD33 SNORD115-5 hsa-miR-758-3p hsa-miR-433-3p hsa-miR-654-5p SNORD114-26 hsa-miR-409-5p hsa-miR-196b-5p SNORD83A SNORD114-1 hsa-miR-369-3p hsa-miR-493-5p hsa-miR-20b-5p MEG8 hsa-miR-487b-3p hsa-miR-127-3p hsa-miR-134-5p hsa-miR-432-5p MT-RNR2 hsa-miR-127-5p hsa-miR-106a-5p hsa-miR-129-2-3p hsa-miR-654-3p hsa-miR-493-3p hsa-miR-372-3p

Table 4.4.3-6 List of down-regulated miRNAs

Table 5.4.3-5 summarizes all miRNAs that were down-regulated as EHMs aged in vitro. However, there were only 2 miRNAs in which expression at day 30 EHMs is more than 1000 copies and fold change to expression on day 90 EHMs is more than 8 fold. In this light, the cut-off copy number was reduced to 500 for day 30 EHMS.

195

Profile 4

Using new cut-off value for copy number of 500, hsa-miR-493-3p and -5p showed the highest fold-change (37.23x and 31.79x, respectively) to expression at day 90 EHMs. Hsa-miR-

493 targets genes which enrich for:

1. Heart morphogenesis: COL5A1 (+3.01x) and NCOR2 (+2.43x) 2. Proteoglycan binding: GPC6 (+5.34x) and COL5A1 (+3.01x) 3. Growth-factor binding: LTBP2 (+4.91x), COL5A1 (+3.01x) and IGF1R (+2.59x) 4. calcium- and calmodulin-dependent protein kinase complex (+2.78x)

Hsa-miR-127-5p and -3p are the next 2 miRNAs showing highest fold change (d30 expression vs d90 expression) at 36.32x and 31.79x respectively. Unfortunately, the genes hsa- miR-127 targets do not intersect with the pool of 4900 mRNAs with known expression profiles in

EHMS at different time points, thus their ontologies could not be derived.

The last miRNA meeting cut off criteria is hsa-miR-411-5p. Interestingly, the top 50 gene ontologies point to embryonic eye development. More relevant gene ontologies of the targeted genes include:

1. Cardiovascular system development: SP2 (+2.32x) 2. TGFβ-receptor signaling pathway: HIPK2 (+3.98x) and SP1 (+2.11x) 3. Sodium-bicarbonate symporter activity of SLC4A7 (+3.66x) (log2 fold change) 4. Iron ion transmembrane transport: SCARA5 (-3.33x) 5. Lipid binding proteins: SYT4 (+6.40x), APOLD1 (+5.12x) and PREX1 (+3.01x). Together with FZD4 (+8.75x), APOLD1 is involved in lipid transport. 6. Wnt signaling pathway, calcium modulating pathway: FDZ4 (+8.75x). Together with HIPK2 (+3.98x), FDZ4 regulates JNK cascade.

196

Profile 7

Using the new cut-off criteria, two miRNAs were further analyzed. The first one is hsa-miR-

184 with 13.02x down-regulation from d30 to d90 EHMs. Genes targeted by hsa-miR-184 enrich for:

1. Extracellular matrix part RUNX1 (+4.79x) and the elastic fiber ELN (+4.32x) 2. Cell proliferation: RUNX1, ELN and NCOR2 (+2.43). NCOR2 also binds Notch. 3. Regulation of store-operated calcium entry: STC2 (+2.02x) 4. Intracellular sterol/cholesterol/lipid transport: NUS1 (+2.41x)

The second miRNA is hsa-miR-302d-3p, with 10.48x down-regulation from d30 to d90 EHMs.

Genes targeted by this miRNA enrich generally for transcriptional regulation:

Transcription factor Transcription regulatory Ion binding binding region DNA binding Up-regulation: Up-regulation: 79 genes, please refer to APBB2 (5.63x) ELK4 (4.12x) appendix 5.5 RUNX1 (4.79x) E2F7 (3.24x) Regulation of transcription, ZFPM (4.21x) TCF7L1 (2.97x) DNA-templated ZFPM1 (3.29x) GATAD2B (2.82x) NFIA (2.94x) RELA (2.49x) 55 genes, please refer to -302d-3p CCND1 (2.69x) ARID5B (2.40x) appendix 5.6 miR - RELA (2.49x) KLF12 (2.36x)

hsa RORA (2.48x) AHR (2.07x) AHR (2.07x) Down-regulation: Down-regulation: TIPARP (2.30x) MED12L (2.84x) FOXF2 (7.45x)

Table 4.4.3-7 Gene ontology for genes targeted by hsa-miR-302d

When miRNA data was analyzed again with Partek flow®, 2 further miRNAs were identified: hsa-miR-4461 and hsa-miR-6087. The read-counts dot plots for these miRNAs are shown below.

197

has-miR-4461 Normalized readscount) (Total Normalized

Figure 4.4.3-6 Dot plot of normalized read counts of hsa-miR-4461 in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range

Calcineurin complex and Ca2+-dependent Elastic fiber Cell junction protein serine/threonine phosphatase activity Up-regulation: Up-regulation: Up-regulation: PPP3CA (+2.43x) MFAP4 (+4.84x) GJC1 (+2.81x) PPP3R1 (+2.26x) SSPN (+2.65x) -4461 ADA (+2.29x) miR

- Sarcolemma Negative regulation of cytokinesis Down-regulation: hsa Up-regulation: GABRB2 (-3.48x) SSPN (+2.65x) Up-regulation: CADPS (-3.67x) PPP3CA (+2.43x) E2F7 (+3.24x) PPP3R1 (+2.26x)

Table 4.4.3-8 Gene ontologies of genes targeted by hsa-miR-4461

Hsa-miR-6087 Normalized readscount) (Total Normalized

Figure 4.4.3-7 Dot plot of normalized read counts of hsa-miR-6087 in EHMs at different time points. Box plots show interquartile range, with line showing median and whiskers symbolizing 10% to 90% range

Unfortunately there is no list of genes targeted by hsa-miR-6087 available in literature yet, so no interpretations of target functional grouping can be made.

198

4.4.4. Electrical stimulation improved force frequency and isoprenaline/carbachol

response of cardiomyocytes in EHMs

Experiment 1 (06.01.2014)

HES2 cardiomyocytes were mixed with human foreskin fibroblasts (HFFs) at 2:1 ratio and casted in “rat molds,” so-called as these molds were originally made for making rat EHM tissue constructs from rat CMs. Culture media used was “cardiac differentiation” medium containing

Biochrom Iscove’s basal media with 20% FBS, 1x NEAA, 1x PSG, ascorbic acid and 2- mercaptoethanol. At end of experiment, the remaining 2 stimulated EHMs were so thin they were not used for contraction experiment.

Stimulation Frequency (Hz) Period (days) 1.5 3 (6 EHMs) 2.0 3 (4 EHMs left) Basic square wave 2.5 3 (2 EHMs left) 5ms, 8V 3.0 3 (2 EHMs left) 2.5 3 (2 EHMs left)

Table 4.4.4-1 Stimulation parameters for first electrical stimulation experiment

A C

0.5cm 0.5cm

B Figure 4.4.4-1 Electrically stimulated EHM rings showed rougher surface and areas of thinning (arrows)

A – After 1 week of stimulation B – After 2 weeks of stimulation C – Control (unstimulated EHM) at same time point as B. 0.5cm

199

EHMs were digested with collagenase I for 1 hour at 370C followed by 0.25%

Trypsin/EDTA at 370C for 5 minutes. EHM2 recovered 3.98E5 cells, out of which 27.4% was viable

(1.091E5 live cells) while EHM3 recovered 2.13E5 cells, out of which 31% was viable (6.61E4 viable cells). To analyze the CM fraction of the isolated cells, alpha-actinin staining was carried out. Isolated cells were fixed in 1% paraformaldehyde (PFA) at room temperature for 15 minutes followed by 90% methanol at -200C for 10 minutes, washed with PBS before staining with alpha- actinin (Sigma) antibody and analyzed with FACS. The data showed that CM fraction in isolated cells were around 30% - 40%. To make one EHM, 3E6 total cells were used: 66.7% of these were

CMs and 33.3% of these were HFFs.

Mouse isotype control EHM2 EHM3 % FITC positive 0.9% 33.1% 37.8% α-actinin positive cell numbers - 36112 25000

Table 4.4.4-2 Alpha-actinin staining result of cells extracted from electrically stimulated EHMs

Experiment 2 (23.04.2014)

EHMs were cast in “mouse molds” on day 27 of differentiation and transferred onto auxotonic stretchers 3 days later. Electrical stimulation was only applied 2 weeks after EHMs were transferred onto auxotonic stretchers to promote cells consolidation and matrix remodeling. EHMs were cultured in serum free maturation media (SFMM).

Stimulation Frequency (Hz) Period (days) 1.5 3 Basic square wave 2.0 2 5ms, 6V 2.5 2 3.0 2

Table 4.4.4-3 Electrical stimulation parameters for experiment 3

At the end of electrical stimulation, media were sampled from all groups for G.

Luciferase expression. The result is summarized in table 4.4.4-4 below. EHMs were also analyzed under fluorescence stereomicroscope to analyze mCherry expression.

200

H9 EHMs H9V15 EHMs H9VC16 EHMs Control Control Stimulated Control Stimulated 1.80E+04 1.25E+06 2.87E+06 3.30E+05 2.49E+06 4.18E+06 2.08E+04 1.53E+06 1.75E+06 1.32E+06 2.73E+06 Average 1.94E+04 1.39E+06 2.31E+06 8.25E+05 3.13E+06 Fold change w.r.t NA 1.66 3.80 control

Table 4.4.4-4 Gaussia Luciferase expression of control EHMs versus electrically stimulated EHMs from 2 different cell lines

There was an increase of 1.66x fold in the luciferase signal in H9V15 EHMs and 3.80x fold in H9VC16 EHMs. This is consistent with earlier data that H9VC16 showed higher expression of reporter genes. This also indicated that MYL2 expression was enhanced with electrical stimulation. This quantitative data was also reflected in the mCherry expression data as illustrated in table 4.4.4-5 below. Electrically stimulated EHMs in both H9V15 and H9VC16 lines showed brighter mCherry signal than the non-electrically stimulated EHMs.

Cell line Non-stimulated Electrically stimulated

H9V15

H9VC16

Table 4.4.4-5 Electrically stimulated V15 and C16 EHMs showed higher mCherry expression than non-stimulated EHMs

201

Shown in figure 4.4.4-2 is a graph showing forces generated by C16 EHMs (control vs electrically stimulated). Throughout all concentrations of extracellular calcium used, electrically stimulated EHMs generated less force than the non-stimulated EHMs. This difference however, was not determined to be statistically significant (P = 0.0595). Forces were transformed with respect to the maximal forces to generate a percentage of maximal force graph. From this graph, the calcium concentration which corresponds to 50% of maximal forces can then be determined

(EC50 value). From figure 4.4.4-3, it appears that the EC50 from both electrically and non- electrically stimulated C16 EHMs did not differ very much; both were around 0.35 mM Ca2+.

1.0

0.8

0.6 NonControl-electrically EHM stimulated C16 EHMs

0.4 ElectricallyE.Stim C16 stimulated C16 EHMs

EHMforces (mN) 0.2

0.0 0.0 1.0 2.0 3.0 4.0 [Ca2+] in mM

Figure 4.4.4-2 Electrically stimulated EHMs contracted with lower forces than non-electrically stimulated EHMs

202

100

80

60

40 ElectricallyC16 E. Stim stimulated C16 EHMs

% of maximum force maximum of % 20 C16Non -controlelectrically stimulated C16 EHMs

0 0 1 2 3 4 [Ca2+] in mM

Figure 4.4.4-3 Electrical stimulation did not affect sensitivity of EHMs to extracellular calcium concentration

110 * * 100

90 ElectricallyC16 E. Stim stimulated C16 EHMs

% of force at Hz 1 NonC16 control-electrically stimulated C16 EHMs

80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 4.4.4-4 Electrically stimulated EHMs showed more positive force-frequency response than non-electrically stimulated EHMs

To assess CM maturation, one can also look at the force-frequency response, as shown in figure 4.4.4-4. Interestingly, as compared to the non-electrically stimulated group, the electrically stimulated group showed more statistically significant positive force frequency relationship. At 3Hz stimulation frequency, the EHM contractile forces from the non-electrically stimulated group plunged to 14% lower (vs 2.1% in electrically stimulated group) than their beating forces at 1 Hz. There was a slight positive force frequency relationship at 2Hz of

203

stimulation in the electrically stimulated group, resulting in an increase of 1.8% the force generated at 1Hz stimulation. This, however, might be contributed to arrhythmic contractions observed at 1Hz. Nonetheless, the fact that there was only a drop of 2.1% in contractile forces when stimulation frequency was increased to 3Hz from 1Hz showed that there was some degree of maturation in the CMs of electrically stimulated EHMs.

Another way to assess CM maturation is by measuring their response to β- adrenoreceptor agonist isoprenaline and cholinergic agonist carbachol. C16 EHMs responses to these drugs are summarized in figure 4.4.4-5. Electrically stimulated group showed higher inotropic response to isoprenaline (131.67% vs 109.51%). This difference however, was not shown to be statistically significant. The carbachol response was also stronger in the electrically stimulated EHMs. Following carbachol treatment, the beating force of electrically stimulated

EHMs decreased by 11.06% in comparison to the non-stimulated group, which decrease was only 6.43%.

Altogether the data demonstrated that although electrical stimulation caused cell death in the EHMs; the maturation level of surviving CMs was much more enhanced that those CMs in non-electrically stimulated group. This was determined by more expression of contractile protein MYL2, more positive force-frequency response, as well as better response to beta adrenergic and muscarinic receptor stimulation. However, this degree of maturation was not sufficient, as CMs from both groups still displayed similar level of extracellular calcium hypersensitivity.

204

Figure 4.4.4-5 C16 EHMs response to isoprenaline and carbachol

4.4.5. Cardiomyocytes in EHMs can tolerate a maximum of 10% increase in strain

Experiment 1 (07.07.2014) - gradual 20% stretch per week

EHMs were casted in “mouse-molds,” transferred onto 60 mm printed stretchers 3 days later and cultured in SFMM. EHMs were kept at 60 mm for 1 week before being transferred to

72 mm stretchers. EHMs were maintained at 72 mm for another week before being transferred further to 84 mm stretchers. The “mechanical-stretch only” group was cultured at 84mm while the “mechanical stretch and electrical stimulation” group was paced on the next day with the following parameters:

Stimulation Frequency (Hz) Period (days) 1.0 2 Basic square wave 1.5 2 5ms, 8V 2.0 2 2.5 4

Table 4.4.5-1 Electrical stimulation parameters for experiment 1

205

All EHMs generated no appreciable force at the end of the experiment and when looked under fluorescence microscope, showed no mCherry positive cells

Cell line Mechanical stretch only Mechanical + Electrical Stimulation

H9

H9VC16 60mm

Cell line Mechanical stretch only Mechanical + Electrical Stimulation H9VC16 84mm

Table 4.4.5-2 mCherry signal from H9 and C16 EHMs (mechanical only vs mechanical + electrical stimulation)

Experiment 2 (22.07.2014) concurrent with experiment 3 – total of 20% stretch

EHMs were casted in “mouse-mold,” transferred to 60 mm printed stretchers 3 days later and cultured in SFMM. 8 days later, EHMs were transferred to 66 mm printed stretchers and maintained on these stretchers for 9 days. Afterwards, EHMs were transferred to 72 mm printed stretchers and electrically stimulated on the next day with the following parameters:

206

Stimulation Frequency (Hz) Period (days) 1.0 2 Basic square wave 1.5 2 5ms, 8V 2.0 2 2.5 2

Table 4.4.5-3 Electrical stimulation parameters for experiment 2

At the end of the experiment, all EHMs with just mechanical stimulation on 72mm stretchers had snapped. 72mm electrically-stimulated EHMs showed only few visible mCherry positive cells (table 4.4.5-4).

Cell line Mechanical stretch only Mechanical + Electrical Stimulation

H9 60mm Only mechanically treated (act as fluorescent (control) bakground control only)

H9 72mm All snapped by the end of experiment (control)

H9V15 60mm (control)

H9V15 All snapped by the end of experiment 72mm

Table 4.4.5-4 mCherry signal from V15 EHMs (non-stimulated vs stimulated)

207

Shown in figure 4.4.5-1 is the G.Luc expression profile of H9 wild type EHMs and H9V15 knock-in EHMs. The profile shows reduction in signal as EHMs were stretched from 60 mm to

72mm (20% stretch). Electrical stimulation also seems to reduce the signal generated, with 60 mm stimulated EHMs generating similar G.Luc signal to 72 mm non-stimulated EHMs.

Altogether, the data indicated that electrical stimulation and passive 20% mechanical stretch either resulted in reduced expression of MYL2, or death in CMs containing MYL2. There are no error bars as biological replicates of EHMs were pooled into 1 well for electrical stimulation. RFU signal was averaged to the number of EHMs in the well.

Figure 4.4.5-1 Gaussia Luciferase expression of stimulated H9 and V15 EHMs

Figure 4.4.5-2 on the next page is similar to figure 4.4.4-2 in that electrically stimulated EHMs beat with lower forces than the non-mechanically stimulated ones. Of note, the red and green line, corresponding to electrically stimulated H9V15 EHMs stretched to 72 mm and electrically stimulated H9 EHMs stretched to 72 mm did not generate any appreciable forces. Interestingly,

H9 EHMs which were over-stretched during the Frank Starling sarcomeric length optimization

208

step of contraction experiment showed the same effect, hinting that stretching EHMs to 72 mm was probably over stretching the CMs, leading to cell death. From the EC50 curve (figure 4.4.5-

3), we see that electrically stimulated EHMs were less sensitive to extracellular calcium concentration (EC50 = 0.8mM Ca2+) in comparison with the non-electrically stimulated EHMs

(EC50 = 0.6mM Ca2+). This is probably indicative of more utilization of CTnI instead of the more calcium sensitive ssTnI (section 4.2.1). Interestingly, the wild type H9 EHMs showed higher contractile forces with even higher EC50 value at 0.9mM Ca2+.

0.5 Non-electrically stimulated H9 EHMs at 60mm 0.4 Electrically stimulated H9 EHMs at 60mm (overstretched during Frank-Starling)

0.3 Non -electrically stimulated V15 EHMs at 60mm Electrically stimulated V15 EHMs at 60mm

Force (mN) 0.2

Electrically stimulated H9 EHMs at 72mm 0.1 Electrically stimulated H9 EHMs at 72mm

0.0 0.0 1.0 2.0 3.0 4.0 [Ca2+] in mM

Figure 4.4.5-2 Electrically stimulated EHMs contracted with lower forces than non-electrically stimulated EHMs. 20% increase in preload was sufficient to abolish contractions in EHMs

100 Non-electrically stimulated H9 EHMs at 60mm Electrically stimulated H9 EHMs at 60mm 2+ (overstretched during Frank-Starling)

Non -electrically stimulated V15 EHMs at 60mm 50 Electrically stimulated V15 EHMs at 60mm

Electrically stimulated H9 EHMs at 72mm % force at 4mM Ca Electrically stimulated H9 EHMs at 72mm

0 0.0 1.0 2.0 3.0 4.0 [Ca2+] in mM

Figure 4.4.5-3 Electrical stimulation lessened calcium hypersensitivity in V15 EHMs at 60mm

209

Again, figure 4.4.5-4 is similar to figure 4.4.4-4 in that electrically stimulated H9V15

EHMs showed much less negative force-frequency response in comparison with the non- electrically stimulated group. For this graph, the statistical significance was calculated in comparison with non-stimulated H9V15 EHMs, not the wild type EHMs. There was an increase in contractile forces by about 20% when stimulation frequency was increased from 1Hz to 2Hz for the electrically stimulated group; whereas there was a drop of about 10% in the non-stimulated group. As stimulation frequency was further increased from 2Hz to 3Hz, the electrically stimulated group exhibited a drop in contractile forces by about 16.5%; whereas the non- stimulated group exhibited 22% drop in forces.

130 ** 120 * 110 EE.lectrically Stim stimulatedV15 60 V15 mm EHMs at 60mm 100

90 NonM.Stim-electrically H9 stimulated 60 mm H9 EHMs at 60mm NonM.-e Stimlectrically V15 stimulated 60 mm V15 EHMs at 60mm 80 % of force at 1 Hz at force of % 70

60 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 4.4.5-4 Electrical stimulation induced a more positive force-frequency response in V15 EHM at 60mm

In addition, electrically stimulated H9V15 EHMs showed more significant response to isoprenaline (+89.2% vs +5.5% in non-stimulated EHMs) and carbachol (-36.5% vs -4.4% in non- stimulated EHMs).

210

*

2+ 200

150

NonM.Stim-electrically H9 60stimulated mm H9 EHMs at 60mm 100 NonM. -eStimlectrically V15 stimulated 60 mm V15 EHMs at 60mm

EE.lectrically Stim V15 stimulated 60 mm V15 EHMs at 60mm 50 % of forces to% Ca force at 0.8mM 0 Isoprenaline Carbachol

Figure 4.4.5-5 Responses to isoprenaline and carbachol in electrically stimulated vs non-stimulated EHMs

* 100 2+ M.StimNon-electrically H9 60 stimulated mm H9 EHMs at 60mm M.Non Stim-electrically V15 stimulated60 mm V15 EHMs at 60mm 50 *

0.8mM Ca E.Electrically Stim V15 stimulated 60 mm V15 EHMs at 60mm Changes in forces as % to force % in forces as at Changes 0 Isoprenaline Carbachol

Figure 4.4.5-6 Electrical stimulation improved response to cardioactive drugs

Taken together, the results from this experiment supported the earlier hypothesis that electrical stimulation served to enhance CM maturation in the EHMs in that it improved calcium handling, force-frequency, as well as responses to isoprenaline and carbachol treatments. From this experiment, it was also observed that 20% strain was detrimental to CMs.

211

Experiment 3 (22.07.2014) – 10% stretch

Experiment 3 was very similar to experiment 2. The only difference being that the 72mm step was omitted: EHMs were maintained at 60mm for 5 days, transferred to and maintained at

66mm for 10 days before starting electrical stimulation. Electrical stimulation parameters were as follows:

Stimulation Frequency (Hz) Period (days) 1.0 1 Basic square wave 1.5 1 5ms, 5V 2.0 1 2.5 2

Table 4.4.5-5 Electrical stimulation parameters for experiment 3

The luciferase data below shows that in absence of electrical stimulation, there was not much difference in the expression level between EHMs stretched with 60mm or 66mm poles. In presence of electrical stimulation, however, a slight drop was observed in luciferase expression.

This difference, however, was not shown to be statistically significant.

Non-electrically stimulated H9 EHMs at 60mm

Non-electrically stimulated C16 EHMs

Electrically stimulated C16 EHMs

Figure 4.4.5-7 Gaussia Luciferase expression of stimulated H9 and C16 EHMs

212

Cell line Non-electrically stimulated Electrically stimulated

H9 60mm (control)

H9VC16 60mm (control)

H9VC16 66mm

Table 4.4.5-6 Electrical stimulation lowered mCherry signal in C16 EHMs. (Dotted white circles indicate position of printed poles)

213

Table 4.4.5-6 shows that mCherry signal was observed in EHMs stretched to 66mm, indicating that stretching to 66mm was not detrimental to CMs in the EHMs. Electrical stimulation, however, seems to cause CM death to a certain extent in that the electrically stimulated EHMs showed less mCherry signal than the non-electrically stimulated ones. It is interesting to note that CMs in electrically stimulated EHMs appeared to be more aligned (red cells appeared to be more anisotropic than those in non-electrically stimulated group), hinting that electrical stimulation might lead to improved sarcomeric organization.

1.0

0.8 NonM. -eStimlectrically H9 60mm stimulated H9 EHMs at 60mm ElectricallyE. Stim H9 stimulated 60mm H9 EHMs at 60mm

0.6 NonM. -eStimlectrically C16 60mmstimulated C16 EHMs at 60mm EE.lectrically Stim C16 stimulated 60mm C16 EHMs at 60mm 0.4

Force (mN) NonM. -eStimlectrically C16 stimulated66mm C16 EHMs at 66mm EE.lectrically Stim C16 stimulated 66mm C16 EHMs at 66mm 0.2

0.0 0.0 1.0 2.0 3.0 4.0 [Ca2+] in mM

Figure 4.4.5-8 Electrical stimulation and 10% increase in preload lowered contraction force in C16 EHMs

Shown above are the contraction force curves for all EHMs: electrically stimulated ones are represented by the darker colors (black for H9; blue for H9VC16 stretched with 60mm wide poles; and red for H9VC16 stretched with 66mm wide poles). The corresponding EHMs which were not electrically stimulated are represented in the lighter colors. From the graph, it can be observed that EHMs stretched with 60mm-wide poles beat with higher forces than those stretched with 66mm-wide poles. It is also apparent that electrically stimulated EHMs beat with

214

lower forces than the non-electrically stimulated EHMs. This trend was reversed for H9 EHMs however; the electrically stimulated EHMs were beating with higher forces than the non- electrically stimulated group.

2+ 100 NonM. -eStimlectrically H9 60mm stimulated H9 EHMs at 60mm E.Electrically Stim H9 stimulated 60mm H9 EHMs at 60mm M.Non Stim-electrically C16 60mm stimulated C16 EHMs at 60mm EE.lectrically Stim C16 stimulated 60mm C16 EHMs at 60mm 50 NonM. Stim-electrically C16 66mm stimulated C16 EHMs at 66mm EE.lectrically Stim C16 stimulated 66mm C16 EHMs at 66mm % forceat 4mM Ca

0 0.0 1.0 2.0 3.0 4.0 [Ca2+] in mM

Figure 4.4.5-9 10% increase in preload further worsened calcium hypersensitivity; 5V stimulation was not sufficient to lessen calcium hypersensitivity.

When % of maximal force was generated, EC50 values of H9VC16 EHMs stretched with

60mm-wide poles did not differ very much between the electrically stimulated and non- stimulated group around 0.7 mM Ca2+. For H9VC16 EHMs stretched with 66mm-wide poles, however, the electrically stimulated EHMs showed higher EC50 value of about 0.48 mM Ca2+ vs the non-electrically stimulated group at around 0.4 mM Ca2+. In general, EHMs stretched wider at 66mm showed more extracellular calcium hypersensitivity than those stretched at 60mm distance. Consistent with the contraction data, electrical stimulation enhanced calcium handling in H9 wild type EHMs, with electrically stimulated EHM having EC50 value of 1.0 mM Ca2+ and non-electrically stimulated EHM having EC50 value of about 0.78 mM Ca2+.

215

160

140 M.Non-e Stimlectrically H9 stimulated H9 EHMs at 60mm

120 ElectricallyE. Stim stimulated H9 60mmH9 EHMs at 60mm 100 NonM.-e Stimlectrically C16 stimulated 60mm C16 EHMs at 60mm 80 (not paced) paced) force at 1Hz EE.lectrically Stim stimulated C16 C1660mm EHMs at 60mm 60 % of % of Non-electrically stimulated C16 EHMs at 66mm 40 M. Stim C16 66mm EE.lectrically Stim stimulated C16 C1666mm EHMs at 66mm 20 1.0 1.5 2.0 2.5 3.0

Frequency (Hz) Figure 4.4.5-10 Force frequency response of electrically stimulated vs non-stimulated EHMs

The force-frequency response was rather puzzling, with electrically stimulated group showing more negative force-frequency response than the non-electrically stimulated group.

The light blue line represents the non-electrically stimulated group at basal stretching distance of 60mm wide: it remains constant at all stimulation frequencies tested. This was because the

EHMs could not be paced in the organ bath during measurement. It is interesting to note that

H9VC16 EHMs (be it electrically stimulated or not; or stretched at 60mm/66mm), all beat with only 30% of its forces at 1Hz with 3Hz stimulation, unlike the results of previous electrical stimulation experiments. At 2Hz, comparing the electrically stimulated EHMs, EHMs stretched at

66mm showed less drop (27%) in contractile forces than the ones stretched at 60mm (50%).

H9 EHMs again supported initial data: the electrically stimulated EHM showed positive force-frequency behavior from 1Hz to 2Hz, and a slight negative behavior from 2Hz to 3Hz.

However, the EHMs still beat with higher force at 3Hz than it did at 1Hz stimulation. The non- electrically stimulated H9 EHM showed no change as stimulation was increase from 1Hz to 2Hz, but slight negative response from 2Hz to 3Hz. For H9 EHMs, the electrically stimulated EHM also 216

fared better during isoprenaline stimulation (with about 150% increase in force) than the non- electrically stimulated EHM (75% increase). With carbachol treatment, electrically- as well as non-electrically stimulated H9 EHMs exhibited about 50% drop in contractile forces. Non- electrically stimulated H9VC16 EHMs at 60mm and 66mm both showed a more moderate increase in contractile force of around 15% after isoprenaline stimulation. The electrically stimulated EHMs, however, showed higher increase, with E. stim 60 mm group showing about

45% increase and the E. stim 66mm group showing about 25% increase. In general, the 60mm groups had better response to carbachol than the 66mm groups. M. stim 60 mm showed 10% drop after carbachol treatment; while E. stim 60 mm group showed 22.5% drop. M. stim 66 mm showed only 5% drop, whereas E. stim 66 mm showed 15% drop following carbachol treatment.

Taken together, for this experiment, 60mm stretch seems to be optimal for the EHMs in terms of MYL2 expression, contractile forces, EC50 values, as well as responses to isoprenaline and carbachol. Force frequency response seems to favor the 66mm group, however. In this experiment, electrical stimulation seemed to fail in maturing the CMs in the EHMs, with the sole exception of H9 EHMs. CMs in electrically stimulated H9 EHMs apparently showed more advance maturation, in terms of contractile forces, EC50 value, force-frequency response as well as responses to isoprenaline and carbachol. The data on H9 was not conclusive however; as there was only 1 EHM per group for H9 EHM (these EHMs were initially only used as negative control for luciferase assay and mCherry signal). Further experiments are needed to substantiate this claim.

217

250 2+

200 NonM. -eStimlectrically H9 stimulated60 mm H9 EHMs at 60mm ElectricallyE. Stim H9 stimulated 60 mm H9 EHMs at 60mm 150 NonM.-e Stimlectrically C16 stimulated 60 mm C16 EHMs at 60mm 100 EE.lectrically Stim C16stimulated 60mm C16 EHMs at 60mm

NonM. -eStimlectrically C16 stimulated 66mm C16 EHMs at 66mm 50 E.Electrically Stim C16 stimulated 66mm C16 EHMs at 66mm % of forces to% Ca force at 0.8mM 0 Isoprenaline Carbachol

Drug added

Figure 4.4.5-11 Responses to isoprenaline and carbachol in electrically stimulated vs non-stimulated EHMs

2+ 150

MechanicalNon-electrically Stim stimulated H9 60 H9 mmEHMs at 60mm 100 ElectricallyElectrical stimulated Stim H9 H9 60 EHMs mm at 60mm

MechanicalNon-electrically Stim stimulated C16 C1660 mmEHMs at 60mm ElectricalElectrically stimulated Stim C16 C16 60 EHMs mm at 60mm 50 NonMechanical-electrically Stim stimulated C16 C1666 EHMsmm at 66mm EElectricallectrically stimulated Stim C16 C16 66 EHMs mm at 66mm

0 Isoprenaline Carbachol % change in forces wrt forces at 0.8mM Ca change in forces wrt% forces at 0.8mM Drug added

Figure 4.4.5-12 %change in forces after isoprenaline and carbachol treatment

218

4.5. Discussion

In vivo, CMs take 6 – 10 years to the reach their adult phenotype (Peters et al., 1994).

This time frame is impractical to adopt in the laboratory. Thus in this chapter, in vitro techniques to enhance CM maturation in a time-effective manner were investigated.

Many investigators in the field have turned to biomimetic approaches to enhance their

CMs maturation. Such approaches are:

1. In vitro aging techniques (Kamakura et al., 2013; Kehat et al., 2002; Lundy et al., 2013; Snir et al., 2003) 2. Varying the mechanical properties of CMs attachment surface (Hazeltine et al., 2012; Rodriguez et al., 2011); 3. Engineering anisotropic CMs attachment surfaces (Au et al., 2009; Bray et al., 2008; Kim et al., 2010b; Wang et al., 2011); 4. Mechanical conditioning/stretching (Fink et al., 2000; Foldes et al., 2011; Tulloch et al., 2011) 5. Electrical stimulation (Deng et al., 2000; Hirt et al., 2014; Martherus et al., 2010; Nunes et al., 2013) 6. Hormonal stimulation (Chattergoon et al., 2012; Foldes et al., 2011; Klein and Ojamaa, 2001; Kruger et al., 2008; Lee et al., 2010; McMullen et al., 2004; Nistri et al., 2012) 7. Micro-RNA (Callis et al., 2009; Fu et al., 2011; van Rooij et al., 2007)and 8. Co-culture with other cell types (Kim et al., 2010a; Tiburcy et al., 2014)

Each approach was either used singularly, or in combination with each other. Studies employing these techniques have reported advanced CM maturation phenotypes, such as enlarged rod-shaped CMs, progressive exit from cell cycle, improved ultrastructural organization, action potential phenotypes, calcium current density and up-regulation of ion channels. Despite these encouraging reports, a few things were always missing: quiescent state of ventricular CMs, adult level contractile forces, t-tubules, as well as adult electrophysiological properties such as

219

fast conduction velocities, fast maximum rate of depolarization and relatively long APD. Thus current techniques were only generating intermediately mature CMs.

In this chapter, two in vitro maturation methods were investigated. The first attempt involved in vitro passive aging for a period of 3 months; this was combined with auxotonic mechanical conditioning and treatment with IGF-1, VEGF and FGF but not electrical stimulation in attempt to mature CMs in EHMs. The RNAs from these EHMs at different time points (day 30,

45, 60 and 90 post-differentiation) were sequenced in order to better understand the transcriptomic events taking place as EHMs aged over time in culture. The second attempt used electrical stimulation in combination with growth factors (as in the first attempt) and auxotonic stretching either at basal width of 60mm or wider (66mm, 72mm or 84mm).

MRNA and miRNA sequencing data from the passive aging experiment showed that down-regulated genes mainly enriched for: (1) Cell proliferation (2) ERKs, PI-3K, Akt and p70S6 kinase (S6K1) signaling cascades as well as (3) Carbohydrate metabolism.

Up-regulated genes, on the other hand, enriched to mainly 10 biological activities:

(1) Cardiac development, morphogenesis and contractile forces; (2) Extracellular matrix development, assembly and remodeling; (3) Ion and lipid binding and transport (calcium ion included); (4) Cytoskeletal reorganization; (5) Wnt, MAPK and I-kappaB kinase/NF-kappaB cascade; (6) Heart rate; (7) Cell-cell junction proteins; (8) Calcium-mediated signaling; (9) Fat metabolism; and (10) Negative regulation of cell proliferation. 220

As such, the RNA-Seq and small RNA-Seq data suggested an improved level of maturation of the CMs in the EHMs over 3 months culture period as seen from reduced proliferation capacity, metabolic substrate preference, structural organization and expression of calcium handling proteins.

Of particular interest are the following up-regulated genes: FOXP1, GATA6, TBX18, and

NOTCH1. In mice, Foxp1 has been reported to control CMs proliferation through different mechanisms in the endocardium and myocardium (Wang and Morrisey, 2010; Zhang et al., 2010).

When expressed in the CMs, Foxp1 restricts CM proliferation autonomously by repressing its direct target Nkx2.5 (Wang and Morrisey, 2010). When expressed in the endocardium however,

Foxp1 enhances CMs proliferation by repressing Sox17, which is a known repressor of Wnt/β- catenin signaling (Zorn et al., 1999). As such, Foxp1 activation causes the derepression of Wnt/β- catenin signaling, which results in increased expression of the important mitogens for proliferation: Fgf16 and Fgf20 (Wang and Morrisey, 2010).

Gata6 is expressed in the mesoderm and the endoderm and appears to be required for the maturation of cardiac phenotype, at least in the Xenopus and murine backgrounds (Peterkin et al., 2003). In these non-myogenic tissues, gata6 regulates the secretion of cardiogenic signals which may in turn, regulate the maintenance but not initial induction of Bmp4 and Nkx2 family expressions (Loose and Patient, 2003; Peterkin et al., 2003). Gata6 loss-of-function resulted in heartless phenotype in Xenopus background and morphogenetic defects in murine background; heartless phenotype in murine background was observed only with concomitant knock-out of gata4 (Nemer and Horb, 2007; Zhao et al., 2008). Tbx18 has been described to be expressed in

CMs derived from proepicardial organ (Cai, et al., 2008). These CMs are migratory and make substantial contribution to CMs in ventricular septum, atrial and ventricular chambers of the

221

heart (Cai et al., 2008). In a separate study, Wiese et al. (2009) found that Tbx18 was required to establish the large head structure of the mouse sinoatrial node from mesenchymal precursors.

Notch1 has been described as a regulator of cardiac stem cells growth and commitment to the

CM lineage (Urbanek et al., 2010). Inhibition of the Notch pathway in newborn mice resulted in impaired cardiomyogenesis and dilated myopathy followed by high mortality (Urbanek et al.,

2010). The up-regulation of FOXP1, GATA6, TBX18 and NOTCH1 could imply the role of these factors in directing CM maturation, i.e. these four factors could be “maturation” factors. It will be interesting to see the impact of adding these four factors into immature population of CMs.

Figure 4.4.1-1 and -2 show that fused EHMs had low maximal diastolic potentials (MDP) averaging at around -70.6mV. This value is rather close to the adult CM MDP which is around

-85mV (Drouin et al., 1998). The maximal rate of depolarization (Vmax) of fused EHMs action potential (AP), however, was still rather slow at 169.66 V/s, almost twice lower than that of the adult CMs (around 300 V/s). In terms of contractile forces, EHMs generated a maximum of 1.96 ±

0.16 mN (mean ± SEM) by day 90 of differentiation (day 72 after EHMs casting). Adult myocardium strip twitch tension has been reported to be around 44 ± 11.7 mN/mm2 (Hasenfuss et al., 1991). Using “mouse molds,” we routinely generated EHMs with total cross sectional area of about 0.5 mm2 (unpublished in-house data). Thus the oldest EHMs (90days of differentiation) beat with a maximum of about 4mN/mm2, or eleven times lower than the reported data by

Hasenfuss et al. The EHMs were also autonomously contracting.

Taken together, the data indicated that in vitro aging, in conjunction with IGF-1, VEGF and FGF treatment matured CMs in the EHMs only to an intermediate stage. Thus another approach was attempted: increasing EHMs preload (mechanical conditioning) and electrical stimulation.

222

The first round of electrical stimulation experiments was carried out in serum-containing media and resulted in progressive thinning of EHMs over time, which eventually led to snapping.

It was postulated that electrical stimulation generated too much reactive oxygen species or other cytotoxic by-products that were detrimental to the cells. However, alpha actinin staining indicated that CMs fraction was still around 32.2% - 36.9%, which is a normally observed percentage of CMs after collagenase I digestion (unpublished in-house data). So it seems that the electrical stimulation could be killing the fibroblasts in the EHMs instead.

The second experiment with electrical stimulation used lower voltage (6V instead of 8V) and shorter stimulation protocol (a total of 9 days stimulation instead of 15 days as in first protocol). With this second attempt, there was still cell death but much less than that observed in the first attempt. In general, the data showed that CM maturity was enhanced in the electrically stimulated group but all EHMs were still showing extracellular calcium hypersensitivity.

In further experiments, electrical stimulation was carried out in parallel with increasing the preload (stretching EHMs to an increased length). It was observed, however, that a gradual increase of 20% strain per week was too much for the CMs; they were overstretched and consequently stopped beating, possibly due to CM rupture following overstretching. Gradual 10% strain per week to a total of 20% strain was also found to be too much for the CMs. Indeed, the

CMs could only tolerate a total strain of 10%.

Without electrical stimulation, EHMs stretched from 60mm to 66mm did not show any decrease in average G. Luc signal. The 66mm-EHMs were also still beating nicely with prominent mCherry expression. With electrical stimulation, however, 66mm EHMs seemed to express less

G. Luc signal. This might be due to CM death caused by electrical stimulation. 223

When only the 60mm EHMs group were analyzed for the electrically stimulated group, experiment batch 22.07.14 EHMs showed more maturity in terms of the degree of calcium hypersensitivity than the EHMs from batch 23.04.14. The difference between these two experiments is in the voltage used during electrical stimulation: batch 22.07 used 8V while batch

23.04 used 6V. In addition, the stimulation duration was less in batch 22.07 at 8 days versus batch 23.04 which was 9 days. However, there was still quite a significant degree of cell death observed with the 22.07.14 electrical stimulation protocol. Thus it was thought that lower stimulation voltage at 5V and shorter stimulation protocol of 5 days would reduce the occurrence of cell death. Five volts of electrical stimulation was observed to be insufficient to pace the EHMs. Despite the low voltage used, some degree of CM death was still observed.

Consequently, in comparison with the 60 mm non-electrically stimulated group, the 60 mm electrically stimulated EHMs beat with weaker forces, showed similar level of extracellular calcium hypersensitivity, and displayed very negative force frequency response. The only sign that the electrically stimulated EHMs were more mature than the non-stimulated ones was the fact that the electrically stimulated EHMs showed better responses to isoprenaline and carbachol treatments. In short, 5V stimulation for 5 days was not enough to induce phenotypical changes in CMs inside EHMs to drive them towards maturation. Protocol 4.4.4.-3 proved to be the best amongst all protocols tested. However, the EHM contractile forces were still very low at an average of around 0.33mN.

From this chapter, it was observed that the in vitro maturation methods tested so far only enhanced the maturation of CMs within EHMs to an intermediate level. A robust protocol which enhances in vitro maturation of CMs within EHMs to the level expected as in vivo remains to be optimized.

224

4.6. References

Angst, B.D., Khan, L.U., Severs, N.J., Whitely, K., Rothery, S., Thompson, R.P., Magee, A.I., and Gourdie, R.G. (1997). Dissociated spatial patterning of gap junctions and cell adhesion junctions during postnatal differentiation of ventricular myocardium. . Circ Res 80, 88–94. Au, H.T.H., Cui, B., Chu, Z.E., Veres, T., and Radisic, M. (2009). Cell culture chips for simultaneous application of topographical and electrical cues enhance phenotype of cardiomyocytes. Lab on a chip 9, 564-575. Audigier, Y. (2006). Apelin receptor. Signaling Gateway - UCSD Molecule Pages. Barth, E., Stammler, G., Speiser, B., and Schaper, J. (1992). Ultrastructural quantitation of mitochondria and myofilaments in cardiac muscle from 10 different animal species including man. Journal of molecular and cellular cardiology 24, 669-681. Bedada, Fikru B., Chan, Sunny S.-K., Metzger, Stefania K., Zhang, L., Zhang, J., Garry, Daniel J., Kamp, Timothy J., Kyba, M., and Metzger, Joseph M. (2014). Acquisition of a Quantitative, Stoichiometrically Conserved Ratiometric Marker of Maturation Status in Stem Cell-Derived Cardiac Myocytes. Stem cell reports 3, 594-605. Bergmann, O., Bhardwaj, R.D., Bernard, S., Zdunek, S., Barnabe-Heider, F., Walsh, S., Zupicich, J., Alkass, K., Buchholz, B.A., Druid, H., et al. (2009). Evidence for cardiomyocyte renewal in humans. Science 324, 98- 102. Bers, D.M. (2002). Cardiac excitation-contraction coupling. Nature 415, 198-205. Bhavsar, P.K., Dhoot, G.K., Cumming, D.V., Butler-Browne, G.S., Yacoub, M.H., and Barton, P.J. (1991). Developmental expression of troponin I isoforms in fetal human heart. . FEBS letters 292, 5–8. Bird, S. (2003). The human adult cardiomyocyte phenotype. Cardiovascular Research 58, 423-434. Borg, T.K., Rubin, K., Lundgren, E., Borg, K., and Obrink, B. (1984). Recognition of extracellular matrix components by neonatal and adult cardiac myocytes. Developmental biology 104, 86-96. Bray, M.A., Sheehy, S.P., and Parker, K.K. (2008). Sarcomere alignment is regulated by myocyte shape. Cell motility and the cytoskeleton 65, 641-651. Callis, T.E., Pandya, K., Seok, H.Y., Tang, R.H., Tatsuguchi, M., Huang, Z.P., Chen, J.F., Deng, Z., Gunn, B., Shumate, J., et al. (2009). MicroRNA-208a is a regulator of cardiac hypertrophy and conduction in mice. The Journal of clinical investigation 119, 2772-2786. Cao, F., Wagner, R.A., Wilson, K.D., Xie, X., Fu, J.D., Drukker, M., Lee, A., Li, R.A., Gambhir, S.S., Weissman, I.L., et al. (2008). Transcriptional and functional profiling of human embryonic stem cell-derived cardiomyocytes. PloS one 3, e3474. Chattergoon, N.N., Giraud, G.D., Louey, S., Stork, P., Fowden, A.L., and Thornburg, K.L. (2012). Thyroid hormone drives fetal cardiomyocyte maturation. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 26, 397-408. Chen, H.S., Kim, C., and Mercola, M. (2009). Electrophysiological challenges of cell-based myocardial repair. Circulation 120, 2496-2508. Deng, X.F., Rokosh, D.G., and Simpson, P.C. (2000). Autonomous and growth factor-induced hypertrophy in cultured neonatal mouse cardiac myocytes. Comparison with rat. Circ Res 87, 781-788. Dolnikov, K., Shilkrut, M., Zeevi-Levin, N., Gerecht-Nir, S., Amit, M., Danon, A., Itskovitz-Eldor, J., and Binah, O. (2006). Functional properties of human embryonic stem cell-derived cardiomyocytes: intracellular Ca2+ handling and the role of sarcoplasmic reticulum in the contraction. Stem cells 24, 236-245. 225

Drouin, E., Lande, G., and Charpentier, F. (1998). Amiodarone reduces transmural heterogeneity of repolarization in the human heart. Journal of the American College of Cardiology 32, 1063-1067. Fijnvandraat, A. (2003). Cardiomyocytes derived from embryonic stem cells resemble cardiomyocytes of the embryonic heart tube. Cardiovascular Research 58, 399-409. Fink, C., Ergun, S., Kralisch, D., Remmers, U., Weil, J., and Eschenhagen, T. (2000). Chronic stretch of engineered heart tissue induces hypertrophy and functional improvement. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 14, 669-679. Foldes, G., Mioulane, M., Wright, J.S., Liu, A.Q., Novak, P., Merkely, B., Gorelik, J., Schneider, M.D., Ali, N.N., and Harding, S.E. (2011). Modulation of human embryonic stem cell-derived cardiomyocyte growth: a testbed for studying human cardiac hypertrophy? Journal of molecular and cellular cardiology 50, 367- 376. Fu, J.D., Rushing, S.N., Lieu, D.K., Chan, C.W., Kong, C.W., Geng, L., Wilson, K.D., Chiamvimonvat, N., Boheler, K.R., Wu, J.C., et al. (2011). Distinct roles of microRNA-1 and -499 in ventricular specification and functional maturation of human embryonic stem cell-derived cardiomyocytes. PloS one 6, e27417. Gerdes, A.M., Kellerman, S.E., Moore, J.A., Muffly, K.E., Clark, L.C., Reaves, P.Y., Malec, K.B., McKeown, P.P., and Schocken, D.D. (1992). Structural remodeling of cardiac myocytes in patients with ischemic cardiomyopathy. Circulation 86, 426–430. Giordano, F.J. (2005). Oxygen, oxidative stress, hypoxia, and heart failure. The Journal of clinical investigation 115, 500-508. Harris, D.A., and Das, A.M. (1991). Control of mitochondrial ATP synthesis in the heart. The Biochemical journal 280 ( Pt 3), 561-573. Hasenfuss, G., Mulieri, L.A., Blanchard, E.M., Holubarsch, C., Leavitt, B.J., Ittleman, F., and Alpert, N.R. (1991). Energetics of isometric force development in control and volume-overload human myocardium: Comparison with animal species. . Circ Res 68 836–846. Hattori, F., Chen, H., Yamashita, H., Tohyama, S., Satoh, Y.S., Yuasa, S., Li, W., Yamakawa, H., Tanaka, T., Onitsuka, T., et al. (2010). Nongenetic method for purifying stem cell-derived cardiomyocytes. Nat Methods 7, 61-66. Hazeltine, L.B., Simmons, C.S., Salick, M.R., Lian, X., Badur, M.G., Han, W., Delgado, S.M., Wakatsuki, T., Crone, W.C., Pruitt, B.L., et al. (2012). Effects of Substrate Mechanics on Contractility of Cardiomyocytes Generated from Human Pluripotent Stem Cells. International Journal of Cell Biology 2012, 13. Henderson, D.J., and Chaudhry, B. (2011). Getting to the heart of planar cell polarity signaling. Birth defects research Part A, Clinical and molecular teratology 91, 460-467. Hirt, M.N., Boeddinghaus, J., Mitchell, A., Schaaf, S., Börnchen, C., Müller, C., Schulz, H., Hubner, N., Stenzig, J., Stoehr, A., et al. (2014). Functional improvement and maturation of rat and human engineered heart tissue by chronic electrical stimulation. Journal of molecular and cellular cardiology 74, 151-161. Hou, J.H., Kralj, J.M., Douglass, A.D., Engert, F., and Cohen, A.E. (2014). Simultaneous mapping of membrane voltage and calcium in zebrafish heart in vivo reveals chamber-specific developmental transitions in ionic currents. Frontiers in physiology 5. http://www.nextprot.org/. Hunkeler, N.M., Kullman, J., and Murphy, A.M. (1991). Troponin I isoform expression in human heart. Circ Res 69, 1409–1414. Huss, J.M., and Kelly, D.P. (2005). Mitochondrial energy metabolism in heart failure: a question of balance. The Journal of clinical investigation 115, 547-555.

226

Kamakura, T., Makiyama, T., Sasaki, K., Yoshida, Y., Wuriyanghai, Y., Chen, J., Hattori, T., Ohno, S., Kita, T., Horie, M., et al. (2013). Ultrastructural Maturation of Human-Induced Pluripotent Stem Cell-Derived Cardiomyocytes in a Long-Term Culture. Circulation Journal 77, 1307-1314. Kehat, I., Gepstein, A., Spira, A., Itskovitz-Eldor, J., and Gepstein, L. (2002). High-resolution electrophysiological assessment of human embryonic stem cell-derived cardiomyocytes: a novel in vitro model for the study of conduction. Circ Res 91, 659-661. Kim, C., Majdi, M., Xia, P., Wei, K.A., Talantova, M., Spiering, S., Nelson, B., Mercola, M., and Chen, H.S. (2010a). Non-cardiomyocytes influence the electrophysiological maturation of human embryonic stem cell-derived cardiomyocytes during differentiation. Stem cells and development 19, 783-795. Kim, D.H., Lipke, E.A., Kim, P., Cheong, R., Thompson, S., Delannoy, M., Suh, K.Y., Tung, L., and Levchenko, A. (2010b). Nanoscale cues regulate the structure and function of macroscopic cardiac tissue constructs. Proceedings of the National Academy of Sciences of the United States of America 107, 565-570. Klein, I., and Ojamaa, K. (2001). Thyroid hormone and the cardiovascular system. The New England journal of medicine 344, 501-509. Kruger, M., Sachse, C., Zimmermann, W.H., Eschenhagen, T., Klede, S., and Linke, W.A. (2008). Thyroid hormone regulates developmental titin isoform transitions via the phosphatidylinositol-3-kinase/ AKT pathway. Circulation research 102, 439-447. Laflamme, M.A., and Murry, C.E. (2011). Heart regeneration. Nature 473, 326-335. Lahmers, S., Wu, Y., Call, D.R., Labeit, S., and Granzier, H. (2004). Developmental control of titin isoform expression and passive stiffness in fetal and neonatal myocardium. Circ Res 94 505–513. Lakatta, E.G., and Sollott, S.J. (2002). Perspectives on mammalian cardiovascular aging: humans to molecules. Comparative biochemistry and physiology Part A, Molecular & integrative physiology 132, 699- 721. Lee, Y.K., Ng, K.M., Chan, Y.C., Lai, W.H., Au, K.W., Ho, C.Y., Wong, L.Y., Lau, C.P., Tse, H.F., and Siu, C.W. (2010). Triiodothyronine promotes cardiac differentiation and maturation of embryonic stem cells via the classical genomic pathway. Molecular endocrinology (Baltimore, Md) 24, 1728-1736. Lieu, D.K., Liu, J., Siu, C.W., McNerney, G.P., Tse, H.F., Abu-Khalil, A., Huser, T., and Li, R.A. (2009). Absence of transverse tubules contributes to non-uniform Ca(2+) wavefronts in mouse and human embryonic stem cell-derived cardiomyocytes. Stem cells and development 18, 1493-1500. Loose, M., and Patient, R. (2003). GATA-6 in Heart Development. Lundy, S.D., Zhu, W.Z., Regnier, M., and Laflamme, M.A. (2013). Structural and functional maturation of cardiomyocytes derived from human pluripotent stem cells. Stem cells and development 22, 1991-2002. Martherus, R.S., Vanherle, S.J., Timmer, E.D., Zeijlemaker, V.A., Broers, J.L., Smeets, H.J., Geraedts, J.P., and Ayoubi, T.A. (2010). Electrical signals affect the cardiomyocyte transcriptome independently of contraction. Physiological genomics 42a, 283-289. McMullen, J.R., Shioi, T., Huang, W.Y., Zhang, L., Tarnavski, O., Bisping, E., Schinke, M., Kong, S., Sherwood, M.C., Brown, J., et al. (2004). The insulin-like growth factor 1 receptor induces physiological heart growth via the phosphoinositide 3-kinase(p110alpha) pathway. The Journal of biological chemistry 279, 4782- 4793. Metzger, J.M., Michele, D.E., Rust, E.M., Borton, A.R., and Westfall, M.V. (2003). Sarcomere thin filament regulatory isoforms. Evidence of a dominant effect of slow skeletal troponin I on cardiac contraction. J Biol Chem 278, 13118–13123. Nemer, M., and Horb, M.E. (2007). The KLF Family of Transcriptional Regulators in Cardiomyocyte Proliferation and Differentiation. Cell Cycle 6, 117-121. 227

Nistri, S., Pini, A., Sassoli, C., Squecco, R., Francini, F., Formigli, L., and Bani, D. (2012). Relaxin promotes growth and maturation of mouse neonatal cardiomyocytes in vitro: clues for cardiac regeneration. Journal of cellular and molecular medicine 16, 507-519. Nuccitelli, R. (1992). Endogenous ionic currents and DC electric fields in multicellular animal tissues. Bioelectromagnetics Suppl 1, 147-157. Nunes, S.S., Miklas, J.W., Liu, J., Aschar-Sobbi, R., Xiao, Y., Zhang, B., Jiang, J., Masse, S., Gagliardi, M., Hsieh, A., et al. (2013). Biowire: a platform for maturation of human pluripotent stem cell-derived cardiomyocytes. Nat Methods 10, 781-787. Opitz, C.A., Leake, M.C., Makarenko, I., Benes, V., and Linke, W.A. (2004). Developmentally regulated switching of titin size alters myofibrillar stiffness in the perinatal heart. Circ Res 94, 967–975. Peterkin, T., Gibson, A., and Patient, R. (2003). GATA-6 maintains BMP-4 and Nkx2 expression during cardiomyocyte precursor maturation. Embo j 22, 4260-4273. Peters, N.S., Severs, N.J., Rothery, S.M., Lincoln, C., Yacoub, M.H., and Green, C.R. (1994). Spatiotemporal relation between gap junctions and fascia adherens junctions during postnatal development of human ventricular myocardium. Circulation 90, 713-725. Petretto, E., Sarwar, R., Grieve, I., Lu, H., Kumaran, M.K., Muckett, P.J., Mangion, J., Schroen, B., Benson, M., Punjabi, P.P., et al. (2008). Integrated genomic approaches implicate osteoglycin (Ogn) in the regulation of left ventricular mass. Nat Genet 40, 546-552. Reiser, P.J., Westfall, M.V., Schiaffino, S., and Solaro, R.J. (1994). Tension production and thin-filament protein isoforms in developing rat myocardium. Am J Physiol 267, H1589–H1596. Roberts, J.T., and Wearn, J.T. Quantitative changes in the capillary-muscle relationship in human hearts during normal growth and hypertrophy. American Heart Journal 21, 617-633. Robertson, C., Tran, D.D., and George, S.C. (2013). Concise review: maturation phases of human pluripotent stem cell-derived cardiomyocytes. Stem Cells 31, 829-837. Rodriguez, A.G., Han, S.J., Regnier, M., and Sniadecki, N.J. (2011). Substrate stiffness increases twitch power of neonatal cardiomyocytes in correlation with changes in myofibril structure and intracellular calcium. Biophysical journal 101, 2455-2464. Sabry, M.A., and Dhoot, G.K. (1989). Identification and pattern of expression of a developmental isoform of troponin I in chicken and rat cardiac muscle. J Muscle Res Cell Motil 10, 85–91. Saggin, L., L., G., Ausoni, S., and Schiaffino, S. ( 1989). Troponin I switching in the developing heart. The Journal of biological chemistry 264, 16299–16302. Sartiani, L., Bettiol, E., Stillitano, F., Mugelli, A., Cerbai, E., and Jaconi, M.E. (2007). Developmental changes in cardiomyocytes differentiated from human embryonic stem cells: a molecular and electrophysiological approach. Stem cells 25, 1136-1144. Satin, J., Kehat, I., Caspi, O., Huber, I., Arbel, G., Itzhaki, I., Magyar, J., Schroder, E.A., Perlman, I., and Gepstein, L. (2004). Mechanism of spontaneous excitability in human embryonic stem cell derived cardiomyocytes. J Physiol 559, 479-496. Schaaf, S., Shibamiya, A., Mewe, M., Eder, A., Stöhr, A., Hirt, M.N., Rau, T., Zimmermann, W.-H., Conradi, L., Eschenhagen, T., et al. (2011). Human Engineered Heart Tissue as a Versatile Tool in Basic Research and Preclinical Toxicology. PloS one 6, e26397. Schaper, J., Meiser, E., and Stämmler, G. (1985). Ultrastructural morphometric analysis of myocardium from dogs, rats, hamsters, mice, and from human hearts. Circ Res 56, 377–391.

228

Schram, G., Pourrier, M., Melnyk, P., and Nattel, S. (2002). Differential distribution of cardiac ion channel expression as a basis for regional specialization in electrical function. Circ Res 90, 939-950. Seddon, A.M., Curnow, P., and Booth, P.J. (2004). Membrane proteins, lipids and detergents: not just a soap opera. Biochimica et biophysica acta 1666, 105-117. Shin, B.K., Wang, H., Yim, A.M., Le Naour, F., Brichory, F., Jang, J.H., Zhao, R., Puravs, E., Tra, J., Michael, C.W., et al. (2003). Global profiling of the cell surface proteome of cancer cells uncovers an abundance of proteins with chaperone function. The Journal of biological chemistry 278, 7607-7616. Siedner, S., Krüger, M., Schroeter, M., Metzler, D., Roell, W., Fleischmann, B.K., Hescheler, J., Pfitzer, G., and Stehle, R. (2003). Developmental changes in contractility and sarcomeric proteins from the early embryonic to the adult stage in the mouse heart. J Physiol 548, 493–505. Snir, M., Kehat, I., Gepstein, A., Coleman, R., Itskovitz-Eldor, J., Livne, E., and Gepstein, L. (2003). Assessment of the ultrastructural and proliferative properties of human embryonic stem cell-derived cardiomyocytes. American journal of physiology Heart and circulatory physiology 285, H2355-2363. Spach, M.S., Heidlage, J.F., Barr, R.C., and Dolber, P.C. (2004). Cell size and communication: role in structural and electrical development and remodeling of the heart. Heart rhythm : the official journal of the Heart Rhythm Society 1, 500-515. Sulkin, M.S., Yang, F., Holzem, K.M., Van Leer, B., Bugge, C., Laughner, J.I., Green, K., and Efimov, I.R. (2014). Nanoscale three-dimensional imaging of the human myocyte. Journal of structural biology 188, 55- 60. Tashiro, A., Masuda, T., and Segawa, I. (1990). Morphometric comparison of mitochondria and myofibrils of cardiomyocytes between hypertrophic and dilated cardiomyopathies. Virchows Archiv A, Pathological anatomy and histopathology 416, 473-478. Tate, C.G. (2010). Practical considerations of membrane protein instability during purification and crystallisation. Methods Mol Biol 601, 187-203. Tiburcy, M., Didie, M., Boy, O., Christalla, P., Doker, S., Naito, H., Karikkineth, B.C., El-Armouche, A., Grimm, M., Nose, M., et al. (2011). Terminal differentiation, advanced organotypic maturation, and modeling of hypertrophic growth in engineered heart tissue. Circulation research 109, 1105-1114. Tiburcy, M., Hudson, J.E., Ziebolz, D., and Zimmermann, W.H. (2014). Abstract 142: Defining the Non- Myocyte Compartment is Key for Enhanced Maturation of Human Engineered Heart Muscle. Circulation research 115, A142-A142. Tulloch, N.L., Muskheli, V., Razumova, M.V., Korte, F.S., Regnier, M., Hauch, K.D., Pabon, L., Reinecke, H., and Murry, C.E. (2011). Growth of engineered human myocardium with mechanical loading and vascular coculture. Circulation research 109, 47-59. Urbanek, K., Cabral-da-Silva, M.C., Ide-Iwata, N., Maestroni, S., Delucchi, F., Zheng, H., Ferreira-Martins, J., Ogorek, B., D'Amario, D., Bauer, M., et al. (2010). Inhibition of notch1-dependent cardiomyogenesis leads to a dilated myopathy in the neonatal heart. Circ Res 107, 429-441. van Rooij, E., Sutherland, L.B., Qi, X., Richardson, J.A., Hill, J., and Olson, E.N. (2007). Control of stress- dependent cardiac growth and gene expression by a microRNA. Science 316, 575-579. Wang, Y., and Morrisey, E.E. (2010). Regulation of cardiomyocyte proliferation by Foxp1. Cell Cycle 9, 4251-4252. Wang, P.Y., Yu, J., Lin, J.H., and Tsai, W.B. (2011). Modulation of alignment, elongation and contraction of cardiomyocytes through a combination of nanotopography and rigidity of substrates. Acta biomaterialia 7, 3285-3293.

229

Warren, C.M., Krzesinski, P.R., Campbell, K.S., Moss, R.L., and Greaser, M.L. (2004). Titin isoform changes in rat myocardium during development. Mech Dev 121, 1301-1312. Xu, X.Q., Soo, S.Y., Sun, W., and Zweigerdt, R. (2009). Global expression profile of highly enriched cardiomyocytes derived from human embryonic stem cells. Stem cells 27, 2163–2174. Yang, X., Pabon, L., and Murry, C.E. (2014). Engineering adolescence: maturation of human pluripotent stem cell-derived cardiomyocytes. Circulation research 114, 511-523. Zak, R. (1974). Development and proliferative capacity of cardiac muscle cells. Circ Res 35, suppl II:17-26. Zhang, Y., Li, S., Yuan, L., Tian, Y., Weidenfeld, J., Yang, J., Liu, F., Chokas, A.L., and Morrisey, E.E. (2010). Foxp1 coordinates cardiomyocyte proliferation through both cell-autonomous and nonautonomous mechanisms. Genes & development 24, 1746-1757. Zhang, D., Shadrin, I.Y., Lam, J., Xian, H.Q., Snodgrass, H.R., and Bursac, N. (2013). Tissue-engineered cardiac patch for advanced functional maturation of human ESC-derived cardiomyocytes. Biomaterials 34, 5813-5820. Zhao, R., Watt, A.J., Battle, M.A., Li, J., Bondow, B.J., and Duncan, S.A. (2008). Loss of both GATA4 and GATA6 blocks cardiac myocyte differentiation and results in acardia in mice. Developmental biology 317, 614-619. Ziman, A.P., Gómez-Viquez, N.L., Bloch, R.J., and Lederer, W.J. (2010). Excitation-contraction coupling changes during postnatal cardiac development. J Mol Cell Cardiol 48, 379-386. Zorn, A.M., Barish, G.D., Williams, B.O., Lavender, P., Klymkowsky, M.W., and Varmus, H.E. (1999). Regulation of Wnt signaling by Sox proteins: XSox17 alpha/beta and XSox3 physically interact with beta- catenin. Molecular cell 4, 487-498.

230

CHAPTER 5

Utility of in vitro Ventricular

Cardiomyocytes Model

231

TABLE OF CONTENT

5.1. Abstract ...... 233

5.2. Introduction and literature review ...... 235

5.3. Materials and Methods ...... 239

5.3.1. Measuring effects of calcium in cardiomyocytes maturity ...... 239 5.3.2. Optimizing protocol to generate ventricular cardiomyocytes ...... 239 5.3.3. Drug screening with knock-in cardiomyocytes ...... 240 5.3.4. Statistical analysis ...... 240

5.4. Results ...... 241

5.4.1. Ventricular cardiomyocytes need calcium for their growth and survival ...... 241 5.4.2. IWR1 did not significantly improve MYL2 positive fraction ...... 246 5.4.3. Effect of select drugs on EHMs maturity ...... 249 Endothelin-1 ...... 251 Noradrenaline ...... 254 IBMX ...... 258 Metformin ...... 261 Gemfibrozil ...... 265

5.5. Discussion ...... 269

5.6. References ...... 278

232

5.1. Abstract

In this chapter, the effects of increasing concentrations of extracellular calcium (Ca2+) on

EHM maturation were investigated. In addition, the utility of IWR1 to induce ventricular subtype specification during cardiomyocytes differentiation was tested against IWP4. Finally, EHMs generated with knock-in cardiomyocytes (CMs) were used as in vitro model for cardiac tissue and their responsiveness to chronic treatment with cardioactive drugs (endothelin-1, noradrenaline, IBMX, metformin and gemfibrozil) were investigated.

MYL2, G. luciferase and mCherry were expressed proportionally to the concentrations of extracellular Ca2+ (up until 1.2mM), beyond which, MYL2 gene expression reached a plateau.

Functional contractile forces were detected when extracellular Ca2+ was ≥ 0.8mM. EHMs cultured in 1.2mM – 2.0mM Ca2+ beat with similar forces (about 0.52 – 0.54 mN) and these forces were about twice the average forces of EHMs cultured in 0.8mM Ca2+ (about 0.26 mN).

EC50 values did not differ significantly in beating EHMs (0.4mM – 0.57mM) and they also exhibited negative force-frequency relationship beyond 2Hz stimulation. Despite the observed increasing response to isoprenaline and carbachol with increasing extracellular Ca2+ concentrations up to 1.6mM, these differences were not statistically significant. Taken together,

10 days in vitro EHMs culture with 1.2mM – 1.6mM extracellular calcium was beneficial to the functionality of the EHMs. However, this alone was not sufficient to drive the cardiomyocytes maturation in these EHMs.

Using the mCherry positive fraction as an indicator for ventricular CMs fraction, it was empirically determined that 5µM IWP4 or IWR1 at 2.5 and 5 µM concentrations did not significantly alter the ventricular CMs fraction. 5 µM IWP4, 2.5 µM IWR1 and 5 µM IWR1 gave rise to 51.7%, 50.8% and 51.7% mCherry positive fractions respectively. 233

The description for each drug tested in this part of the study will be further elaborated in the introduction and discussion part of this chapter. It was observed that EHMs treated with endothelin-1 (ET-1) expressed more luciferase, followed by those treated with IBMX and noradrenaline (NA). However, treatment with NA right-shifted the EHMs’ EC50 values the most

(0.8 – 0.95 mM Ca2+), followed by treatment with ET-1 (0.7 – 0.75 mM Ca2+) and IBMX (0.62 –

0.72 mM Ca2+). The responses to isoprenaline were generally dose-dependent for EHMs treated with ET-1 and IBMX but not for NA whereby the intermediate concentration of 100nM showed the maximal response. EHMs treated with Metformin expressed higher level of luciferase than those treated with Gemfibrozil. Metformin treated EHMs also expressed more luciferase with higher doses of Metformin, which was opposite in trend to the Gemfibrozil treated EHMs. EHMs treated with Gemfibrozil, however, showed a more right-shifted EC50 values and better isoprenaline response than the Metformin treated ones (EC50 = about 0.7 – 0.8mM vs 0.67 –

0.73mM Ca2+).

In all of the EHMs tested so far, mCherry signal was a good indicator for contractile function.

G. Luc signals gave quantitative measurements which were more sensitive than mCherry. With prior calibrations, G.Luc read-out could be correlated with contractile forces under normal conditions. With drug treatments, however, there could be other factors outside MYL2 gene expression that could potentially regulate contraction of the EHMs, thus G.Luc and mCherry signals might not correlate significantly with the corresponding strength of EHM contractions.

Nevertheless, the MYL2 knock-in line generated in this study and described in this thesis presents itself as a useful tool for identifying ventricular CMs and for predicting their functionality under normal conditions.

234

5.2. Introduction and literature review

From the previous chapter, it was understood that mechanical conditioning, electrical stimulation, biochemical stimulation and in vitro aging were not sufficient to drive hESCs-derived

CM maturation towards the adult phenotype. Thus one more approach was tested, i.e. to vary extracellular calcium concentration in the EHM culture media.

In addition, this part of the thesis aims to look at the different functionalities of the knock-in cell line. One immediate use of the cell line was in the development of a ventricular subtype specific differentiation protocol. Recently, the Wnt inhibitor IWR-1 has been shown to play a key role in ventricular subtype CM specification from HES-2, H7 and H1 hESCs lines

(Karakikes et al., 2014) as well as hiPSCs line and H9DF hESCs line (Weng et al., 2014). In addition, retinoid signaling has also been demonstrated to regulate this process (Zhang et al., 2011).

Other than small molecules, miRNAs have also been used to direct CM differentiation and specification towards ventricular CMs. One such miRNA is miR-499; this was used to enhance ventricular specification of hESCs-derived CMs while miR-1 was used to promote electrophysiological maturation of these ventricular CMs (Fu et al., 2011). Very recently, one group combined these 2 miRNAs with miRNAs 133 and 208 to reprogram non-myocytes to cardiomyocyte-like cells in situ, these reprogrammed cells were reported to significantly improve cardiac function following myocardial infarction in an animal model (Jayawardena et al., 2014).

These CM-like cells were also found to have morphological and functional properties of mature adult ventricular CMs (Jayawardena et al., 2014). In this study, this miRNA approach was not attempted. Instead, the role of IWR1 in CM subtype specification during hESCs cardiac differentiation will be investigated.

235

Finally, the functional potential of the knock-in line as a drug screening platform was investigated in a pilot study of drug screening using noradrenaline (also known as norepinephrine), endothelin-1, metformin, IBMX and Gemfibrozil. Noradrenaline (NA) has been reported to increase ventricular mass; ratios of ventricular weight to ventricular volume and body weight (around 50%); as well as ratio of ventricular RNA: protein levels (Kaddoura et al.,

1996; Newling et al., 1989). NA was also demonstrated to exert its hypertrophic effect through increasing endogenous Endothelin-1 mRNA levels (Kaddoura et al., 1996). However, high concentrations of NA at 50μM and 100μM have been reported to cause a shift from hypertrophy to apoptosis (Jain et al., 2014). At these high concentrations, NA altered cellular and nuclear morphology as well as mitochondrial membrane potential thus increasing the production of reactive oxygen species in the CMs which subsequently led to caspase dependent activation of cell death (Jain et al., 2014).

3-Isobutyl-l-methylxanthine (IBMX) has been shown to be a potent competitive nonselective phosphodiesterase (PDE) inhibitor (Beavo et al., 1970; McNeill et al., 1973;

Peytremann et al., 1973; Terasaki and Brooker, 1977) and nonselective adenosine receptor antagonist (Daly et al., 1987). PDE inhibitors such as IBMX block the PDE-mediated hydrolysis of the 3’-phosphodiester bond in cyclic nucleotides such as cyclic adenosine monophosphate

(cAMP) thus generating the corresponding inactive form (Essayan, 2001). Accumulation of cAMP in the cell leads to the activation of protein kinases; inhibition of TNF-alpha (Deree et al., 2008;

Marques et al., 1999) and synthesis of leukotriene (Peters-Golden et al., 2005); as well as reduced inflammation and innate immunity (Peters-Golden et al., 2005). Increased cAMP levels also leads to an increase in junctional conductance in the heart (De Mello, 1989) and protein

2+ 2+ phosphorylation in sarcolemmal slow-Ca channel leading to increased Ca inward current (Isi)

236

(Reuter, 1983; Reuter and Scholz, 1977) and increased contractile force. Being an adenosine receptor antagonist, the effects of IBMX on isometric force of contraction closely mimics those of isoprenaline (Korth, 1978). IBMX-induced positive inotropic response begins at 10 µM to reach a maximum of 400% increase from basal level of contraction with treatment at 300 µM;

Isoprenaline treatment resulted in similar inotropic response but at much lower concentrations at 1nM – 100nM (Brückner et al., 1985). In the above-mentioned studies, however, IBMX was only applied for a short term (minutes). In this study, the effect of chronic (10 days) EHMs treatment with IBMX was investigated.

Metformin has been a conventional therapy for patients with type 2 diabetes (Inzucchi,

2005). Patients who have both diabetes and heart failure exhibit preserved systolic function but impaired left ventricular relaxation (Piccini et al., 2004) which eventually result in maladaptive ventricular remodeling (Parsonage et al., 2002). This could be due to hypertension, diabetic microcirculatory disease, oxidative stress, and the deleterious influences of hyperglycemia and insulin resistance on inflammatory mediators, growth factors, vascular endothelial function, and the renin-angiotensin system (Inzucchi, 2005). The effect of chronic Metformin treatment on generated from normal knock-in CMs was investigated in this chapter.

Gemfibrozil belongs to fibrates group of drugs; it is a potent lipid regulating drug whose major effects are to increase plasma high density lipoproteins (HDL) and to decrease plasma triglycerides (TG) (Saku et al., 1985). Gemfibrozil is also a ligand for peroxisome proliferator– activated receptors isoform alpha (PPARα) (Forman et al., 1997), which modulates genes involved in the β-oxidative degradation of fatty acids (Takano et al., 2000). In addition,

Gemfibrozil has been shown to ameliorate cardiac remodeling events in obese and hypertensive rats through its antioxidant activity (Singh et al., 2011). This antioxidant property, along with its

237

anti-inflammatory property (Adabag et al., 2009) has sparked an interest whether Gemfibrozil could prevent or delay the development of atrial fibrillation in heart disease patients.

Unfortunately, it was not found to reduce 4-year incidence of atrial fibrillation in men who have coronary heart disease (Adabag et al., 2009). In this study, the effects of Gemfibrozil treatment for 10 days on normal EHMs were investigated.

238

5.3. Materials and Methods

5.3.1. Measuring effects of calcium in cardiomyocytes maturity

Day 15 H9VC16 cardiomyocytes were mixed with HFFs (75%: 25%) to form EHMs. EHMs were cultured in normal SFMM medium from the day of casting for three days before they were transferred to auxotonic stretchers. On the next day, the medium was changed to SFMM medium but the base medium of SFMM was changed to DMEM without calcium instead of RPMI

1640. EHMs were cultured in DMEM SFMM without calcium for two days before starting the calcium response experiment.

Six different concentrations of calcium were applied to the EHMs (0mM – 2.0mM) for ten days before analysis; thus analysis was carried out on day 30 of differentiation. Analysis of expression of mCherry (mChr) and Gaussia luciferase (G.Luc), as well as contraction experiments were carried out to determine the effects of varying calcium concentration on cardiomyocytes in tissue-mimetic three dimensional EHM format.

5.3.2. Optimizing protocol to generate ventricular cardiomyocytes

Two Wnt inhibitors, IWP4 and IWR1, were tested for their capacity to differentiate knock-in hESCs to ventricular cardiomyocyte (vCMs). On day 30 of differentiation, efficiency of the differentiation protocol was assessed through the expression of knocked-in reporter genes.

In brief, the beating layer of cardiomyocytes was dissociated into single cells and counted. Live cells were analyzed by FACS to determine the percentage of mCherry-positive cells in the population. In addition, G.Luc expression was normalized to cell number to get RFU per cell.

239

5.3.3. Drug screening with knock-in cardiomyocytes

Cardiomyocytes were trypsinized and mixed with human foreskin fibroblasts (HFF) in a collagen matrix on day 15 of differentiation at (3:1) ratio to make EHMs. On day 18 of differentiation, EHMs were transferred onto auxotonic stretchers and maintained in culture for seven days. After one week maintenance on auxotonic stretchers, five different drugs were tested on these EHMs. Drug-containing SFMM media were changed daily and G.Luc secreted into the medium was assayed at different time points (1, 3, 7, and 10 days of drug treatment) to assess acute (≤7 days) and chronic (>7 days) drug response. The response to drug treatment was also assessed through the analysis of contractile force, reporter gene expression (proxy for MYL2 expression), force frequency measurement, as well as isoprenaline and carbachol response.

5.3.4. Statistical analysis

Data is presented as means ± standard error mean (SEM). Statistical analyses were performed with Student’s t-test (2-tailed) or ANOVA when applicable, using post-hoc Dunnett's

Multiple Comparison test or Bonferroni's Multiple Comparison test. Results were considered to be significant at p ≤ 0.05 (*); very significant at p ≤ 0.01(**) and extremely significant at p ≤

0.001 (***).

240

5.4. Results

5.4.1. Ventricular cardiomyocytes need calcium for their growth and survival

0mM Ca2+ 1.2mM Ca2+

0.4mM Ca2+ 1.6mM Ca2+

0.8mM Ca2+ 2.0mM Ca2+

Table 5.4.1-1 mCherry expression increased in EHMs cultured in increasing calcium concentrations (up to 1.6mM)

From table 5.4.1-1 above, we can see that mCherry positive cells are sensitive to external calcium concentration. When EHMs were cultured in SFMM containing 0mM Ca2+, they did not beat and there were no observable mCherry expression under fluorescence microscope.

EHMs cultured in 0.4mM Ca2+ were beating albeit very weakly; this could be explained by the presence of only few bright mCherry positive cells in these EHMs (table 5.4.1-1). In general, as

241

the external Ca2+ concentration was increased from 0.4mM – 2.0mM, more mCherry positive cells were observed in the corresponding EHMs.

** ** 10.0 *** * *** *** 8.0 ) 5 6.0 ***

4.0 RFU (x10 RFU ** 2.0

0.0 0.4 0.8 1.2 1.6 2.0 Calcium concentration (mM)

Figure 5.4.1-1 More luciferase was expressed by EHMs cultured in higher concentrations of calcium (up to 1.6mM)

Figure 5.4.1-1 above shows that there was a gradual increase in the expression of luciferase as extracellular Ca2+ concentration was increased, reaching a plateau at 1.6mM –

2.0mM Ca2+. 0.4mM – 2.0mM Ca2+ significantly increased the luciferase expression in the EHMs in comparison with 0mM Ca2+ (the level of statistical significance is shown as * above the respective bars). However, in comparison with EHMs cultured in 0.4mM Ca2+, only 1.2mM –

2.0mM Ca2+ significantly increased the level of luciferase expression. Beyond 0.8mM Ca2+, the increase in luciferase expression was not shown to be statistically significant. Interestingly, the increase of Gaussia luciferase expression from 0.0mM – 1.2mM Ca2+ was found to be linear and could be fitted to the equation = (584000) + 43930 with = 0.9989 (when the mean of 2 each group is considered) or 푦 = 0.9274 (when푥 all replicates푅 were considered individually). 2 Shown below are the forces generated푅 by the EHMs cultured in different Ca2+ concentration for

10 days.

242

0.7

0.6 2+ 0.5 0mM Ca 0.4mM Ca2+ 0.4 0.8mM Ca2+ 0.3 1.2mM Ca2+ Force (mN) 0.2 1.6mM Ca2+ 2+ 0.1 2.0mM Ca

0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.1-2 Forces of EHMs cultured in different Ca2+ concentrations

From the graph above, 0 – 0.4mM Ca2+ seems to generate EHMs with negligible forces.

This is reminiscent of the earlier data whereby EHMs cultured in Ca2+ free media did not appear to be beating and that there were no observable mCherry positive cells observed in these EHMs.

The EHMs cultured in 0.4mM Ca2+ were beating very weakly with few mCherry positive cells observed under the fluorescence microscope. Interestingly, EHMs cultured in 0.4mM Ca2+ were showing luciferase expression about 4.96x the level of expression of EHMs cultured in Ca2+ free medium. Yet this number of ventricular CMs appeared to be insufficient to generate appreciable contractions in the EHMs.

The graph also shows that EHMs cultured in 1.2 – 2.0mM Ca2+ beat with similar forces

(around 0.52 – 0.54 mN) and these forces were about twice the average forces of EHMs cultured in 0.8mM Ca2+ (around 0.26 mN). Interestingly, in EHMs cultured in 1.2 – 2.0mM Ca2+, there was an observed drop in forces when extracellular Ca2+ went beyond 2.8mM. This drop appears more clearly in figure 5.4.1-3. The dotted line in this graph indicates the highest forces observed in all

EHMs which were generally observed at 2.8mM Ca2+. It is apparent that beyond 2.8mM Ca2+, the 243

EHM contractile forces drop steadily for EHMs cultured in 1.2mM – 2.0mM Ca2+. This drop is most pronounced in EHMs cultured in 2.0mM Ca2+. From this graph, we can also determine that concentration of Ca2+ which gave rise to 50% of maximal forces (EC50). These respective EC50 values are presented in table 5.4.1-2.

120

2+ 100

80 0.4mM Ca2+ 60 0.8mM Ca2+ 40 1.2mM Ca2+ 2+

% force at 2.8mM Ca 1.6mM Ca 20 2.0mM Ca2+ 0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0

[Ca2+] in mM Figure 5.4.1-3 Forces of EHMs treated with different calcium concentrations as percentage of forces at 2.8mM Ca2+

0.4mM Ca2+ 0.8mM Ca2+ 1.2mM Ca2+ 1.6mM Ca2+ 2.0mM Ca2+ EC50 (mM) 0.25 0.53 0.55 0.4 0.57

Table 5.4.1-2 EC50 values of EHMs cultured in different calcium concentrations

EHMs cultured in 0.8mM, 1.2mM and 2.0mM Ca2+ had similar EC50 at about 0.5mM but those cultured in 1.6mM Ca2+ had lower EC50 at 0.4mM and those cultured in 0.4mM Ca2+ had lowest EC50 values of 0.25mM. When force frequency response was analyzed, the following chart was obtained (figure 5.4.1-4). The chart show similar responses of EHMs cultured in media containing Ca2+: there was slight increase in forces when frequency of stimulation was increased from 1 Hz to 2 Hz. However, a drop in forces was observed when this stimulation frequency was

244

further increased from 2Hz to 3 Hz. EHMs cultured in 2.0mM Ca2+ showed the highest increase in forces (about 7.9%) when the stimulation frequency was increased from 1Hz to 2Hz. 0.4mM and

0.8mM Ca2+ groups followed next with a force increase of about 3.5-4.0% while the 1.2mM and

1.6mM Ca2+ group did not show much change in force with only about 1.4-1.5% increase. When the stimulation frequency was increased from 2Hz to 3Hz, similar trend was observed. 2mM Ca2+

EHM group showed the least drop in forces at about 9.2%, followed by the 0.8mM group at

13.2%. The largest drops in forces were shown by 1.2mM and 1.6mM groups at 15.7% and 16.4% respectively.

Force frequency response 120

0mM Ca2+ 100 0.4mM Ca2+ 0.8mM Ca2+ 2+ 80 1.2mM Ca % of force at 1 Hz % 1.6mM Ca2+ 2.0mM Ca2+ 60 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.1-4 Force frequency response of EHMs cultured in different concentrations of calcium

Responses to Isoprenaline improved in EHMs cultured in increasing Ca2+ concentrations from 0.8mM – 1.6mM. Only EHMs cultured in 2.0mM Ca2+ do not follow this trend, showing level of response similar to EHMs cultured in 0.8mM Ca2+. This trend was also observed for carbachol response. % change of force after treatment with Isoprenaline, as well as after carbachol treatment was plotted to give figure 5.4.1-6.

245

Isoprenaline, carbachol response 2+

140

120

100 0.4mM Ca2+ 80 0.8mM Ca2+ 60 1.2mM Ca2+

40 1.6mM Ca2+ 2.0mM Ca2+ 20

0 % of forces force to at 0.8mM Ca Isoprenaline Carbachol

Figure 5.4.1-5 EHMs response to Isoprenaline and carbachol stimulation

2+

40 0.4mM Ca2+ 0.8mM Ca2+

20 1.2mM Ca2+ 1.6mM Ca2+ 2.0mM Ca2+ 0 Isoprenaline Carbachol

-20 % change of force at force change% of 0.8mM Ca

Figure 5.4.1-6 Carbachol response of EHMs cultured in different calcium concentrations

5.4.2. IWR1 did not significantly improve MYL2 positive fraction

Hudson 2D protocol (for details please see chapter 3) was employed to test the efficacy of different Wnt inhibitors in directing hESCs towards ventricular cardiomyocytes. The Wnt inhibitors used were IWP4 (5µM) and IWR1 (2.5µM or 5µM) on day 3-12 of differentiation.

Shown in figure 5.4.2-1 is the luciferase expression per cell from flasks treated with different Wnt inhibitors. The result indicated that 2.5µM IWR1 gave the highest RFU per total cell number (about 0.2 RFU/cell) in comparison with 5µM IWP4 (about 0.15 RFU/cell) and 5µM

246

IWR1 (about 0.11 RFU/cell). The differences in the values across time-points, however, were not statistically significant.

1

0.1 55uM µM IWP4 IWP4

2.52.5uM µM IWR1 IWR1

55uM µM IWR1 IWR1

RFU per cell 0.01 NegativeH9 control

0.001 18 21 24 27 30 day post differentiation

Figure 5.4.2-1 Different Wnt inhibitors did not significantly affect luciferase expression per cell (proxy for ventricular cardiomyocytes fraction). Negative control was the luciferase expression from wild type H9-derived cardiomyocytes

Figure 5.4.2-2 shows that 5µM IWP4 gave the highest fraction of CMs positive for mCherry expression (about 55.3%) in comparison with 2.5µM IWR1 (about 50.8%) and 5µM

IWR1 (about 51.7%). However, these differences were not found to be statistically significant.

Mcherry positive fraction was normalized to alpha-actinin positive fraction (cardiomyocytes) to plot figure 5.4.2-3. 2.5µM IWR1 gave the highest mCherry to α-actinin-positive ratio and this was significantly higher than the 5µM IWP4 group. However, the result was puzzling as the mCherry(+) to α-actinin(+) ratio was more than 100%.

247

Figure 5.4.2-2 Different Wnt inhibitors did not significantly affect mCherry(+) fraction (vCM population)

mCherry (+)ve fraction normalized to alpha actinin (+)ve fraction 250 *

200

150 194.8% H9 CMs 2.52.5uM µM IWR1IWR1 100 55uM µM IWR1 IWR1 101.5% 107.0% 5uM5 µM IWP4 IWP4 50 % of mCherry positive CMs positive mCherry of % 0

Figure 5.4.2-3 Percentage of MYL2 positive cardiomyocytes in flasks differentiated with different Wnt inhibitors

248

5.4.3. Effect of select drugs on EHMs maturity

Control Gemfibrozil Metformin 10 µM 2 µM

100 µM 20 µM

1000 µM 200 µM

249

IBMX Endothelin-1 Nor-adrenaline 1 µM 0.1 nM 10 nM

10 µM 1 nM 100 nM

100 µM 10 nM 1000 nM

Table 5.4.3-1 mCherry expression from EHMs cultured in different drugs for 10 days

250

Endothelin-1

Shown below is the luciferase expression of EHMs treated with Endothelin-1, at different time-points. There is general up-regulation of luciferase expression in EHMs treated with 1nM and 10nM ET-1 for the acute phase of drug treatment, with 1nM showing the highest level of expression followed by 10nM group. This trend, however, changes for the chronic phase, with

1nM group showing the highest drop in luciferase expression, followed by 10nM and 0.1nM group. In fact, the 0.1nM group showed very little changes in luciferase expression throughout the study.

10.0

8.0 ) 5 Control 6.0 ET-1 0.1nM 4.0 ET-1 1nM RFU (x10 RFU ET-1 10nM 2.0

0 1 2 3 4 5 6 7 8 9 10 Days of treatment with drugs

Figure 5.4.3-1 Luciferase expression of EHMs treated with different concentrations of Endothelin-1

Forces of EHMs cultured in different [Endothelin-1] 1.0 0.9 0.8 0.7 Control 0.6 0.1nM Endothelin-1 0.5 1nM Endothelin-1 0.4 10nM Endothelin-1

Force (mN) Force 0.3 0.2 0.1 0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.3-2 Forces of EHMs treated with different Endothelin-1 concentrations 251

Up to 2.0mM Ca2+, EHMs in all groups showed increasing forces with increasing Ca2+ concentration, with 1nM group generating the highest forces of contraction, followed by 0.1nM and 10nM group. This data is not correlated with the luciferase data. However, from 2.0 –

2.8mM Ca2+, a slight dip in forces was observed for EHMs in 0.1-1nM groups, owing to arrhythmic contractions. The EHMs in these groups recovered to the level of contraction before the dip once extracellular Ca2+ concentration was increased to 3.2mM. Due to this dip, EC50 calculations using normalization to 4.0mM Ca2+ cannot be done, since in this experiment, the forces at 4.0mM Ca2+ were not the maximal forces. Therefore, a straight line was drawn to indicate the 50% percentile of forces and the corresponding Ca2+ concentrations recorded as

EC50 values. From the graph, treatment with Endothelin-1 seems to shift EC50 values to the right slightly (by 0.5-1mM Ca2+), indicating that the drug helped to reduce the EHMs sensitivity to extracellular Ca2+ concentrations. EC50 value of EHMs treated with 0.1nM ET-1 was the largest at

0.75mM, followed by 10nM Endothelin-1 group at 0.73nM and 1nM at 0.7mM. The control

EHMs EC50 value was at 0.65mM.

140

120

Control 0.1nM Endothelin-1 100 1nM Endothelin-1 % force at 2Hz 10nM Endothelin-1 80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.3-3 Force frequency response of EHMs treated by different concentrations of Endothelin-1

252

EHMs from all groups exhibited similar force-frequency trend: there is a slight increase in force as stimulation frequency was increased from 1 – 2 Hz, and a slight decrease as frequency was further increased to 3 Hz. This does not indicate positive force frequency response (1-2Hz), however, as all EHMs could not be paced at 1Hz. The lowest frequency at which all EHMs could be paced was 2 Hz. Thus all EHMs still exhibited negative force frequency response, with 0.1nM group showing almost identical response to 10nM group and 1nM group showing almost identical response to the control group.

There was also no apparent difference in the responses to Isoprenaline from EHMs treated with all concentration of ET-1 to control group. Increase in forces after Isoprenaline addition, as well as the difference between Isoprenaline-induced and carbachol-induced forces, did not differ very much between different groups of EHMs.

Isoprenaline carbachol response of EHMs treated with [Endothelin-1] 2+ 220 200 180 160 140 120 100 Control 80 60 0.1nM Endothelin-1 40 1nM Endothelin-1 20 10nM Endothelin-1 0 % force to force at 0.8mM at Ca force to force % Isoprenaline Carbachol Drug added

Figure 5.4.3-4 Isoprenaline-carbachol response of EHMs treated with different concentrations of Endothelin-1

253

Reduction in Isoprenaline-induced forces with Carbachol 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 -0.02 0 0.1 1 10

Reduction in forces (mN) forces in Reduction -0.04 [Endothelin-1] in nM

Figure 5.4.3-5 Increase in force after Figure 5.4.3-6 Reduction in force after Isoprenaline treatment carbachol treatment

Noradrenaline

EHMs treated with 10-100nM Noradrenaline (NA) showed up-regulation of G.Luc for the first two days of treatment, while those treated with 1000nM showed a slight down-regulation.

Beyond 48 hours of treatment, there was a down-regulation of luciferase expression in all EHMs.

This down-regulation continued until day seven for 10nM NA while in EHMs treated with higher concentrations of NA at 100 and 1000nM, there is an up-regulation, with 1000nM group expressing higher level of luciferase than 100nM group. Beyond seven days of treatment, luciferase expression dropped for 100nM and 1000nM groups to reach similar level whereas, for the 10nM group, the expression increased slightly. After 10 days of treatment with NA, EHMs from all groups expressed similar levels of luciferase. From this data, one could expect similar levels of forces generated by all EHMs 10 days after drug treatment.

254

10.0

8.0 ) 5 6.0 Control 4.0 NA 10nM

RFU (x10 RFU NA 100nM 2.0 NA 1000nM

0 1 2 3 4 5 6 7 8 9 10 Days of treatment with drugs

Figure 5.4.3-7 Luciferase expression of EHMs treated with different concentrations of Noradrenaline

Indeed, as evident from graph 6.4.3-9, the forces of EHMs treated with different concentrations of NA beat at very similar level to one another at about 0.57mN and this force was lower than the control group at about 0.88mN. From 0 – 2.4mM calcium, EHMs treated with

10nM and 100nM NA beat at almost identical forces. The EHMs treated with 1000nM NA beat at higher forces <2.4mM calcium but at 2.4mM, they beat at almost identical force as the EHMs from other groups. There was a dip in forces when Ca2+ was further increased from 2.4 – 2.8mM, with the 10nM NA-treated group showing the largest dip in force.

Similar to treatment with ET-1, treatment of EHMs with NA seems to reduce EHMs sensitivity to extracellular calcium concentrations. EHMs treated with 10nM and 100nM NA showed the lowest sensitivity to extracellular calcium, with EC50 value of about 0.95mM. This was followed by the 1000nM NA group with EC50 value of 0.8mM. The control group has EC50 value of 0.65mM.

255

Forces of EHMs cultured in different [Noradrenaline] 1.0 0.9 0.8 Control 0.7 10nM Noradrenaline 0.6 100nM Noradrenaline 0.5 1000nM Noradrenaline 0.4

Force (mN) Force 0.3 0.2 0.1 0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.3-8 Forces of EHMs treated with different concentrations of Noradrenaline

Force frequency response of EHMs treated with different [Noradrenaline]

140

120 Control 10nM Noradrenaline 100 100nM Noradrenaline % force at 2Hz 1000nM Noradrenaline 80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.3-9 Force frequency response of EHMs treated with different concentrations of Noradrenaline

Force frequency response of EHMs treated with NA was similar to those treated with ET-

1. All EHMs could not be paced at 1Hz. Thus all the forces were normalized to the forces generated at 2Hz of stimulation. Here, we see that EHMs treated with 10nM and 1000nM NA showed negative force frequency behavior, with 1000nM group showing exactly the same

256

behavior as the control group. EHMs treated with 100nM NA showed a mixed response at 3Hz of stimulation, some replicates showed negative while others showed positive force frequency behavior. Figure 5.4.3-11 shows that Isoprenaline and carbachol responses of EHMs treated with

10nM and 1000nM NA did not differ very much from the control group. Only the 100nM NA group seems to stand out, showing higher response to isoprenaline and carbachol. To further analyze this, the increase in forces after isoprenaline addition, as well as the difference between isoprenaline-induced and carbachol-induced forces was tabulated and the difference analyzed with ANOVA using Bonferroni's multiple comparison test. The result indicated that there is a statistically significant difference in the response to isoprenaline in 100nM NA-treated EHMs in comparison with control group. This difference however, was not observed in the response to carbachol.

Isoprenaline carbachol response of EHMs treated with [Noradrenaline] 2+ 220 200 180 160 140 120 Control 100 80 10nM Noradrenaline 60 100nM Noradrenaline 40 20 1000nM Noradrenaline 0 % force to force at 0.8mM at Ca force to force % Isoprenaline Carbachol Drug added

Figure 5.4.3-10 Noradrenaline-treated EHMs response to Isoprenaline and Carbachol

257

Increase in force (w.r.t forces at 0.8mM Ca2+) Reduction in after Isoprenaline treatment Isoprenaline-induced forces with Carbachol 120 * 0.14 100 0.12

80 0.10

60 0.08 0.06 40 0.04

% increase of force of increase % 20 0.02 0 0.00 0 10 100 1000 -0.02 0 10 100 1000

[Noradrenaline] in nM (mN) forces in Reduction -0.04 [Noradrenaline] in nM Figure 5.4.3-12 Response to Isoprenaline Figure 5.4.3-11 Reduction in forces after Carbachol stimulation in Noradrenaline-treated EHMs stimulation

IBMX

Gaussia Luciferase expression 10.0

8.0 ) 5 6.0 Control IBMX 1uM 4.0 IBMX 10uM

RFU (x10 RFU IBMX 100uM 2.0

0 1 2 3 4 5 6 7 8 9 10 Days of treatment with drugs

Figure 5.4.3-13 Luciferase expression of EHMs treated with different concentrations of IBMX

There is general up-regulation in G.Luc expression for the first 72 hours for all IBMX concentrations tested: 10µM group showed the highest level of increase, followed by the 1µM and 100µM group. The higher concentrations groups, however, showed reduced expression of

G.Luc from day 3-7 of drug treatment whereas the 1µM group did not show much change. This

258

resulted in a dose-dependent expression of G.Luc after seven days of drug treatment: 1µM showing the highest luciferase expression, followed by 10µM and 100µM. It is worthy of note however, that all IBMX-treated EHMs expressed lower luciferase than the control group. Beyond day 7, luciferase expression dropped for control, 1µM and 10µM groups; whereas for the 100µM group, there was a slight increase in luciferase expression. Nevertheless, after 10 days of drug treatment, similar trend was observed as that after seven days of drug treatment: there was

IBMX dose-dependent luciferase expression, with 1µM group showing highest expression of luciferase, followed by 10µM group and 100µM.

Forces of EHMs cultured in different [IBMX] 1.0 0.9 0.8 Control 0.7 1uM IBMX 0.6 10uM IBMX 0.5 100uM IBMX 0.4

Force (mN) Force 0.3 0.2 0.1 0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.3-14 Forces of EHMs treated with different concentrations of IBMX

The contraction force data above seems to recapitulate the G.Luc expression data, whereby EHMs treated with the least dose of IBMX (1µM) beat with the highest forces, followed by the higher doses of 10µM and 100µM. The 10µM group showed higher variability in the beating forces and also a dip in forces owing to arrhythmia when extracellular calcium was increased from 1.2mM – 1.6mM. Arrhythmia ceased when calcium was further increased from

259

1.6mM – 2.0mM and forces returned to the level of 1.2mM Ca2+ when extracellular calcium was about 2.4mM. The slight dip in forces owing to arrhythmia occurred at higher concentration of extracellular calcium for the 1µM and 100µM IBMX group (2.0mM – 2.8mM).

Only EHMs treated with 1µM IBMX showed right-shifted EC50 value, from about

0.65mM of the control group to about 0.72mM. EHMs treated with 10µM and 100µM IBMX had

EC50 value of about 0.62mM and about 0.63mM respectively. There is a generally negative force frequency response for EHMs treated with all concentrations of IBMX, with the 100µM group showing the most negative response, followed by the 1µM group and 10µM group.

Force frequency response of EHMs treated with different [IBMX] 120

110

100 Control 1uM IBMX 10uM IBMX 90 % force at 2Hz 100uM IBMX

80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.3-15 Force frequency response of EHMs treated with different concentrations of IBMX

There were also no statistically significant differences in the responses to isoprenaline from all EHM groups. However, there were statistically significant differences in the responses to carbachol from the 100µM group, in comparison with the control and 1µM group. The negative value for the 100µM group was due to the highly arrhythmic contractions after isoprenaline treatment.

260

Isoprenaline carbachol response of EHMs treated with [IBMX] 2+ 220 200 180 160 140 120 100 Control 80 60 1uM IBMX 40 10uM IBMX 20 100uM IBMX 0 % force to force at 0.8mM at Ca force to force % Isoprenaline Carbachol Drug added

Figure 5.4.3-16 Isoprenaline and Carbachol response from EHMs treated with different concentrations of IBMX

Increase in force Reduction in (w.r.t forces at 0.8mM Ca2+) Isoprenaline-induced forces with after Isoprenaline treatment Carbachol 120

0.14 * 100 0.12 * 80 0.10 0.08 60 0.06 40 0.04

0.02 force of increase % 20 0.00 0 -0.02

Reduction in forces (mN) forces in Reduction 0 1 10 -0.04 0 1 10 100 100 [IBMX] in uM [IBMX] in uM Figure 5.4.3-17 Reduction in contratile forces in Figure 5.4.3-18 Response to Isoprenaline IBMX-treated EHMs after Carbachol treatment stimulation in IBMX-treated EHMs

Metformin

Unlike EHMs treated with NA, EHMs treated with Metformin did not show a dip in G.Luc expression three days after treatment with the drug. There was only a slight dip 48 hours after treatment in EHMs treated with 2µM Metformin. EHMs treated with higher concentrations of

261

Metformin (20µM and 200µM) showed continued increase of G.Luc expression in the first 48 hours of treatment with the drug. This increase of G.Luc expression was maintained from day 2 –

7 of drug treatment, with more G.Luc being expressed in EHMs treated with higher Metformin concentrations. From day 7-10 post treatment, there was a drop in luciferase expression, with

EHMs treated with 2µM Metformin showing same rate of reduction in expression as the control group; and the 20µM and 200µM EHM groups showing higher rate of reduction.

Gaussia Luciferase expression 10.0

8.0 ) 5 6.0 Control Met 2uM 4.0 Met 20uM RFU (x10 RFU Met 200uM 2.0

0 1 2 3 4 5 6 7 8 9 10 Days of treatment with drugs

Figure 5.4.3-19 Luciferase expression of EHMs treated with different concentrations of Metformin

Surprisingly, despite consistently showing higher G.Luc expression in the first week of drug treatment, EHMs treated with 200µM Metformin beat with lower forces than EHMs treated with lower dosage of the drug. Also, despite the low levels of G.Luc expression throughout 10 days of treatment, EHMs treated with 2µM Metformin beat with the highest forces among all Metformin-treated EHMs, beating almost with similar forces as the control group. EHMs treated with 200µM Metformin beat with very similar forces to EHMs treated with

20µM Metformin.

262

Forces of EHMs cultured in different [Metformin] 1.0 0.9 0.8 0.7 Control 0.6 2uM Metformin 0.5 20uM Metformin 0.4 200uM Metformin

Force (mN) Force 0.3 0.2 0.1 0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.3-20 Forces of EHMs treated with different concentrations of Metformin

EC50 values of Metformin-treated EHMs were also shifted to the right (by about 0.02 –

0.08mM increase) of the control group (about 0.65mM), with 2µM Metformin-treated EHMs having EC50 value of about 0.70mM; 20µM Metformin-treated EHMs at about 0.73mM and

200µM Metformin-treated EHMs at about 0.67mM.

Force frequency response of EHMs treated with different [Metformin] 120

110

100 Control 2uM Metformin 20uM Metformin 90 % force at 2Hz 200uM Metformin

80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.3-21 Force frequency response of EHMs treated with different concentrations of Metformin

263

The force frequency response in EHMs treated with Metformin is similar to EHMs treated with NA: there is a generally negative force frequency response, in control, 2µM and

200µM Metformin-treated groups. 2µM Metformin-treated group has the highest drop in forces as frequency of stimulation was increased to 3Hz, followed by the 200µM group and the control group. There is a mixed response in the 20µM Metformin-treated group, resulting in the average of slightly positive force-frequency response. There were no significant differences in the responses to Isoprenaline and carbachol between all groups of EHMs: control, 2µM, 20µM and

200µM Metformin.

Isoprenaline carbachol response of EHMs treated with [Metformin] 2+ 220 200 180 160 140 120 Control 100 80 2uM Metformin 60 40 20uM Metformin 20 200uM Metformin 0 % force to force at 0.8mM at Ca force to force % Isoprenaline Carbachol Drug added

Figure 5.4.3-22 Metformin-treated EHMs response to Isoprenaline and Carbachol

264

Increase in force 2+ (w.r.t forces at 0.8mM Ca ) Reduction in after Isoprenaline treatment 120 Isoprenaline-induced forces with Carbachol 100 0.14 80 0.12 60 0.10 0.08 40 0.06

% increase of force of increase % 20 0.04 0 0.02 0 2 20 200 0.00 [Metformin] in uM -0.02 0 2 20 200

Figure 5.4.3-24 Response to Isoprenaline stimulation Figure (mN) forces in Reduction 5.4.3-0.04-23 Reduction[Metformin] in forces in MetforminuM - in Metformin-treated EHMs treated EHMs stimulated with Carbachol

Gemfibrozil

EHMs treated with Gemfibrozil showed a dose-dependent G.Luc expression for the first seven days of drug application, with the smallest dosage of 10µM resulting in EHMs secreting the most G.Luc, followed by 100 µM and 1000µM. In fact, the signal from 1000µM group is more than twice lower the 100µM group. For the 10µM and 100µM groups, after initial increase in

G.Luc expression after 48 hours of treatment, there was a decrease in G.Luc expression between

48 – 72 hours of drug treatment. This resulted in little difference in the G.Luc expression level on the third day, in comparison with the first day. For the 1000µM group, however, there was a drop in expression from 24 – 48 hours of drug treatment and this drop continued further from

48 – 72 hours of treatment. From day 3-7, there was a general increase in G.Luc expression for all EHMs from all Gemfibrozil treated groups. From day 7 – 10, there was a general decrease in

G.Luc expression for all EHMs from all experimental groups, with the 10µM group showing the sharpest rate of decrease. Similar to result obtained on the 7th day of treatment, the 1000 µM group showed the least level of luciferase expression on the 10th day of treatment.

265

10.0

8.0 ) 5 6.0 Control 4.0 Gem 10uM

RFU (x10 RFU Gem 100uM Gem 1000uM 2.0

0 1 2 3 4 5 6 7 8 9 10 Days of treatment with drugs

Figure 5.4.3-25 Gaussia Luciferase expression of EHMs treated with different concentrations of Gemfibrozil

The G.Luc data was somewhat recapitulated in the contractile force data. EHMs treated with 100 µM Gemfibrozil beat with the highest forces, although there was an arrhythmic drop in force from 2.0 – 2.4 mM Ca2+. Beyond 2.4 mM Ca2+, arrhythmia stopped and the force recovered to the level at 2.0 mM Ca2+ when extracellular Ca2+ concentration was increased to 3.2 mM.

There was also a very slight drop in forces owing to arrhythmia from 2.0 – 2.4 mM Ca2+ in the

1000 µM group. The EC50 values of Gemfibrozil-treated EHMs were generally right-shifted, indicating that treatment with the drug resulted in lowered extracellular Ca2+ sensitivity. The extent of shift, however, was not dosage dependent as EC50 (10µM) is about 0.8mM, EC50 (100µM) is about 0.7mM and EC50 (1000µM) is about 0.75mM.

Increasing frequency of stimulation from 2Hz – 3Hz resulted in no change of contractile forces in the EHMs treated with 10 µM and 100 µM Gemfibrozil. This is in contrast with the control and 1000 µM groups, which showed negative force-frequency relationship.

266

1.0 0.9 0.8 0.7 Control 0.6 10uM Gemfibrozil 0.5 100uM Gemfibrozil 0.4 1000uM Gemfibrozil

Force (mN) Force 0.3 0.2 0.1 0.0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0 [Ca2+] in mM

Figure 5.4.3-26 Contractile forces o EHMs treated with different concentrations of Gemfibrozil

Force frequency response of EHMs treated with different [Gemfibrozil] 120

110

100 Control 10uM Gemfibrozil 100uM Gemfibrozil 90 % force at 2Hz 1000uM Gemfibrozil

80 1.0 1.5 2.0 2.5 3.0 Frequency (Hz)

Figure 5.4.3-27 Force frequency response of Gemfibrozil-treated EHMs

267

Isoprenaline carbachol response of EHMs treated with [Gemfibrozil] 2+ 220 200 180 160 140 120 100 Control 80 10uM Gemfibrozil 60 40 100uM Gemfibrozil 20 1000uM Gemfibrozil 0 % force to force at 0.8mM at Ca force to force % Isoprenaline Carbachol Figure 5.4.3-28 Responses of Gemfibrozil -treated EHMs to Isoprenaline and Carbachol

Increase in force Reduction in 2+ (w.r.t forces at 0.8mM Ca ) Isoprenaline-induced forces with after Isoprenaline treatment Carbachol 120 0.14 0.12 100 0.10 80 0.08 60 0.06

40 0.04 0.02

% increase of force of increase % 20 0.00 0 -0.02 0 10 100 1000

0 10 100 1000 (mN) forces in Reduction -0.04 [Gemfibrozil] in uM [Gemfibrozil] in uM Figure 5.4.3-30 Increase in Gemfibrozil-treated Figure 5.4.3-29 Reduction in Gemfibrozil-treated EHM forces after isoprenaline stimulation EHM force after carbachol stimulation

Despite the apparent difference in responses to carbachol treatment between the 100µM and 1000µM groups, there were no statistically significant differences in the responses to isoprenaline and carbachol stimulations in the experimental groups.

268

5.5. Discussion

5.5.1.

In this chapter, the effect of extracellular calcium on CM maturation in EHM format was investigated. From section 5.4.1, it is apparent that ventricular CMs (mCherry bright cells) need extracellular calcium for proper growth and survival. EHMs cultured in calcium free media showed no observable mCherry (mChr) expression and no observable beating. Despite the presence of a few ventricular CMs in EHMs cultured in 0.4mM Ca2+, the contractile forces were negligible for these EHMs. Increasing the extracellular Ca2+ to 0.8mM appeared to increase MYL2 expression by 2-fold in comparison to 0.4mM Ca2+, and 10-fold in comparison with 0mM Ca2+.

This was evident from the G.Luc expression level that was about 1.93x and about 9.53x the level observed from EHMs cultured in 0.4mM Ca2+ and calcium free media respectively. Increasing the extracellular Ca2+ further to 1.2mM further improved MYL2 expression by about 1.47 fold leading to maximal contraction force of 0.54 ± 0.008 mN (mean ± SEM) vs 0.27 ± 0.042 mN in

EHMs cultured in 0.8mM Ca2+. It is interesting to note that the observed contractile forces were not proportional to MYL2 expression levels.

Increasing the extracellular Ca2+ further to 1.6mM and 2.0mM did not seem to benefit the EHMs further in terms of MYL2 expression levels and contractile forces. There was only about 1.13-fold more MYL2 in EHMs cultured in 1.6mM and 2.0mM Ca2+ and this did not translate to higher contractile forces. In fact, culturing EHMs in 2.0mM Ca2+ seemed to have a negative effect – these EHMs beat arrhythmically with reduced forces when extracellular Ca2+ was > 2.8mM in the organ bath (figure 5.4.1-2); they also showed poorer response to isoprenaline treatment (figure 5.4.1-7). Considering the CMs inside the EHMs were only 30 days old of differentiation, their calcium handling machinery would not have been adult-stage yet; 269

which means that their contractility depends mostly on calcium fluxes between the cytosol and extracellular fluid (Katz and Lorell, 2000). In view of the report that in the failing hearts, fetal transcriptomes are re-activated (Katz and Lorell, 2000), the effects seen here could parallel to the decreased response to beta-adrenergic receptor stimulation seen in heart failure (Wynne et al., 1993), possibly due to changes in the excitation-contraction machinery instead of limitation of cAMP generation (Laurent et al., 2001).

The second part of this chapter looked into whether IWR1 conferred ventricular subtype specificity during cardiac differentiation as has been reported before (Karakikes et al., 2014;

Weng et al., 2014). The knock-in cell line is an excellent tool for these types of studies as mCherry expression faithfully represents MYL2 gene expression. Thus the ventricular CMs fraction can be easily identified by the mCherry positive cell population. This approach eliminates false positives or negatives from conventional FACS staining approach. Indeed, CMs from the same population fixed in different fixatives such as 90% methanol, 70% ethanol or 4% paraformaldehyde have showed different levels of MYL2 positive fraction, leading to overestimation or underestimation of the true MYL2 positive fraction in the cell suspension (own unpublished data). Therefore, as a measure of success for different subtype differentiation protocols, live-cell FACS was used to determine the absolute mCherry positive fraction from the

CMs cell suspension.

Using this method there was no statistically significant difference between flasks differentiated with 5 µM IWP4 or IWR1 at 2.5 and 5 µM concentrations. Although 2.5µM IWR1 was shown to generate significantly more mCherry positive cells when data was normalized to α- actinin positive fraction, this data was not reliable as the ratio went above 100%, indicating that the α-actinin positive fraction was underestimated during FACS. This is not surprising since FACS

270

necessitates CM monolayer trypsinization before fixation and staining. A number of CMs might have been killed during the trypsinization process leading to an underestimation of α-actinin positive fraction. Furthermore, each flask was divided into 2 fractions after trypsinization (one half was used for mCherry live FACS and the other for fixation and staining with α-actinin). The cell suspension may have been divided unequally during this process. Furthermore, Karakikes

(2014) and Weng (2014) employed hypoxic conditions either for maintaining undifferentiated hPSCs (Karakikes et al., 2014) or during differentiation between day 0 – 8 (Weng et al., 2014). It is possible that the transition between hypoxic to normoxic or vice versa enhanced the ventricular subtype specification. Further experiments are necessary before any conclusions can be made.

Finally the third part of this chapter looks into the different effects that certain cardio- active select drugs (noradrenaline, endothelin-1, IBMX, Metformin and Gemfibrozil) might have on EHMs generated from knock-in CMs. Since NA, ET-I and IBMX have been reported to have hypertrophic effects on the heart (Bupha-Intr et al., 2012; Kaddoura et al., 1996; Korth, 1978;

Newling et al., 1989; Reuter, 1983; Reuter and Scholz, 1977) the analysis for these 3 drugs will be combined.

Combining the luciferase graphs of ET-1, NA and IBMX, the data suggest that ET-1 produced most pronounced response after 7 days of treatment, followed by IBMX and NA (figure

5.5.1-1). When we look at 5.5.1-1 and -2 side by side, it is apparent that luciferase expression cannot be the proxy of contractile forces as the 2 graphs do not correlate. Thus contractile forces might not be proportional to the level of MYL2 expression alone. The EC50 values also indicated that treatment with NA rendered the EHMs less sensitive to extracellular calcium concentrations, followed by treatment with ET-1 and IBMX. Finally, the responses to isoprenaline are generally

271

dose-dependent for ET-1 and IBMX, but not for NA for which 100nM showed the maximal response (figure 5.5.1-3). Chronic treatment with NA has been reported to result in desensitization of adrenoreceptors resulting in reduced response during isoprenaline stimulation, similar to that observed in heart failure (Wynne et al., 1993). This dose-dependent reduced response to isoprenaline stimulation was also observed in EHMs treated with ET-1 and IBMX

(figure 5.5.1-3), possibly also owing to adrenoreceptor desensitization mechanism.

G.Luc expression after 10 days of drug treatment 6.0

) 5

4.0 Control Noradrenaline

RFU (x10

2.0 IBMX Endothelin-1

10 100 1 1 10 100 0.1 1 10 nM nM µM µM µM µM nM nM nM

Figure 5.5.1-1 Combined luciferase expression of EHMs treated with different hypertrophic-inducing drugs

Highest contractile forces of drug-treated EHMs

0.8

Control 0.6

Noradrenaline 0.4 IBMX Force (mN)Force Endothelin-1 0.2

0.0 10 100 1 1 10 100 0.1 1 10 nM nM µM µM µM µM nM nM nM

Figure 5.5.1-2 Combined EHM contractile forces treated with different hypertrophic-inducing drugs

272

Control [ET-1] [NA] [IBMX] [EC50 values [EC50 values (mM Ca2+)] [EC50 values (mM Ca2+)] [EC50 values (mM Ca2+)] (mM Ca2+)] 0.1nM (0.75) 10 nM (0.95) 1 µM (0.72) (0.65) 1 nM (0.7) 100 nM (0.95) 10 µM (0.62) 10 nM (0.73) 1000 nM (0.8) 100 µM (0.63)

Table 5.5.1-1 Summary of EC50 values

EHMs response to Isoprenaline

* * 100 Control

Noradrenaline

50 IBMX

Endothelin-1

% increase of force of increase % 0 10 100 1 1 10 100 0.1 1 10 nM nM µM µM µM µM nM nM nM

Figure 5.5.1-3 Summary of isoprenaline response

p 80 EHMs response to Carbachol

60

40 *** Control Noradrenaline

20 * IBMX

% increase of force of increase % Endothelin-1 0 10 100 1 1 10 100 0.1 1 10 nM nM µM µM µM µM nM nM nM

Figure 5.5.1-4 Summary of carbachol response 273

Combining the luciferase graph of EHMs treated with Metformin and Gemfibrozil, it is apparent that EHMs treated with Metformin expressed higher level of luciferase than those treated with Gemfibrozil. Metformin displayed a very interesting trend. From day 1 to day 7, luciferase expression was proportional to the concentration of Metformin in the culture media

(EHMs treated with higher concentrations of Metformin expressed more luciferase). But by the

10th day of treatment, the luciferase expression level for 2 µM Metformin-treated EHMs was very similar to the 200 µM Metformin-treated ones; with those treated with 20 µM Metformin expressing slightly lower level of luciferase. It can be inferred from this data that MYL2 gene expression should be rather similar in these EHMs on day 10.

G.Luc expression after 10 days of drug treatment

6.0 ) 5

4.0

Control

RFU (x10 Gemfibrozil 2.0 Metformin

10 100 1 2 20 200 µM µM mM µM µM µM

Figure 5.5.1-5 Summary of luciferase expression from EHMs treated with Metformin and Gemfibrozil

Gemfibrozil, on the other hand displayed a more straightforward trend: EHMs treated with increasing dose of Gemfibrozil expressed decreasing amount of luciferase for a good 9 days of treatment. On the 10th day however, the level of luciferase expression for EHMs treated with

10 µM and 100 µM fell to the same level. Thus it could be implied that chronic treatment with

274

increasing concentrations of Gemfibrozil led to decreasing expression of MYL2 gene, at least when Gemfibrozil concentration was >100 µM.

Yet the EHMs beating force profiles were quite different from expected. EHMs treated with 2µM Metformin beat with the strongest forces, almost similar to the control level, whereby those treated with the higher concentrations of Metformin (20 µM and 200 µM) beat with lower forces. Although the mean forces of EHMs treated with 20 µM and 200 µM Metformin appeared to be similar in magnitude, the variation EHMs within the 200 µM group was larger.

Highest contractile forces of drug-treated EHMs 1.0

0.8

0.6

Control 0.4

Force (mN)Force Gemfibrozil

0.2 Metformin

0.0 10 100 1 2 20 200 µM µM mM µM µM µM

Figure 5.5.1-6 Summary of forces for EHMs treated with Metformin and Gemfibrozil

Gemfibrozil contractile forces also did not seem to follow the luciferase expression trend.

From 0 mM – 1.6 mM Ca2+, 100 µM Gemfibrozil-treated EHMs beat with stronger forces than their 10 µM counterparts. From 2.0mM – 3.2mM Ca2+, however, 100 µM Gemfibrozil-treated

EHMs exhibited arrhythmia, which was not observed with EHMs treated with 10 µM, or 1000

µM of the drug.

275

Control [Metformin] [Gemfibrozil] [EC50 values (mM Ca2+)] [EC50 values (mM Ca2+)] [EC50 values (mM Ca2+)] 2 µM (0.70) 10 µM (0.8) (0.65) 20 µM (0.73) 100 µM (0.7) 200 µM (0.67) 1000 µM (0.75)

Table 5.5.1-2 Summary of EC 50 values of EHMs treated with Metformin and Gemfibrozil

In general, EHMs treated with Gemfibrozil showed a more right-shifted EC50 values and better isoprenaline response (table 5.5.1-2) than the Metformin treated ones (EC50 = about 0.7

– 0.8mM vs 0.67 – 0.73mM Ca2+ respectively).

EHMs response to Isoprenaline

100 Control Gemfibrozil

Metformin

50

% increase of force of increase % 0 10 100 1 2 20 200 µM µM mM µM µM µM

Figure 5.5.1-7 Summary of isoprenaline response from EHMs treated with Metformin and Gemfibrozil

80 EHMs response to Carbachol

60

Control 40 Gemfibrozil

Metformin 20

% increase of force of increase % 0 10 100 1 2 20 200 µM µM mM µM µM µM

Figure 5.5.1-8 Summary of carbachol response from EHMs treated with Metformin and Gemfibrozil

276

The effects seen with Gemfibrozil treatment could stem from the fact that this drug is a peroxisome proliferator–activated receptor alpha (PPARα) agonist. Chronic PPARα stimulation has been associated with increased MYL2 transcription as well as cardiac hypertrophy (Finck et al., 2002; Hamano et al., 2001) and moderate systolic dysfunction (Finck, 2007). Furthermore, in presence of high concentrations of circulating lipids, chronic stimulation of PPARα could result in uncontrolled fatty acid oxidation, toxic lipid intermediate accumulation, and reactive oxygen species accumulation which might lead to cellular degeneration, necrosis and cardiomyopathic remodeling (Finck, 2007; Pruimboom-Brees et al., 2006). Furthermore, since Gemfibrozil has been reported to inhibit mitochondrial respiration (Nadanaciva et al., 2007); this might lead to mitochondrial dysfunction and subsequent cell death. Altogether, these might explain the functional decline in EHMs treated with increasing doses of Gemfibrozil.

277

5.6. References

Adabag, A.S., Mithani, S., Al Aloul, B., Collins, D., Bertog, S., and Bloomfield, H.E. (2009). Efficacy of gemfibrozil in the primary prevention of atrial fibrillation in a large randomized controlled trial. American Heart Journal 157, 913-918. Beavo, J.A., Rogers, N.L., Crofford, O.B., Hardman, J.G., Sutherland, E.W., and Newman, E.V. (1970). Effects of xanthine derivatives on lipolysis and on adenosine 3',5'-monophosphate phosphodiesterase activity. Molecular pharmacology 6, 597-603. Brückner, R., Gramann, S., Nose, M., Schmitz, W., and Scholz, H. (1985). Isoprenaline-like effects of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine on mechanical, biochemical and electrophysiological parameters in the mammalian heart. Experientia 41, 732-734. Bupha-Intr, T., Haizlip, K.M., and Janssen, P.M. (2012). Role of endothelin in the induction of cardiac hypertrophy in vitro. PloS one 7, e43179. Daly, J.W., Jacobson, K.A., and Ukena, D. (1987). Adenosine receptors: development of selective agonists and antagonists. Progress in clinical and biological research 230, 41-63. De Mello, W.C. (1989). Effect of isoproterenol and 3-isobutyl-1-methylxanthine on junctional conductance in heart cell pairs. Biochimica et biophysica acta 1012, 291-298. Deree, J., Martins, J.O., Melbostad, H., Loomis, W.H., and Coimbra, R. (2008). Insights into the regulation of TNF-alpha production in human mononuclear cells: the effects of non-specific phosphodiesterase inhibition. Clinics (Sao Paulo, Brazil) 63, 321-328. Essayan, D.M. (2001). Cyclic nucleotide phosphodiesterases. The Journal of allergy and clinical immunology 108, 671-680. Finck, B.N. (2007). The PPAR regulatory system in cardiac physiology and disease. Cardiovasc Res 73, 269- 277. Finck, B.N., Lehman, J.J., Leone, T.C., Welch, M.J., Bennett, M.J., Kovacs, A., Han, X., Gross, R.W., Kozak, R., Lopaschuk, G.D., et al. (2002). The cardiac phenotype induced by PPARalpha overexpression mimics that caused by diabetes mellitus. The Journal of clinical investigation 109, 121-130. Forman, B.M., Chen, J., and Evans, R.M. (1997). Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors alpha and delta. Proceedings of the National Academy of Sciences of the United States of America 94, 4312-4317. Fu, J.D., Rushing, S.N., Lieu, D.K., Chan, C.W., Kong, C.W., Geng, L., Wilson, K.D., Chiamvimonvat, N., Boheler, K.R., Wu, J.C., et al. (2011). Distinct roles of microRNA-1 and -499 in ventricular specification and functional maturation of human embryonic stem cell-derived cardiomyocytes. PloS one 6, e27417. Hamano, T., Kobayashi, K., Sakairi, T., Hayashi, M., and Mutai, M. (2001). Peroxisome proliferator- activated receptor alpha (PPAR alpha) agonist, WY-14,643, increased transcription of myosin light chain-2 in cardiomyocytes. The Journal of toxicological sciences 26, 275-284. Inzucchi, S.E. (2005). Metformin and Heart Failure: Innocent until proven guilty. Diabetes Care 28, 2585- 2587. Jain, A., Atale, N., Kohli, S., Bhattacharya, S., Sharma, M., and Rani, V. (2014). An assessment of norepinephrine mediated hypertrophy to apoptosis transition in cardiac cells: A signal for cell death. Chemico-biological interactions.

278

Jayawardena, T.M., Finch, E.A., Zhang, L., Zhang, H., Hodgkinson, C., Pratt, R.E., Rosenberg, P.B., Mirotsou, M., and Dzau, V.J. (2014). MicroRNA Induced Cardiac Reprogramming In vivo: Evidence for Mature Cardiac Myocytes and Improved Cardiac Function. Circ Res. Kaddoura, S., Firth, J.D., Boheler, K.R., Sugden, P.H., and Poole-Wilson, P.A. (1996). Endothelin-1 is involved in norepinephrine-induced ventricular hypertrophy in vivo. Acute effects of bosentan, an orally active, mixed endothelin ETA and ETB receptor antagonist. Circulation 93, 2068-2079. Karakikes, I., Senyei, G.D., Hansen, J., Kong, C.W., Azeloglu, E.U., Stillitano, F., Lieu, D.K., Wang, J., Ren, L., Hulot, J.S., et al. (2014). Small molecule-mediated directed differentiation of human embryonic stem cells toward ventricular cardiomyocytes. Stem Cells Transl Med 3, 18-31. Katz, A.M., and Lorell, B.H. (2000). Regulation of cardiac contraction and relaxation. Circulation 102, Iv69- 74. Korth, M. (1978). Effects of several phosphodiesterase-inhibitors on guinea-pig myocardium. Naunyn- Schmiedeberg's archives of pharmacology 302, 77-86. Laurent, C.E., Cardinal, R., Rousseau, G., Vermeulen, M., Bouchard, C., Wilkinson, M., Armour, J.A., and Bouvier, M. (2001). Functional desensitization to isoproterenol without reducing cAMP production in canine failing cardiocytes. American journal of physiology Regulatory, integrative and comparative physiology 280, R355-364. Marques, L.J., Zheng, L., Poulakis, N., Guzman, J., and Costabel, U. (1999). Pentoxifylline inhibits TNF-alpha production from human alveolar macrophages. American journal of respiratory and critical care medicine 159, 508-511. McNeill, J.H., Brenner, M.J., and Muschek, L.D. (1973). Interaction of four methylxanthine compounds and norepinephrine on cardiac phosphorylase activation and cardiac contractility. Recent advances in studies on cardiac structure and metabolism 3, 261-273. Nadanaciva, S., Dykens, J.A., Bernal, A., Capaldi, R.A., and Will, Y. (2007). Mitochondrial impairment by PPAR agonists and statins identified via immunocaptured OXPHOS complex activities and respiration. Toxicology and applied pharmacology 223, 277-287. Newling, R.P., Fletcher, P.J., Contis, M., and Shaw, J. (1989). Noradrenaline and cardiac hypertrophy in the rat: changes in morphology, blood pressure and ventricular performance. Journal of hypertension 7, 561- 567. Parsonage, W., Hetmanski, D., and Cowley, A. (2002). Differentiation of the metabolic and vascular effects of insulin in insulin resistance in patients with chronic heart failure. The American journal of cardiology 89, 696-703. Peters-Golden, M., Canetti, C., Mancuso, P., and Coffey, M.J. (2005). Leukotrienes: underappreciated mediators of innate immune responses. Journal of immunology (Baltimore, Md : 1950) 174, 589-594. Peytremann, A., Nicholson, W.E., Liddle, G.W., Hardman, J.G., and Sutherland, E.W. (1973). Effects of methylxanthines on adenosine 3',5'-monophosphate and corticosterone in the rat adrenal. Endocrinology 92, 525-530. Piccini, J.P., Klein, L., Gheorghiade, M., and Bonow, R.O. (2004). New insights into diastolic heart failure: role of diabetes mellitus. The American Journal of Medicine 116, 64-75. Pruimboom-Brees, I., Haghpassand, M., Royer, L., Brees, D., Aldinger, C., Reagan, W., Singh, J., Kerlin, R., Kane, C., Bagley, S., et al. (2006). A critical role for peroxisomal proliferator-activated receptor-alpha nuclear receptors in the development of cardiomyocyte degeneration and necrosis. The American journal of pathology 169, 750-760.

279

Reuter, H. (1983). calcium channel modulation by neurotransmitters, enzymes and drugs. Nature 301, 569-574. Reuter, H., and Scholz, H. (1977). A study of the ion selectivity and the kinetic properties of the calcium dependent slow inward current in mammalian cardiac muscle. The Journal of physiology 264, 17-47. Saku, K., Gartside, P.S., Hynd, B.A., and Kashyap, M.L. (1985). Mechanism of action of gemfibrozil on lipoprotein metabolism. Journal of Clinical Investigation 75, 1702-1712. Singh, R., Singh, A.P., Singh, M., and Krishan, P. (2011). Impact of obesity on hypertension-induced cardiac remodeling: role of oxidative stress and its modulation by gemfibrozil treatment in rats. Free radical biology & medicine 50, 363-370. Takano, H., Nagai, T., Asakawa, M., Toyozaki, T., Oka, T., Komuro, I., Saito, T., and Masuda, Y. (2000). Peroxisome proliferator-activated receptor activators inhibit lipopolysaccharide-induced tumor necrosis factor-alpha expression in neonatal rat cardiac myocytes. Circulation research 87, 596-602. Terasaki, W.L., and Brooker, G. (1977). Cardiac adenosine 3':5'-monophosphate. Free and bound forms in the isolated rat atrium. The Journal of biological chemistry 252, 1041-1050. Weng, Z., Kong, C.W., Ren, L., Karakikes, I., Geng, L., He, J., Chow, M.Z., Mok, C.F., Keung, W., Chow, H., et al. (2014). A simple, cost-effective but highly efficient system for deriving ventricular cardiomyocytes from human pluripotent stem cells. Stem cells and development 23, 1704-1716. Wynne, D.G., Poole-Wilson, P.A., and Harding, S.E. (1993). Incomplete reversal of beta-adrenoceptor desensitization in human and guinea-pig cardiomyocytes by cyclic nucleotide phosphodiesterase inhibitors. British journal of pharmacology 109, 1071-1078. Zhang, Q., Jiang, J., Han, P., Yuan, Q., Zhang, J., Zhang, X., Xu, Y., Cao, H., Meng, Q., Chen, L., et al. (2011). Direct differentiation of atrial and ventricular myocytes from human embryonic stem cells by alternating retinoid signals. Cell Res 21, 579-587.

280

CHAPTER 6

General Conclusion

and

Future Directions

281

Human pluripotent stem cells (hPSCs) hold great potential in cardiovascular research.

However, these potentials cannot yet be fully realized due to a number of persistent technical challenges. Published methods of cardiomyocyte (CM) differentiation from hPSCs were found here to be non-robust and non-reproducible, necessitating rounds of optimizations in this and other laboratories before success can be anticipated from the protocols. In addition, current differentiation protocols generate a mixture of phenotypically immature CMs displaying nodal- lie, atrial-like and ventricular-like action potentials (Mummery et al., 2012). These fetal-stage

CMs might not be optimal for downstream applications as they behave differently (reviewed in

Chapter 2 section 2.2.3) from their adult counterparts (reviewed in Chapter 4 section 4.2.1). The current study has three main objectives: to develop a robust and reproducible CM differentiation protocol; to generate an MYL2-knock-in reporter line which would facilitate the enrichment for the ventricular CM subtype and to improve the maturation of hESCs-derived CMs in vitro.

The MYL2-knock-in reporter line was successfully generated through TALEN-mediated homologous recombination (HR). TALEN technology was empirically demonstrated to improve

HR efficiency from a reported ratio of 26:1 stable transfected clones to HR events (Zwaka and

Thomson, 2003), to about 18.8:1 in H9 background and about 3.7:1 in induced PSCs MBJ6 background. The reporter line was also shown to be pluripotent with no gross chromosomal aberrations. In addition, functional validation experiments demonstrated that the reporter line was functional as designed. Each knocked-in reporter (mCherry, neomycin resistance and

Gaussia luciferase) was expressed in significant correlation to MYL2 gene (p ≤ 0.05 for G. Luc; p ≤

0.01 for mCherry and p ≤ 0.001 for NeoR), as well as to each other (table 4.4.4-1). Ventricular

CMs (vCMs) fraction could be monitored in real time through the qualitative mCherry signal as

282

well as quantitative G. luciferase signal. Furthermore, the vCM fraction could be enriched through G418 selection of up to 50µg/ml and live cell FACS for mCherry positive cell population.

Removal of the floxed selection cassette PGK-Hygromycin was shown to enhance mCherry reporter expression, as evident from the significantly higher mean mCherry fluorescence intensity from H9VC16 CMs versus that from H9V15 CMs (figure 3.4.6-7). The utility of the cell line was also tested in pilot studies to develop a vCM differentiation protocol, and in drug screening using the knock-in CMs in EHMs format. The novel MYL2-knock-in reporter line was demonstrated to provide a simple yet effective read-out for MYL2 positive fraction, eliminating potential under- or overestimation seen with conventional immunohistochemistry methods. From the drug screening study, it was learned that G.luc and mCherry reporter expressions could be used to indicate the drug effects on CM viability as well as MYL2 gene expression, yet they cannot be used as proxy for EHM contractile forces.

Through nine iterations, a set of robust and reproducible monolayer and embryoid body- based differentiation protocols were finally obtained. All differentiated embryoid bodies (EBs) beat by day 9 of differentiation, giving a differentiation efficiency of 100%. Monolayer based differentiation gave beating monolayers of CMs with >80% success rate from all differentiation efforts. Furthermore, in this study, a simple yet efficient CM trypsinization protocol has been successfully developed to improve CM viability post digestion and consequently CM yield.

In vitro aging, in conjunction with auxotonic stretching and IGF1, VEGF and FGF treatment enhanced the CM maturation in the EHMs as determined by the mRNA and miRNA profiles. Electrical stimulation also seemed to improve CM maturation in the EHMs, as evident from the higher level of expression as well as enhanced organization and alignment of MYL2; decrease in hypersensitivity to extracellular calcium; enhanced response to isoprenaline and 283

carbachol stimulation and improvement in force-frequency behavior. However, increasing preload by 10% did not seem to enhance CM maturation. Including calcium at 1.2mM – 1.6mM was demonstrated to be beneficial for force development of the CMs, yet this approach also did not enhance maturation of CMs in the EHMs.

Despite the encouraging level of maturation seen in in vitro aging as well as electrical stimulation experiments, additional work is still required to further enhance the maturation level in these CMs as they are still far from the maturation level seen in adult CMs in vivo. So far, it seems that the electrical stimulation approach is the most promising. However, the electrical stimulation parameters need to be further optimized to minimize cell death and maximize CM maturation. In addition, further work is also required to assess CM maturation from electrophysiological and ultrastructural angles.

There are a number of publications on electrical stimulation for enhancement of CM maturation in vitro yet each approach uses different stimulation parameters as well as maturation parameters (Deng et al., 2000; Hirt et al., 2014; Martherus et al., 2010; Nunes et al.,

2013). It would be useful for the field to develop a gold standard protocol to measure CM maturation. Once this gold standard is established, one can then work on optimizing in vitro CM maturation protocol. With a CM maturation protocol at hand, the engineering of a biomimetic ventricular cardiac tissue model with properties close to that of the ventricular myocardium of the heart in vivo would be feasible. Then, not only drug screening assays for cardiotoxicity will more representative of the in vivo data, cell therapy for the treatment of heart failure patients will also be closer to reality. The novel MYL2 knock-in reporter line described here in this thesis, however, provides a starting point from which a homogenous population of ventricular cardiomyocytes can be obtained for these studies.

284

References

Deng, X.F., Rokosh, D.G., and Simpson, P.C. (2000). Autonomous and growth factor-induced hypertrophy in cultured neonatal mouse cardiac myocytes. Comparison with rat. Circ Res 87, 781-788.

Hirt, M.N., Boeddinghaus, J., Mitchell, A., Schaaf, S., Börnchen, C., Müller, C., Schulz, H., Hubner, N., Stenzig, J., Stoehr, A., et al. (2014). Functional improvement and maturation of rat and human engineered heart tissue by chronic electrical stimulation. Journal of molecular and cellular cardiology 74, 151-161.

Martherus, R.S., Vanherle, S.J., Timmer, E.D., Zeijlemaker, V.A., Broers, J.L., Smeets, H.J., Geraedts, J.P., and Ayoubi, T.A. (2010). Electrical signals affect the cardiomyocyte transcriptome independently of contraction. Physiological genomics 42a, 283-289.

Mummery, C.L., Zhang, J., Ng, E.S., Elliott, D.A., Elefanty, A.G., and Kamp, T.J. (2012). Differentiation of human embryonic stem cells and induced pluripotent stem cells to cardiomyocytes: a methods overview. Circulation research 111, 344-358.

Nunes, S.S., Miklas, J.W., Liu, J., Aschar-Sobbi, R., Xiao, Y., Zhang, B., Jiang, J., Masse, S., Gagliardi, M., Hsieh, A., et al. (2013). Biowire: a platform for maturation of human pluripotent stem cell-derived cardiomyocytes. Nat Methods 10, 781-787.

Zwaka, T.P., and Thomson, J.A. (2003). Homologous recombination in human embryonic stem cells. Nature Biotechnology 21, 319 - 321.

285

APPENDICES

286

Appendix from Chapter 4

1. Appendix 4.1 Expression of Gaussia luciferase from CMs cultured in monolayer format across different time points.

Appendix 4 - 1 G.Luc expression of 2D CMs

2. Appendix 4.2 qPCR data from EHMs at different time points. Each reading was normalized to GAPDH followed by their corresponding expression on day 18, when cardiomyocytes were trypsinized to generate EHMs

Appendix 4 - 2 C16 time point qPCR data relative to GAPDH and d18

287

Appendix from Chapter 5

Appendix 5 - 1 Gene expression profile 3 (ratio is ratio of log2expression at different time points)

p-value p-value p-value p-value p-value p-value Ratio (Day Ratio (Day Ratio (Day Ratio (Day Ratio (Day Ratio (Day (Day 45 (Day 60 (Day 30 (Day 45 (Day 30 external_gene_id (D30 vs. 30 vs. Day 45 vs. Day 60 vs. Day 30 vs. Day 45 vs. Day 30 vs. Day vs. Day vs. Day vs. Day vs. Day vs. Day D90) 90) 90) 90) 60) 60) 45) 90) 90) 60) 60) 45) ABCB1 1.14E-05 4.61661 4.55E-04 2.71356 1.20E-03 2.40381 5.18E-03 1.92054 5.32E-01 1.12886 1.64E-02 1.70131 AC002480.3 6.10E-06 5.2387 4.85E-05 3.7524 2.80E-05 4.07968 2.16E-01 1.2841 6.69E-01 0.919777 1.07E-01 1.39609 AC007308.6 2.21E-05 12.2754 1.55E-04 7.44393 3.18E-03 3.82299 4.73E-03 3.21093 6.90E-02 1.94715 1.59E-01 1.64904 AC009518.2 2.85E-05 6.53199 1.13E-04 4.9785 5.12E-03 2.60488 4.08E-03 2.5076 2.72E-02 1.91122 3.09E-01 1.31204 ACAP1 7.86E-06 5.5113 2.19E-04 3.23229 6.57E-04 2.77604 5.88E-03 1.98531 4.66E-01 1.16436 2.26E-02 1.70507 ACP5 1.24E-06 4.36822 5.85E-05 2.66034 3.95E-05 2.78063 9.40E-03 1.57094 7.64E-01 0.956738 5.44E-03 1.64198 ACSF2 1.31E-05 2.49519 6.17E-04 1.7904 1.25E-04 2.03356 1.00E-01 1.227 2.88E-01 0.880425 1.41E-02 1.39365 ACTBL2 1.06E-06 12.9633 3.21E-04 3.92055 3.80E-05 5.7898 7.36E-03 2.23899 1.41E-01 0.677148 5.00E-04 3.30651 ALOX12B 3.18E-05 9.09617 1.26E-04 6.58968 9.75E-03 2.77637 2.41E-03 3.27629 1.57E-02 2.37349 3.10E-01 1.38037 ANKEF1 8.71E-09 4.68473 1.75E-06 2.50893 1.17E-06 2.60747 5.79E-05 1.79665 6.86E-01 0.962208 3.28E-05 1.86722 ANKRD1 6.10E-06 4.01505 6.78E-06 3.95386 7.50E-05 2.87502 6.05E-02 1.39653 7.14E-02 1.37525 9.25E-01 1.01548 APOA1 1.82E-05 3.72316 3.92E-05 3.34686 6.89E-04 2.35139 2.04E-02 1.58338 6.17E-02 1.42335 5.43E-01 1.11243 ARHGDIG 4.03E-06 5.54545 4.80E-05 3.7114 2.08E-04 3.01984 8.10E-03 1.83634 2.97E-01 1.22901 5.65E-02 1.49416 ATP1B4 2.81E-07 4.64269 3.82E-08 6.44844 1.09E-05 2.87848 3.49E-03 1.6129 6.30E-05 2.24022 2.73E-02 0.719971 ATP4A 5.95E-07 15.9215 4.32E-05 5.84238 6.18E-05 5.4485 1.31E-03 2.92218 7.86E-01 1.07229 2.09E-03 2.72517 ATP6V1G2 7.47E-06 3.15114 4.41E-05 2.57626 3.72E-04 2.08516 1.08E-02 1.51122 1.45E-01 1.23552 1.63E-01 1.22315 BAIAP3 1.22E-05 3.82974 2.09E-03 2.04857 2.50E-04 2.59018 3.81E-02 1.47856 1.85E-01 0.790897 3.11E-03 1.86947 BEX1 5.81E-06 6.56057 4.01E-05 4.60161 1.90E-03 2.56432 1.12E-03 2.55841 2.00E-02 1.79448 1.27E-01 1.42571 BNIP3P1 4.88E-06 3.11241 1.02E-05 2.86069 3.82E-05 2.48358 1.04E-01 1.2532 2.91E-01 1.15184 5.22E-01 1.08799 C19orf67 2.31E-06 15.8353 4.88E-05 7.33253 3.11E-05 8.13028 4.03E-02 1.94769 7.26E-01 0.90188 2.13E-02 2.15959 C4orf26 2.38E-05 11.6733 1.99E-04 6.85999 7.12E-03 3.20158 2.24E-03 3.64611 3.97E-02 2.14269 1.32E-01 1.70165 C6orf52 1.35E-07 5.35412 6.95E-07 4.16806 6.16E-05 2.43107 8.78E-05 2.20237 1.75E-03 1.7145 8.30E-02 1.28456 CADM4 9.37E-08 2.58254 5.29E-07 2.22671 3.14E-07 2.32344 1.68E-01 1.11152 5.65E-01 0.958367 6.30E-02 1.1598 CCDC24 1.25E-05 4.54457 1.68E-04 3.09285 1.95E-04 3.03097 5.39E-02 1.49938 9.16E-01 1.02041 6.49E-02 1.46938 288

CDH13 8.75E-06 4.1952 5.41E-06 4.52033 1.64E-04 2.80775 3.88E-02 1.49415 1.78E-02 1.60995 6.71E-01 0.928074 CHRNA3 5.88E-08 10.0702 3.98E-05 3.27104 7.46E-06 4.14258 2.43E-04 2.43089 1.84E-01 0.789612 3.19E-05 3.07859 CLYBL 1.36E-06 3.22288 1.77E-05 2.44645 4.32E-05 2.24985 9.74E-03 1.43248 4.82E-01 1.08738 3.57E-02 1.31737 COX7A1 2.06E-06 15.2092 2.11E-06 15.1012 7.31E-04 4.04377 6.10E-04 3.76115 6.35E-04 3.73444 9.80E-01 1.00715 CTB-50L17.16 4.12E-06 5.55631 2.09E-05 4.23456 4.47E-04 2.73948 3.27E-03 2.02823 4.18E-02 1.54575 1.79E-01 1.31213 CTD-2619J13.17 1.35E-06 12.3388 2.30E-05 6.45722 5.69E-04 3.5842 4.01E-04 3.44254 3.64E-02 1.80158 2.38E-02 1.91085 DDX3Y 3.78E-09 34.0993 1.59E-08 21.7597 6.15E-06 5.51093 1.54E-06 6.18758 2.28E-05 3.94847 4.29E-02 1.56708 DIO2 1.86E-06 6.03355 1.41E-06 6.35077 6.44E-04 2.5304 5.92E-04 2.38443 3.79E-04 2.50979 7.84E-01 0.95005 DMKN 2.62E-07 3.22638 5.47E-07 2.9716 3.60E-06 2.45895 1.80E-02 1.3121 7.91E-02 1.20848 4.18E-01 1.08574 EIF1AY 5.32E-08 27.2418 3.67E-07 15.4491 6.25E-05 4.93401 1.69E-05 5.52123 5.28E-04 3.13114 3.53E-02 1.76333 ENO3 1.16E-06 4.22773 1.47E-06 4.08923 2.72E-04 2.20847 7.07E-04 1.91432 1.05E-03 1.85161 8.16E-01 1.03387 ENSG00000240742 1.26E-05 8.01131 2.23E-04 4.48549 2.73E-03 2.9207 2.44E-03 2.74294 1.25E-01 1.53576 4.62E-02 1.78605 EPCAM 3.14E-09 6.26896 3.44E-08 4.31559 2.04E-07 3.42244 8.49E-05 1.83172 4.11E-02 1.26097 3.36E-03 1.45263 ETFB 5.39E-06 2.82595 5.25E-05 2.2512 6.39E-05 2.21123 6.15E-02 1.278 8.82E-01 1.01807 7.98E-02 1.25531 FAM167A 9.16E-08 9.16323 2.24E-06 5.01394 6.99E-06 4.1902 6.65E-04 2.18682 3.05E-01 1.19659 4.09E-03 1.82755 FAM19A5 3.64E-07 7.59408 3.41E-06 5.03988 9.89E-06 4.24882 6.83E-03 1.78734 3.50E-01 1.18618 3.89E-02 1.5068 FBXO2 5.64E-06 11.0022 2.66E-05 7.62067 1.22E-05 9.1197 5.07E-01 1.20642 5.25E-01 0.835627 2.06E-01 1.44373 FDXR 1.21E-05 3.80416 9.98E-04 2.20585 9.13E-05 2.90573 1.31E-01 1.30919 1.23E-01 0.759138 7.05E-03 1.72458 FSD1 7.36E-08 5.75801 3.28E-06 3.31841 3.06E-06 3.34723 1.49E-03 1.72023 9.48E-01 0.991391 1.33E-03 1.73517 GAPDH 7.10E-06 2.92059 1.62E-05 2.66796 6.58E-05 2.31626 9.09E-02 1.26091 2.83E-01 1.15184 4.85E-01 1.09469 GAPDHP60 9.88E-06 2.98236 8.66E-05 2.36129 3.66E-05 2.57958 2.96E-01 1.15614 5.17E-01 0.915378 1.05E-01 1.26302 GAPDHP70 3.39E-06 4.62319 5.67E-05 3.09532 8.16E-05 2.95562 2.05E-02 1.5642 7.85E-01 1.04726 3.37E-02 1.49361 GBP2 1.28E-09 3.3382 1.47E-07 2.15286 1.47E-07 2.15314 1.66E-05 1.55039 9.98E-01 0.99987 1.65E-05 1.55059 H2AFY2 1.88E-06 2.21845 2.32E-05 1.84302 3.59E-06 2.10807 5.42E-01 1.05236 1.26E-01 0.874266 4.30E-02 1.20371 HMGB3P24 2.14E-06 9.0969 6.78E-05 4.58336 1.47E-04 4.0188 4.25E-03 2.26359 5.75E-01 1.14048 1.18E-02 1.98477 HOXB1 3.47E-06 12.5006 2.43E-05 7.79493 2.08E-04 4.97263 6.26E-03 2.51388 1.29E-01 1.56757 1.12E-01 1.60368 HSPA8P16 5.41E-06 6.53032 2.11E-05 5.06759 1.25E-03 2.68959 1.57E-03 2.428 1.26E-02 1.88415 2.59E-01 1.28865 IL17REL 8.79E-08 10.6704 4.46E-06 4.94586 1.71E-05 4.00055 1.81E-04 2.66723 2.58E-01 1.2363 1.20E-03 2.15744 IL1A 5.27E-10 13.0307 4.96E-09 8.00911 5.20E-07 3.77256 4.78E-07 3.45408 5.50E-05 2.12299 1.76E-03 1.62699 INADL 5.98E-10 2.51444 5.91E-10 2.51665 5.02E-09 2.12814 2.62E-03 1.18152 2.53E-03 1.18256 9.84E-01 0.999124 KBTBD8 5.30E-08 3.81304 5.15E-06 2.33469 6.21E-06 2.29625 2.53E-04 1.66055 8.65E-01 1.01674 3.30E-04 1.63321

289

KLHDC9 1.33E-08 4.73245 2.38E-06 2.54493 8.84E-07 2.80865 2.30E-04 1.68495 3.33E-01 0.906103 5.27E-05 1.85956 KCNV2 3.82E-06 7.094 7.45E-06 6.21401 1.57E-03 2.62459 7.19E-04 2.7029 2.01E-03 2.36761 5.49E-01 1.14161 KRT8P10 5.23E-06 5.16695 6.58E-05 3.47761 2.14E-04 2.95818 1.20E-02 1.74667 4.02E-01 1.17559 5.63E-02 1.48577 LBX1-AS1 1.30E-05 16.0649 6.14E-04 5.86141 1.09E-03 5.14409 7.16E-03 3.12298 7.13E-01 1.13945 1.40E-02 2.74079 LIN28B 4.18E-08 7.98286 1.94E-07 5.98674 6.38E-06 3.5167 1.49E-04 2.26998 3.72E-03 1.70237 7.30E-02 1.33342 LINC00880 2.44E-06 9.03613 6.20E-05 4.73182 4.36E-04 3.4189 1.42E-03 2.64299 1.85E-01 1.38402 1.69E-02 1.90965 LYPD6B 1.01E-06 6.26932 6.96E-06 4.5081 7.35E-06 4.46915 7.99E-02 1.4028 9.61E-01 1.00872 8.69E-02 1.39068 MASP1 1.96E-08 5.21344 5.05E-07 3.33345 2.60E-06 2.77449 1.07E-04 1.87906 1.15E-01 1.20146 1.57E-03 1.56398 MGARP 1.93E-05 3.12479 5.12E-05 2.77946 2.62E-04 2.32439 7.09E-02 1.34435 2.52E-01 1.19578 4.46E-01 1.12424 MRAP2 4.76E-05 4.9846 2.90E-05 5.4615 1.21E-03 2.93665 4.25E-02 1.69737 2.10E-02 1.85977 7.00E-01 0.912679 MSX2 3.66E-07 5.16021 1.10E-05 3.18311 5.56E-05 2.63626 6.08E-04 1.95739 2.10E-01 1.20743 5.82E-03 1.62112 MT1E 1.05E-05 6.03937 1.17E-05 5.91063 5.62E-04 3.1018 1.11E-02 1.94705 1.33E-02 1.90555 9.23E-01 1.02178 MT1G 7.32E-06 20.7795 9.02E-06 19.4478 8.79E-03 3.38685 3.41E-04 6.13534 4.58E-04 5.74215 8.56E-01 1.06847 MT1X 5.21E-06 8.92182 1.74E-05 6.85436 2.34E-04 4.17513 1.07E-02 2.1369 7.03E-02 1.64171 3.09E-01 1.30163 NAP1L2 2.91E-06 3.53265 8.45E-05 2.40098 9.86E-06 3.03733 2.85E-01 1.16308 1.08E-01 0.790491 1.51E-02 1.47133 OGDHL 8.14E-06 4.37152 4.91E-04 2.50889 3.41E-04 2.62057 1.37E-02 1.66816 8.08E-01 0.957384 8.79E-03 1.74241 P2RY14 2.36E-05 4.65296 1.57E-06 7.72124 4.72E-03 2.17779 3.39E-03 2.13656 6.73E-05 3.54546 3.06E-02 0.602618 PDE6B 3.86E-06 3.13647 1.75E-05 2.6476 3.14E-05 2.49318 9.39E-02 1.25802 6.41E-01 1.06194 2.03E-01 1.18465 PDHA2 1.92E-06 3.77241 1.50E-05 2.91645 6.01E-04 2.00249 6.73E-04 1.88386 1.80E-02 1.45641 8.38E-02 1.29349 PFKP 1.62E-05 2.65793 1.21E-04 2.18051 1.11E-04 2.19773 1.55E-01 1.2094 9.51E-01 0.992165 1.40E-01 1.21895 PGK1 7.34E-07 3.37134 2.70E-06 2.90062 6.65E-06 2.6374 5.15E-02 1.27828 4.16E-01 1.0998 2.08E-01 1.16228 PLEKHB1 5.59E-06 3.52621 1.38E-04 2.41935 1.42E-04 2.41276 2.28E-02 1.46148 9.85E-01 1.00273 2.36E-02 1.4575 PLS1 1.80E-09 4.46696 1.20E-08 3.4952 1.06E-06 2.24563 2.44E-06 1.98917 1.38E-04 1.55644 9.08E-03 1.27802 PODXL2 1.58E-06 4.28701 2.31E-05 2.99744 3.81E-05 2.82448 1.53E-02 1.5178 6.91E-01 1.06124 3.17E-02 1.43022 PPL 1.72E-06 5.07621 1.96E-04 2.6296 2.18E-05 3.46861 4.02E-02 1.46347 1.19E-01 0.758113 2.03E-03 1.93041 PROM1 1.22E-05 5.75012 6.76E-05 4.25336 3.59E-04 3.27193 2.44E-02 1.75741 2.51E-01 1.29995 1.91E-01 1.3519 PRRX2 4.33E-07 5.57336 4.36E-06 3.89389 4.59E-04 2.22723 7.93E-05 2.50237 3.47E-03 1.74831 3.66E-02 1.43131 PTPRR 7.34E-07 3.37246 1.01E-06 3.24521 2.63E-05 2.309 6.27E-03 1.46058 1.15E-02 1.40546 7.39E-01 1.03921 RP11-433O3.1 6.31E-08 7.47562 4.36E-06 3.7382 5.07E-06 3.66168 4.87E-04 2.04158 8.90E-01 1.0209 6.11E-04 1.99979 RAB38 9.51E-08 4.13154 1.49E-07 3.88804 7.75E-06 2.48665 6.19E-04 1.66149 1.58E-03 1.56356 5.83E-01 1.06263 RASL10B 2.80E-07 3.53243 1.22E-05 2.35907 3.99E-05 2.12483 5.62E-04 1.66246 3.46E-01 1.11024 2.92E-03 1.49739

290

RND2 2.50E-08 7.33715 2.20E-07 5.02414 6.30E-07 4.28369 1.88E-03 1.71281 2.54E-01 1.17285 1.56E-02 1.46038 RP11-15G8.1 3.65E-06 9.05731 2.01E-05 6.28843 3.44E-04 3.75087 3.67E-03 2.41472 5.43E-02 1.67653 1.57E-01 1.44031 RP11-224O19.2 1.76E-06 5.71049 1.07E-06 6.24995 1.30E-04 2.97883 3.53E-03 1.91703 1.46E-03 2.09813 6.18E-01 0.913686 RP11-44D5.1 6.20E-10 8.4361 5.42E-09 5.69702 1.91E-07 3.44214 2.20E-06 2.45083 3.82E-04 1.65508 2.38E-03 1.48079 RP11-451J24.1 1.49E-05 3.31339 1.71E-04 2.48089 7.19E-04 2.13461 1.43E-02 1.55223 3.42E-01 1.16222 8.23E-02 1.33557 RP11-516A11.1 5.63E-06 3.82797 1.51E-04 2.53902 1.34E-04 2.57339 2.49E-02 1.48752 9.32E-01 0.986644 2.12E-02 1.50766 RP11-545E17.3 3.23E-07 4.59722 1.13E-06 3.8419 8.85E-07 3.97379 2.86E-01 1.15689 8.00E-01 0.966811 1.95E-01 1.1966 RP11-6F2.5 3.09E-06 9.07699 1.76E-04 4.10474 9.21E-05 4.59504 1.49E-02 1.97539 6.42E-01 0.893297 6.36E-03 2.21134 RPS4Y1 1.03E-10 49.2956 4.71E-10 29.5483 1.12E-07 7.49891 1.02E-07 6.5737 2.46E-06 3.94034 7.03E-03 1.66831 SDSL 1.04E-05 4.88471 1.23E-03 2.45577 3.07E-04 2.93864 2.32E-02 1.66224 3.72E-01 0.835683 4.44E-03 1.98908 SELP 1.80E-05 11.2539 1.11E-03 4.37196 1.34E-04 6.88294 1.44E-01 1.63504 1.74E-01 0.635188 1.15E-02 2.57411 SLC31A1P1 1.74E-05 2.34529 1.87E-04 1.91669 7.50E-05 2.06358 2.67E-01 1.13652 5.14E-01 0.92882 9.26E-02 1.22361 SPESP1 4.51E-07 10.4255 2.24E-06 7.35279 3.45E-04 3.1201 1.13E-04 3.3414 1.62E-03 2.35659 1.19E-01 1.4179 SPOCD1 1.36E-06 6.25492 9.10E-06 4.50737 1.47E-05 4.18019 4.73E-02 1.49633 6.84E-01 1.07827 9.67E-02 1.38771 SRP14-AS1 5.43E-06 3.20257 6.04E-05 2.44862 2.99E-05 2.63508 1.68E-01 1.21536 5.90E-01 0.929237 6.70E-02 1.30791 TCTE1 1.38E-05 14.2269 2.96E-04 6.46904 2.63E-03 4.00936 2.91E-03 3.54841 1.78E-01 1.61348 3.72E-02 2.19922 TM7SF2 1.88E-05 3.68764 1.24E-04 2.86928 2.14E-04 2.68421 8.70E-02 1.37382 7.01E-01 1.06894 1.66E-01 1.28522 TMEM38A 3.42E-06 3.98764 8.04E-06 3.54284 3.06E-05 2.98861 8.02E-02 1.33428 2.80E-01 1.18545 4.46E-01 1.12555 TOM1L1 2.06E-06 3.15889 1.49E-06 3.28163 1.77E-05 2.50289 7.44E-02 1.2621 4.25E-02 1.31114 7.53E-01 0.962599 TPST2P1 3.57E-06 6.3466 3.92E-04 2.96834 7.23E-05 3.79339 2.64E-02 1.67307 2.47E-01 0.782503 3.02E-03 2.1381 TREML3P 7.04E-07 7.67641 4.90E-06 5.32442 2.85E-05 3.99267 5.14E-03 1.92262 1.54E-01 1.33355 7.73E-02 1.44173 TSPAN8 6.99E-06 8.18671 1.78E-04 4.32413 8.29E-04 3.3353 3.73E-03 2.45457 3.12E-01 1.29648 2.45E-02 1.89326 TXLNG2P 9.57E-06 40.0012 3.90E-05 23.7001 6.31E-03 5.02966 6.88E-04 7.95306 5.08E-03 4.71207 2.64E-01 1.68781 USP9Y 5.40E-11 20.6186 8.89E-11 17.9851 5.35E-07 3.53664 7.70E-09 5.82999 1.80E-08 5.08537 2.53E-01 1.14643 UTY 3.37E-08 25.5746 1.69E-07 15.9971 1.02E-04 4.12208 5.00E-06 6.2043 7.61E-05 3.88083 5.81E-02 1.5987 VAMP8 5.68E-06 2.53686 7.02E-06 2.48528 5.60E-05 2.0665 7.96E-02 1.22761 1.10E-01 1.20265 8.50E-01 1.02075 WFDC1 3.06E-05 16.3989 1.92E-05 18.9769 6.24E-04 7.03703 4.90E-02 2.33037 2.49E-02 2.69672 7.10E-01 0.864148 ZFY 7.32E-09 47.044 2.55E-08 30.5435 2.41E-06 8.88194 1.51E-05 5.29659 2.10E-04 3.43883 8.46E-02 1.54023 ZIC1 4.08E-06 21.1973 5.74E-05 9.91206 5.78E-04 5.66583 2.39E-03 3.74125 1.25E-01 1.74945 4.52E-02 2.13854 ZNF157 5.45E-06 5.6896 1.17E-04 3.45512 4.29E-04 2.87913 5.42E-03 1.97615 3.75E-01 1.20005 2.81E-02 1.64672 ZNF541 5.41E-06 13.492 1.31E-04 6.22665 5.52E-04 4.62286 3.97E-03 2.91854 3.35E-01 1.34692 2.39E-02 2.16682

291

Appendix 5 - 2 Gene expression profile 18 (ratio is ratio of log2expression at different time points)

Fold- Fold- Fold- Fold- Fold- p-value p-value Fold-Change external Change p-value (D45 Change p-value (D60 Change p-value (D30 Change Change p-value (D30 (D30 vs. (D45 vs. (D30 vs. gene id (D30 vs. vs. D90) (D45 vs. vs. D90) (D60 vs. vs. D60) (D30 vs. (D45 vs. vs. D45) D90) D60) D45) D90) D90) D90) D60) D60)

COL6A6 1.84E-08 -46.12 5.34E-08 -31.80 1.74E-05 -6.87 9.45E-06 -6.71 6.80E-05 -4.63 1.61E-01 -1.45 ADAMTS5 9.41E-09 -7.15 1.10E-08 -6.94 1.70E-06 -3.27 4.04E-05 -2.19 5.67E-05 -2.12 8.09E-01 -1.03 SMOC2 4.58E-08 -10.41 7.80E-08 -9.26 4.66E-05 -3.17 1.75E-05 -3.29 4.47E-05 -2.92 4.93E-01 -1.12 RORB 1.24E-06 -10.35 1.40E-06 -10.06 7.73E-04 -3.10 2.56E-04 -3.34 3.11E-04 -3.25 9.02E-01 -1.03 SOX9 5.94E-09 -6.73 1.04E-09 -9.45 2.01E-06 -2.94 1.24E-05 -2.29 4.35E-07 -3.21 1.06E-02 1.40 RP11- 152P17.2 1.10E-09 -9.06 2.86E-09 -7.50 1.97E-06 -2.91 4.97E-07 -3.11 2.99E-06 -2.58 1.12E-01 -1.21 CILP2 1.04E-07 -14.36 3.33E-07 -10.79 8.09E-04 -2.73 5.37E-06 -5.27 3.12E-05 -3.96 1.87E-01 -1.33 TGFBR3 1.33E-09 -6.44 9.70E-10 -6.81 8.40E-07 -2.72 1.75E-06 -2.37 9.32E-07 -2.50 5.60E-01 1.06 SEMA3F 1.82E-06 -5.60 3.14E-06 -5.10 2.72E-04 -2.69 1.51E-03 -2.08 3.80E-03 -1.89 6.01E-01 -1.10 NCAM2 4.84E-10 -10.99 2.55E-09 -7.78 5.47E-06 -2.63 5.09E-08 -4.17 8.50E-07 -2.96 9.43E-03 -1.41 B4GALT1 3.19E-08 -5.09 2.74E-08 -5.21 6.18E-06 -2.62 9.00E-05 -1.94 6.63E-05 -1.99 8.34E-01 1.02 HMGB1P21 5.88E-06 -5.40 6.18E-06 -5.36 9.09E-04 -2.55 2.55E-03 -2.12 2.76E-03 -2.10 9.65E-01 -1.01 NFATC2 3.78E-05 -5.45 2.50E-05 -5.90 4.32E-03 -2.50 7.35E-03 -2.18 4.07E-03 -2.36 7.44E-01 1.08 ANGPTL1 7.34E-05 -5.57 6.10E-06 -9.53 7.80E-03 -2.48 9.68E-03 -2.25 2.95E-04 -3.84 6.23E-02 1.71 HSPA5 1.08E-07 -5.29 7.16E-08 -5.66 3.94E-05 -2.48 9.67E-05 -2.13 4.59E-05 -2.28 6.07E-01 1.07 ADAMTS4 1.63E-07 -6.26 2.56E-07 -5.78 1.55E-04 -2.43 4.62E-05 -2.58 9.93E-05 -2.38 5.98E-01 -1.08 TTYH2 1.16E-08 -4.96 1.91E-09 -6.67 5.19E-06 -2.40 1.49E-05 -2.07 5.36E-07 -2.78 1.22E-02 1.35 COLEC12 1.01E-09 -8.84 2.75E-09 -7.26 1.11E-05 -2.40 9.87E-08 -3.69 5.18E-07 -3.03 9.42E-02 -1.22 CTD-2369P2.2 2.08E-09 -4.46 2.21E-09 -4.42 5.50E-07 -2.40 7.35E-06 -1.86 8.39E-06 -1.84 9.15E-01 -1.01 C6 5.29E-07 -7.35 8.08E-08 -11.00 8.95E-04 -2.38 5.82E-05 -3.08 3.40E-06 -4.61 4.50E-02 1.50 BAI3 1.46E-07 -8.19 4.65E-08 -10.48 6.12E-04 -2.34 1.11E-05 -3.50 1.99E-06 -4.48 1.65E-01 1.28 292

RNF24 5.21E-11 -5.41 4.98E-10 -3.93 8.01E-08 -2.33 3.80E-08 -2.32 4.58E-06 -1.69 3.62E-04 -1.38 KIAA1211 3.95E-08 -4.06 2.64E-08 -4.29 6.31E-06 -2.33 1.35E-04 -1.74 5.98E-05 -1.84 5.81E-01 1.06 CBLN1 7.54E-07 -5.01 1.10E-06 -4.72 2.86E-04 -2.32 3.22E-04 -2.16 6.06E-04 -2.03 6.97E-01 -1.06 HSD11B1 2.93E-06 -4.49 1.96E-06 -4.78 4.87E-04 -2.32 1.70E-03 -1.93 8.76E-04 -2.06 7.00E-01 1.07 PXDNL 1.74E-05 -5.35 2.40E-05 -5.05 5.34E-03 -2.24 1.93E-03 -2.39 3.07E-03 -2.26 7.92E-01 -1.06 DCLK1 4.58E-07 -5.64 6.07E-07 -5.38 5.20E-04 -2.22 7.69E-05 -2.54 1.21E-04 -2.42 7.59E-01 -1.05 GAS7 1.99E-09 -4.63 3.74E-09 -4.24 1.90E-06 -2.20 1.54E-06 -2.11 5.28E-06 -1.93 2.93E-01 -1.09 TBX18 2.94E-06 -5.25 2.87E-05 -3.68 1.75E-03 -2.19 4.29E-04 -2.40 1.35E-02 -1.68 6.84E-02 -1.43 MBNL3 3.99E-06 -4.69 3.90E-06 -4.71 1.44E-03 -2.17 8.45E-04 -2.16 8.15E-04 -2.17 9.83E-01 1.00 GAREM 1.46E-10 -4.03 3.53E-10 -3.61 9.32E-08 -2.13 2.52E-07 -1.89 1.68E-06 -1.69 8.15E-02 -1.12 BNC2 7.83E-07 -4.26 1.91E-07 -5.29 3.47E-04 -2.10 2.79E-04 -2.03 2.78E-05 -2.52 1.36E-01 1.24 IL32 1.17E-06 -4.23 1.03E-05 -3.16 5.77E-04 -2.06 3.13E-04 -2.06 1.05E-02 -1.54 6.15E-02 -1.34 DOCK9 5.31E-09 -4.02 2.86E-08 -3.27 5.15E-06 -2.03 3.43E-06 -1.98 9.17E-05 -1.61 2.58E-02 -1.23 LHFPL2 2.81E-10 -5.45 5.72E-10 -4.89 2.49E-06 -2.03 3.76E-08 -2.69 1.25E-07 -2.41 1.70E-01 -1.11 KLF2 5.08E-06 -4.56 4.21E-06 -4.70 2.96E-03 -2.02 5.80E-04 -2.26 4.40E-04 -2.33 8.65E-01 1.03

SCARF1 4.29E-06 -3.74 1.07E-05 -3.32 1.07E-03 -2.00 1.29E-03 -1.87 5.35E-03 -1.66 4.27E-01 -1.13

293

Appendix 5 - 3 Gene expression profile 23 (ratio is ratio of log2expression at different time points) Fold- Fold- Fold- Fold- Fold- Fold- p-value p-value p-value p-value p-value p-value external Change Change Change Change Change Change (D30 vs. (D45 vs. (D60 vs. (D30 vs. (D45 vs. (D30 vs. gene id (D30 vs. (D45 vs. (D60 vs. (D30 vs. (D45 vs. (D30 vs. D90) D90) D90) D60) D60) D45) D90) D90) D90) D60) D60) D45)

ABCC9 6.43E-08 -11.39 1.88E-07 -8.95 2.29E-06 -5.51 1.77E-03 -2.07 1.94E-02 -1.62 2.01E-01 -1.27 AC011526.1 3.50E-06 -4.18 1.13E-05 -3.54 8.50E-05 -2.75 2.05E-02 -1.52 1.30E-01 -1.29 3.09E-01 -1.18 ADAMTS2 6.25E-07 -3.79 4.18E-06 -3.00 6.82E-06 -2.84 3.64E-02 -1.33 6.59E-01 -1.06 7.99E-02 -1.26 ADAMTS7 5.07E-07 -4.81 1.33E-06 -4.16 7.72E-06 -3.28 1.98E-02 -1.46 1.18E-01 -1.27 3.25E-01 -1.16 ANKHD1 1.70E-06 -2.69 9.77E-06 -2.28 1.19E-05 -2.24 9.55E-02 -1.20 8.66E-01 -1.02 1.27E-01 -1.18 APBA1 4.21E-10 -3.22 1.08E-09 -2.92 3.53E-08 -2.16 1.14E-05 -1.49 1.40E-04 -1.35 9.15E-02 -1.10 APOLD1 1.57E-08 -5.12 1.07E-08 -5.44 1.09E-06 -2.94 2.40E-04 -1.74 1.02E-04 -1.85 5.77E-01 1.06 ARHGAP23 7.37E-09 -5.49 1.39E-08 -4.97 8.08E-08 -3.87 5.36E-03 -1.42 3.08E-02 -1.28 3.47E-01 -1.10 ARHGAP28 5.75E-10 -5.09 2.51E-09 -4.12 3.00E-07 -2.45 1.06E-06 -2.08 2.66E-05 -1.68 1.85E-02 -1.23 ARHGEF28 6.52E-09 -9.66 1.45E-07 -5.39 7.91E-07 -4.15 5.41E-05 -2.32 7.61E-02 -1.30 1.07E-03 -1.79 AVPR1A 6.66E-08 -84.64 5.44E-07 -37.06 8.73E-07 -31.38 1.09E-02 -2.70 6.18E-01 -1.18 2.72E-02 -2.28 CHST2 7.69E-07 -3.73 3.20E-06 -3.12 8.47E-05 -2.22 1.39E-03 -1.68 1.76E-02 -1.41 1.75E-01 -1.19 CILP 9.98E-11 -44.44 4.37E-10 -27.34 4.00E-09 -14.71 1.43E-05 -3.02 1.69E-03 -1.86 7.96E-03 -1.63 CLEC2B 1.05E-07 -4.08 2.42E-06 -2.80 1.87E-06 -2.87 7.63E-03 -1.42 8.02E-01 1.03 4.84E-03 -1.46 COL1A1 7.58E-08 -16.63 5.98E-07 -9.93 2.15E-06 -7.52 2.87E-03 -2.21 2.09E-01 -1.32 3.06E-02 -1.68 COL5A1 2.79E-05 -3.01 2.35E-05 -3.07 6.52E-04 -2.13 4.09E-02 -1.41 3.17E-02 -1.44 8.89E-01 1.02 CRIM1 1.69E-07 -4.41 5.22E-06 -2.86 8.36E-06 -2.71 1.81E-03 -1.62 6.78E-01 -1.05 3.79E-03 -1.54 CSPG4 1.23E-06 -3.61 1.56E-05 -2.68 2.35E-05 -2.57 1.97E-02 -1.41 7.40E-01 -1.04 3.61E-02 -1.35 CTDSP2 7.01E-08 -2.58 4.12E-07 -2.22 1.31E-06 -2.03 5.71E-03 -1.27 2.35E-01 -1.09 5.32E-02 -1.16 DAB2 1.36E-09 -4.51 6.44E-09 -3.67 6.45E-07 -2.30 2.47E-06 -1.96 7.19E-05 -1.60 2.08E-02 -1.23 DACH1 1.93E-07 -10.66 1.05E-06 -7.40 6.61E-06 -5.24 3.45E-03 -2.04 9.90E-02 -1.41 8.30E-02 -1.44 DOCK6 8.87E-08 -6.74 1.81E-07 -5.92 7.97E-06 -3.36 5.14E-04 -2.00 2.26E-03 -1.76 3.90E-01 -1.14 DTNBP1 4.63E-07 -3.59 2.21E-06 -2.98 6.77E-06 -2.64 2.03E-02 -1.36 3.12E-01 -1.13 1.27E-01 -1.21 294

ECM2 1.43E-06 -11.99 3.08E-06 -9.95 2.19E-04 -4.19 1.30E-03 -2.86 4.79E-03 -2.38 4.64E-01 -1.21 EFCAB14 6.97E-11 -4.54 2.69E-10 -3.80 1.89E-09 -3.04 2.57E-05 -1.49 2.80E-03 -1.25 1.06E-02 -1.20 EFNA5 2.12E-08 -3.06 2.33E-07 -2.43 2.18E-06 -2.03 1.63E-04 -1.51 3.35E-02 -1.19 9.05E-03 -1.26 EIF4EBP2 3.22E-10 -2.48 1.66E-09 -2.18 3.95E-09 -2.05 6.18E-04 -1.21 1.57E-01 -1.06 8.29E-03 -1.14 ELK3 1.68E-07 -3.67 8.83E-07 -3.01 2.11E-05 -2.21 4.82E-04 -1.66 1.23E-02 -1.36 8.45E-02 -1.22 ENPEP 1.52E-08 -4.82 1.14E-07 -3.65 5.18E-07 -3.05 8.04E-04 -1.58 1.01E-01 -1.20 1.80E-02 -1.32 EPAS1 3.60E-09 -7.40 9.59E-09 -6.20 1.01E-06 -3.19 1.06E-05 -2.32 8.91E-05 -1.94 1.38E-01 -1.19 F2RL2 1.40E-08 -8.29 4.92E-08 -6.52 2.78E-06 -3.52 4.79E-05 -2.36 7.14E-04 -1.85 9.92E-02 -1.27 FAM101B 2.69E-07 -3.24 7.79E-06 -2.30 2.41E-05 -2.09 1.01E-03 -1.55 3.52E-01 -1.10 5.34E-03 -1.41 FAM110B 5.03E-10 -5.50 9.65E-09 -3.64 5.40E-07 -2.40 4.19E-07 -2.29 2.47E-04 -1.52 2.66E-04 -1.51 FAM13C 2.17E-07 -5.39 6.46E-07 -4.53 2.32E-05 -2.83 6.98E-04 -1.90 5.86E-03 -1.60 2.37E-01 -1.19 FAM89A 7.61E-07 -4.46 1.63E-06 -3.99 8.04E-06 -3.23 4.09E-02 -1.38 1.56E-01 -1.24 4.44E-01 -1.12 FGD5 4.67E-10 -6.84 4.01E-10 -7.03 2.71E-08 -3.71 2.39E-05 -1.84 1.60E-05 -1.89 7.60E-01 1.03 FMNL3 1.72E-10 -5.96 2.03E-09 -4.13 7.18E-08 -2.74 4.86E-07 -2.17 1.90E-04 -1.50 4.42E-04 -1.44 GLIS2 1.58E-06 -3.28 9.38E-06 -2.68 3.32E-05 -2.36 1.87E-02 -1.39 3.09E-01 -1.14 1.20E-01 -1.22 GPR153 3.12E-08 -5.82 3.77E-08 -5.64 4.76E-06 -2.91 1.22E-04 -2.00 1.81E-04 -1.94 7.96E-01 -1.03 GPRC5B 4.34E-08 -4.07 5.58E-08 -3.94 3.84E-06 -2.46 3.32E-04 -1.66 5.69E-04 -1.60 7.40E-01 -1.03 HECA 7.97E-10 -3.75 7.49E-09 -2.92 2.64E-07 -2.14 2.47E-06 -1.75 4.74E-04 -1.36 2.26E-03 -1.29 HEG1 1.69E-08 -5.28 6.69E-08 -4.30 2.11E-06 -2.82 1.04E-04 -1.87 2.23E-03 -1.53 7.98E-02 -1.23 HNRNPA3 5.38E-10 -3.13 1.25E-08 -2.34 3.70E-09 -2.59 4.42E-03 -1.21 7.52E-02 1.11 1.77E-04 -1.34 HSPE1P26 3.76E-05 -3.05 3.60E-05 -3.06 1.53E-03 -2.02 2.32E-02 -1.51 2.19E-02 -1.52 9.74E-01 1.01 INF2 2.35E-07 -2.77 2.22E-07 -2.79 3.70E-06 -2.17 1.37E-02 -1.28 1.20E-02 -1.29 9.44E-01 1.01 ITGA8 9.27E-08 -3.93 1.36E-06 -2.86 7.78E-07 -3.03 2.93E-02 -1.30 5.71E-01 1.06 1.03E-02 -1.38 ITGA9 6.21E-07 -3.45 2.15E-05 -2.35 5.45E-06 -2.70 5.16E-02 -1.28 2.54E-01 1.14 6.09E-03 -1.46 ITPRIPL2 1.12E-06 -3.42 1.01E-05 -2.66 1.44E-04 -2.07 1.40E-03 -1.65 5.73E-02 -1.29 5.83E-02 -1.28 KCNJ8 9.37E-07 -5.94 7.27E-06 -4.24 2.07E-05 -3.63 1.42E-02 -1.64 3.84E-01 -1.17 7.03E-02 -1.40 KIAA1462 1.84E-08 -5.36 2.02E-08 -5.28 4.95E-07 -3.39 1.39E-03 -1.58 1.76E-03 -1.56 8.93E-01 -1.02 295

KIAA0922 5.93E-06 -2.77 1.12E-04 -2.09 1.46E-04 -2.04 2.41E-02 -1.36 8.47E-01 -1.02 3.43E-02 -1.33 KIAA2018 2.38E-08 -2.76 5.34E-07 -2.11 3.22E-07 -2.20 6.24E-03 -1.25 5.78E-01 1.04 2.30E-03 -1.30 LMBR1L 8.20E-07 -2.71 1.31E-06 -2.59 1.05E-05 -2.15 3.10E-02 -1.26 7.32E-02 -1.20 6.33E-01 -1.05 LRRC32 8.40E-08 -4.60 1.57E-06 -3.13 5.15E-06 -2.75 8.73E-04 -1.67 2.73E-01 -1.14 6.29E-03 -1.47 MAML2 1.19E-07 -2.99 1.39E-06 -2.36 2.62E-06 -2.24 5.58E-03 -1.34 5.38E-01 -1.06 1.73E-02 -1.27 MEGF9 1.54E-08 -3.41 1.68E-08 -3.38 1.48E-06 -2.19 1.45E-04 -1.56 1.75E-04 -1.54 9.03E-01 -1.01 MFHAS1 5.36E-08 -2.83 5.81E-08 -2.81 7.49E-07 -2.23 8.67E-03 -1.27 1.05E-02 -1.26 9.16E-01 -1.01 MICALL1 1.81E-08 -3.27 1.06E-08 -3.48 1.20E-06 -2.20 2.88E-04 -1.49 8.63E-05 -1.59 4.36E-01 1.06 MXRA5 8.02E-10 -9.81 1.91E-08 -5.44 2.59E-07 -3.73 2.58E-06 -2.63 5.53E-03 -1.46 2.26E-04 -1.80 NOTCH1 2.86E-07 -8.49 2.86E-06 -5.46 2.72E-06 -5.51 3.51E-02 -1.54 9.61E-01 1.01 3.21E-02 -1.56 OLFML2A 1.94E-05 -3.08 1.11E-04 -2.52 5.66E-04 -2.13 2.83E-02 -1.45 2.75E-01 -1.18 1.96E-01 -1.22 PCDHGA10 4.51E-09 -3.54 6.47E-08 -2.66 5.83E-07 -2.20 3.72E-05 -1.61 2.23E-02 -1.21 2.20E-03 -1.33 PCDHGB6 6.34E-07 -2.50 7.88E-06 -2.03 8.27E-06 -2.02 2.75E-02 -1.24 9.67E-01 -1.00 2.97E-02 -1.23 PCDHGC3 2.51E-08 -5.48 3.97E-08 -5.09 4.04E-07 -3.66 4.48E-03 -1.50 1.42E-02 -1.39 5.30E-01 -1.08 PDGFRB 1.88E-06 -4.43 1.60E-05 -3.28 1.09E-04 -2.61 4.99E-03 -1.70 1.57E-01 -1.26 7.36E-02 -1.35 PLAGL1 4.60E-08 -6.56 1.93E-07 -5.14 4.79E-06 -3.27 2.72E-04 -2.01 5.83E-03 -1.57 9.26E-02 -1.28 PXDN 4.39E-08 -4.84 6.13E-08 -4.61 6.26E-06 -2.62 1.69E-04 -1.85 3.38E-04 -1.76 6.61E-01 -1.05 RAPH1 1.43E-09 -6.01 3.97E-09 -5.10 5.25E-08 -3.57 1.36E-04 -1.68 2.47E-03 -1.43 9.80E-02 -1.18 ROBO4 3.16E-08 -15.73 1.12E-07 -11.44 1.14E-06 -6.94 1.13E-03 -2.27 2.19E-02 -1.65 1.18E-01 -1.37 RP1-79C4.4 3.20E-06 -7.94 9.17E-06 -6.40 6.04E-05 -4.52 2.81E-02 -1.76 1.47E-01 -1.41 3.53E-01 -1.24 RP11-238F2.1 5.32E-07 -38.03 1.00E-06 -30.38 8.07E-05 -8.48 7.58E-04 -4.49 2.40E-03 -3.59 5.05E-01 -1.25 RP11-458N5.1 1.40E-05 -3.51 3.19E-04 -2.40 2.31E-04 -2.49 5.18E-02 -1.41 8.26E-01 1.04 3.48E-02 -1.46 RP11-483C6.1 4.57E-06 -4.69 4.99E-06 -4.62 3.48E-04 -2.58 5.24E-03 -1.82 6.07E-03 -1.79 9.36E-01 -1.01 RP11-488L18.10 1.33E-06 -4.20 1.96E-05 -2.95 5.99E-04 -2.06 3.72E-04 -2.04 2.69E-02 -1.43 2.96E-02 -1.42 RP5-1085F17.3 4.65E-06 -3.24 1.47E-05 -2.83 3.58E-04 -2.05 5.21E-03 -1.58 3.26E-02 -1.38 3.26E-01 -1.14 RUNX1 1.37E-07 -4.79 1.15E-07 -4.92 4.39E-06 -3.02 3.22E-03 -1.58 2.21E-03 -1.63 8.31E-01 1.03 SASH1 1.75E-08 -4.42 1.42E-07 -3.37 1.08E-05 -2.19 1.43E-05 -2.02 8.83E-04 -1.54 1.58E-02 -1.31 296

SCAMP4 1.11E-07 -3.74 2.87E-07 -3.33 1.02E-06 -2.88 2.65E-02 -1.30 1.87E-01 -1.15 2.73E-01 -1.12 SEPP1 1.08E-07 -5.37 1.50E-06 -3.65 5.65E-05 -2.41 6.43E-05 -2.23 8.00E-03 -1.52 1.21E-02 -1.47 SF3A2 4.68E-06 -3.01 3.89E-05 -2.40 4.31E-05 -2.38 8.30E-02 -1.26 9.37E-01 -1.01 9.51E-02 -1.25 SGK223 1.12E-06 -7.73 1.06E-05 -5.05 2.07E-05 -4.51 1.97E-02 -1.71 5.79E-01 -1.12 5.42E-02 -1.53 SH2B3 1.86E-09 -3.88 2.30E-09 -3.77 2.85E-07 -2.31 1.44E-05 -1.68 2.35E-05 -1.63 7.11E-01 -1.03 SLC43A3 8.92E-08 -5.28 4.72E-07 -4.11 2.60E-05 -2.54 1.11E-04 -2.08 2.73E-03 -1.62 7.01E-02 -1.29 SLC7A1 1.74E-05 -3.55 8.74E-05 -2.88 3.12E-04 -2.48 4.95E-02 -1.43 3.77E-01 -1.16 2.25E-01 -1.23 SORBS3 1.21E-06 -3.21 3.15E-06 -2.88 1.83E-05 -2.41 2.76E-02 -1.33 1.41E-01 -1.20 3.61E-01 -1.11 SPON1 4.46E-06 -9.28 3.99E-06 -9.52 7.16E-05 -5.19 3.88E-02 -1.79 3.23E-02 -1.83 9.20E-01 1.03 STAB1 1.71E-05 -3.82 4.59E-05 -3.32 1.53E-03 -2.17 7.34E-03 -1.76 3.09E-02 -1.53 4.37E-01 -1.15 SULT1C4 6.92E-07 -3.09 1.35E-05 -2.29 4.03E-05 -2.09 3.06E-03 -1.48 3.84E-01 -1.10 1.52E-02 -1.35 TDO2 4.94E-06 -6.34 4.22E-06 -6.54 4.53E-05 -4.28 8.45E-02 -1.48 6.57E-02 -1.53 8.86E-01 1.03 TENC1 3.06E-07 -5.12 4.44E-07 -4.83 9.32E-06 -3.17 5.27E-03 -1.62 1.12E-02 -1.52 6.79E-01 -1.06 TMTC2 1.83E-10 -2.69 1.75E-09 -2.23 7.88E-09 -2.01 2.04E-05 -1.34 2.69E-02 -1.11 8.56E-04 -1.21 TP53I11 1.05E-08 -5.49 4.46E-08 -4.40 1.23E-06 -2.92 8.20E-05 -1.88 2.25E-03 -1.51 5.92E-02 -1.25 TRAM2 1.21E-10 -4.01 1.74E-10 -3.83 8.45E-09 -2.54 5.61E-06 -1.58 1.53E-05 -1.51 4.28E-01 -1.05 TSPAN14 6.39E-10 -3.47 1.01E-09 -3.30 1.38E-08 -2.54 2.38E-04 -1.37 9.89E-04 -1.30 3.90E-01 -1.05 UNC5C 4.27E-08 -4.42 2.10E-07 -3.57 8.67E-06 -2.39 1.04E-04 -1.85 2.70E-03 -1.49 6.54E-02 -1.24 UNKL 3.56E-05 -2.72 2.95E-05 -2.78 3.58E-04 -2.15 1.15E-01 -1.27 8.91E-02 -1.30 8.83E-01 1.02 USP31 5.65E-09 -3.27 1.90E-07 -2.33 2.54E-07 -2.27 2.50E-04 -1.44 7.36E-01 -1.02 4.28E-04 -1.40 VSTM4 1.14E-06 -2.88 2.70E-06 -2.63 3.82E-05 -2.07 8.43E-03 -1.39 3.97E-02 -1.27 4.03E-01 -1.09 VWF 1.12E-07 -11.85 2.45E-07 -9.88 2.25E-06 -6.26 6.33E-03 -1.89 3.51E-02 -1.58 3.58E-01 -1.20 WASF2 1.25E-10 -3.27 1.41E-10 -3.23 6.13E-09 -2.28 1.13E-05 -1.44 1.62E-05 -1.42 7.82E-01 -1.01 ZC3HAV1L 1.50E-07 -3.47 3.73E-07 -3.12 1.87E-06 -2.63 1.70E-02 -1.32 1.12E-01 -1.19 3.03E-01 -1.11 ZNF469 1.27E-07 -14.81 4.82E-06 -6.49 2.48E-06 -7.41 7.10E-03 -2.00 5.39E-01 1.14 2.34E-03 -2.28 ZNF697 5.14E-09 -4.36 8.89E-08 -3.07 5.20E-06 -2.11 3.18E-06 -2.07 9.71E-04 -1.46 1.57E-03 -1.42 ZNF704 9.93E-09 -3.63 5.02E-08 -3.02 2.01E-07 -2.63 1.69E-03 -1.38 1.05E-01 -1.15 3.81E-02 -1.20 297

Appendix 5 - 4 Gene expression profile 24 (ratio is ratio of log2expression at different time points)

Fold- Fold- Fold- Fold- Fold-Change p-value Fold-Change external p-value (D30 p-value (D45 Change p-value (D60 Change p-value (D30 Change Change p-value (D30 (D30 vs. (D45 vs. (D30 vs. gene id vs. D90) vs. D90) (D45 vs. vs. D90) (D60 vs. vs. D60) (D30 vs. (D45 vs. vs. D45) D90) D60) D45) D90) D90) D60) D60) ABCA6 2.81E-08 -5.60 1.80E-09 -9.35 2.00E-08 -5.93 6.34E-01 1.06 2.36E-03 -1.58 1.03E-03 1.67 ADNP 1.26E-08 -2.38 2.46E-08 -2.26 3.97E-08 -2.17 1.23E-01 -1.09 5.12E-01 -1.04 3.42E-01 -1.06 AGO2 9.32E-08 -3.16 8.40E-08 -3.20 5.46E-07 -2.63 5.85E-02 -1.20 4.63E-02 -1.22 8.95E-01 1.01 ARHGAP17 2.40E-06 -2.72 1.26E-06 -2.91 2.78E-06 -2.68 8.90E-01 -1.01 4.46E-01 -1.09 5.30E-01 1.07 ARHGAP32 3.47E-07 -2.28 2.25E-07 -2.37 2.35E-07 -2.36 6.56E-01 1.03 9.60E-01 -1.00 6.20E-01 1.04 ARHGAP6 4.53E-08 -2.82 1.25E-08 -3.23 2.44E-08 -3.01 4.04E-01 1.07 3.45E-01 -1.07 9.05E-02 1.15 ARHGDIA 6.15E-06 -2.49 3.82E-06 -2.61 3.14E-05 -2.15 1.89E-01 -1.16 9.21E-02 -1.21 6.66E-01 1.05 ARHGEF15 3.20E-05 -5.39 2.18E-05 -5.80 2.07E-04 -3.89 1.85E-01 -1.39 1.12E-01 -1.49 7.60E-01 1.08 ARHGEF17 1.14E-05 -2.45 3.53E-06 -2.76 3.22E-05 -2.23 4.01E-01 -1.10 7.86E-02 -1.24 3.09E-01 1.12 ARID1B 2.80E-10 -2.56 3.94E-11 -3.08 3.74E-10 -2.49 5.53E-01 -1.03 3.02E-04 -1.24 8.02E-04 1.21 ASXL2 4.64E-08 -2.15 4.21E-08 -2.17 2.62E-08 -2.25 4.40E-01 1.04 5.18E-01 1.04 8.97E-01 1.01 ATP13A1 6.04E-06 -2.16 8.24E-07 -2.57 3.26E-06 -2.28 5.74E-01 1.05 1.91E-01 -1.13 7.43E-02 1.19 BCL2 1.23E-07 -5.64 4.61E-07 -4.57 6.05E-07 -4.39 8.83E-02 -1.29 7.68E-01 -1.04 1.45E-01 -1.23 BMP8A 1.40E-07 -7.67 2.38E-05 -3.33 3.88E-07 -6.32 2.52E-01 -1.21 2.10E-03 1.90 2.92E-04 -2.30 BRD3 5.37E-07 -2.34 7.40E-08 -2.81 4.04E-07 -2.40 7.56E-01 1.02 6.25E-02 -1.17 3.58E-02 1.20 C15orf39 1.23E-06 -4.11 1.85E-07 -5.51 9.93E-07 -4.24 8.28E-01 1.03 8.37E-02 -1.30 5.72E-02 1.34 C18orf54 9.91E-06 -3.46 9.05E-06 -3.50 1.29E-05 -3.34 8.21E-01 -1.04 7.61E-01 -1.05 9.37E-01 1.01 CAMK4 1.72E-06 -3.25 8.31E-06 -2.72 1.09E-05 -2.64 1.11E-01 -1.23 8.13E-01 -1.03 1.64E-01 -1.19 CASC7 7.76E-09 -3.83 2.92E-08 -3.26 2.99E-08 -3.26 6.79E-02 -1.18 9.76E-01 -1.00 7.16E-02 -1.17 CASP2 1.41E-07 -2.46 3.35E-07 -2.29 7.98E-08 -2.59 4.86E-01 1.05 1.06E-01 1.13 3.22E-01 -1.08 CBL 4.56E-10 -2.98 1.61E-09 -2.64 1.90E-09 -2.60 1.93E-02 -1.15 7.72E-01 -1.01 3.28E-02 -1.13

298

CCBE1 2.77E-06 -3.87 7.68E-07 -4.67 1.70E-06 -4.15 6.38E-01 1.07 4.25E-01 -1.13 2.16E-01 1.21 CD36 8.24E-09 -36.43 6.70E-09 -39.12 1.16E-08 -32.44 6.00E-01 -1.12 4.03E-01 -1.21 7.47E-01 1.07 CDC25B 8.47E-08 -2.82 3.01E-08 -3.14 1.81E-07 -2.62 3.61E-01 -1.08 3.82E-02 -1.20 1.89E-01 1.11 CENPP 1.51E-05 -2.34 2.01E-05 -2.28 3.79E-05 -2.16 4.64E-01 -1.09 6.16E-01 -1.06 8.12E-01 -1.03 CIC 4.01E-06 -3.85 1.12E-06 -4.64 3.02E-06 -4.00 7.93E-01 1.04 3.40E-01 -1.16 2.32E-01 1.21 CNOT6 1.52E-07 -2.18 4.89E-07 -2.00 8.95E-08 -2.27 5.16E-01 1.04 6.60E-02 1.13 1.98E-01 -1.09 COL5A3 1.97E-05 -3.34 5.97E-06 -3.93 2.46E-05 -3.24 8.56E-01 -1.03 2.44E-01 -1.21 3.19E-01 1.18 COL6A3 6.44E-06 -2.92 7.36E-06 -2.88 1.57E-05 -2.65 4.50E-01 -1.10 5.21E-01 -1.09 9.07E-01 -1.01 CRAMP1L 1.32E-06 -2.58 5.79E-07 -2.80 4.96E-07 -2.85 3.17E-01 1.10 8.67E-01 1.02 4.00E-01 1.08 CREB3L1 7.69E-08 -7.65 3.27E-07 -5.85 1.51E-07 -6.73 4.11E-01 -1.14 3.75E-01 1.15 1.03E-01 -1.31 CREB3L2 4.83E-11 -4.04 2.44E-11 -4.43 8.08E-11 -3.79 2.37E-01 -1.07 1.19E-02 -1.17 1.07E-01 1.10 CRTC3 2.51E-09 -2.19 3.74E-10 -2.55 4.04E-09 -2.12 4.30E-01 -1.03 9.60E-04 -1.21 3.87E-03 1.17 CTD-3099C6.9 2.87E-06 -2.48 2.33E-04 -1.75 4.28E-06 -2.39 7.10E-01 -1.04 7.72E-03 1.37 3.93E-03 -1.42 DIAPH1 1.11E-07 -2.25 4.37E-08 -2.43 8.93E-08 -2.29 7.89E-01 1.02 3.63E-01 -1.06 2.47E-01 1.08 DYRK1B 4.26E-06 -2.52 1.57E-06 -2.78 1.88E-06 -2.73 4.44E-01 1.08 8.61E-01 -1.02 3.52E-01 1.10 EBF1 1.60E-07 -11.65 5.81E-07 -8.69 9.04E-07 -7.91 7.35E-02 -1.47 6.42E-01 -1.10 1.62E-01 -1.34 EGFR 2.76E-06 -2.70 1.70E-06 -2.84 1.63E-06 -2.86 6.12E-01 1.06 9.66E-01 1.00 6.42E-01 1.05 EHD2 8.55E-08 -4.14 1.78E-07 -3.75 1.73E-07 -3.76 3.95E-01 -1.10 9.73E-01 1.00 3.77E-01 -1.10 ELMO1 1.47E-07 -3.38 1.91E-07 -3.28 9.74E-07 -2.74 5.37E-02 -1.23 9.20E-02 -1.19 7.59E-01 -1.03 ENG 5.41E-06 -4.20 8.97E-06 -3.90 4.91E-06 -4.26 9.30E-01 1.01 5.95E-01 1.09 6.56E-01 -1.08 EPN1 2.73E-05 -2.16 2.15E-05 -2.20 2.78E-05 -2.15 9.87E-01 -1.00 8.36E-01 -1.02 8.49E-01 1.02 FAM122C 1.27E-05 -2.42 4.66E-06 -2.68 1.47E-05 -2.39 9.03E-01 -1.01 3.27E-01 -1.12 3.87E-01 1.10 FAM174B 4.36E-06 -3.15 2.13E-06 -3.43 7.10E-06 -2.97 6.63E-01 -1.06 2.79E-01 -1.16 5.04E-01 1.09 FAM193A 9.18E-07 -2.22 4.45E-07 -2.35 1.39E-06 -2.14 6.72E-01 -1.03 2.46E-01 -1.10 4.46E-01 1.06 FAM214A 5.07E-07 -2.16 1.09E-06 -2.04 1.11E-06 -2.04 4.15E-01 -1.06 9.87E-01 -1.00 4.24E-01 -1.06 FBXL7 1.72E-08 -3.63 7.23E-08 -3.08 2.21E-08 -3.52 7.20E-01 -1.03 1.29E-01 1.15 7.00E-02 -1.18 FTX 1.58E-05 -3.07 5.08E-05 -2.68 6.53E-05 -2.61 2.77E-01 -1.18 8.50E-01 -1.03 3.62E-01 -1.15 299

GPC4 6.19E-08 -9.06 1.89E-08 -11.81 2.03E-07 -7.13 1.62E-01 -1.27 9.07E-03 -1.66 1.25E-01 1.30 GSK3B 2.37E-09 -2.02 1.16E-07 -1.63 1.54E-09 -2.08 4.49E-01 1.03 3.83E-05 1.28 1.16E-04 -1.24 HEBP1 1.25E-05 -2.11 1.01E-06 -2.64 7.21E-06 -2.21 6.36E-01 1.05 8.10E-02 -1.19 3.47E-02 1.25 HIPK2 2.49E-09 -3.98 5.74E-10 -4.88 1.50E-09 -4.26 3.79E-01 1.07 9.30E-02 -1.15 1.87E-02 1.23 HMBOX1 6.58E-06 -2.22 9.70E-05 -1.81 9.41E-06 -2.16 7.54E-01 -1.03 8.46E-02 1.19 4.87E-02 -1.23 HNRNPD 1.15E-07 -2.46 7.50E-08 -2.56 1.54E-07 -2.40 7.24E-01 -1.03 3.80E-01 -1.07 5.92E-01 1.04 HSPG2 1.75E-05 -6.46 1.54E-05 -6.64 1.39E-05 -6.77 8.48E-01 1.05 9.34E-01 1.02 9.14E-01 1.03 HUWE1 1.15E-07 -1.94 8.67E-09 -2.34 7.05E-08 -2.01 5.41E-01 1.03 1.16E-02 -1.17 3.81E-03 1.21 INO80D 1.76E-08 -2.43 1.37E-06 -1.79 7.65E-08 -2.16 6.50E-02 -1.12 6.52E-03 1.21 2.18E-04 -1.36 INSR 2.32E-07 -2.38 2.50E-07 -2.36 8.16E-07 -2.15 1.83E-01 -1.11 2.09E-01 -1.10 9.32E-01 -1.01 JMY 3.70E-08 -2.36 7.29E-09 -2.72 7.24E-08 -2.23 3.84E-01 -1.06 6.69E-03 -1.22 3.37E-02 1.15 KCNMB2 8.51E-06 -3.21 5.21E-06 -3.41 1.94E-05 -2.91 4.93E-01 -1.10 2.72E-01 -1.17 6.64E-01 1.06 KCNQ3 9.10E-06 -3.15 6.22E-06 -3.31 1.36E-05 -3.01 7.34E-01 -1.05 5.04E-01 -1.10 7.39E-01 1.05 KIAA0430 7.52E-08 -2.53 5.41E-08 -2.61 1.12E-07 -2.44 6.16E-01 -1.04 3.61E-01 -1.07 6.71E-01 1.03 KMT2A 1.07E-09 -2.82 1.23E-09 -2.78 1.06E-09 -2.82 9.81E-01 1.00 7.90E-01 1.01 8.08E-01 -1.01 KMT2C 5.34E-09 -2.59 5.20E-08 -2.15 1.37E-08 -2.39 1.66E-01 -1.08 8.22E-02 1.11 5.97E-03 -1.20 LAMC3 2.85E-07 -4.43 7.91E-08 -5.40 1.21E-07 -5.05 3.20E-01 1.14 6.00E-01 -1.07 1.42E-01 1.22 LARP1 5.71E-08 -2.74 7.33E-09 -3.41 3.25E-08 -2.90 4.55E-01 1.06 4.72E-02 -1.18 1.19E-02 1.24 LATS1 1.17E-09 -2.52 8.48E-09 -2.15 6.54E-09 -2.20 1.20E-02 -1.15 6.86E-01 1.02 5.75E-03 -1.17 MAML1 2.20E-08 -2.33 1.81E-09 -2.93 1.43E-08 -2.42 5.35E-01 1.04 5.71E-03 -1.21 1.88E-03 1.25 MAP1B 1.14E-06 -2.22 1.13E-06 -2.23 2.36E-06 -2.10 4.79E-01 -1.06 4.69E-01 -1.06 9.86E-01 1.00 MAPK7 8.35E-05 -2.16 5.76E-05 -2.23 3.58E-05 -2.33 5.24E-01 1.08 7.15E-01 1.05 7.82E-01 1.03 MAST4 3.39E-09 -2.88 1.62E-08 -2.49 1.22E-08 -2.55 6.33E-02 -1.13 6.72E-01 1.03 2.95E-02 -1.16 MBTPS2 6.00E-08 -2.11 6.05E-08 -2.11 7.46E-08 -2.08 7.79E-01 -1.02 7.87E-01 -1.02 9.91E-01 -1.00 MEX3A 5.08E-08 -2.45 5.81E-08 -2.42 4.14E-07 -2.08 2.47E-02 -1.18 3.41E-02 -1.17 8.60E-01 -1.01 MGRN1 1.17E-05 -3.51 2.20E-06 -4.46 4.70E-06 -3.98 4.28E-01 1.14 4.81E-01 -1.12 1.49E-01 1.27 MIR3682 1.33E-05 -3.68 1.68E-05 -3.56 1.58E-05 -3.60 8.85E-01 -1.02 9.57E-01 1.01 8.42E-01 -1.03 300

MIR4435-1HG 2.90E-07 -2.09 1.00E-05 -1.67 1.68E-07 -2.18 5.27E-01 1.04 1.37E-03 1.30 4.20E-03 -1.25 MYO10 3.28E-09 -5.42 4.85E-09 -5.10 2.47E-08 -4.04 8.83E-03 -1.34 2.88E-02 -1.26 5.22E-01 -1.06 NDST1 1.68E-07 -3.52 1.85E-07 -3.48 1.10E-06 -2.84 5.72E-02 -1.24 7.00E-02 -1.22 9.09E-01 -1.01 NFAT5 2.26E-09 -2.99 2.43E-08 -2.40 6.96E-09 -2.68 8.46E-02 -1.12 8.32E-02 1.12 2.94E-03 -1.25 NOTCH2NL 3.51E-07 -2.53 2.94E-07 -2.58 4.90E-07 -2.46 7.12E-01 -1.03 5.71E-01 -1.05 8.41E-01 1.02 NOTCH3 1.40E-06 -5.14 1.22E-05 -3.70 1.19E-06 -5.28 8.70E-01 1.03 4.91E-02 1.43 6.58E-02 -1.39 NOTCH4 7.59E-09 -7.63 6.07E-09 -7.97 8.80E-08 -4.99 4.79E-03 -1.53 2.56E-03 -1.60 7.26E-01 1.04 NOVA1 8.17E-06 -2.08 6.87E-07 -2.58 4.85E-06 -2.17 6.44E-01 1.04 7.51E-02 -1.19 3.28E-02 1.24 NR2C2 7.83E-10 -2.20 2.00E-09 -2.06 8.04E-10 -2.20 9.59E-01 -1.00 1.17E-01 1.07 1.07E-01 -1.07 OAS3 5.69E-06 -2.94 4.89E-07 -3.99 5.12E-06 -2.97 9.23E-01 1.01 3.58E-02 -1.34 3.01E-02 1.36 OGFRL1 2.05E-07 -3.04 1.40E-07 -3.17 5.59E-07 -2.74 2.71E-01 -1.11 1.34E-01 -1.16 6.55E-01 1.04 OLFML1 9.37E-06 -3.16 1.90E-06 -3.89 1.20E-05 -3.07 8.31E-01 -1.03 1.13E-01 -1.27 1.61E-01 1.23 OLFML2B 3.73E-07 -7.42 1.36E-07 -9.15 5.78E-07 -6.82 6.33E-01 -1.09 1.17E-01 -1.34 2.53E-01 1.23 PABPC1P3 3.16E-05 -2.49 3.63E-06 -3.16 3.43E-05 -2.47 9.48E-01 -1.01 7.55E-02 -1.28 8.45E-02 1.27 PABPC1P4 1.35E-06 -2.46 1.26E-07 -3.13 6.68E-07 -2.63 4.72E-01 1.07 7.76E-02 -1.19 2.10E-02 1.27 PABPC3 3.56E-07 -2.76 2.37E-07 -2.88 2.11E-06 -2.34 8.34E-02 -1.18 3.67E-02 -1.23 6.45E-01 1.04 PAK2 8.71E-10 -2.21 5.76E-10 -2.28 1.40E-09 -2.14 3.87E-01 -1.04 1.15E-01 -1.07 4.34E-01 1.03 PCBP2 7.87E-09 -2.90 3.51E-09 -3.15 6.72E-09 -2.94 8.05E-01 1.02 3.05E-01 -1.07 2.11E-01 1.09 PCDHB13 2.24E-06 -3.00 8.44E-07 -3.36 9.55E-07 -3.31 4.02E-01 1.10 8.98E-01 -1.02 3.38E-01 1.12 PCDHB16 4.14E-07 -3.28 1.78E-07 -3.64 4.14E-07 -3.28 9.98E-01 -1.00 3.43E-01 -1.11 3.44E-01 1.11 PCDHB9 4.95E-06 -3.11 5.90E-06 -3.05 6.17E-06 -3.03 8.44E-01 -1.03 9.69E-01 -1.01 8.75E-01 -1.02 PCDHGB7 1.29E-05 -2.60 2.96E-05 -2.39 3.84E-05 -2.34 3.85E-01 -1.11 8.39E-01 -1.02 5.00E-01 -1.09 PCGF2 9.19E-05 -2.60 1.95E-05 -3.12 5.22E-05 -2.77 6.74E-01 1.07 4.43E-01 -1.12 2.45E-01 1.20 PDK4 2.74E-05 -3.28 1.34E-04 -2.69 1.22E-04 -2.72 2.72E-01 -1.20 9.44E-01 1.01 2.45E-01 -1.22 PEAK1 1.55E-09 -3.42 1.82E-08 -2.65 7.44E-09 -2.89 2.14E-02 -1.18 1.97E-01 1.09 1.90E-03 -1.29 PEAR1 3.00E-06 -3.52 3.25E-07 -4.83 1.37E-06 -3.91 4.51E-01 1.11 1.45E-01 -1.23 3.84E-02 1.37 PIP5K1C 1.00E-07 -4.07 6.68E-08 -4.31 1.30E-07 -3.93 7.48E-01 -1.04 4.08E-01 -1.10 6.06E-01 1.06 301

PLEKHH2 1.61E-07 -3.72 1.26E-07 -3.84 1.12E-06 -2.95 5.03E-02 -1.26 2.91E-02 -1.30 7.63E-01 1.03 PLXNA2 1.22E-07 -2.94 3.02E-08 -3.44 1.16E-07 -2.96 9.50E-01 1.01 9.96E-02 -1.16 8.94E-02 1.17 PREX2 8.73E-07 -6.91 2.13E-06 -5.85 1.86E-06 -5.99 4.51E-01 -1.15 8.96E-01 1.02 3.79E-01 -1.18 PRKD2 5.25E-07 -2.66 2.50E-07 -2.87 2.12E-06 -2.34 1.70E-01 -1.14 4.08E-02 -1.23 4.14E-01 1.08 PRRC2C 8.01E-09 -2.48 2.98E-08 -2.23 3.00E-08 -2.23 7.31E-02 -1.11 9.89E-01 -1.00 7.48E-02 -1.11 PTPN12 2.19E-09 -2.77 7.22E-09 -2.49 7.22E-09 -2.49 6.93E-02 -1.11 1.00E+00 1.00 6.93E-02 -1.11 RAB11B 1.28E-05 -2.31 1.22E-06 -2.91 6.33E-06 -2.46 5.45E-01 1.07 1.37E-01 -1.18 4.79E-02 1.26 RBM33 5.35E-07 -2.43 6.97E-06 -1.98 8.38E-07 -2.33 6.35E-01 -1.04 6.12E-02 1.18 2.59E-02 -1.23 RBMS2 1.44E-09 -3.06 9.96E-10 -3.19 5.80E-10 -3.38 1.08E-01 1.10 3.13E-01 1.06 5.01E-01 1.04 RFX7 9.33E-08 -2.85 3.01E-07 -2.54 5.71E-07 -2.40 5.38E-02 -1.19 4.84E-01 -1.06 1.79E-01 -1.12 RGMB 1.11E-08 -3.31 4.29E-09 -3.71 4.94E-08 -2.82 5.38E-02 -1.17 3.56E-03 -1.31 1.54E-01 1.12 RMI1 9.81E-08 -3.02 8.09E-08 -3.08 9.34E-07 -2.42 2.35E-02 -1.25 1.51E-02 -1.27 8.08E-01 1.02 RNF152 1.41E-07 -3.78 5.99E-07 -3.17 8.94E-07 -3.03 5.72E-02 -1.25 6.73E-01 -1.05 1.19E-01 -1.19 RP11-175O19.4 6.00E-07 -2.36 1.09E-05 -1.89 7.46E-07 -2.31 8.16E-01 -1.02 2.43E-02 1.22 1.59E-02 -1.25 RP11-214K3.21 2.40E-05 -3.94 1.86E-04 -2.96 1.25E-04 -3.12 2.26E-01 -1.26 7.76E-01 1.05 1.44E-01 -1.33 RP11-252A24.7 1.59E-05 -2.24 3.97E-06 -2.54 6.65E-06 -2.42 4.59E-01 1.08 6.43E-01 -1.05 2.39E-01 1.14 RP11-575G13.2 3.09E-05 -4.09 7.18E-06 -5.19 3.51E-05 -4.00 9.19E-01 -1.02 2.04E-01 -1.30 2.39E-01 1.27 RP11-690G19.3 1.35E-06 -2.52 7.23E-07 -2.68 4.10E-06 -2.28 2.97E-01 -1.10 1.07E-01 -1.17 5.23E-01 1.06 RP11-720L2.3 2.86E-05 -3.85 2.01E-05 -4.06 6.47E-05 -3.42 5.32E-01 -1.13 3.70E-01 -1.19 7.76E-01 1.05 RP11-730G20.2 2.20E-07 -2.63 7.72E-08 -2.91 3.34E-07 -2.53 6.32E-01 -1.04 9.94E-02 -1.15 2.18E-01 1.11 RP11-73M18.8 2.46E-05 -2.85 9.20E-05 -2.47 7.90E-05 -2.51 3.76E-01 -1.14 9.11E-01 1.02 3.22E-01 -1.16 RP11-91J3.1 2.47E-05 -5.24 3.35E-05 -4.97 6.68E-05 -4.40 4.47E-01 -1.19 5.99E-01 -1.13 8.09E-01 -1.06 RPS6KA2 5.09E-08 -2.80 7.77E-09 -3.43 9.33E-08 -2.64 4.41E-01 -1.06 4.48E-03 -1.30 1.86E-02 1.22 SCN7A 4.82E-07 -9.97 3.14E-08 -20.10 2.94E-07 -11.19 5.83E-01 1.12 1.52E-02 -1.80 5.54E-03 2.02 SCUBE2 1.64E-05 -2.40 7.91E-06 -2.57 2.02E-05 -2.35 8.63E-01 -1.02 4.35E-01 -1.09 5.39E-01 1.07 SEMA5A 1.29E-06 -2.29 3.70E-05 -1.79 1.01E-06 -2.34 8.02E-01 1.02 6.97E-03 1.31 1.10E-02 -1.28 SETBP1 8.19E-07 -2.23 1.06E-06 -2.19 7.73E-07 -2.24 9.51E-01 1.00 7.46E-01 1.03 7.92E-01 -1.02 302

SETD5 1.42E-06 -2.42 1.09E-06 -2.48 1.51E-06 -2.41 9.46E-01 -1.01 7.43E-01 -1.03 7.94E-01 1.02 SFRP2 6.16E-08 -5.11 1.97E-07 -4.29 7.76E-08 -4.93 7.65E-01 -1.04 2.68E-01 1.15 1.69E-01 -1.19 SH3D19 2.09E-06 -2.65 1.11E-05 -2.26 6.33E-06 -2.38 3.15E-01 -1.11 6.24E-01 1.05 1.48E-01 -1.17 SHANK3 1.92E-07 -6.09 9.43E-08 -6.93 1.80E-07 -6.16 9.42E-01 1.01 4.36E-01 -1.13 3.96E-01 1.14 SKI 7.31E-07 -4.34 1.40E-07 -5.63 3.54E-07 -4.84 4.37E-01 1.12 2.90E-01 -1.16 8.15E-02 1.30 SLC26A7 1.36E-06 -9.33 6.88E-06 -6.61 6.48E-06 -6.69 1.59E-01 -1.39 9.58E-01 1.01 1.46E-01 -1.41 SNX29 5.87E-08 -2.00 7.50E-09 -2.33 2.73E-08 -2.11 3.14E-01 1.05 7.69E-02 -1.10 1.19E-02 1.16 SRRM1 4.16E-06 -1.99 2.68E-05 -1.76 1.98E-06 -2.10 4.88E-01 1.06 3.84E-02 1.20 1.31E-01 -1.13 SYVN1 1.66E-08 -4.16 1.25E-08 -4.33 2.19E-08 -4.01 6.94E-01 -1.04 4.21E-01 -1.08 6.74E-01 1.04 TAF1C 9.15E-06 -2.41 5.07E-06 -2.55 1.59E-05 -2.29 6.40E-01 -1.05 3.32E-01 -1.11 6.04E-01 1.06 TATDN2 1.92E-09 -2.50 1.54E-09 -2.54 1.72E-08 -2.10 4.22E-03 -1.19 2.12E-03 -1.21 6.99E-01 1.02 TET3 6.27E-09 -3.65 4.79E-09 -3.77 4.03E-08 -2.96 1.77E-02 -1.23 8.02E-03 -1.28 6.65E-01 1.03 TGFB1 4.99E-05 -2.17 3.54E-06 -2.81 2.41E-05 -2.32 5.71E-01 1.07 1.17E-01 -1.21 4.33E-02 1.29 TIAL1 2.12E-08 -2.59 4.76E-08 -2.41 2.19E-08 -2.58 9.64E-01 -1.00 2.97E-01 1.07 2.78E-01 -1.07 TMEM131 8.52E-08 -2.17 4.34E-08 -2.28 6.05E-08 -2.22 6.61E-01 1.03 6.62E-01 -1.03 3.88E-01 1.05 TTYH3 1.11E-05 -3.10 7.88E-06 -3.23 6.38E-06 -3.32 6.32E-01 1.07 8.52E-01 1.03 7.69E-01 1.04 UBN2 1.34E-06 -2.22 2.37E-05 -1.81 1.05E-06 -2.27 8.08E-01 1.02 1.45E-02 1.26 2.26E-02 -1.23 ULK1 2.99E-06 -3.34 4.38E-07 -4.33 1.22E-06 -3.75 3.87E-01 1.12 2.92E-01 -1.15 7.01E-02 1.29 USP42 1.02E-08 -2.23 1.66E-09 -2.59 4.73E-09 -2.37 2.41E-01 1.06 9.63E-02 -1.09 1.09E-02 1.16 WDR26 1.48E-08 -2.35 8.86E-08 -2.06 4.19E-08 -2.17 1.63E-01 -1.08 3.39E-01 1.06 2.98E-02 -1.14 WNK1 1.07E-08 -2.69 4.98E-08 -2.35 2.01E-08 -2.54 3.64E-01 -1.06 2.27E-01 1.08 4.78E-02 -1.14 WWC3 5.70E-08 -2.75 5.66E-08 -2.75 3.32E-07 -2.34 5.07E-02 -1.17 4.98E-02 -1.17 9.92E-01 1.00 ZFP36L1 4.63E-08 -3.78 2.09E-08 -4.20 9.10E-08 -3.47 3.90E-01 -1.09 6.85E-02 -1.21 2.85E-01 1.11 ZFPM1 7.54E-07 -3.29 1.07E-06 -3.16 1.03E-06 -3.17 7.50E-01 -1.04 9.63E-01 1.01 7.15E-01 -1.04 ZNF154 3.15E-05 -2.36 1.79E-04 -2.02 4.67E-05 -2.28 7.60E-01 -1.04 3.34E-01 1.13 2.13E-01 -1.17 ZNF318 1.72E-07 -2.01 8.42E-08 -2.11 1.44E-07 -2.03 8.27E-01 1.01 5.12E-01 -1.04 3.87E-01 1.05

ZNF37BP 4.44E-06 -2.61 5.32E-05 -2.08 1.32E-05 -2.35 3.50E-01 -1.11 2.77E-01 1.13 5.76E-02 -1.25 303

Appendix 5 - 5 Gene expression profile 4 (ratio is ratio of log2expression at different time points)

Fold- Fold- Fold- Fold- Fold- Fold- p-value p-value p-value p-value p-value p-value Change Change Change Change Change Change external_gene_id (d30 vs (d45 vs (d60 vs (d30 vs (d45 vs (d30 vs (d30 vs (d45 vs (d60 vs (d30 vs (d45 vs (d30 vs d90) d90) d90) d60) d60) d45) d90) d90) d90) d60) d60) d45)

MMP3 7.95E-07 96.25 4.85E-05 19.29 5.96E-03 4.77 2.12E-05 20.17 7.35E-03 4.04 3.08E-03 4.99 HSPA6 3.13E-08 69.84 4.56E-08 60.08 1.35E-02 2.50 1.74E-07 27.95 2.79E-07 24.04 6.12E-01 1.16 SLC16A6 1.68E-06 50.71 1.88E-04 10.41 1.71E-02 3.31 2.54E-05 15.32 1.43E-02 3.15 2.05E-03 4.87 AIM1L 8.30E-08 43.04 1.02E-05 10.13 1.44E-04 5.59 1.61E-05 7.70 5.81E-02 1.81 3.18E-04 4.25 NLGN4Y 1.81E-09 36.83 1.42E-08 19.46 6.90E-06 4.99 3.80E-07 7.38 1.59E-05 3.90 5.82E-03 1.89 CLDN6 3.83E-08 32.62 1.14E-06 12.16 7.57E-04 3.33 1.29E-06 9.81 2.27E-04 3.66 1.80E-03 2.68 IL1RN 2.39E-06 31.82 5.75E-04 6.43 1.25E-02 3.19 5.27E-05 9.96 7.85E-02 2.01 1.00E-03 4.95 EREG 4.61E-11 31.19 2.06E-09 11.20 3.22E-07 4.43 8.71E-09 7.04 1.51E-05 2.53 5.99E-06 2.78 CTSF 2.10E-10 28.11 4.82E-09 12.05 1.75E-06 4.06 3.05E-08 6.92 9.66E-06 2.97 9.17E-05 2.33 KIRREL3 2.60E-07 27.93 1.17E-05 9.56 6.87E-04 4.06 2.45E-05 6.88 1.04E-02 2.36 2.66E-03 2.92 INSC 2.12E-06 26.96 7.34E-05 9.48 2.08E-02 2.69 2.96E-05 10.04 3.39E-03 3.53 1.05E-02 2.84 ANKRD66 2.05E-06 26.72 6.58E-04 5.52 3.43E-02 2.41 1.91E-05 11.08 3.17E-02 2.29 6.78E-04 4.84 HNRNPA1P45 8.74E-07 26.67 3.05E-04 5.63 3.91E-04 5.34 2.94E-04 4.99 8.68E-01 1.05 3.84E-04 4.74 HNF4A 1.72E-05 26.09 4.57E-04 9.25 1.84E-02 3.49 5.49E-04 7.47 4.01E-02 2.65 3.07E-02 2.82 CRHBP 8.60E-10 25.57 5.28E-08 8.98 7.68E-05 2.81 2.23E-08 9.09 1.36E-05 3.19 3.51E-05 2.85 ZNF667-AS1 9.28E-06 24.69 3.42E-04 8.35 4.54E-03 4.38 9.17E-04 5.63 1.22E-01 1.91 1.68E-02 2.96 ZFP28 3.70E-06 23.76 8.37E-05 9.54 7.04E-03 3.43 1.60E-04 6.93 1.29E-02 2.78 2.30E-02 2.49 PRKY 1.84E-06 23.51 1.54E-05 12.52 4.43E-03 3.44 8.90E-05 6.84 1.96E-03 3.64 7.50E-02 1.88 KRT42P 6.19E-06 23.51 4.30E-04 7.00 1.20E-02 3.25 2.03E-04 7.23 5.86E-02 2.15 6.66E-03 3.36 RP11-744K17.9 7.88E-06 23.39 3.91E-04 7.51 2.08E-02 2.96 1.70E-04 7.91 2.91E-02 2.54 1.10E-02 3.11 FRG1B 1.81E-06 22.51 4.32E-05 9.23 3.49E-03 3.53 1.08E-04 6.37 1.12E-02 2.61 1.65E-02 2.44 ZNF229 2.40E-08 21.36 2.90E-07 11.07 3.09E-04 3.10 1.22E-06 6.88 6.11E-05 3.57 7.89E-03 1.93 304

HECW1 2.81E-05 21.14 3.68E-04 9.55 6.86E-02 2.46 2.91E-04 8.59 7.31E-03 3.88 8.04E-02 2.22 ZNF506 2.95E-05 20.61 1.30E-04 12.86 1.62E-02 3.53 1.30E-03 5.83 9.48E-03 3.64 2.77E-01 1.60 GRK1 7.42E-06 20.59 1.29E-03 5.17 6.05E-03 3.66 4.98E-04 5.62 3.53E-01 1.41 2.49E-03 3.98 KRBOX1 3.80E-06 20.37 7.47E-04 5.17 1.77E-02 2.70 7.35E-05 7.56 7.38E-02 1.92 1.59E-03 3.94 COL26A1 2.49E-08 20.25 7.78E-07 8.58 2.54E-05 4.45 1.10E-05 4.55 7.38E-03 1.93 1.27E-03 2.36 CAT 2.18E-08 20.21 1.38E-07 12.38 1.80E-04 3.25 1.58E-06 6.23 3.03E-05 3.81 3.04E-02 1.63 AC006262.6 9.48E-06 20.19 3.42E-04 7.34 7.21E-03 3.62 5.95E-04 5.58 7.59E-02 2.03 1.74E-02 2.75 SHD 6.97E-06 19.95 1.38E-04 8.57 2.50E-03 4.34 1.11E-03 4.60 7.68E-02 1.98 3.39E-02 2.33 GRIA1 4.73E-08 19.67 2.20E-06 7.67 5.46E-04 2.99 2.22E-06 6.59 9.35E-04 2.57 9.46E-04 2.56 TRIM4 8.74E-06 19.53 2.87E-04 7.38 2.55E-02 2.69 1.62E-04 7.26 1.59E-02 2.74 1.91E-02 2.65 TMCO5B 3.24E-06 19.51 3.36E-05 10.11 2.72E-03 3.77 3.22E-04 5.17 1.04E-02 2.68 6.42E-02 1.93 KDM5D 2.78E-09 19.45 1.47E-08 12.64 1.21E-04 2.73 9.39E-08 7.13 1.15E-06 4.63 2.10E-02 1.54 CAMKV 3.41E-06 19.29 5.35E-05 9.02 3.03E-03 3.70 3.13E-04 5.22 1.77E-02 2.44 3.69E-02 2.14 LYPD4 6.80E-06 19.18 5.31E-04 5.99 1.05E-02 3.13 2.60E-04 6.13 8.55E-02 1.91 6.01E-03 3.20 RP11-329E24.6 2.64E-06 19.18 4.34E-04 5.28 3.65E-03 3.44 1.80E-04 5.57 1.96E-01 1.54 1.63E-03 3.63 FGF10-AS1 2.91E-06 19.16 1.45E-05 12.12 4.22E-03 3.39 1.82E-04 5.66 1.89E-03 3.58 1.73E-01 1.58 ZNF558 2.72E-07 18.91 3.31E-06 9.83 3.50E-03 2.68 7.08E-06 7.06 2.61E-04 3.67 2.24E-02 1.92 LINC00839 1.66E-07 18.70 6.13E-06 7.64 9.34E-04 3.10 9.54E-06 6.03 2.65E-03 2.47 2.82E-03 2.45 AC008132.13 7.69E-07 18.60 1.44E-04 5.28 9.67E-03 2.50 1.39E-05 7.43 1.89E-02 2.11 7.24E-04 3.52 GDF2 6.08E-07 17.54 2.59E-04 4.41 4.81E-03 2.69 1.72E-05 6.53 8.29E-02 1.64 2.50E-04 3.98 TRH 9.52E-09 17.46 1.15E-06 6.01 2.09E-03 2.12 1.01E-07 8.25 8.76E-05 2.84 7.22E-05 2.91 SIGLEC9 9.41E-07 16.89 4.98E-05 6.41 1.27E-02 2.36 1.48E-05 7.15 3.38E-03 2.71 4.07E-03 2.64 RP11-258B16.1 4.69E-07 16.70 1.14E-04 4.80 1.21E-03 3.20 3.71E-05 5.22 1.32E-01 1.50 4.00E-04 3.48 TFAP2C 1.15E-05 16.33 5.34E-04 5.98 7.06E-02 2.10 9.26E-05 7.78 1.11E-02 2.85 1.37E-02 2.73 RP11-388N2.1 5.09E-06 16.02 6.09E-05 8.31 2.58E-02 2.39 7.86E-05 6.71 2.12E-03 3.48 5.98E-02 1.93 AOC4P 1.11E-05 15.87 1.22E-04 8.28 1.12E-02 3.04 4.85E-04 5.22 1.30E-02 2.72 8.08E-02 1.92 LINC00955 1.74E-05 15.33 3.78E-03 3.99 2.51E-02 2.67 4.11E-04 5.75 2.74E-01 1.50 2.77E-03 3.84 305

DMBX1 2.21E-05 15.30 3.88E-04 7.05 6.80E-02 2.20 2.18E-04 6.97 7.87E-03 3.21 5.41E-02 2.17 RP1-29C18.8 1.80E-05 15.26 1.12E-03 5.25 3.75E-02 2.45 2.94E-04 6.22 5.34E-02 2.14 1.13E-02 2.91 C9orf117 4.25E-07 15.23 1.45E-05 6.65 1.53E-03 2.94 2.61E-05 5.18 5.73E-03 2.26 5.19E-03 2.29 RGS13 4.39E-08 15.14 4.59E-06 5.53 3.82E-03 2.12 5.49E-07 7.14 3.89E-04 2.61 2.61E-04 2.74 SSX5 1.36E-05 15.08 5.88E-04 5.73 3.80E-02 2.37 2.02E-04 6.35 2.48E-02 2.42 1.59E-02 2.63 KRT14 1.11E-05 14.83 2.52E-04 6.62 2.94E-03 3.88 1.87E-03 3.82 1.33E-01 1.71 3.27E-02 2.24 SLC2A2 1.37E-05 14.73 3.00E-05 11.81 1.73E-02 2.77 4.26E-04 5.32 1.24E-03 4.27 5.25E-01 1.25 ZNF562 8.73E-09 14.54 4.64E-08 9.80 4.77E-05 3.07 1.05E-06 4.74 1.74E-05 3.19 3.22E-02 1.48 TYW3 3.12E-06 14.46 1.89E-05 9.09 9.69E-03 2.63 9.46E-05 5.51 1.18E-03 3.46 1.33E-01 1.59 FOLR2 2.36E-06 14.29 1.75E-04 5.25 1.85E-02 2.29 3.77E-05 6.25 1.21E-02 2.30 4.04E-03 2.72 QPCT 2.80E-06 14.21 2.95E-05 7.89 6.04E-03 2.80 1.24E-04 5.08 3.60E-03 2.82 6.04E-02 1.80 HSD3BP2 5.31E-06 13.81 1.52E-04 6.11 2.01E-02 2.39 1.04E-04 5.77 9.21E-03 2.55 1.91E-02 2.26 ZNF667 1.96E-05 13.76 1.13E-03 4.99 7.20E-02 2.08 1.77E-04 6.61 2.65E-02 2.40 1.27E-02 2.76 RP11-1018N14.2 3.44E-07 13.64 1.47E-04 3.94 6.95E-03 2.23 5.94E-06 6.12 2.76E-02 1.77 1.75E-04 3.46 STAU2-AS1 8.56E-06 13.63 1.67E-04 6.50 4.21E-02 2.16 1.01E-04 6.30 4.68E-03 3.00 3.65E-02 2.10 HOXA11-AS 6.65E-06 13.53 1.44E-04 6.36 8.38E-03 2.85 3.12E-04 4.76 2.20E-02 2.24 2.96E-02 2.13 AC005077.12 1.53E-05 13.53 1.64E-04 7.32 5.22E-03 3.45 1.64E-03 3.92 4.34E-02 2.12 8.98E-02 1.85 RP11-66B24.2 2.79E-05 13.47 5.46E-04 6.22 6.64E-02 2.17 2.99E-04 6.21 1.22E-02 2.87 5.03E-02 2.17 CRYZ 2.84E-07 13.43 4.73E-06 7.06 7.86E-03 2.15 4.20E-06 6.24 2.05E-04 3.28 1.35E-02 1.90 RP11-521O16.2 5.04E-06 13.27 2.13E-04 5.49 9.19E-03 2.70 1.94E-04 4.92 3.28E-02 2.03 1.14E-02 2.42 AQP8 3.28E-06 13.00 2.57E-04 4.85 8.24E-03 2.61 1.17E-04 4.98 4.68E-02 1.86 4.38E-03 2.68 HOXA10 1.90E-07 12.86 5.87E-07 9.82 2.43E-03 2.39 5.59E-06 5.37 2.84E-05 4.10 2.19E-01 1.31 ALOX5 8.53E-06 12.74 7.73E-04 4.50 1.98E-03 3.74 1.92E-03 3.41 5.55E-01 1.20 5.54E-03 2.83 RP11-423G4.7 2.17E-07 12.67 8.59E-05 3.88 2.19E-03 2.44 7.29E-06 5.19 4.84E-02 1.59 1.41E-04 3.27 STC1 1.84E-07 12.62 6.71E-06 5.83 2.44E-04 3.24 3.63E-05 3.89 1.53E-02 1.80 3.06E-03 2.17 TUSC5 8.44E-06 12.50 5.96E-05 7.68 4.69E-03 3.14 7.71E-04 3.97 1.27E-02 2.44 1.33E-01 1.63 CPNE7 6.58E-06 12.45 2.73E-04 5.24 5.03E-03 3.01 5.02E-04 4.13 8.44E-02 1.74 1.29E-02 2.38 306

LYZL4 5.84E-06 12.15 4.67E-04 4.54 5.26E-03 2.92 4.04E-04 4.17 1.49E-01 1.56 5.48E-03 2.67 AC092071.1 3.19E-05 11.67 3.61E-04 6.35 6.05E-02 2.14 3.89E-04 5.45 8.21E-03 2.96 9.83E-02 1.84 FOXQ1 3.12E-05 11.55 5.64E-04 5.66 6.45E-02 2.10 3.54E-04 5.49 1.31E-02 2.69 5.68E-02 2.04 HOXA11 1.16E-07 11.47 3.13E-06 5.78 3.62E-03 2.11 2.11E-06 5.43 2.33E-04 2.74 3.89E-03 1.98 APOA2 1.51E-05 11.40 2.33E-03 3.70 1.61E-02 2.57 5.21E-04 4.43 2.61E-01 1.44 3.80E-03 3.08 DLX6 1.80E-05 11.03 9.51E-05 7.33 9.25E-03 2.87 1.15E-03 3.85 1.14E-02 2.56 2.13E-01 1.51 ZNF680 6.16E-06 10.68 6.38E-05 6.27 2.13E-02 2.20 1.20E-04 4.85 2.76E-03 2.84 7.61E-02 1.70 GPR128 1.95E-09 10.67 1.54E-08 7.02 1.06E-04 2.20 6.07E-08 4.86 1.38E-06 3.20 6.07E-03 1.52 FGFBP1 1.37E-05 10.66 1.80E-03 3.70 3.31E-02 2.18 2.31E-04 4.89 1.02E-01 1.70 4.37E-03 2.88 VAC14-AS1 1.05E-05 10.62 5.38E-04 4.46 9.93E-03 2.63 5.04E-04 4.04 9.21E-02 1.70 1.18E-02 2.38 UBTFL8 1.15E-05 10.47 2.08E-03 3.48 4.00E-02 2.07 1.56E-04 5.07 9.79E-02 1.69 2.89E-03 3.01 CTC-559E9.5 1.25E-05 10.29 2.12E-04 5.42 9.22E-03 2.67 6.90E-04 3.85 3.26E-02 2.03 4.84E-02 1.90 FSTL5 4.16E-06 10.29 1.64E-05 7.51 2.36E-02 2.07 6.49E-05 4.96 4.03E-04 3.62 2.46E-01 1.37 RP11-343H19.1 1.18E-07 10.23 1.08E-05 4.36 5.68E-05 3.41 7.51E-05 3.00 1.98E-01 1.28 6.00E-04 2.35 BTBD7P1 9.64E-07 10.15 1.75E-04 3.74 1.73E-03 2.66 7.97E-05 3.82 1.51E-01 1.41 8.52E-04 2.72 KRT222 3.38E-07 10.09 1.52E-04 3.34 1.07E-04 3.51 2.41E-04 2.87 8.01E-01 -1.05 1.63E-04 3.02 CCDC60 6.50E-06 10.08 3.72E-04 4.30 8.51E-03 2.52 2.98E-04 4.01 7.06E-02 1.71 8.59E-03 2.35 RP11-311F12.1 1.10E-05 10.03 3.56E-04 4.69 3.27E-02 2.10 1.75E-04 4.77 1.59E-02 2.23 2.08E-02 2.14 PRSS21 1.92E-05 10.03 2.63E-04 5.50 1.55E-02 2.52 7.50E-04 3.98 2.44E-02 2.18 7.02E-02 1.82 SIGLEC11 6.90E-06 9.91 1.53E-04 5.07 6.03E-03 2.66 4.50E-04 3.73 3.41E-02 1.91 2.89E-02 1.95 RPL7P38 1.18E-05 9.89 1.21E-03 3.74 7.12E-03 2.73 8.26E-04 3.62 2.91E-01 1.37 5.48E-03 2.64 RP11-399K21.11 1.25E-05 9.88 2.86E-04 4.96 2.62E-02 2.20 2.55E-04 4.49 1.54E-02 2.26 3.35E-02 1.99 TPM3P1 6.64E-07 9.41 2.70E-05 4.59 6.01E-03 2.12 1.64E-05 4.44 3.18E-03 2.16 4.98E-03 2.05 CTC-378H22.2 3.89E-06 9.85 7.15E-05 5.30 1.92E-03 2.99 6.02E-04 3.29 4.39E-02 1.77 3.10E-02 1.86 AP1M2 1.11E-05 9.84 6.68E-04 4.12 5.02E-03 2.87 1.07E-03 3.42 2.25E-01 1.43 9.88E-03 2.39 PRODH 1.50E-06 9.72 5.59E-04 3.22 1.15E-03 2.90 2.33E-04 3.36 6.44E-01 1.11 4.88E-04 3.01 AC113167.2 1.44E-06 9.71 1.85E-04 3.81 8.01E-03 2.19 3.81E-05 4.43 3.18E-02 1.74 1.61E-03 2.55 307

RP11-60L3.1 1.40E-08 9.54 3.65E-07 5.18 1.80E-05 3.01 5.62E-06 3.17 2.78E-03 1.72 1.27E-03 1.84 RP11-930P14.1 5.85E-06 9.51 4.66E-04 3.92 5.08E-03 2.64 4.18E-04 3.60 1.53E-01 1.48 5.51E-03 2.43 CTD-2337J16.1 5.63E-06 9.39 2.43E-04 4.34 1.24E-02 2.28 1.72E-04 4.13 2.81E-02 1.91 1.15E-02 2.16 ADIRF 4.22E-06 9.31 3.46E-05 5.93 1.38E-02 2.18 1.06E-04 4.27 1.90E-03 2.72 9.48E-02 1.57 LRRC15 2.05E-06 9.28 9.55E-04 3.01 1.15E-03 2.93 3.77E-04 3.17 9.08E-01 1.03 4.55E-04 3.08 AC007403.3 1.75E-06 9.16 7.96E-05 4.34 2.79E-03 2.52 1.25E-04 3.63 3.34E-02 1.72 6.64E-03 2.11 COMT 6.11E-08 9.09 1.87E-06 4.82 4.37E-04 2.35 3.86E-06 3.87 9.13E-04 2.05 2.23E-03 1.88 ANGPT4 5.93E-06 9.01 5.27E-04 3.72 1.54E-02 2.18 1.52E-04 4.13 5.70E-02 1.71 4.87E-03 2.42 TUBBP5 8.38E-06 8.85 1.59E-03 3.20 4.00E-03 2.76 8.99E-04 3.21 5.75E-01 1.16 2.38E-03 2.76 CTSV 1.61E-09 8.73 4.40E-07 3.54 1.05E-06 3.19 1.95E-06 2.74 3.63E-01 1.11 5.69E-06 2.47 HOXD10 1.92E-05 8.69 7.74E-05 6.36 3.00E-02 2.13 3.96E-04 4.08 2.48E-03 2.99 2.90E-01 1.37 AP000696.2 1.87E-05 8.61 4.09E-03 2.97 4.00E-02 2.02 2.91E-04 4.27 1.93E-01 1.47 2.84E-03 2.90 RP11-382E9.1 2.10E-05 8.45 9.84E-04 3.84 3.05E-02 2.12 4.40E-04 3.99 5.72E-02 1.81 1.67E-02 2.20 BX255923.3 6.14E-06 8.31 3.18E-04 3.88 1.31E-02 2.18 1.85E-04 3.82 3.71E-02 1.78 9.52E-03 2.14 CTC-327F10.1 1.33E-05 8.25 2.03E-03 3.13 2.53E-02 2.09 2.86E-04 3.95 1.51E-01 1.50 3.66E-03 2.63 S100A6 1.26E-08 8.13 1.65E-07 5.13 4.17E-05 2.52 2.14E-06 3.23 1.97E-04 2.04 4.83E-03 1.58 TMEM179 1.28E-06 7.96 4.83E-04 2.92 5.81E-03 2.11 4.23E-05 3.77 1.40E-01 1.38 4.41E-04 2.73 RP11-909M7.3 6.35E-07 7.94 1.71E-05 4.37 1.74E-03 2.31 4.23E-05 3.44 6.31E-03 1.89 8.95E-03 1.82 ANKRD22 1.11E-05 7.89 1.39E-04 4.74 2.58E-02 2.02 2.21E-04 3.90 6.24E-03 2.35 6.88E-02 1.67 PRSS33 3.87E-06 7.74 6.65E-04 3.11 4.84E-03 2.34 2.42E-04 3.30 2.32E-01 1.33 1.87E-03 2.49 RP11-27M9.1 2.05E-06 7.21 3.91E-04 3.00 4.49E-04 2.94 1.03E-03 2.45 9.26E-01 1.02 1.20E-03 2.40 CRABP1 1.06E-06 7.70 3.73E-05 4.06 2.16E-03 2.32 7.48E-05 3.32 1.55E-02 1.75 7.55E-03 1.90 KB-1568E2.1 1.18E-05 7.70 6.97E-04 3.54 1.49E-02 2.19 4.04E-04 3.52 8.19E-02 1.62 1.04E-02 2.17 CHRNA9 1.02E-06 7.67 9.92E-05 3.48 2.65E-03 2.25 5.95E-05 3.40 4.75E-02 1.54 1.90E-03 2.20 CYP3A7 1.64E-05 7.65 2.59E-04 4.40 4.97E-03 2.67 1.92E-03 2.86 8.03E-02 1.65 5.72E-02 1.74 AC079776.1 1.17E-07 7.49 9.37E-06 3.64 1.49E-04 2.58 2.77E-05 2.90 4.87E-02 1.41 7.26E-04 2.05 LGI1 9.09E-06 7.48 1.52E-03 2.97 3.63E-03 2.61 1.12E-03 2.87 5.96E-01 1.14 2.82E-03 2.51 308

C1orf167 9.74E-06 7.43 4.47E-04 3.63 2.09E-02 2.02 2.25E-04 3.67 3.41E-02 1.80 1.31E-02 2.05 RP11-492E3.2 1.05E-06 7.36 1.14E-04 3.34 1.50E-03 2.38 1.04E-04 3.09 1.05E-01 1.40 1.68E-03 2.20 MEG9 5.20E-07 7.10 3.26E-04 2.64 1.34E-04 2.95 3.98E-04 2.40 5.40E-01 -1.12 1.51E-04 2.69 KRT8P35 1.93E-05 7.01 2.91E-04 4.15 1.14E-02 2.29 1.03E-03 3.06 3.86E-02 1.81 6.27E-02 1.69 PRR7-AS1 1.54E-05 6.89 3.34E-03 2.68 1.32E-02 2.18 6.50E-04 3.16 4.19E-01 1.23 2.65E-03 2.57 AQP10 8.99E-07 6.87 9.11E-05 3.24 6.74E-04 2.51 1.75E-04 2.74 1.85E-01 1.29 1.68E-03 2.12 ZBBX 9.07E-07 6.79 9.51E-05 3.20 3.02E-04 2.75 4.13E-04 2.46 4.22E-01 1.16 1.62E-03 2.12 ADCY8 9.45E-07 6.61 5.31E-04 2.55 1.43E-03 2.26 9.20E-05 2.92 5.27E-01 1.12 2.43E-04 2.60 RP11-460N16.1 2.39E-06 6.57 9.19E-06 5.13 3.00E-03 2.23 1.86E-04 2.94 1.38E-03 2.30 2.34E-01 1.28 SORL1 1.31E-05 6.50 2.52E-03 2.66 1.65E-02 2.04 4.19E-04 3.19 2.75E-01 1.31 2.81E-03 2.44 THPO 1.30E-05 6.45 2.76E-03 2.62 6.29E-03 2.33 1.08E-03 2.77 6.22E-01 1.13 2.54E-03 2.46 GAS8 8.92E-06 6.42 1.10E-03 2.86 2.97E-03 2.49 1.34E-03 2.58 5.42E-01 1.15 3.93E-03 2.24 AC006994.2 3.78E-06 6.37 1.65E-04 3.39 1.08E-03 2.61 1.06E-03 2.44 2.26E-01 1.30 9.74E-03 1.88 RN7SL357P 4.84E-08 6.23 1.39E-06 3.72 2.09E-04 2.14 4.86E-06 2.92 1.30E-03 1.74 2.18E-03 1.67 HSPA1A 7.69E-07 6.09 1.28E-04 2.84 2.62E-03 2.02 3.96E-05 3.02 6.64E-02 1.41 8.37E-04 2.15 PDZD7 8.22E-06 5.96 3.05E-03 2.37 2.02E-03 2.51 1.78E-03 2.38 7.99E-01 -1.06 1.15E-03 2.51 RP11-342M1.3 3.35E-06 5.84 2.55E-04 2.97 7.10E-04 2.60 1.37E-03 2.25 4.98E-01 1.14 4.57E-03 1.97 CLDND2 7.49E-06 5.83 6.22E-04 2.88 3.25E-03 2.31 9.41E-04 2.53 3.11E-01 1.25 5.81E-03 2.03 TRIM55 3.89E-07 5.17 3.93E-05 2.76 4.76E-04 2.13 6.47E-05 2.43 9.48E-02 1.30 1.07E-03 1.87 HSPA1B 2.29E-06 5.13 1.00E-03 2.25 7.14E-04 2.34 7.43E-04 2.19 8.28E-01 -1.04 5.18E-04 2.27 RALYL 1.23E-06 5.00 1.19E-05 3.57 1.60E-03 2.02 1.23E-04 2.47 4.01E-03 1.76 5.42E-02 1.40 TTC39A 1.82E-08 4.72 6.18E-07 3.01 2.05E-05 2.15 7.82E-06 2.19 6.49E-03 1.40 9.01E-04 1.57 MLPH 2.11E-07 4.71 9.39E-05 2.26 7.06E-05 2.33 1.68E-04 2.03 8.33E-01 -1.03 1.22E-04 2.08 GPC2 5.80E-07 4.69 7.18E-05 2.53 7.51E-04 2.00 7.90E-05 2.34 1.24E-01 1.26 1.03E-03 1.85 ZFR2 6.94E-06 4.46 3.29E-04 2.62 1.65E-03 2.18 1.71E-03 2.05 3.08E-01 1.20 1.11E-02 1.70 WNT5B 3.03E-07 4.20 1.65E-05 2.59 1.27E-04 2.13 1.64E-04 1.97 1.35E-01 1.22 2.19E-03 1.62

PRDX2P4 1.20E-07 3.72 1.06E-04 1.91 3.33E-05 2.09 1.43E-04 1.78 3.90E-01 -1.10 4.00E-05 1.95 309

Appendix 5 - 6 Genes targeted by miRNA hsa-miR-302d which enrich for “ion binding”

Fold-Change Fold-Change Fold-Change Target Fold-Change Target Fold-Change Fold-Change Target gene (D30 vs. (D30 vs. D90) (D30 vs. D90) gene (D30 vs. D90) gene (D30 vs. D90) (D30 vs. D90) (Description) D90) (Description) (Description) BCL6B -25.60 Day 30 down vs Day 90 GATAD2B -2.82 Day 30 down vs Day 90 ZNFX1 -2.19 Day 30 down vs Day 90 ARHGAP29 -13.69 Day 30 down vs Day 90 RPS6KA2 -2.80 Day 30 down vs Day 90 EPHA5 -2.18 Day 30 down vs Day 90 RORB -10.35 Day 30 down vs Day 90 CYBRD1 -2.73 Day 30 down vs Day 90 CNOT6 -2.18 Day 30 down vs Day 90 PLSCR4 -9.92 Day 30 down vs Day 90 FOXP1 -2.73 Day 30 down vs Day 90 ZDHHC9 -2.16 Day 30 down vs Day 90 FGD5 -6.84 Day 30 down vs Day 90 UNKL -2.72 Day 30 down vs Day 90 ASXL2 -2.15 Day 30 down vs Day 90 TSHZ3 -6.43 Day 30 down vs Day 90 GALNT10 -2.65 Day 30 down vs Day 90 SNX30 -2.15 Day 30 down vs Day 90 KCNJ8 -5.94 Day 30 down vs Day 90 SNRK -2.64 Day 30 down vs Day 90 AEBP2 -2.15 Day 30 down vs Day 90 GPC6 -5.34 Day 30 down vs Day 90 LATS2 -2.54 Day 30 down vs Day 90 FBXO11 -2.13 Day 30 down vs Day 90 HEG1 -5.28 Day 30 down vs Day 90 BMPR2 -2.50 Day 30 down vs Day 90 CTDSPL -2.12 Day 30 down vs Day 90 KLF3 -5.09 Day 30 down vs Day 90 RELA -2.49 Day 30 down vs Day 90 UBE2J1 -2.10 Day 30 down vs Day 90 RUNX1 -4.79 Day 30 down vs Day 90 RORA -2.48 Day 30 down vs Day 90 MDM4 -2.10 Day 30 down vs Day 90 MBNL3 -4.69 Day 30 down vs Day 90 MEX3A -2.45 Day 30 down vs Day 90 ZC3H12C -2.06 Day 30 down vs Day 90 FAT4 -4.57 Day 30 down vs Day 90 CDK19 -2.42 Day 30 down vs Day 90 TBC1D8B -2.05 Day 30 down vs Day 90 ZNF697 -4.36 Day 30 down vs Day 90 IRF2BP2 -2.39 Day 30 down vs Day 90 IP6K1 -2.04 Day 30 down vs Day 90 BNC2 -4.26 Day 30 down vs Day 90 KLF12 -2.36 Day 30 down vs Day 90 UBE2R2 -2.04 Day 30 down vs Day 90 TRPS1 -4.24 Day 30 down vs Day 90 MBNL2 -2.34 Day 30 down vs Day 90 HK1 2.15 Day 30 up vs Day 90 ZFPM2 -4.21 Day 30 down vs Day 90 ZBTB44 -2.33 Day 30 down vs Day 90 TIPARP 2.30 Day 30 up vs Day 90 TET3 -3.65 Day 30 down vs Day 90 RPS6KA3 -2.32 Day 30 down vs Day 90 ZFHX4 2.64 Day 30 up vs Day 90 TGFBR2 -3.63 Day 30 down vs Day 90 TNS1 -2.31 Day 30 down vs Day 90 PFKP 2.66 Day 30 up vs Day 90 NR2F2 -3.59 Day 30 down vs Day 90 RNF38 -2.26 Day 30 down vs Day 90 MYLK 2.71 Day 30 up vs Day 90 PEAK1 -3.42 Day 30 down vs Day 90 ABL2 -2.26 Day 30 down vs Day 90 EPHA7 3.23 Day 30 up vs Day 90 ULK1 -3.34 Day 30 down vs Day 90 TIMP3 -2.26 Day 30 down vs Day 90 SYT14 5.52 Day 30 up vs Day 90 ZFPM1 -3.29 Day 30 down vs Day 90 PPP3R1 -2.26 Day 30 down vs Day 90 CYBB 8.40 Day 30 up vs Day 90 CREB5 -3.09 Day 30 down vs Day 90 GAB2 -2.23 Day 30 down vs Day 90 TRHDE 8.71 Day 30 up vs Day 90 ASAP1 -2.96 Day 30 down vs Day 90 PAK2 -2.21 Day 30 down vs Day 90 FSTL5 10.29 Day 30 up vs Day 90 PKD2 -2.89 Day 30 down vs Day 90 NR2C2 -2.20 Day 30 down vs Day 90 ZBTB7A -2.85 Day 30 down vs Day 90 PHLPP2 -2.19 Day 30 down vs Day 90

310

Appendix 5 - 7 Genes targeted by miRNA hsa-miR-302d which enrich for “Regulation of transcription, DNA-templated”

Fold-Change Fold-Change Fold-Change Fold-Change Fold-Change Target Target Target Fold-Change (D30 vs. (D30 vs. D90) (D30 vs. (D30 vs. D90) (D30 vs. gene gene gene (D30 vs. D90) (Description) D90) (Description) D90) (Description) D90) BCL6B -25.5965 Day 30 down vs Day 90 TCF7L1 -2.97118 Day 30 down vs Day 90 HMBOX1 -2.22313 Day 30 down vs Day 90 RORB -10.3475 Day 30 down vs Day 90 NFIA -2.93904 Day 30 down vs Day 90 NR2C2 -2.2011 Day 30 down vs Day 90 NFIB -8.09729 Day 30 down vs Day 90 PKD2 -2.88517 Day 30 down vs Day 90 EPHA5 -2.18236 Day 30 down vs Day 90 TSHZ3 -6.42572 Day 30 down vs Day 90 ZBTB7A -2.85341 Day 30 down vs Day 90 CNOT6 -2.18194 Day 30 down vs Day 90 APBB2 -5.62895 Day 30 down vs Day 90 FOXP1 -2.72602 Day 30 down vs Day 90 ASXL2 -2.15156 Day 30 down vs Day 90 PRRX1 -5.56372 Day 30 down vs Day 90 FOXJ2 -2.71467 Day 30 down vs Day 90 AEBP2 -2.14562 Day 30 down vs Day 90 RUNX1 -4.78903 Day 30 down vs Day 90 TP53INP2 -2.70208 Day 30 down vs Day 90 MDM4 -2.09508 Day 30 down vs Day 90 ZNF697 -4.35972 Day 30 down vs Day 90 CCND1 -2.68979 Day 30 down vs Day 90 AHR -2.06683 Day 30 down vs Day 90 BNC2 -4.25529 Day 30 down vs Day 90 BMPR2 -2.50171 Day 30 down vs Day 90 NFYA -2.02035 Day 30 down vs Day 90 TRPS1 -4.24474 Day 30 down vs Day 90 RELA -2.49423 Day 30 down vs Day 90 SETD7 -2.01176 Day 30 down vs Day 90 ZFPM2 -4.20865 Day 30 down vs Day 90 RORA -2.47789 Day 30 down vs Day 90 TMEM100 2.04343 Day 30 up vs Day 90 ELK4 -4.12117 Day 30 down vs Day 90 INO80D -2.43012 Day 30 down vs Day 90 ELAVL2 2.22349 Day 30 up vs Day 90 TET3 -3.64749 Day 30 down vs Day 90 FOXO3 -2.40099 Day 30 down vs Day 90 BAMBI 2.58174 Day 30 up vs Day 90 NR2F2 -3.59439 Day 30 down vs Day 90 ARID5B -2.39598 Day 30 down vs Day 90 POLR3G 2.70467 Day 30 up vs Day 90 RGMB -3.31117 Day 30 down vs Day 90 IRF2BP2 -2.39273 Day 30 down vs Day 90 DKK1 3.0768 Day 30 up vs Day 90 ZFPM1 -3.28907 Day 30 down vs Day 90 KLF12 -2.36267 Day 30 down vs Day 90 FOXF2 7.45493 Day 30 up vs Day 90 E2F7 -3.24071 Day 30 down vs Day 90 MAML1 -2.33308 Day 30 down vs Day 90 ALX4 12.5154 Day 30 up vs Day 90 CREB5 -3.09017 Day 30 down vs Day 90 ZBTB44 -2.32786 Day 30 down vs Day 90 TP53INP1 -3.02878 Day 30 down vs Day 90 RPS6KA3 -2.31955 Day 30 down vs Day 90

311

Appendix 5 - 8 List of miRNA that are differentially expressed in d90 and d30 EHMs. Down-regulations are shown in red

Log 2 Log 2 Log 2 Log 2 Log 2 Fold Fold Log 2 Fold P-value Fold P-value Fold P-value P-value P-value Fold P-value miRNA Gene change change change (d90 vs. change (d90 vs. change (d90 vs. (d60 vs. (d60 vs. change (d45 vs. symbol (d90 (d60 (d45 vs. d30) (d90 vs. d45) (d90 vs. d60) d30) d45) (d60 vs. d30) vs. vs. d30) d30) d45) d45) d60) d30) hsa-let-7a-5p 2.16E-02 13.29 4.48E-02 1.75 3.34E-01 1.02 8.39E-02 12.27 1.80E-01 1.66 6.17E-01 11.54 hsa-let-7b-5p 7.85E-01 -1.27 2.97E-01 -1.62 1.00E+00 -0.94 7.42E-01 -0.33 1.96E-01 -1.61 3.90E-01 0.35 hsa-let-7c-5p 1.53E-03 3.13 9.21E-03 2.07 3.92E-02 0.66 1.04E-01 2.47 4.49E-01 2.66 3.52E-01 1.06 hsa-let-7f-5p 2.13E-02 4.94 2.75E-02 3.80 4.90E-01 0.41 5.97E-04 4.53 5.77E-04 10.52 7.91E-01 1.14 hsa-let-7g-5p 1.24E-01 2.41 2.01E-02 5.53 1.97E-02 4.28 4.00E-01 -1.87 9.84E-01 2.37 3.96E-01 -3.11 hsa-miR-1 3.12E-02 4.53 7.97E-01 -0.32 5.85E-01 -0.24 2.31E-02 4.76 6.88E-01 -1.06 1.01E-02 4.85 hsa-miR-100-5p 2.09E-01 2.58 5.00E-01 1.11 9.86E-01 0.79 1.25E-01 1.79 3.90E-01 1.25 4.57E-01 1.47 hsa-miR-1225-3p 6.38E-01 -10.44 2.79E-02 -12.77 2.84E-01 -11.77 5.13E-01 1.33 1.45E-01 -2.00 5.08E-02 2.33 hsa-miR-125b-5p 9.63E-02 2.83 4.17E-02 4.91 9.09E-01 0.98 3.38E-02 1.84 8.51E-03 15.25 7.32E-01 -2.09 hsa-miR-126-3p 4.85E-01 10.96 4.51E-01 -1.16 3.23E-01 -1.11 5.24E-02 12.07 7.77E-01 -1.04 9.49E-02 12.12 hsa-miR-133a-3p 8.09E-01 1.50 9.85E-02 -1.91 3.97E-01 -0.57 2.12E-01 2.07 3.03E-01 -2.54 3.26E-02 3.42 hsa-miR-143-3p 7.70E-01 0.06 7.82E-02 -0.98 4.63E-01 -0.72 6.30E-01 0.78 1.91E-01 -1.20 9.77E-02 1.04 hsa-miR-146b-5p 1.31E-02 13.60 9.15E-03 13.60 3.18E-01 0.44 5.46E-02 13.16 3.81E-02 9162.23 1.00E+00 0.00 hsa-miR-181a-5p 7.30E-02 2.79 3.17E-01 1.03 7.90E-01 0.48 6.15E-02 2.31 3.59E-01 1.47 3.01E-01 1.76 hsa-miR-181b-5p 4.81E-01 -0.42 3.29E-01 1.07 1.02E-01 2.23 3.05E-01 -2.65 4.13E-01 -2.22 7.88E-01 -1.50 hsa-miR-1973 1.32E-02 -13.78 4.41E-01 -11.33 4.65E-01 -10.70 2.72E-02 -3.08 9.52E-01 -1.55 3.30E-02 -2.45 hsa-miR-199a-3p 4.35E-02 13.20 2.41E-01 2.26 8.74E-01 0.22 2.42E-02 12.98 2.07E-01 4.09 2.63E-01 10.94 hsa-miR-199b-3p 2.55E-02 13.31 4.68E-02 3.24 5.90E-01 0.49 3.47E-02 12.83 6.78E-02 6.73 6.61E-01 10.07 hsa-miR-199b-5p 4.63E-01 12.18 3.19E-01 -0.22 3.71E-01 0.09 6.03E-02 12.08 8.90E-01 -1.24 4.82E-02 12.39 hsa-miR-208b-3p 4.31E-01 12.18 9.51E-01 1.12 1.67E-01 -0.74 1.37E-02 12.92 7.71E-02 3.65 3.84E-01 11.05 hsa-miR-21-5p 8.06E-01 1.11 4.64E-01 0.53 6.74E-01 0.38 4.39E-01 0.73 6.92E-01 1.11 2.63E-01 0.58 312

hsa-miR-210-3p 3.39E-01 0.88 4.08E-01 -0.63 9.07E-01 0.20 3.06E-01 0.68 2.43E-01 -1.78 4.09E-02 1.51 hsa-miR-222-3p 9.91E-01 -0.43 1.76E-01 -0.82 6.14E-01 -0.63 5.57E-01 0.20 2.85E-01 -1.14 1.23E-01 0.39 hsa-miR-26a-5p 4.23E-02 13.82 2.56E-01 1.88 8.13E-01 1.01 2.88E-02 12.81 2.62E-01 1.83 2.40E-01 11.94 hsa-miR-26b-5p 6.54E-01 0.42 4.67E-01 1.02 8.83E-01 0.73 7.07E-01 -0.30 4.69E-01 1.22 7.71E-01 -0.59 hsa-miR-302a-5p 1.00E-01 -2.04 2.15E-01 -1.40 4.89E-01 1.88 6.87E-03 -3.92 1.93E-02 -9.66 5.64E-01 -0.65 hsa-miR-30a-5p 4.57E-01 11.35 6.85E-01 -0.95 4.69E-01 -1.21 8.87E-02 12.56 6.98E-01 1.20 1.80E-01 12.30 hsa-miR-30c-5p 7.47E-01 2.89 9.39E-01 1.49 1.78E-01 -1.10 5.75E-02 3.99 8.20E-02 6.02 7.65E-01 1.40 hsa-miR-30d-5p 1.84E-01 1.82 4.47E-01 0.63 2.58E-01 0.93 7.41E-01 0.89 2.22E-02 -1.23 1.65E-02 1.19 hsa-miR-30e-5p 6.42E-01 -10.22 1.23E-01 -12.80 1.60E-01 -11.92 2.96E-01 1.71 8.48E-01 -1.83 2.28E-01 2.58 hsa-miR-320a 6.80E-01 -0.50 2.75E-01 2.34 8.19E-01 -0.14 4.48E-01 -0.36 1.04E-01 5.58 4.53E-01 -2.84 hsa-miR-3609 1.76E-02 2.63 4.01E-01 1.65 6.77E-01 1.17 1.59E-02 1.46 5.93E-01 1.39 5.38E-02 0.98 hsa-miR-3687 6.39E-01 -10.22 1.40E-01 -12.12 1.43E-01 -12.51 2.68E-01 2.30 9.75E-01 1.32 2.62E-01 1.90 hsa-miR-425-5p 1.13E-01 2.45 3.74E-01 2.30 3.66E-01 1.63 3.64E-01 0.82 9.93E-01 1.60 3.66E-01 0.14 hsa-miR-4461 6.54E-03 -3.23 1.86E-03 -2.66 1.22E-01 -1.47 1.05E-01 -1.76 3.62E-02 -2.28 7.38E-01 -0.57 hsa-miR-4485 1.27E-01 -2.14 3.60E-02 -2.82 1.49E-02 -3.26 3.27E-01 1.13 6.71E-01 1.36 5.59E-01 0.68 hsa-miR-4521 2.70E-02 13.17 9.79E-02 3.22 3.74E-01 0.65 8.83E-02 12.52 3.26E-01 5.91 4.21E-01 9.96 hsa-miR-5096 6.84E-02 -3.60 1.19E-01 -3.08 9.77E-02 -3.01 7.34E-01 -0.59 6.52E-01 -1.05 8.84E-01 -0.52 hsa-miR-6087 4.20E-04 -5.73 6.10E-03 -4.12 1.48E-02 -3.58 9.06E-02 -2.15 6.38E-01 -1.45 2.07E-01 -1.62 hsa-miR-619-5p 2.58E-02 -14.07 4.15E-02 -13.53 1.30E-02 -14.27 8.82E-01 0.20 5.70E-01 1.67 7.10E-01 -0.55 hsa-miR-6723-5p 3.40E-02 -1.09 5.78E-01 0.41 9.68E-01 1.48 1.25E-02 -2.57 5.21E-01 -2.10 5.44E-02 -1.50 hsa-miR-7641 2.01E-02 -1.61 2.39E-02 -1.48 6.27E-01 -0.31 2.27E-02 -1.30 2.61E-02 -2.26 8.19E-01 -0.13 hsa-miR-99a-5p 6.13E-01 -0.71 4.07E-01 -0.49 5.01E-01 0.59 8.78E-01 -1.30 6.69E-02 -2.11 1.26E-01 -0.22 hsa-miR-99b-5p 5.22E-01 -2.31 1.48E-01 -2.35 1.63E-01 -2.48 4.12E-01 0.17 9.31E-01 1.09 3.74E-01 0.05

313

Appendix 5 - 9 Dot plots of different miRNAs that are expressed differentially in d90 and d30 EHMs

hsa-miR-30e-5p hsa-miR-26b- hsa-miR-99a-5p

hsa-miR-30c-5p hsa-miR-99b-5p hsa-miR-30d-5p

hsa-miR-100- hsa-miR-125b-5p hsa-miR-146b-5p 5

314

hsa-miR-126-3p hsa-miR-133a-3p hsa-miR-143-3p

hsa-let-7a-5p hsa-let-7f-5p hsa-let-7b-5p

hsa-let-7g-5p hsa-miR-619-5p hsa-miR-302a-5p

315

hsa-miR-26a- hsa-miR-208b-3p hsa-miR-425-5p

hsa-miR-181b- hsa-miR-199a-3p hsa-miR-199b-5p

hsa-miR- hsa-miR-1973 hsa-miR-4521

316

hsa-miR-4461 hsa-miR-6723-5p hsa-miR-3609

hsa-miR-3687 hsa-miR-21-5p hsa-miR-7641

hsa-miR-199b-3p hsa-miR-5096 hsa-miR-181a-5p

317

hsa-miR-30a-5p hsa-miR-6087

318

Miscellaneous Appendix - 1 List of primers

1. MYL2f : 5’-CTG TAT GAG CAA GCA ATG GTG CCG-3’ 2. MYL2r : 5’-GTG ACA GAA AAC AGA GGC CTC CCC-3’ 3. SeqPrmr1 : 5’-CAC AGC AGG TGG AAA GTG AA-3’ 4. SeqPrmr2 : 5’-ATG CAC TTT TCC TTC CTT GC-3’ 5. SeqPrmr3 : 5’-TAC TCG GGG GAG AGA GAT GA-3’ 6. SeqPrmr4 : 5’-TCC GCT TCC TTC TTG TCA CT-3’ 7. SeqPrmr5 : 5’-AGT TCA CTG CAG CCT CGA C-3’ 8. SeqPrmf1 : 5’-TGG CAA GGG CAA TAT AAG TG-3’ 9. SeqPrmf2 : 5’-TTA TAC CCC CAC ACC AAT CC-3’ 10. SeqPrmf3 : 5’-TGG GTG GCT ATG GGT ACT TC-3’ 11. SeqPrmf4 : 5’-GCC TGG ATG TGA GTT GAC CT-3’ 12. Phlox5 : 5’-aga Gat AAC TTC GTA TAG CAT ACA TTA TAC GAA GTT ATG CTG CAG GAA TTC TAC CGG GTA G-3’ 13. Phlox3 : 5’-tct cAT AAC TTC GTA TAA TGT ATG CTA TAC GAA GTT ATA AGC AAT AGT AAA GAG CCT TCA G-3’ 14. SAC-Primer1-1 : 5'-CCA CGG AGA AGA GAA GGA CG-3’ 15. SAC-Primer1-2 : 5’-GCA CAT CAT CAC CCA CGG AGA AGA GAA-3’ 16. SAC-Primer2-1 : 5’-TCC GCC TCA GAA GCC ATA GA-3’ 17. SAC-Primer2-2 : 5’-CAG CTG GTT CTT TCC GCC TC-3’ 18. SAC-Primer3-1 : 5’-AGA GAT AAC TTC GTA TAG CAT ACA T-3’ 19. SAC-Primer3-2 : 5’-AAG AAC CAG CTG AGA GAT AAC TTC GTA TAG CAT ACA T-3’ 20. SAC-Primer4-1 : 5’-TCT CAT AAC TTC GTA TAA TGT ATG C-3’ 21. SAC-Primer4-2 : 5’-ACC ACT CTG CAA TCT CAT AAC TTC GTA TAA TGT ATG C-3’ 22. SAC-Primer5-1 : 5’- CCC TGC CCT CAT CTC TCT C-3’ 23. SAC-Primer5-2 : 5’-TTG CAG AGT GGT CCC TGC CCT CAT CTC TCT C-3’ 24. SAC-Primer6-1 : 5’-ACC AGG TTC TTG TAG TCC AAG TT-3’ 25. SAC-Primer6-2 : 5’- GTG ATG ATG TGC ACC AGG TTC TTG TAG TC-3’ 26. Reverse Primer C : 5’– AAC TGA GGG GAC AGG ATG TC – 3’ 27. Forward primer Phlox-M : 5’– CTC GTC CGA GGG CAA AGG AA – 3’ 28. Forward primer L-Phlox : 5’– TTG CCC CTC CCC CGT GCC TT – 3’ 29. Reverse primer L-Phlox : 5’– CAA AGG CCT ACC CGC TTC CA – 3’ 30. Seq-hyg-f1 : 5’– CAT TCT GCA CGC TTC AAA AG – 3’ 31. Seq-hyg-f2 : 5’– CGA GGA CTG CCC CGA AGT CC – 3’ 32. Seq-hyg-r1 : 5’– TGA CAG GAA AAA CTT CCA TTT T – 3’ 33. Seq-hyg-r2 : 5’– GGC CCA AAG CAT CAG CTC – 3’ 34. TAL-5 : 5’– CAC CAT GGG AAA ACC TAT TCC TAA TCC – 3’ 35. TAL-3 : 5’– TTA TCA GAA GTT GAT CTC GCC GTT – 3’ 36. MYL2KI-F :5’-GGT GCG GTA GTG AGT TTG GT-3’ 319

37. MYL2KI-R :5’-CTT CAG CTT CAG CCT CTG CT-3’ 38. Nde508F :5’-TGC ACC TGT GGT CTC AGC TAC-3’ 39. Nde508R :5’-TGG TTC AGG AGG ACT CTG ATT T-3’ 40. Nde500F :5’-AGG CAA AGG CAG ATG AGC TA-3’ 41. Nde500R :5’-GGG ACT CTT GCA GCC ATA AC-3’ 42. Puro Phlox 5 : 5'- AGA GAT AAC TTC GTA TAG CAT ACA TTA TAC GAA GTT ATG CTG CAG GAA TTC TAC CGG GTA GGG GA-3' 43. Puro Phlox 3 : 5'- TCT CAT AAC TTC GTA TAA TGT ATG CTA TAC GAA GTT ATG GCC ACA ACT CCT CAT AAA GA-3' 44. Cherry-F : 5’- CCCCGTAATGCAGAAGAAGA-3’ 45. Cherry-R : 5’- TTGACCTCAGCGTCGTAGTG-3’ 46. KIgDNA-F : 5’- GTCCTTAGCACGTGTTGCTG-3’ 47. KIgDNA-R : 5’- AGGAAAGGACAGTGGGAGTG-3’ 48. KIhomo-F : 5’- CATTGCAGGTTGACCAGATG-3’ 49. KIhomo-R : 5’- CAAGGCCCTCTAAATGCTTG- 3’ 50. MYL-tcp-F : 5’- ATTCTTCTCGGGAGGCAGTG-3’ 51. MYL-spl-seq-R : 5’- AGGTCCAGGGTTCTCCTCC -3’ 52. qPCR-KI-R : 5’- GCCGTCCTCGAAGTTCATCA -3’ 53. MYL2-KI-R : 5’- CTTCAGCTTCAGCCTCTGCT-3’ 54. MYL-tcp-R : 5’- GCAGGAGCAAGGTGAGATGA-3’ 55. Neo-F : 5’-ATGCCTGCTTGCCGAATATC-3’ 56. Neo-R : 5’-TCAGCAATATCACGGGTAGCC-3’ 57. Luc-F :5’-TCGTCGACATTCCTGAGATTCC-3’ 58. Luc-R :5’-TCTTGAGCAGGTCAGAACACTG-3’ 59. Seq-puro-F :5’-CTG GAC CAC GCC GGA GAG-3’ 60. Seq-puro-R : 5’-GAT GTG GCG GTC CGG ATC-3’

320

Miscellaneous Appendix - 2 Materials

1. Cell lines Name Cat. No Company HES3 and H9 WiCell Research Institute, Inc MBJ5 and MBJ6 In-house (by Manasi and Dr. Vinod Verma) EmbryoMax® Primary Mouse Embryo Fibroblasts, Not PMEF-CFL Millipore Mytomycin C Treated, Strain CF1, passage 3 MEF (DR4) Mouse Embryonic fibroblasts SCRC-1045 ATCC HFF-1 Human Foreskin Fibroblasts SCRC-1041 ATCC EmbryoMax® Primary Mouse Embryo Fibroblasts, PMEF-HL Millipore Hygromycin resistant, passage 3 Normal Human Dermal Fibroblasts (neonatal skin) CC-2509 Lonza Puromycin-Resistant Mouse Embryonic Fibroblasts, Day 00325 Stemcell technologies E13.5

2. Cell culture and differentiation components Name Cat. No Company Dulbecco’s Modified Eagle’s Medium (DMEM) 11965-092 Gibco, Life Technologies Fetal Bovine Serum 10270-106 Gibco, Life Technologies DMEM:F12 10565-018 and 11330-032 Gibco, Life Technologies KnockOut™ Serum Replacement (KOSR) 10828-028 Gibco, Life Technologies BD Matrigel hESC-qualified Matrix 354277 BD Biosciences BD Matrigel Matrix Growth Factor Reduced 354230 BD Biosciences Non-Essential Amino Acids solution (100x) 11140-050 Gibco, Life Technologies Penicillin-Streptomycin-Glutamine (100X) 10378-016 Gibco, Life Technologies Human recombinant bFGF 13256029 Gibco, Life Technologies RPMI 1640 Medium, GlutaMAX™ Supplement 61870-036 Gibco, Life Technologies Sodium Pyruvate (100 mM) 11360-070 Gibco, Life Technologies Penicillin-Streptomycin (10,000 U/mL) 15140-122 Gibco, Life Technologies B-27® Supplement (50X), serum free 17504-044 Gibco, Life Technologies B-27® Supplement, minus insulin A1895601 Gibco, Life Technologies Iscove’s liquid medium without L-Glutamine F 0465 Biochrom, Merck Millipore Recombinant Albumin A9731 Sigma Bovine Serum Albumin A3311 Sigma L-Glutamine (200 mM) 25030-081 Gibco, Life Technologies mTeSR™1 05850 Stemcell Technologies Geneticin® Selective Antibiotic (G418 Sulfate) 10131-027 Gibco, Life Technologies G418 Sulfate 345812 Calbiochem Puromycin Dihydrochloride A11138-03 Gibco, Life Technologies Hygromycin B 400050 CAlbiochem BMS453 3409 Tocris Bioscience rh-Dkk1 1096-DK R&D systems bFGF 130-093-841 Miltenyi Biotech Activin-A 338-AC R&D systems BMP4 314-BP R&D systems IGF-1 AF-100-11 Peprotech VEGF AF-100-20 Peprotech FGF AF-100-18B Peprotech TGF-β1 AF-100-21C Peprotech Lipid Mixture 1, Chemically Defined L0288 Sigma Type I, Bovine Soluble Collagen CS019 Collagen Solutions LLC DMEM (high glucose powder) 12100-061 Gibco, Life Technologies

321

3. Chemicals, small molecules and enzymes Name Cat. No Company Collagenase I C0130 Sigma NAOH CaCl2 Y27632 04-0012-10 Stemgent 2-Mercaptoethanol (55mM) 21985-023 Gibco, Life Technologies L-Ascorbic acid 2-phosphate sesquimagnesium salt hydrate A8960-5G Sigma DPBS, no calcium, no magnesium 14190-144 Gibco, Life Technologies DPBS, calcium, magnesium 14040-133 Gibco, Life Technologies DPBS, calcium, magnesium, glucose, pyruvate 14287-080 Gibco, Life Technologies Goat Serum, New Zealand origin 16210-064 Gibco, Life Technologies IWP4 04-0036 Stemgent IWR1 I0161-5MG Sigma CHIR99021 04-0004 Stemgent Dimethyl sulfoxide (DMSO) D2650 Sigma Roti ® -Histofix 4% P087.5 Carl Roth TRIzol® Reagent 15596-026 Ambion, Life Technologies Sodium Lactate Isoprenaline hydrochloride I5627 Sigma Gemfibrozil G9518 Sigma 0.25% Trypsin with EDTA 25300-054 Gibco, Life Technologies DNAse 260913 Calbiochem Hoechst 33342, Trihydrochloride, Trihydrate H3570 Invitrogen, Life Technologies Triton X-100 T8787 Sigma Expand High FidelityPLUS PCR System 03300242001 Roche Expand Long Template PCR System 11681834001 Roche Taq DNA Polymerase, native 18038042 Invitrogen, Life Technologies 10x DreamTaq Buffer B65 Thermo Scientific X-Gal Solution, ready-to-use R0941 Thermo Scientific DNAse I recombinant, RNAse free 04716728001 Roche Phusion Green Hot Start II High-Fidelity DNA Polymerase (2 F-537L Thermo Scientific U/µL) DpnI R0176S NEB EcoRI-HF® R3101S NEB NheI-HF® R3131S NEB SalI-HF® R3138S NEB NdeI R0111S NEB NotI-HF® R3189S NEB HindIII R0104S NEB SpeI R0133S NEB Mitomycin-C Kyowa Hakko Kirin BglII R0144S NEB

4. Antibodies and Kits Name Cat. No Company IgG1 control MAB002 R&D systems α-Actinin A7811 Sigma SOX2 MAB2018 R&D OCT4 AF1759 R&D NANOG R&D SSEA-4 AB16287 Abcam TRA-1-60 AB16288 Abcam 322

Goat anti mouse Alexa 488 A11001 Invitrogen, Life technologies Goat anti rabbit Alexa 546 A11010 Invitrogen, Life technologies Goat anti mouse Alexa 546 A11003 Invitrogen, Life technologies Goat anti rabbit Alexa 488 A11008 Invitrogen, Life technologies DNeasy Blood & Tissue Kit (250) 69506 Qiagen First Strand cDNA Synthesis Kit K1612 Thermo scientific iQ™ SYBR® Green Supermix 170-8880 Biorad QIAquick Gel Extraction Kit 28706 Qiagen QIAquick PCR Purification Kit 28106 Qiagen TOPO® TA Cloning® Kit for Subcloning, with One Shot® K4500-01 Invitrogen, Life Technologies TOP10 chemically competent E. coli cells RNeasy Mini Kit 74106 Qiagen Wizard® SV Gel and PCR Clean-Up System A9282 Promega QIAprep Spin Miniprep Kit 27106 Qiagen PureLink® HiPure Plasmid Filter Maxiprep Kit K2100-17 Invitrogen, Life Tehcnologies

5. Equipments Name Company Kryo 750 Plus Planer CASY cell counter and analyzer Roche Gene Pulser Xcell™ Electroporation Systems Biorad Veriti® Thermal Cycler Applied Biosystems, Life Technologies Stratalinker® UV Crosslinker Stratagene

6. Software Name Company Partek® Flow® and Genomics Suite® Partek Inc. Graphpad Prism Graphpad Software

323