Title: Building blocks of the human heart

Cover Page

The handle http://hdl.handle.net/1887/78948 holds various files of this Leiden University dissertation.

Author: Giacomelli, E.

Title: Building blocks of the human heart

Issue Date: 2019-10-01

BUILDING BLOCKS OF THE HUMAN HEART

Elisa Giacomelli

2019

Colophon

Building blocks of the human heart PhD thesis

This thesis was prepared at the Department of Anatomy & Embryology of the Leiden University Medical Center, Leiden, The Netherlands.

Copyright © Elisa Giacomelli, Leiden, the Netherlands, 2019.

All rights reserved. No parts of this book may be reproduced or transmitted, in any forms and by any means, without permission in writing from the author.

The copyright of the articles that have been published has been transferred to the respective journals.

The research described in this thesis as well as its publications was supported by the following grants: European Research Council (ERCAdG 323182 STEMCARDIOVASC) and Transnational Research Project on Cardiovascular Diseases (JTC2016_FP-40-021 ACM-HF).

Financial support by the Dutch Heart Foundation for the publication of this thesis is gratefully acknowledged.

ISBN: 978-94-6380-428-8

Cover: human heart model made by LEGO. Made by Nathan Sawaya, who kindly

gave the permission for the use of this image for this book.

Building blocks of the human heart

Proefschrift

ter verkrijging van

de graad van Doctor aan de Universiteit Leiden, op gezag van Rector Magnificus prof.mr. C.J.J.M Stolker,

volgens besluit van het College voor Promoties te verdedigen op dinsdag 1 oktober 2019 klokke 12.30 uur

door

Elisa Giacomelli

geboren te Pietrasanta, Italie in 1990

Promotor

Prof.dr. C.L. Mummery

Co-Promotoren Dr. M. Bellin Dr. V.V. Orlova

Promotiecommissie Prof.dr. M.J.T.H Goumans Prof.dr. D.E Atsma Prof.dr. P.C.J.J. Passier

Prof.dr. J. van der Velden (VUmc)

Prof.dr. P. van der Meer (UMCG)

Ai miei genitori

Stefano e Luisella

Chapter 1

Introduction of the thesis

Chapter 2

Human heart disease: lessons from human pluripotent stem cell-derived cardiomyocytes

Chapter 3

Three-dimensional cardiac microtissues composed of cardiomyocytes and endothelial cells co-differentiated from human pluripotent stem cells

Chapter 4

Co-differentiation of human pluripotent stem cells-derived cardiomyocytes and endothelial cells from cardiac mesoderm provides a three-dimensional model of cardiac microtissue

Chapter 5

Human pluripotent stem cell differentiation into functional epicardial progenitor cells

Chapter 6

Cardiac-but not dermal fibroblasts induce structural and functional maturation of hiPSC-derived cardiomyocytes in 3D microtissues

Chapter 7

General Discussion

Appendix Summary Samenvatting Curriculum Vitae List of publications Acknowledgments

9

23

83

123

163

195

267 283 284 286 288 290 292

CONTENTS

CHAPTER 1 1

Introduction of the thesis

Cell communication in the heart 1

The heart is a complex organ broadly structured into four chambers (two atria and two ventricles) that pumps oxygenated and deoxygenated blood throughout the body in an intricate system of arteries, veins and capillaries (Thiriet, 2007).

Oxygen-poor blood enters the right atrium and flows through the right ventricle; the right ventricle pumps deoxygenated blood to the lungs, where it becomes oxygenated; oxygenated blood returns to the heart, specifically to the left atrium, through the pulmonary veins. From the left atrium, oxygenated blood flows to the left ventricle from where it is distributed via the aorta (the main and largest artery) throughout the body (Thiriet, 2007). These diverse functions need to be carried out in synchrony and this is regulated by cells of the conduction system. Each chamber is composed of different cell types, some with their own chamber-specific identity. The electrical cardiac conduction system triggers and orchestrates the excitation-contraction function of the heart ventricles; it generates the cardiac impulse and conducts it from the atria to the ventricles, making the heart chambers contract sequentially (Moorman and Christoffels, 2003). The principle cell types that make up the (mammalian) heart are illustrated in Figure 1.

Cardiac muscle cells, or cardiomyocytes (CMs), the fundamental contractile cells of the myocardium, occupy approximately 75% of the myocardial tissue volume; however they account for less than a third of the total cell number.

The remaining non-CM fraction includes many additional cell types, including endothelial cells (ECs), cardiac fibroblasts (CFs), smooth muscle cells, other connective tissue cells and even immune system–related cells and neurons (Armour et al., 1997; Hulsmans et al., 2017; Tirziu et al., 2010; Xin et al., 2013).

Atrial and ventricular CMs form the myocardium, the muscle walls of the heart.

ECs form the endocardium, the interior lining of blood vessels and cardiac

valves. Given the high energy and oxygen demands of the heart, every CM

is in contact with minimally one EC. CFs form the main cell population in

the adult heart (up to 50%). Smooth muscle cells contribute to the coronary

arteries and inflow and outflow vasculature. The epicardium gives rise to the

precursors of CFs and smooth muscle cells and covers the surface of the heart

during development migrating from the proepicardial organ to form a single

epithelial layer (Brade et al., 2013). In the adult heart, it is the epicardial cells

that give rise to the CFs, which develop after myocardial infarction and form

the scar tissue that prevents heart rupture (Furtado et al., 2016). Pacemaker

cells and Purkinje fibres in the conduction system are specialized CMs that generate and conduct electrical impulses. The sinoatrial node (SAN), composed of pacemaker cells, resides in the right atrium and generates impulses to initiate the contraction of the heart. The atrioventricular node (AVN), located between the atria and ventricles, conducts electrical impulses from the atria to the ventricles (Tirziu et al., 2010). Cardiac-tissue resident macrophages are spindle-like cells found abundantly in the AVN that modulate CM electrical activity through electric coupling (Harari et al., 2017; Hulsmans et al., 2017);

they also play a role in atherosclerotic development and promote both injury and repair after myocardial infarction (Johnson and Camelliti, 2018). The heart

Figure 1. Principle cell types found in the mammalian heart

also possesses an intrinsic cardiac nervous system (ICNS) that controls heart 1

rate, atrial and ventricular refractoriness, cardiac contractility and conduction and blood flow (Zipes et al., 2017). The ICNS is composed of sensory (afferent), interconnecting (local circuit), and motor (adrenergic and cholinergic efferent) neurons that, communicating with intrathoracic extracardiac ganglia, control cardiac function under the influence of the central nervous system (CNS) and circulating catecholamines (Zipes et al., 2017). Cardiac ganglia are located at specific atrial regions: around the SAN, roots of caval and pulmonary veins, and near the AVN (Armour et al., 1997; Zipes et al., 2017); they are also found scattered through the ventricles, although in a smaller number compared to the atria (Armour et al., 1997; Zipes and Jalife, 2013).

These distinct cell populations are not isolated from one another within the heart but instead communicate physically via a variety of soluble endocrine, paracrine and autocrine factors. The principle factors identified, mostly from transgenic or knockout model systems in the mouse, are summarized in Figure 2 (simplified from (Tirziu et al., 2010)).

Figure 2. Schematic representation of cell-to-cell interactions in the mammalian heart

This cell-to-cell interaction and cross-talk contributes to structural, electrical, mechanical, and metabolic properties of the functional heart.

In the human heart, dialogue between CMs, cardiac ECs, and CFs is essential

to ensure proper cardiac function: on the one hand, cardiac ECs supply

oxygen and free fatty acids to the CMs and release paracrine factors (NRG-

1, PDGF-B, NO, ET-1) that support CM metabolism, survival and contractile

function (Brutsaert, 2003; Tirziu et al., 2010); on the other hand, CFs produce

extracellular matrix, which promotes tissue and matrix stretch, and form direct gap junctions with the CMs, influencing their electrophysiological properties and providing a substrate for electrical conduction between separated CMs over extended distances (Kakkar and Lee, 2010). Thus, to develop a model that fully recapitulates the complexity of this organ, the influence of these specific features of cell-to-cell cross talk needs to be taken into account.

Miniaturized heart tissues for disease modeling: using hiPSCs to study multilineage cardiovascular disorders

It is well established that human induced pluripotent stem cell derived- cardiomyocytes (hPSC-CMs) have great potential for studying cardiac developmental processes and mechanisms underlying cardiac disorders, as they capture the genetic background of the patient from which they were originated and incorporate aspects of the normal and pathological physiology of the heart tissue. However, one of their drawbacks that they also have in common with other differentiated derivatives of hPSCs, is their developmental immaturity, as they show a much closer resemblance to first- and second trimester human embryos than to adult tissues (Pavlovic et al., 2018).

In addition, other cardiac-specific cell types such as those described earlier, are essential for studying complex multilineage cardiovascular diseases in which not only the CMs but also other cardiac-specific cell types may be affected and play a role. For many diseases of the heart, it is still unclear what the cell- or tissue of origin for the disease is. One example is arrhythmogenic cardiomyopathy (ACM), a rare genetic disease predominately associated with mutations in desmosomal genes (PKG and PKP2 among others) and characterized by arrhythmias and fibro-fatty replacement of the myocardium (Lazzarini et al., 2015; Sommariva et al., 2017). The origin of ACM, particularly the fibro-fatty deposits in the heart, is still largely unknown. In vitro models (mouse and human; primary and hPSC- derived) investigated so far have been derived from either the CM- or the non- CM- (bucca mucosal cells, fibro-adipocytes progenitors, human embryonic kidney 293 cells, cardiac fibroblasts) compartment (Caspi et al., 2013; Cerrone et al., 2014;

El-Battrawy et al., 2018; Kim et al., 2014; Sommariva et al., 2016). Mouse models

have greatly enhanced our understanding of ACM and of the desmosome, such

as that PKG and PKP2 deficiency causes heart rupture in cardiac development,

and that PKP2 mutations are shown to cause gap junction remodeling (Awad et

al., 2008). Nevertheless, mouse models did not reveal the extensive cardiac fibro- 1

fatty deposits typical of ACM patients (Cerrone et al., 2012; Krusche et al., 2011), thus suggesting that human (cell-based) models might be better for investigating causative pathways linked to lipid metabolism. In addition, multicellular models that incorporate both the CM- and non-CM compartment will be extremely valuable in distinguishing the “culprit” cells from their “victims” and in understanding how these compartments singularly or synergistically contribute to different aspects of a complex disease pathogenesis in conditions such as ACM.

Other examples include (rare) metabolic syndromes and mitochondrial diseases that are often associated with young-onset cardiac failure or even death, yet which cell type in the heart causes the condition is still often unclear. Given the important (metabolic) roles of CFs and cardiac ECs in the heart, their contribution needs to be taken into account.

In addition, it has been proposed that in vitro three-dimensional (3D) culture systems better recapitulate the complexity of natural tissues compared to two- dimensional (2D) cultures, and, specifically for the heart, better mimic the real myocardial environment rather than 2D cultures on plastic (McDonald et al., 1972;

Veerman et al., 2015)

(Laschke and Menger, 2017). As an example, the development of 3D tissues from hiPSC-CMs carrying PRKAG2 cardiomyopathy revealed key links between metabolic sensing by AMPK and CM survival, metabolism and TGFb signalling that were not observed previously in 2D cultures of hiPSC-CMs (Hinson et al., 2016).

Thus, the development of miniaturized 3D multicellular models of the heart could allow in-depth mechanistic assessments of the nature of the disease, the identification of potential drug targets, as well as tissue-level validation of the effect of novel therapeutic compounds.

The need of culture maturation systems for cardiac disease modeling

As mentioned earlier, a largely unresolved issue in using hiPSC-CMs as preclinical

cardiac disease models and drug discovery platforms is their immature state

and similarity to fetal- rather than adult cells (reviewed in (Veerman et al.,

2015)). This presents hurdles to using these models for studying adult-onset

cardiac genetic diseases in which expression of the gene (or splice variant) of interest only occurs during postnatal heart development. Two examples are ion channel-related diseases, one caused by an imprinted gene (KCNQ1), the other by a postnatally expressed splice variant (SCN5A).

Mutations in the KCNQ1 gene, a voltage-gated potassium channel, are associated with congenital long QT syndrome type 1 (LQT1), a cardiac disorder associated with severe cardiac arrhythmias which causes sudden death especially in young individuals (Bokil et al., 2010). The KCNQ1 gene is initially imprinted (expressed from only one allele) but becomes bi-allelic (loss of imprinting) during postnatal heart development (Korostowski et al., 2011). However, this imprinting is not lost in fetal- and immature hiPSC-CMs.

The cardiac sodium channel Nav1.5, encoded by SCN5A, mediates the cardiac sodium current (I

) crucial for the rapid depolarization of CM action potential and impulse propagation in the heart (Gellens et al., 1992). Mutations in SCN5A have been associated with a broad spectrum of inherited cardiac rhythm disorders, such as long QT syndrome type 3 (LQT3), Brugada syndrome (BrS), and cardiac conduction disease (CCD) (reviewed in (Zimmer and Surber, 2008)).

Several SCN5A splice variants are expressed in the heart and in various other tissues including brain, dorsal root ganglia, breast cancer cells and neuronal stem cell lines (reviewed in (Schroeter et al., 2010)). Particularly in the heart,

“fetal” and “adult” splice variants have been described (Chioni et al., 2005). The

“adult” isoform differs from the“ fetal” SCN5A isoform in the alternate usage of exon 6: splicing of exon 6 occurs in a mutually exclusive manner, with inclusion of either the adult exon 6b or the fetal exon 6a (Chioni et al., 2005). The fetal splice isoform of SCN5A is predominantly expressed before birth and is gradually replaced by the adult isoform postnatally. As a result, “adult” splicing variants are not expressed in fetal- and immature hiPSC-CFs.

Based on these considerations, the development of a culture system for

maturation in which the CMs resemble those in postnatal heart is required to

reveal aspects of these and other adult-onset disease phenotypes that have

not been yet characterized.

Aim and outline of the thesis 1

Using a building-block approach, this thesis describes stepwise progression in the complex interactions between distinct cell populations present in the human heart. The overall aim of this thesis was to induce structural, electrical, mechanical and metabolic maturation of hiPSC-CMs by 1) developing a miniaturized 3D cardiac model using small cell numbers and to make it cost effective in scaling production and thus amenable to screening compound libraries, and 2) developing a multi-cell type cardiac model using multiple cardiac cell types, to make it possible to distinguish “culprit” cells from their

“victims” in complex multi-lineage cardiovascular disorders.

In chapter 2, we reviewed the development of CM induction from human pluripotent stem cells (hPSCs), the progress in cardiac disease modeling using hiPSC-CMs, and the challenges associated with understanding complex cardiac diseases.

In chapter 3, we established conditions for simultaneous differentiation of CMs and cardiac-specific ECs from common hPSC-cardiac mesoderm progenitors.

We then described the development of a bi-culture system, termed “cardiac microtissue”, that integrates both cell types in 3D, using just 5000 cells per tissue.

Finally, we demonstrated that presence of cardiac ECs was essential to induce CM maturation in this system.

In chapter 4, we firstly described in depth the protocol for co-differentiation of CMs and cardiac ECs from cardiac mesoderm using both human embryonic stem cells (hESCs) and hiPSCs, and provided details for the enrichment of both cell populations from heterogeneous-differentiated cultures as well as cell maintenance, characterization, dissociation and cryopreservation. Secondly, we described the detailed bench protocol for generation of cardiac microtissues, and we provided guidelines for their culture and characterization for downstream applications.

In chapter 5, by adapting the protocol we developed in chapter 3 for simultaneous

differentiation of CMs and cardiac ECs from common cardiac mesoderm, we

found that RA and BMP4 synergistically promote the formation of epicardial cells

in both hESCs and hiPSCs. As epicardial cells have the ability to undergo EMT and

give rise to CFs, this work provided the foundation for chapter 6.

In chapter 6, we established a 3D triple-cell type model of cardiac microtissue by adding hiPSC-CFs (derived from hiPSC-epicardium) to the bi-cell type culture model developed in chapter 3. We demonstrated that inclusion of CFs was crucial for inducing (post-natal) structural, electrical, mechanical and metabolic maturation of hiPSC-CMs in microtissues containing hiPSC-ECs. We found that primary adult cardiac- but not skin fibroblasts could replace hiPSC- CFs, and that three-cell-type crosstalk between the three major cardiac cell components was essential for metabolic maturation of hiPSC-CMs. Finally, we showed that features of the arrhythmogenic cardiomyopathy (ACM) phenotype were uniquely recapitulated by inclusion of patient hiPSC-derived cardiac fibroblasts. This work provided proof of concept that tissue- and organ- specific cells differentiated from hiPSCs following developmental principles are essential to 1) mediate maturation and 2) uncover mechanistic insights in multi-lineage disease phenotypes, specifically in the heart.

In chapter 7, we discussed the significance of findings presented in this thesis

and limitations and advantages of our model compared to existing systems

developed by other groups. Scope for future work in the field is also proposed.

References 1

Armour, J.A., Murphy, D.A., Yuan, B.-X., MacDonald, S., Hopkins, D.A., 1997. Gross and microscopic anatomy of the human intrinsic cardiac nervous system. Anat. Rec. 247, 289–298. doi:10.1002/

(SICI)1097-0185(199702)247:2<289::AID-AR15>3.0.CO;2-L

Awad, M.M., Calkins, H., Judge, D.P., 2008. Mechanisms of disease: molecular genetics of arrhythmogenic right ventricular dysplasia/cardiomyopathy. Nat Clin Pract Cardiovasc Med 5, 258–

267. doi:10.1038/ncpcardio1182

Bokil, N.J., Baisden, J.M., Radford, D.J., Summers, K.M., 2010. Molecular genetics of long QT syndrome.

Molecular Genetics and Metabolism 101, 1–8. doi:10.1016/j.ymgme.2010.05.011

Brade, T., Pane, L.S., Moretti, A., Chien, K.R., Laugwitz, K.L., 2013. Embryonic heart progenitors and cardiogenesis. Cold Spring Harbor Perspectives in Medicine 3, a013847–a013847. doi:10.1101/

cshperspect.a013847

Brutsaert, D.L., 2003. Cardiac Endothelial-Myocardial Signaling: Its Role in Cardiac Growth, Contractile Performance, and Rhythmicity. Physiological Reviews 83, 59–115. doi:10.1152/physrev.00017.2002 Caspi, O., Huber, I., Gepstein, A., Arbel, G., Maizels, L., Boulos, M., Gepstein, L., 2013. Modeling of

Arrhythmogenic Right Ventricular Cardiomyopathy With Human Induced Pluripotent Stem Cells.

Circulation: Cardiovascular Genetics 6, 557–568. doi:10.1161/CIRCGENETICS.113.000188

Cerrone, M., Lin, X., Zhang, M., Agullo-Pascual, E., Pfenniger, A., Chkourko Gusky, H., Novelli, V., Kim, C., Tirasawadichai, T., Judge, D.P., Rothenberg, E., Chen, H.S.V., Napolitano, C., Priori, S.G., Delmar, M., 2014.

Missense Mutations in Plakophilin-2 Cause Sodium Current Deficit and Associate With a Brugada Syndrome Phenotype. Circulation 129, 1092–1103. doi:10.1161/CIRCULATIONAHA.113.003077 Cerrone, M., Noorman, M., Lin, X., Chkourko, H., Liang, F.-X., van der Nagel, R., Hund, T., Birchmeier,

W., Mohler, P., van Veen, T.A., van Rijen, H.V., Delmar, M., 2012. Sodium current deficit and arrhythmogenesis in a murine model of plakophilin-2 haploinsufficiency. Cardiovascular Research 95, 460–468. doi:10.1093/cvr/cvs218

Chioni, A.-M., Fraser, S.P., Pani, F., Foran, P., Wilkin, G.P., Diss, J.K.J., Djamgoz, M.B.A., 2005. A novel polyclonal antibody specific for the Nav1.5 voltage-gated Na+ channel “neonatal” splice form.

Journal of Neuroscience Methods 147, 88–98. doi:10.1016/j.jneumeth.2005.03.010

El-Battrawy, I., Zhao, Z., Lan, H., Cyganek, L., Tombers, C., Li, X., Buljubasic, F., Lang, S., Tiburcy, M., Zimmermann, W.-H., Utikal, J., Wieland, T., Borggrefe, M., Zhou, X.-B., Akin, I., 2018. Electrical dysfunctions in human-induced pluripotent stem cell-derived cardiomyocytes from a patient with an arrhythmogenic right ventricular cardiomyopathy. EP Europace 20, f46–f56. doi:10.1093/

europace/euy042

Furtado, M.B., Nim, H.T., Boyd, S.E., Rosenthal, N.A., 2016. View from the heart: cardiac fibroblasts in development, scarring and regeneration. Development 143, 387–397. doi:10.1242/dev.120576 Gellens, M.E., George, A.L., Chen, L.Q., Chahine, M., Horn, R., Barchi, R.L., Kallen, R.G., 1992. Primary

structure and functional expression of the human cardiac tetrodotoxin-insensitive voltage- dependent sodium channel. Proc Natl Acad Sci USA 89, 554–558.

Harari, E., Guo, L., Smith, S.L., Braumann, R.E., Virmani, R., Finn, A.V., 2017. Heart-resident macrophages:

are they involved in the rhythm of every beat? J Thorac Dis 9, 2264–2267. doi:10.21037/jtd.2017.07.43 Hinson, J.T., Chopra, A., Lowe, A., Sheng, C.C., Gupta, R.M., Kuppusamy, R., O’Sullivan, J., Rowe, G.,

Wakimoto, H., Gorham, J., Zhang, K., Musunuru, K., Gerszten, R.E., Wu, S.M., Chen, C.S., Seidman, J.G., Seidman, C.E., 2016. Integrative Analysis of PRKAG2 Cardiomyopathy iPS and Microtissue Models Identifies AMPK as a Regulator of Metabolism, Survival, and Fibrosis. CellReports 17, 3292–3304.

doi:10.1016/j.celrep.2016.11.066

Hulsmans, M., Clauss, S., Xiao, L., Aguirre, A.D., King, K.R., Hanley, A., Hucker, W.J., Wülfers, E.M., Seemann, G., Courties, G., Iwamoto, Y., Sun, Y., Savol, A.J., Sager, H.B., Lavine, K.J., Fishbein, G.A., Capen, D.E., Da Silva, N., Miquerol, L., Wakimoto, H., Seidman, C.E., Seidman, J.G., Sadreyev, R.I., Naxerova, K., Mitchell, R.N., Brown, D., Libby, P., Weissleder, R., Swirski, F.K., Kohl, P., Vinegoni, C., Milan, D.J., Ellinor, P.T., Nahrendorf, M., 2017. Macrophages Facilitate Electrical Conduction in the Heart. Cell 169, 510–522.e20. doi:10.1016/j.cell.2017.03.050

Johnson, R.D., Camelliti, P., 2018. Role of Non-Myocyte Gap Junctions and Connexin Hemichannels in Cardiovascular Health and Disease: Novel Therapeutic Targets? IJMS 19, 866. doi:10.3390/

ijms19030866

Kakkar, R., Lee, R.T., 2010. Intramyocardial Fibroblast Myocyte Communication. Circulation Research 106, 47–57. doi:10.1161/CIRCRESAHA.109.207456

Kim, C., Wong, J., Wen, J., Wang, S., Wang, C., Spiering, S., Kan, N.G., Forcales, S., Puri, P.L., Leone, T.C., Marine, J.E., Calkins, H., Kelly, D.P., Judge, D.P., Chen, H.-S.V., 2014. Studying arrhythmogenic right ventricular dysplasia with patient-specific iPSCs. Nature 494, 105–110. doi:10.1038/nature11799 Korostowski, L., Raval, A., Breuer, G., Engel, N., 2011. Enhancer-driven chromatin interactions during

development promote escape from silencing by a long non-coding RNA. Epigenetics Chromatin 4, 21. doi:10.1186/1756-8935-4-21

Krusche, C.A., Holthöfer, B., Hofe, V., van de Sandt, A.M., Eshkind, L., Bockamp, E., Merx, M.W., Kant, S., Windoffer, R., Leube, R.E., 2011. Desmoglein 2 mutant mice develop cardiac fibrosis and dilation.

Basic Res Cardiol 106, 617–633. doi:10.1007/s00395-011-0175-y

Laschke, M.W., Menger, M.D., 2017. Life is 3D: Boosting Spheroid Function for Tissue Engineering.

Trends Biotechnol. 35, 133–144. doi:10.1016/j.tibtech.2016.08.004

Lazzarini, E., Jongbloed, J.D.H., Pilichou, K., Thiene, G., Basso, C., Bikker, H., Charbon, B., Swertz, M., van Tintelen, J.P., van der Zwaag, P.A., 2015. The ARVD/C Genetic Variants Database: 2014 Update.

Human Mutation 36, 403–410. doi:10.1002/humu.22765

McDonald, T.F., Sachs, H.G., DeHaan, R.L., 1972. Development of sensitivity to tetrodotoxin in beating chick embryo hearts, single cells, and aggregates. Science 176, 1248–1250.

Moorman, A.F.M., Christoffels, V.M., 2003. Development of the cardiac conduction system: a matter of chamber development. Novartis Found. Symp. 250, 25–34– discussion 34–43– 276–9.

Pavlovic, B.J., Blake, L.E., Roux, J., Chavarria, C., Gilad, Y., 2018. A Comparative Assessment of Human and Chimpanzee iPSC-derived Cardiomyocytes with Primary Heart Tissues. Sci Rep 8, 15312.

doi:10.1038/s41598-018-33478-9

Schroeter, A., Walzik, S., Blechschmidt, S., Haufe, V., Benndorf, K., Zimmer, T., 2010. Structure and function of splice variants of the cardiac voltage-gated sodium channel Nav1.5. Journal of Molecular and Cellular Cardiology 49, 16–24. doi:10.1016/j.yjmcc.2010.04.004

Sommariva, E., Brambilla, S., Carbucicchio, C., Gambini, E., Meraviglia, V., Russo, Dello, A., Farina, F.M., Casella, M., Catto, V., Pontone, G., Chiesa, M., Stadiotti, I., Cogliati, E., Paolin, A., Ouali Alami, N., Preziuso, C., d’Amati, G., Colombo, G.I., Rossini, A., Capogrossi, M.C., Tondo, C., Pompilio, G., 2016.

Cardiac mesenchymal stromal cells are a source of adipocytes in arrhythmogenic cardiomyopathy.

European Heart Journal 37, 1835–1846. doi:10.1093/eurheartj/ehv579

Sommariva, E., Stadiotti, I., Perrucci, G.L., Tondo, C., Pompilio, G., 2017. Cell models of arrhythmogenic cardiomyopathy: advances and opportunities. Disease Models & Mechanisms 10, 823–835.

doi:10.1242/dmm.029363

Thiriet, M., 2007. Biology and Mechanics of Blood Flows. Springer Science & Business Media, New York, NY. doi:10.1007/978-0-387-74849-8

Tirziu, D., Giordano, F.J., Simons, M., 2010. Cell Communications in the Heart. Circulation 122, 928–937.

doi:10.1161/CIRCULATIONAHA.108.847731

Veerman, C.C., Kosmidis, G., Mummery, C.L., Casini, S., Verkerk, A.O., Bellin, M., 2015. Immaturity of Human Stem-Cell-Derived Cardiomyocytes in Culture: Fatal Flaw or Soluble Problem? Stem Cells and Development 24, 1035–1052. doi:10.1089/scd.2014.0533

Xin, M., Olson, E.N., Bassel-Duby, R., 2013. Mending broken hearts: cardiac development as a basis

1

for adult heart regeneration and repair. Nat. Rev. Mol. Cell Biol. 14, 529–541. doi:10.1038/nrm3619 Zimmer, T., Surber, R., 2008. SCN5A channelopathies – An update on mutations and mechanisms.

Progress in Biophysics and Molecular Biology 98, 120–136. doi:10.1016/j.pbiomolbio.2008.10.005 Zipes, D.P., Jalife, J., 2013. Cardiac Electrophysiology: from Cell to Bedside. Elsevier Health Sciences.

2

CHAPTER 2

Human heart disease: lessons from human pluripotent stem cell-derived cardiomyocytes

Elisa Giacomelli¹, Christine L Mummery^1,2, Milena Bellin¹

1 Department of Anatomy and Embryology, Leiden University Medical Center, Einthovenweg 20, 2333 ZC Leiden, The Netherlands

2 Department of Applied Stem Cell Technologies, University of Twente, Building Zuidhorst, 7500 AE Enschede, The Netherlands

Published in Cellular and Molecular Life Sciences; 2017 Oct; 74(20):3711-3739. doi: 10.1007/s00018- 017-2546-5. Epub 2017 Jun 1.

Abstract

Technical advances in generating and phenotyping cardiomyocytes from

human pluripotent stem cells (hPSC-CMs) are now driving their wider

acceptance as in vitro models to understand human heart disease and discover

therapeutic targets that may lead to new compounds for clinical use. Current

literature clearly shows that hPSC-CMs recapitulate many molecular, cellular,

and functional aspects of human heart pathophysiology and their responses to

cardioactive drugs. Here, we provide a comprehensive overview of hPSC-CMs

models that have been described to date and highlight their most recent and

remarkable contributions to research on cardiovascular diseases and disorders

with cardiac traits. We conclude discussing immediate challenges, limitations

and emerging solutions.

2 Introduction

Human embryonic stem cells, derived from early human embryos, and human induced pluripotent stem cells, derived by reprogramming somatic cells, (hESCs and hiPSCs, respectively, and collectively called hPSCs) can self-renew and differentiate into all cell types of the human body, including cardiomyocytes (Kehat et al., 2001; Takahashi et al., 2007; Thomson, 1998). They have potential applications in regenerative medicine but are also becoming a useful tool in cardiovascular research. Most particularly, they offer new opportunities to develop in vitro models of human cardiac development and cardiovascular diseases, as they are able to capture much of the normal and pathological physiology of the human heart, including aspects of congenital defects. In addition, hPSC-derived cardiomyocytes (hPSC-CMs) may be used in cardiac safety pharmacology, drug screening and drug discovery, to predict the effects of candidate drugs and new compounds and to identify key target pathways in disease. Whilst hESCs can now readily be engineered to carry specific disease mutations, the derivation of hiPSCs from virtually any patient of interest offers some advantages over hESCs for disease modeling, since hiPSCs incorporate individual complex genetic backgrounds of the patients from which they were originated. For this reason, expectations are high on their contribution to precision medicine where the goal is to prevent disease development and find personalized treatments that take genetic variability of patients into account (Collins and Varmus, 2015).

In this review we provide comprehensive coverage of hPSC models of human heart disease.

Generation of hiPSCs and hESCs for cardiac disease modelling

The need for more robust cell models for human disease, including cardiovascular disorders, has led to increasing interest in hPSCs.

hESCs were the first human pluripotent stem cells described. They were derived

from the inner cell mass of blastocyst-stage embryos in 1998 by Thomson

(Thomson, 1998). These cells could differentiate toward cell lineages of all three

germ layers yet be maintained in a state of self-renewal indefinitely in their

undifferentiated state. Multiple hESC lines have been used successfully for studying

genetic disorders most often through specific gene knockdown or deletion using homologous recombination (Urbach, 2004) or lentiviral transduction (Tulpule and Daley, 2009). Furthermore, in the case of some potentially fatal or untreatable conditions, hESCs have also been derived from preimplantation embryos genetically diagnosed as defective by single blastomere sampling during Preimplantation Genetic Diagnosis (PGD). Disorders that have been studied using PGD-hESC include a number of severe congenital disorders such as fragile X syndrome (Eiges et al., 2007), Turner syndrome (Urbach and Benvenisty, 2009) and trisomy 21 (Bittles et al., 2007). However, hESC lines to investigate multifactorial and complex diseases may not be available through PGD because they may not be considered sufficiently severe (Lengerke and Daley, 2009), as is the case for many cardiac diseases. Thus, even though hESCs are useful when there is pre-existing knowledge on the specific mutations causing the disease and the mutations can be introduced into an otherwise healthy line, hiPSC are preferred where the entire genetic background is relevant.

The use of patient somatic cells to derive hiPSC is also preferable in some countries since it circumvents ethical issues that surround the destruction of human embryos for research purposes. The advent of hiPSCs has also superseded efforts to derive cloned embryos by somatic cell nuclear transfer and isolate individual hESC lines from them (Bellin et al., 2012; Takahashi et al., 2007; Yu et al., 2007). Many methods have now been described that allow somatic cell reprogramming (Raab et al., 2014).

The first and still among the most efficient methods described overexpresses the

reprogramming factors c-MYC, SOX2, KLF4 and OCT3/4 after retroviral or lentiviral

transduction of dermal fibroblasts. This results in the integration of reprogramming

genes into the genome and subsequent reactivation of the endogenous

counterparts (Takahashi and Yamanaka, 2006; Yu et al., 2007). Alternative non-

integrating reprogramming methods are now more widely used and include the

use of Sendai viruses (Ban et al., 2011), plasmids (Okita et al., 2008), and modified

RNA (Warren et al., 2010). Small molecules have also been used but have relatively

lower efficiencies (Huangfu et al., 2008). Somatic cell sources currently used for

reprogramming not only include the original dermal fibroblasts isolated from

skin biopsies, but also blood cells (Loh et al., 2009), keratinocytes from plucked

hair (Aasen et al., 2008), and exfoliated renal tubular epithelial cells obtained from

urine (Zhou et al., 2011). Many patient-specific lines have been described that are

suitable for cardiovascular disease modeling and are proving of particular value for

studying disorders of unknown or complex genetic origin, as will be discussed in

this review.

2 Differentiation into cardiomyocytes

In vitro differentiation of hPSCs into cardiomyocytes mimics the sequential stages of embryonic cardiac development (Mummery et al., 2012). In the vertebrate embryo, the heart is one of the first organs to develop; after gastrulation, anterior migrating mesodermal cells intercalate between the ectoderm and the endoderm germ layers in the primitive streak to start generating the heart (KAUFMAN and Navaratnam, 1981; Kirby and Waldo, 2002). Cardiac progenitor cells derive from two small tracts of epiblast cells of the developing primitive streak and take residence in the lateral plate mesoderm (Abu-Issa and Kirby, 2007). Signals from the surrounding tissues, such as growth factors of the WNT, BMP, and TGF-β families are critical to promote the specification of myocardial fate. Accordingly, many of the successful protocols developed to induce cardiomyogenesis in hPSCs are based on activating and inhibiting these signaling pathways. As an example, stimulation of extraembryonic ectoderm via BMP signalling (by BMP4) and posterior primitive streak via WNT signalling (by CHIR99201) during the first 24 h of differentiation promotes the exit from self-renewal and the induction of cardiac mesoderm (Rao et al., 2015). Moreover, inhibitors of WNT signalling, such as IWR-1, IWP-3 and XAV939, have been shown to induce cardiogenesis when added after mesoderm formation (Elliott et al., 2011; Karakikes et al., 2014; Willems et al., 2011), while SB-431542, an inhibitor of the TGF-β pathway, promotes cardiogenesis when its addition occurs after mesoderm specification (Acimovic et al., 2014). Current methods for cardiac differentiation of hPSCs rely on three different approaches that are summarized in Table 1, embryoid body formation, co-cultures, and monolayer culture (Mummery et al., 2012).

Functional cardiomyocytes can be generated from hPSCs as three-dimensional spheroid-like aggregates termed embryoid bodies (EBs), referring to their similarity with early post implantation embryos. Protocols to form EBs were originally developed using fetal bovine serum supplemented culture medium but a variety of serum-free, defined media formulations are now available.

Methods to form EBs from hPSCs range from an enzymatic partial dissociation of

hPSC colonies, to precise control of cell number and size by forced aggregation

in microwells, to microwells in which hPSC colonies are first expanded to a

defined size, to micropatterned substrates (Mummery et al., 2012).

Differentiation Culture conditions Limits Efficiency¹ Reference EBs Serum-based media Low efficiency

Serum media

5-15% (Kehat et al., 2001)

RPMI + B27 supplement ActivinA + BMP4

Medium efficiency Batch-to-batch variability of growth factors

Chemically undefined

“B27”

60% (Kattman et al., 2011)

Bioreactor suspension culture RPMI + B27 supplement Small molecules

Chemical undefined

“B27”

90% (V. C. Chen et al., 2012)

Inductive co- culture

Serum-based media Feeder layer Mouse END-2 cells

Low efficiency Serum media Requirement for mouse feeder cells

35% (Mummery et al., 2012)

Monolayer culture

RPMI + B27 supplement ActivinA + BMP4

Low efficiency Batch-to-batch variability of growth factors

Chemically undefined

“B27”

35% (Laflamme et al., 2007)

RPMI + B27 supplement Matrigel Sandwich ActivinA + BMP4

Batch-to-batch variability of Matrigel and growth factors Chemically undefined

“B27”

90% (J. Zhang et al., 2012)

RPMI + B27 supplement Small molecules

Chemically undefined

“B27”

90% (Lian et al., 2012)

RPMI + human albumin L-ascorbic acid 2-phosphate Small molecules Na⁺ lactate

85%

95%

(Burridge et al., 2014)

Table 1. Methods for differentiating hPSCs into cardiomyocytes. (Modified from (Mathur et al., 2015))

1 Efficiency was calculated from flow cytometry data as the number of cells positive for cardiac troponin T (cTnT), MLC-2a and MLC-2v, by immunostaining for MHC-b or by determining the percentage of EBs containing contracting areas.

2

Alternatively, early studies also used inductive co-culture of mechanically passaged hESCs with visceral endodermal-like END2 cells derived from mouse P19 embryonal carcinoma cells (Mummery et al., 2002). Notably, visceral endoderm plays a key role in the induction of cardiogenic precursors cells in development.

For ease of use though, monolayer differentiation protocols have been preferred.

Benefits compared to the EB and co-culture systems include higher efficiencies and easy monitoring of outcome. Refinements over the last decade now support the generation of differentiated cell populations containing 85% cardiomyocytes;

multiple methods have been described in which cardiomyocytes can be enriched to 95% using for example selection in sodium (Na

) lactate containing medium (Burridge et al., 2014; Mathur et al., 2015) or on the basis of cell surface markers like SIRPA and VCAM1 (Dubois et al., 2011; Elliott et al., 2011).

Cardiomyocytes derived under all these culture conditions beat spontaneously, express sarcomeric proteins and ion channels, and exhibit cardiac-type action potentials (APs) and calcium (Ca

) transients. Furthermore, they show similar functional properties to the cardiomyocytes in the developing heart, such as comparable dose-dependent response to cardiac drugs in terms of beating frequency and contractility, b-adrenoreceptor responses, action potential (AP) morphologies and excitation-contraction coupling mechanisms (Khan et al., 2013). Although opportunities still remain for improvement of reproducibility in cardiac differentiation between individual hPSC lines, reduction in the cost of reagents and in batch-to-batch variability, and of the yield and purity of required cardiomyocyte types, several protocols now support robust cardiac differentiation and some of these are available commercially as kits.

Characterization of cardiomyocyte phenotype

The use of hPSC-CMs as a platform to model cardiovascular disorders requires

their rigorous molecular and functional characterization. In order to maximize

their potential applications in cardiovascular medicine, a qualitative comparison

with adult (or fetal) primary human cardiomyocytes is advisable. Parameters used

to characterize the cardiomyocyte phenotype are listed in Table 2 and include

size and morphology, sarcomere structure, electrophysiological properties,

Ca

handling and contractile force, responses to b-adrenergic stimulation,

mitochondrial function and metabolic profile, and conduction velocity.

Features Measured parameters Human adult cardiomyocyte Size and

Morphology

Shape (rod, round) Size (mm)

Cell capacitance (pF)

Elongated Rod shaped

~65 % mononucleated

Sarcomeres Alignment

Organization (Z lines, H zone, I bands, A bands)

Molecular composition (MYH7:MYH6, MYL2:MYL7, TNNI3:TNNI1)

Organized and aligned

MHC-b predominant isoform in the ventricle

MLC-2a predominant isoform in the atrium

Electrophysiological properties

AP (APA, RMP, Vmax, APD) Ion current densities and gating properties (I_Na, I_CaL, I_CaT, I_to, I_Kur, I_Kr, I_Ks, I_K1, I_K,Ach, I_K,ATP, I_f)

Typical atrial, ventricular, pacemaker, and Purkinje AP shapes (Antzelevitch and Dumaine, 2011; Nerbonne, 2005) Distinct ion current densities and function in atrial, ventricular, pacemaker, and Purkinje cardiomyocytes (Antzelevitch and Dumaine, 2011; Nerbonne, 2005) Ca²⁺ handling and

contractile force

Ca²⁺ transients Force of contraction Ca²⁺ sparks and Ca²⁺ waves

Efficient Ca²⁺ transient induction by Ca²⁺

influx through L-type Ca²⁺ channels (Ca²⁺- induced Ca²⁺-release) (Bers, 2002) 10-50 mN/mm² (ventricular myocytes) (Mulieri et al., 1992)

Positive force-frequency relationship (Bowditch phenomenon) (Wiegerinck et al., 2009)

Low rate of spontaneous Ca²⁺ release Response to

b-adrenergic stimulation (cascade of events)

Chronotropic effect Inotropic effect Lusitropic effect

Positive chronotropic, inotropic and lusitropic effects

Mitochondrial function and metabolic profile

Oxygen consumption Glycolysis and ATP measurements

Mitochondrial membrane potential

Redox state

Intramitochondrial pH ROS generation

Mitochondria occupies one third of the total volume of CMs

ATP production occurs mainly through oxidative metabolism of fatty acids

Conduction velocity Conduction velocity maps Expression level of ion channels and gap junction proteins Localization, density and composition of gap junction proteins

Generation of the electrical signal through Na⁺ channels and propagation through gap junctions

Localization of gap junction proteins at cell borders

Table 2. Key features and assays used to characterize the human cardiomyocyte phenotype.

2 Size and morphology

In the adult heart, cardiomyocytes are elongated and rod shaped, and ~65 % of them are mononucleated and this percentage does not change significantly throughout life (Bird, 2003; Mollova et al., 2013). Further, adult cardiomyocytes align longitudinally in the heart and are connected by intercalated discs that facilitate the electrical conduction and muscle contraction (Peters et al., 1993).

To date, despite the high differentiation efficiencies now achievable, hPSC- CMs remain small in size and round in shape (Snir et al., 2003) suggesting an immature or fetal phenotype. Several strategies have been used to mature hPSC-CMs. These include prolonged time in culture (>50 days), where hPSC- CMs become more elongated and less rounded (Snir et al., 2003) and advanced engineering approaches such as 3D platforms, either as “biowires”, or EHTs (engineered heart tissues), which allow the generation of hiPSC-CMs with improved ultrastructural and electrophysiological properties (Mannhardt et al., 2016; Nunes et al., 2013). Examples of improved ultrastructural properties included cardiomyocyte anisotropy with Z bands frequently visible and aligned, pronounced presence of H zones and I bands, scattered presence of T-tubule- like structures (Mannhardt et al., 2016; Nunes et al., 2013). These methods as well as other maturation strategies are summarized in the “Conclusions” section of this review.

Sarcomere structure

Human adult cardiomyocytes are characterized by organized and aligned sarcomeres (Bird, 2003), the smallest contractile units of striated muscles.

Sarcomeres are composed of contractile proteins, including actin and myosin,

which generate the force of contraction, and thin filament proteins, which

calibrate the force generated by contractile proteins. In the adult ventricle, the

b isoform of the protein Myosin Heavy Chain (MHC-b), encoded by the gene

MYH7, is predominant compared to the atrial a isoform MHC-a, encoded by

MYH6 (Reiser et al., 2001); in addition, the isoform Myosin Light Chain 2v (MLC-

2v), encoded by the gene MYL2, is predominant compared with the MLC-2a,

encoded by MYL7, which is instead the primary human atrial isoform. Similarly,

a genetic switch between the troponin I fetal (TNNI1) and adult isoforms

(TNNI3) in the human heart characterizes the transition from fetal to post-natal

development (Bhavsar et al., 1991).

Sarcomeres in hPSC-CMs are less organized than in adult cardiomyocytes and MHC-a and MLC-2a are generally highly expressed, while MHC-b and MLC- 2v display relatively low level of expression (Xu et al., 2009). Also the TNNI1:

TNNI3 protein isoform ratio reflects a fetal stage, even after long term culture (Bedada et al., 2014). This is partly due to hPSC-CMs being more similar to fetal cardiomyocytes but also to the heterogeneous nature of the hPSC-CMs population, which consists on a mixture of ventricular-, atrial- and nodal-like cells. Recent engineering approaches have attempted to improve sarcomere organization and myofibril alignment in hPSC-CMs, to allow study of their structural and contractile properties, such as actin-myosin cross-bridge cycling, myofibril tension and kinetics of activation and relaxation. Examples include the work of Salick and colleagues in which hESC-CMs were seeded onto controlled 2-dimensional micropatterned rectangles made with high-resolution photolithography and microcontact printing (Salick et al., 2014), and the work of Pioner and colleagues in which hiPSC-CMs were seeded on nanogrooved surfaces and cultured long term (80-100 days) (Pioner et al., 2016). Importantly, the latter study demonstrated that myofibril tension and kinetics were similar between long-term cultures of hiPSC-CMs and second trimester human fetal ventricular cardiomyocytes. Importantly, the fetal sarcomeric properties of hPSC-CMs may represent an obstacle to faithfully recapitulating cardiomyopathy-associated phenotypes that are linked to sarcomere protein mutations. For example, the force of contraction was decreased in hiPSC-CMs with MYBPC3-mutations compared with wild-type cells, while HCM due to sarcomeric mutations is usually associated with hypercontractility (Birket et al., 2015; Spudich, 2014).

Electrophysiological properties

Electrophysiological properties of adult cardiomyocytes can be described by their AP profile, which is widely considered specific for each cardiomyocyte subtype (atrial, ventricular, pacemaker, and Purkinje). However, independent of subtype, AP always starts with a rapid influx of Na

as a rapid depolarizing current (I

), termed “AP upstroke” (phase 0). Afterwards, phase 1 of the AP is characterized by a transient repolarizing current (I

) of efflux of potassium (K

), followed by the inward Ca

current (I

) through the L-type depolarization- activated Ca

channels, which is called the plateau phase of the AP (phase 2).

Next, two K

currents (I

and I

) drive the repolarizing phase 3 of the AP. Hence,

in adult atrial and ventricular cardiomyocytes, the presence of a rectifying K

current (I

) stabilizes the resting membrane potential (RMP) at -85 mV; this is

termed phase 4 of the AP.

2

hPSC-CMs are more depolarized compared to adult cardiomyocytes: RMP is less negative (-50/-60mV), Na

channels are fewer and phase 0 of the AP is slow.

In addition, hPSC-CMs exhibit spontaneous contractile activity, due to the absence or very low expression of I

, and the presence of a funny current (I

), which is a pacemaker Na

/K

hyperpolarizing current (Bers, 2002; Veerman et al., 2015).

Despite the differences with adult cardiomyocytes (reviewed in (Hoekstra et al., 2012) and (Barbuti et al., 2016)), hPSC-CMs offer the opportunity to study some developmental- and disease-relevant cardiac properties. As an example, arrhythmogenic diseases of the heart have been successfully recapitulated using patient hiPSC-CMs, displaying significant AP changes, such as AP prolongation in the long-QT syndrome (Sinnecker et al., 2012). In addition, in 2013, the US Food and Drug Administration (FDA) chose hiPSC-CMs as cell type of choice for testing cardiac effects of novel compounds (Veerman et al., 2015).

Ca

handling and contractile force (excitation-contraction coupling) The process termed “excitation-contraction coupling” (ECC) consists of the repeated contraction and relaxation of the chambers of the heart, in which Ca

is perhaps the most important ion involved. Ca

that enters the cell during the plateau phase of the AP enhances Ca

release from the sarcoplasmic reticulum (SR) through ryanodine receptor (RYR2) channels. This causes an increase in intracellular Ca

, which binds to the myofilament protein troponin C, activating the mechanism of the contraction. For relaxation, Ca

instead dissociates from troponin C and leaves the cytosol through four different systems: SR Ca

- ATPase (SERCA2a); sarcolemmal Na

/Ca

exchange (NCX); sarcolemmal Ca

-ATPase; mitochondrial Ca

uniport (Bers, 2002). T-tubules are invaginations in the cell membrane located where L-type Ca

channels and RYR2 channels are close to each other and represent one of the most important components of the Ca

handling system, contributing to ECC (Ferrantini et al., 2013). To date, although hPSC-CMs express NCX at comparable levels of adult cardiomyocytes (Fu et al., 2011), the SR is still poorly developed and T-tubules have been rarely described. Consequently, Ca

handling kinetics, as well as ECC are overall slow in hPSC-CMs (Karakikes et al., 2015).

Responses to b-adrenergic stimulation

Sympathetic stimulation of the heart through b-adrenergic receptor agonists,

such as epinephrine, activates a membrane stimulatory GTP-binding protein,

which stimulates adenylyl cyclase to produce cyclic adenosine monophosphate (cAMP), which in turns leads to the subsequent activation of Protein Kinase A (PKA), therefore potentiating the cardiac Ca

transients. In response to b-adrenergic stimulation, adult cardiomyocytes display positive chronotropic (increase in beating frequency), positive inotropic (increase in contractility), and positive lusitropic (acceleration of relaxation) effects (Bers, 2002). Although hPSC-CMs as well as fetal cardiomyocytes do exhibit chronotropic responses to b-adrenergic stimulation (T. D. Chang and Cumming, 1972; Pillekamp et al., 2012), they do not show an increase in contraction, or acceleration in the relaxation period (Brito-Martins et al., 2008), unless when incorporated in human engineered heart tissues as shown by Mannhardt (Mannhardt et al., 2016). These considerations need to be taken into account when hPSC-CMs are used for testing the efficiency of b-adrenergic drugs on the cardiovascular system.

Mitochondrial function and metabolic profile

Due to its incessant contraction, the heart has an extremely high energy demand compared to other tissues of the human body (Bers, 2002).

Mitochondrial biogenesis increases over time during heart development, so that in adult cardiomyocytes one third of the cell volume is indeed occupied by mitochondria (Barth, 1992). Due to this change during development, glucose and lactate represent the predominant substrates for the majority of ATP production in fetal cardiomyocytes, while adult cardiomyocytes mainly use fatty acids (Lopaschuk et al., 1992; Lopaschuk and Jaswal, 2010). Although hPSC-CMs still display an immature phenotype, they also use fatty acids for the majority of ATP production and mitochondrial density increases over time, recapitulating to a certain extent the development of the human heart (Birket et al., 2013;

Veerman et al., 2015). For this reason, cardiomyocytes derived from hPSCs have been successfully used to recapitulate and study key aspects of mitochondrial and metabolic diseases in humans, as Drawnel and colleagues have recently showed by modelling diabetic cardiomyopathy and phenotypically screening drugs for a complication of type 2 diabetes (Drawnel et al., 2014).

Conduction velocity

While the parameters above can be evaluated in single cells, the conduction

velocity can only be measured in monolayer cultures. Major factors contribute

to determine the conduction velocity of cardiomyocytes: propagation of the

electrical signal through Na

channels (Kléber and Rudy, 2004); localization

2

of Na

channels and gap junction proteins (Jansen et al., 2010); localization, density and composition of gap junction proteins (S.-C. Chen et al., 2006); cell size (Wiegerinck, 2006). Although the composition of gap junction proteins is similar in hPSC-CMs and adult cardiomyocytes, Na

channels and gap junctions need to be distributed at the edges of two adjacent cells (adult cardiomyocytes) (Vreeker et al., 2014), rather than all around the cell circumference (fetal and hPSC-CMs). This, together with a reduced availability of Na

channels due to a hyperpolarized RMP and cell size, contributes to the slow conduction velocity observed in hPSC-CMs (Veerman et al., 2015). Of note though, several groups have addressed this issue by repolarizing the RMP through overexpression or electronic enhancement of I

as a robust method to obtain more physiological electrical behaviour, including increased Na

channel availability and improved Ca

transients profile (Meijer van Putten et al., 2015; Rocchetti et al., 2017;

Vaidyanathan et al., 2016; Veerman et al., 2016). Importantly, I

-enhanced hiPSC-CMs displayed a stable RMP in absence of spontaneous beating activity, allowing more accurate quantitative analysis of AP in comparing healthy and diseased myocytes (Meijer van Putten et al., 2015; Rocchetti et al., 2017;

Vaidyanathan et al., 2016; Veerman et al., 2016). In addition, increased cell size, membrane capacitance and DNA synthesis were also observed (Vaidyanathan et al., 2016).

Existing hiPSC-models of cardiovascular and non- cardiovascular diseases with cardiac traits

To date, hiPSC-CMs have been successfully used not only to recapitulate, but

also to better understand and elucidate the disease-relevant cellular and

molecular pathological mechanisms of several cardiovascular diseases. They

remain one of the few opportunities to study the heart against a background of

human gene expression. Below, as well as in Figure 1 and Table 3, we list most

of the hiPSC cardiac models to date and provide specific examples.

Figure 1. Schematic representation of cardiomyocyte structure and relevant cellular and molecular components that are mutated in cardiac diseases. This schematic shows the cardiac proteins encoded by mutated genes for which hiPSCs have been generated and reviewed here.

Disease-genes of interest, which are also listed in Table 3, are located in different compartments of the cardiomyocyte, such as the extracellular matrix, sarcoplasmic reticulum (SR), cytoskeleton, sarcomere, desmosome, lysosome, mitochondrion and the nucleus.

2 Arrhythmias and channelopathies:

Familial Long-QT syndrome (LQTS)

Long-QT syndrome (LQTS) is a potentially life-threatening arrhythmia characterized by a prolongation in the ventricular repolarization component (QT interval) of the electrocardiogram (ECG) (Crotti et al., 2008). Patients affected by LQTS experience polymorphic ventricular tachycardia with a characteristic shape of the ECG also termed “Torsades de Pointes”, syncope and sudden cardiac death. LQTS includes hereditary variants: the autosomal dominant form or Romano-Ward syndrome and the recessive form or Jervell and Lange-Nielsen syndrome (JLNS) (Anton Jervell, 1957; OC, 1964; ROMANO et al., 1963; Schwartz et al., 1975). LQTS is associated with more than 500 mutations in 16 different genes encoding cardiac ion channel proteins and their auxiliary subunits or modulating proteins and displays a wide range of phenotypes even within members of the same family (Giudicessi and Ackerman, 2013; Schwartz, 2013).

LQT1 - LQT1 patients harbor mutations in the KCNQ1 gene, which encodes the K

channel K

7.1 mediating the repolarizing current I

of the AP (Morita et al., 2008). To date, several LQT1 hiPSC lines have been generated and characterized from patients carrying distinct mutations in the KCNQ1 gene, such as R190Q (Moretti et al., 2010; Y. Wang et al., 2014), G269S and G345E (Liang et al., 2013;

Y. Wang et al., 2014), P631fs/33 (Egashira et al., 2012), and a novel heterozygous exon 7 deletion (ex7Del) (Ma et al., 2015).

In 2010, Moretti and colleagues used retroviral vectors to generate patient-

specific hiPSCs from members of a family affected by the autosomal dominant

missense mutation R190Q in the KCNQ1 gene and differentiated the patient-

derived cells into functional cardiomyocytes that recapitulated in vitro

the electrophysiological features of the LQT1 disease phenotype and the

therapeutic approach of b-blockade (Moretti et al., 2010). In the same study,

hiPSC-CMs helped demonstration of a dominant negative trafficking defect of

the mutated channel. Similarly, Egashira et al. identified the same molecular

mechanism as being responsible of a LQT1 phenotype in P631fs/33-KCNQ1

mutated hiPSC-CMs (Egashira et al., 2012). In another study, Liang and colleagues

generated a library of hiPSC-CMs from healthy individuals and patients with

different hereditary cardiac disorders, including LQT1, for recapitulating and

predicting drug-induced arrhythmia. Interestingly, these cells displayed a

broad spectrum of cardiotoxicity effects suggesting that diseases specific

hiPSC-CMs may accurately predict adverse drug-induced cardiotoxicity (Liang

et al., 2013). Further, in 2014, Wang et al. generated hiPSCs by overexpressing ion channel genes with dominant negative mutations causing LQT1 (G269S, G345E, and R190Q). In order to achieve stable transgene expression, these genes were integrated into the AAVS1 safe harbor locus by using the Zinc Finger Nuclease (ZFN) technology. Next, transgene cells and isogenic unedited controls were differentiated into cardiomyocytes and recapitulated the LQT1 disease phenotype showing a prolongation in the AP duration (APD) (Y. Wang et al., 2014).

Disease Gene Mutation Reference

Arrhythmias and channelopathies

LQT1 KCNQ1 R190Q (Moretti et al., 2010)

LQT1 KCNQ1 P631fs/33 (Egashira et al., 2012)

LQT1 KCNQ1 Ex7Del (Ma et al., 2015)

LQT1/JLNS KCNQ1 R594Q (M. Zhang et al., 2014)

E160fs+138X

LQT1/LQT2 KCNQ1 G269S (Y. Wang et al., 2014)

KCNQ1 G345E

KCNQ1 R190Q

KCNH2 A614V

LQT2 KCNH2 G1681A (Matsa et al., 2014; 2011)

LQT2 KCNH2 R176W (Lahti et al., 2012)

LQT2 KCNH2 A561V (Mehta et al., 2014)

LQT2 KCNH2 N996I (Bellin et al., 2013)

LQT2 KCNH2 A614V (Itzhaki et al., 2012b)

LQT2 KCNH2 A561P (Jouni et al., 2015)

LQT2/LQT3 KCNH2 A422T (Spencer et al., 2014)

SCN5A N406K

LQT2 TBX20 R311C (Caballero et al., 2017)

LQT3 SCN5A V1763M (Ma et al., 2013b)

LQT3 SCN5A V240M (Fatima et al., 2013)

R535Q

LQT3 SCN5A F1473C (Terrenoire et al., 2012)

KCNH2 K897T

LQT3 SCN5A R1644H (Malan et al., 2016)

Table 3. >

2

LQT8/TS CACNA1C G1216A (Yazawa et al., 2012)

LQT14 CALM1 F142L (Rocchetti et al., 2017)

LQT15 CALM2 D130G (Limpitikul et al., 2016)

BrS/LQT3 SCN5A 1795insD (Bezzina et al., 1999;

Davis et al., 2012; van den Berg MP et al., 2003;

Veerman et al., 2016)

BrS/LQT3 SCN5A E1784K (Okata et al., 2016)

BrS SCN5A R620H/R811H (Liang et al., 2016)

4189delT

CPVT RYR2 M4109R (Itzhaki et al., 2012a)

CPVT RYR2 F2483I (Fatima et al., 2011; X. H.

Zhang et al., 2013)

CPVT RYR2 P2328S (Kujala et al., 2012)

CPVT RYR2 S406L (Jung et al., 2012)

CPVT RYR2 P2328S

EX3del T2538R L4115F Q4201R V4653F

(Penttinen et al., 2015)

CPVT RYR2 L3741P (Preininger et al., 2016)

CPVT RYR2 I4587V (Sasaki et al., 2016)

CPVT RYR2 E2311D (Di Pasquale et al., 2013)

CPVT CASQ2 G112+5X (Lodola et al., 2016)

CPVT/LQTS TECRL SRD5A2L2 (Devalla et al., 2016)

c.331+1G>A Cardiomyopathies

BTHS TAZ 517delG (G. Wang et al., 2014)

BTHS TAZ Gly197Val (Dudek et al., 2013)

EX2Del Arg57Leu

Leopard PTPN11 T468M (Carvajal-Vergara et al.,

2010)

ARVC PKP2 Gly828Gly (Kim et al., 2014)

R672fsX683

ARVC PKP2 L614P (Ma et al., 2013a)

ARVC PKP2 A324fs335X (Caspi et al., 2013)

ARVC SCN5A R1898H (Riele et al., 2017)

Table 3. >

DCM TNNT2 R173W (Sun et al., 2012; Wu et al., 2015)

DCM LMNA R225X (Siu et al., 2012)

DCM TTN W976R (Hinson et al., 2015)

A22352fs P2258fs

DCM DES A285V (Tse et al., 2013)

DCM RBM20 R636S (Wyles et al., 2016)

HCM MYBPC3 C2373dipG (Birket et al., 2015)

HCM MYH7 Arg663His (Lan et al., 2013)

HCM BRAF T599R (Cashman et al., 2016)

HCM BRAF T599R (Josowitz et al., 2016)

Q257R

DCM/HCM ALPK3 W1264X (Phelan et al., 2016)

HCM PRKAG2 N488I (Hinson et al., 2016)

R531Q

LQT1 KCNQ1 G269S (Liang et al., 2013)

HCM MYH7 R663H

DCM TNNT2 R173W

HCM MYH7 R442G (Han et al., 2014)

HCM MYBPC3 Arg91Cys (Tanaka et al., 2014)

N/A

Gly999/Gln1004del

HLHS N/A N/A (Jiang et al., 2014)

IHD/CAD ALDH2 ALDH2*2 (Ebert et al., 2014)

Cardiometabolic diseases

PD GAA Ex18Del (Raval et al., 2015)

1441delT/

TRP746TER

PD GAA Arg266Cys/M439K (Sato et al., 2015)

PD GAA D645E/D645E (Huang et al., 2011)

D645E/2040-1G

PD GAA Ex18del (Higuchi et al., 2014)

Danon LAMP2 129-130 insAT (Hashem et al., 2015)

IVS-1 c.64+1 G>A

Fabry GLA W162X (Kawagoe et al., 2013)

Table 3. >

2

Fabry GLA W162X/R220X (Itier et al., 2014)

Fabry GLA IVS4+919 G>A (Chien et al., 2016; Chou

et al., 2017)

Diabetic cardiomyopathy N/A N/A (Drawnel et al., 2014)

Non-cardiovascular diseases with cardiac traits

DMD DMD Ex50Del (Macadangdang et al.,

2015)

DMD DMD Ex45-52del (Lin et al., 2015)

ATTR TTR L55P (Leung et al., 2013)

Table 3. Existing hiPSC-models of cardiovascular diseases and disorders with cardiac traits.

A search for original articles published up to February 2017 was performed using PubMed Advanced Search Builder using the following criteria: (i) (human induced pluripotent stem cells) AND (cardiac disease model) NOT review; (ii) (human induced pluripotent stem cells) AND (cardiomyocytes) NOT review; (iii) (human induced pluripotent stem cells) AND (cardiomyocytes) AND (mechanistic insight) NOT review. References on cardiac regeneration were manually excluded. References from some of the most comprehensive reviews of the field were screened and manually added when not present in the above-mentioned search. Limitation of this review relates to selection bias.

LQT2 - LQT2 patients carry mutations in the KCNH2 gene, also termed human ether-a-go-go related gene (hERG), which encodes the K

channel mediating the repolarizing current I

of the AP (Curran et al., 1995). A panel of LQT2- diseased hiPSCs carrying the following hERG mutations have been generated and characterized: G1681A (Matsa et al., 2014; 2011), A614V (Itzhaki et al., 2012b; Y. Wang et al., 2014), R176W (Lahti et al., 2012), N996I (Bellin et al., 2013), A561V (Mehta et al., 2014), A422T (Spencer et al., 2014), and A561P (Jouni et al., 2015).

By performing multi-electrode array, patch-clamp electrophysiology and drug testing, Matsa et al. demonstrated that hiPSC-CMs from two patients carrying the G1681A KCNH2 mutation showed prolonged APs but displayed different drug-induced sensitivity (Matsa et al., 2014; 2011). Two independent laboratories applied similar strategies for modeling LQT2 by generating hiPSCs from patients carrying the missense A614V (Itzhaki et al., 2012b) and R176W (Lahti et al., 2012) mutations on the hERG channel. However, despite the novelty of using patient hiPSC-CMs for modeling LQT2, these studies were performed under genetically non-defined conditions and therefore genetic background variations were not taken into account. To address this limitation, we modeled LQT2 syndrome by generating hiPSCs from a patient carrying the

Table 3. >