cardiomyocyte differentiation of hipscs in 2d and 3d ... · derivados de células estaminais...
TRANSCRIPT
Cardiomyocyte Differentiation of hiPSCs in 2D and 3D Culture Systems:
A Comparative Study and Transcriptomic Analysis
João Pedro Brinquete Cotovio
Thesis to obtain the Master of Science Degree in
Biotechnology
Supervisors
Professor Dr. Leonilde de Fátima Morais Moreira
Doctor Carlos André Vitorino Rodrigues
Examination Committee
Chairperson: Professor Dr. Arsénio do Carmo Sales Mendes Fialho
Supervisor: Doctor Carlos André Vitorino Rodrigues
Member of the Committee: Doctor Evguenia Pavlovna Bekman
October 2017
i
ii
iii
iv
I would like to thank to everyone that contribute to this thesis and to all those who have shared this
adventure with me. Particularly, I would like to thank:
To Professor Joaquim Cabral for receiving me in SCERG, for all the lessons and for continuing to
drive bioengineering in our country and beyond.
To Professor Margarida, my supervisor and to whom I owe my position in this project, for believing
in me from the first moment and for all the kind words and guidance during these months.
To Professor Leonilde, also my supervisor, for planting the idea that gave rise to this project and
for all the guidance in the transcriptomic analysis which made such a great contribution to this work.
To Carlos, more than a supervisor, for always making time to listening to me and for supporting
me in all the decisions even when things were not going so well. Thank you for making me see things
from a more positive perspective. At first everything seemed cloudy, but I think we have come a long way.
I hope I did not disappoint you.
To Professor Tiago who has always contributed with ideas and suggestions that somehow
contributed to the final form of this work. I hope you are proud of my heatmaps that have no gray
boundaries.
To Mariana Branco, my mentor, for everything she taught me not only in the basic stem cell culture
but mainly in cardiac differentiation. Thank you for all the friendship and patience, I know you needed a
lot of this last one, but I think it was worth it. From the first heart beat to your departure, I will never forget
your teachings and I hope you are proud of my work. In fact, I think we are a brilliant cardiac team.
To Teresa, who I consider my second mother mentor, for always listening to me, for all the answers
and for being there in moments of despair. Your help went beyond what was demanded of you, not only
in the molecular biology field but during all this work. I just apologize for one thing, I still cannot measure
everything by eye, but I promise I will work on that. Also, I cannot fail to thank you for your friendship
and for all the laughs that I will never forget.
To Ana Rita, for always being available to help and for being such a therapist during dark times.
Thank you for your friendship and for all the time that we spent together in that lab, and believe me, it
was a lot. Not just for the time in the lab, but for the times in between and outside of it too. Although I
have a current issue with the shuttle, I will not forget our travels on it listening to pimba.
To Tiago Dias, for fruitful discussions. To Carmen Claudia Miranda for all the senior advices and
for the fun. To Mariana Pina, Kikas and Leonor, my master colleagues and friends, for making this season
a happy season (I do not know how to translate this but, “é rir para não chorar”). Mariana, I have two
promises for you: I will never leave tubes in the centrifuge again; I will never forget the value 3.24 (is this
really the value?). To Diogo Pinto, Carina, Queen Sara Bukar and André Branco for being in the other
side; thank God there is a wall between; besides that, thank you for all that moments that made me laugh
v
out loud, seriously. To Diogo Nogueira, Jorge Pascoal, Marta Costa and Joana Serra who were also very
helpful during this time. To all the other SCERG colleagues, cell donors, mice and rabbits who sacrificed
themselves for the sake of this manuscript.
To Bioimaging Facility, Histology and Comparative Pathology Laboratory, Helena Cabaço and
Sandra Vaz from IMM. Specially to Sandra Vaz for all the help in calcium handling analysis.
To Edward O. Wilson for the letter.
To my friends of every journey from North to South, even China. To those who dreamed together
with me in Alentejo; we have been in this for so long. Specially to my housemates, they have been through
a lot. To those I was fortunate enough to meet in Lisbon; From the Club of Malate to the family that the
Master in Biotechnology gave to me. To my cdlnhs; this thesis is also for you. I could name you all, but
you know how much I am thankful to each one of you.
To Dre, my partner in crime. I do not know how to thank you for everything that we have been
through these last two years. Thank you for every moment and for all the support. I deeply appreciate
your friendship and it was a pleasure to share not only that laminar flow hood but this journey with you.
This adventure would not have been the same. We did it, Dre.
To Mom, Dad and Sister. I would not be here without you. This thesis is also yours. Above all things,
thank you for believing in me and for being part of what I am today. My thanks will never be enough,
and words cannot express it. To my Grandmothers and to my all family for being there in the good and
bad moments and for the huge support through these years; I wish we had spent more time together in
this last one. Specially, I would like to thank to my Grandfathers who could not see me becoming a
master; I want you to know that I will always march away from the sounds of the guns, I will not join the
fray, I will always make a fray of my own.
____________________
This work was supported by a MIT-Portugal grant from the test-bed project CardioStem and by an iBB
grant from the project BB6.
vi
Cardiovascular diseases (CVDs) are the number one human cause of death worldwide,
representing 31% of global death with unprecedented costs associated. Accordingly, cardiomyocytes
derived from human induced pluripotent stem cells (hiPSCs) represent a promising cell source for heart
regeneration, disease modelling and drug screening. However, similarities and differences between
different culture platforms for cardiomyocyte differentiation, either two-dimensional (2D) or three-
dimensional (3D), are still poorly explored. To unveil how the particular microenvironment of each
platform can be translated into distinct cardiomyocyte properties, in this work, a comparative study
between cardiomyocyte differentiation of hiPSCs cultured as 2D adherent monolayer and 3D aggregates
was developed. In addition, a transcriptomic analysis throughout sequential stages of cardiomyocyte
differentiation for 3D culture system was performed. Although both platforms originate cardiomyocyte-
like features, distinct molecular signatures between them were found. The results support a faster
progression in cardiac development for 3D aggregates that ultimately exhibit a maturation status that
closely resembles that of adult cardiomyocytes. It is proposed that stronger Wnt signalling induction in 3D
culture is due to larger amounts of E-cadherin (E-CAD) on 3D aggregates, leading to higher pools of
β-catenin upon adherens junction dissociation during epithelial-to-mesenchymal transition (EMT). These
findings elevate the considerable potential of 3D culture system for cardiomyocyte differentiation and aid
in the basic understanding of mechanisms governing either pluripotent spheroids or monolayer during
mesoderm commitment, demonstrating that cardiomyocyte differentiation is indeed culture
shape-dependent.
KEYWORDS
induced pluripotent stem cells; cardiomyocytes; cardiomyocyte differentiation; 3D culture; 2D culture;
transcriptomic analysis.
vii
viii
As doenças cardiovasculares são a principal causa de morte em todo o mundo, representando
31% das mortes a nível global com grandes custos associados. Consequentemente, cardiomiócitos
derivados de células estaminais pluripotentes induzidas humanas são tidos como uma fonte celular
promissora para a regeneração cardíaca, modelação de doenças e teste de fármacos. No entanto, as
semelhanças e diferenças entre as diferentes plataformas de cultura para a diferenciação de
cardiomiócitos, tanto bidimensionais como tridimensionais, foram ainda pouco exploradas. Para revelar
a forma como o microambiente característico de cada uma das plataformas pode ser traduzido em
propriedades distintas nos cardiomiócitos, foi desenvolvido um estudo comparativo entre a diferenciação
em monocamada 2D e agregados 3D. Além disso, foi realizada uma análise transcriptómica ao longo
da diferenciação no sistema 3D. Embora ambas as plataformas apresentem propriedades características
de cardiomiócitos, os resultados suportam uma progressão mais rápida no desenvolvimento cardíaco
em agregados 3D que, em última instância, exibem um estado de maturação mais elevado. Propõe-se
que o facto da indução da sinalização Wnt ser mais forte em 3D se deva a maiores quantidades de E-
caderina nos agregados 3D, levando a uma maior acumulação de β-catenina após a dissociação de
junções aderentes durante a transição epitelial-mesenquimal. Estas descobertas elevam o potencial
considerável do sistema de cultura 3D para a diferenciação de cardiomiócitos e auxiliam na
compreensão básica dos mecanismos que regem esferóides e monocamadas pluripotentes durante o
comprometimento para com a mesoderme, demonstrando que a diferenciação de cardiomiócitos é de
facto dependente do formato da cultura.
PALAVRAS-CHAVE
células estaminais pluripotentes induzidas; cardiomiócitos; diferenciação de cardiomiócitos; cultura 3D;
cultura 2D; análise transcriptómica.
ix
x
Acknowledgements .......................................................................................................................... iv
Resumo .......................................................................................................................................... viii
Table of Contents ............................................................................................................................. x
List of Tables .................................................................................................................................. xiv
List of Abbreviations ...................................................................................................................... xviii
1 Introduction ............................................................................................................................... 1
1.1 Stem Cells ......................................................................................................................... 1
1.1.1 Pluripotent Stem Cells ................................................................................................. 2
1.1.1.1 Embryonic Stem Cells ............................................................................................. 2
1.1.1.2 Induced Pluripotent Stem Cells ................................................................................ 3
1.1.2 Applications of Stem Cells .......................................................................................... 4
1.2 Origin and Fates of Cardiovascular Cells through Development .......................................... 5
1.2.1 From Mesoderm to the Heart ...................................................................................... 6
1.2.2 The Cardiomyocyte Lineage ........................................................................................ 8
1.3 Cardiovascular Disease and Therapeutic Strategies ............................................................. 9
1.3.1 Cardiac Regeneration ................................................................................................. 9
1.3.1.1 Cell-Based Therapies for Cardiac Regeneration ..................................................... 10
1.3.2 Modelling Cardiovascular Diseases and Drug Screening ............................................ 13
1.4 Cardiomyocyte Differentiation from Pluripotent Stem Cells ................................................. 14
1.4.1 Methods for Cardiomyocyte Differentiation ................................................................ 14
1.4.1.1 Three-dimensional Embryoid Bodies and Aggregates ............................................. 14
1.4.1.2 Two-dimensional Adherent Monolayer ................................................................... 15
1.4.1.3 Scalable Cardiac Differentiation ............................................................................ 17
1.4.2 Other Methods and Cell Sources ............................................................................... 18
1.4.3 Cardiomyocyte Subtype Specification ........................................................................ 18
1.4.4 Cardiomyocyte Characterization and Maturation ....................................................... 19
1.4.4.1 Methodologies for Cardiomyocyte Characterization ............................................... 19
xi
1.4.4.2 Maturation of Human Pluripotent Stem Cell - Derived Cardiomyocytes .................... 20
1.5 Mechanotransduction, Forces and Geometry Sensing ........................................................ 22
2 Aim of Studies ......................................................................................................................... 25
3 Materials and Methods ............................................................................................................ 27
3.1 Expansion of Human Induced Pluripotent Stem Cells ......................................................... 27
3.1.1 Cell Lines ................................................................................................................. 27
3.1.2 Adhesion Substrate ................................................................................................... 27
3.1.3 Culture Medium ....................................................................................................... 27
3.1.4 Cell Thawing and Maintenance ................................................................................. 27
3.1.5 Cell Passaging ......................................................................................................... 28
3.1.6 Cell Cryopreservation ............................................................................................... 28
3.1.7 Cell Counting .......................................................................................................... 28
3.2 Cardiomyocyte Differentiation .......................................................................................... 28
3.2.1 2D Adherent Monolayer ........................................................................................... 28
3.2.2 3D Aggregates ......................................................................................................... 28
3.2.3 Aggregate Size Analysis ............................................................................................ 29
3.2.4 Cardiac Differentiation via GSK3 inhibitor and Wnt inhibitor ...................................... 29
3.3 Cell Characterization ....................................................................................................... 30
3.3.1 Flow Cytometry ........................................................................................................ 30
3.3.1.1 Cardiac Troponin T .............................................................................................. 30
3.3.1.2 Octamer-Binding Transcription Factor 4 ................................................................ 30
3.3.2 Hematoxylin and Eosin Staining ................................................................................ 31
3.3.3 Immunocytochemistry ............................................................................................... 31
3.3.4 RNA Sequencing and Transcriptomic Analysis ............................................................ 32
3.3.5 Quantitative Real-Time Polymerase Chain Reaction .................................................... 32
3.3.6 Gene Expression Analysis and Protein-Protein Association .......................................... 33
3.3.7 Calcium Imaging ..................................................................................................... 33
4 Results and Discussion ............................................................................................................. 35
4.1 Cardiomyocyte Differentiation in 2D and 3D Culture Systems ............................................ 35
4.2 Transcriptomic Analysis of 3D Cardiomyocyte Differentiation ............................................. 39
xii
4.3 Expression of Regulatory and Structural Genes in 3D Culture ............................................. 44
4.4 Similar and Different Molecular Features in 2D and 3D Cultures ........................................ 47
4.5 Epithelial-to-Mesenchymal Transition in Cardiomyocyte Differentiation ............................... 49
4.6 Maturation and Functional Analysis of de novo Cardiomyocytes ......................................... 52
5 Conclusion and Future Trends .................................................................................................. 57
6 References .............................................................................................................................. 59
7 Supplemental Data .................................................................................................................. 71
xiii
xiv
Table 1 | List of clinical trials using cell therapy approach for heart regeneration in acute myocardial
infarction and heart failure (Adapted from Cahill et al., 2017). ......................................................... 11
Table 2 | Summary of differential properties of cardiomyocytes dependent on the maturation status. .. 21
Table 3 | List of primary and secondary antibodies used for immunocytochemistry. ........................... 32
Table 4 | Genes (and Assay ID) used for quantitative real-time polymerase chain reaction. ................ 33
Table 5 | Gene ontology analysis of enriched transcripts at day 20 of differentiation. ........................ 42
xv
xvi
Figure 1 | Clinical application of induced pluripotent stem cells and their derivatives. .......................... 5
Figure 2 | Cardiac development in the mouse embryo: specification and progression of the cardiac cell
lineages. .......................................................................................................................................... 7
Figure 3 | Methods for cardiomyocyte differentiation of pluripotent stem cells. ................................... 16
Figure 4 | Characterization of hiPSC cardiomyocyte differentiation in 2D and 3D culture systems. ...... 36
Figure 5 | Transcriptomic analysis of cardiomyocyte differentiation in 3D culture and protein-protein
interaction. ..................................................................................................................................... 40
Figure 6 | Gene expression profile of regulatory and structural genes at sequential stages in 3D culture.
..................................................................................................................................................... 44
Figure 7 | Relative expression profiles and immunofluorescent staining for pluripotency, mesendoderm,
cardiac progenitor and cardiomyocyte markers during hiPSCs differentiation. .................................... 46
Figure 8 | Gene expression analysis between 2D and 3D culture systems during cardiomyocyte
differentiation. ................................................................................................................................ 48
Figure 9 | Epithelial-mesenchymal transition and cell-cell interaction during cardiomyocyte differentiation.
‡ .................................................................................................................................................... 50
Figure 10 | Morphological and structural features in cardiomyocyte maturation and calcium handling
analysis. ......................................................................................................................................... 54
xvii
xviii
2D two-dimensional
3D three-dimensional
ACTN1 actinin alpha 1
AP action potential
asRNA antisense RNA
BMMNC Bone marrow-derived mononuclear
stem cell
BMP bone morphogenetic protein
BPM beats per minute
Cas9 CRISPR associated protein 9
CD31 cluster of differentiation 31
CER1 cerberus 1
CNCC cardiac neural crest cell
CPC cardiac progenitor cell
CRISPR clustered regularly interspaced short
palindromic repeats
CVD cardiovascular disease
CVPC multipotent cardiovascular progenitor
cell
CX40 connexin 40
CX43 connexin 43
DAPI 4’,6-diamidino-2-phenylindole
DKK1 dickkopf Wnt signaling pathway
inhibitor 1
DMEM dulbecco’s modified eagle medium
DMSO dimethylsulfoxide
EB embryoid bodie
EC embryonal carcinoma
E-CAD E-cadherin
ECM extracellular matrix
EMT epithelial-to-mesenchymal transition
EOMES eomesodermin
ESC embryonic stem cell
ESRG sense intronic RNA embryonic stem
cell related
FBS fetal bovine serum
FGF fibroblast growth factor
FGS fetal goat serum
FHF first heart field
Fura-2 AM fura-2 acetoxymethyl ester
GAPDH glyceraldehyde-3-phosohate
dehydrogenase
GATA4 GATA binding protein 4
GO gene ontology
GSK3β glycogen synthase kinase 3 beta
GYPA glycophorin A
H&E hematoxylin and eosin
HAND2 heart and neural crest derivatives
expressed 2
HCN4 hyperpolarization-activated cyclic
nucleotide-gated K+4
hESC human embryonic stem cell
hESC-CM human ESC-derived cardiomyocyte
hiPSC human induced pluripotent stem cell
hiPSC-CM human iPSC-derived cardiomyocyte
hMSC human mesenchymal stem cell
HSPB7 heat shock protein family B (small)
member 7
ICM inner cell mass
IGF-1 insulin-like growth factor 1
ISL1 LIM homeodomain transcription factor
Islet1
iPSC induced pluripotent stem cell
IRX4 iroquois homeobox 4
IWP-4 Wnt production 4
IWR-1 Wnt response 1
KLF4 kruppel like factor 4
L1TD1 LINE1 type transposase domain
containing 1
LEFTY1 left right determination factor 1
xix
LIN28 lin-28 homolog A
LINC00678 long intergenic non-coding RNA 678
lincRNA long intergenic non-coding RNA
LQTS long QT syndromes
MCL Markov cluster algorithm
MEA microelectrode arrays
mESC mouse embryonic stem cell
MESP1 mesoderm posterior 1
MYBPC3 myosin-binding protein C
MYC MYC proto-oncogene, bHLH
transcription factor
MYH myosin heavy chain
MYL myosin light chain
N-CAD N-cadherin
NKX2-5 NK2 homeobox 5
NPPA natriuretic peptide A
OCT4 octamer-binding transcription factor 4
PBS phosphate-buffered saline
PCA principal component analysis
PEO proepicardial organ
PGRN pluripotency gene regulatory network
PSC pluripotent stem cell
qRT-PCR quantitative real-time polymerase
chain reaction
RALDH2 retinaldehyde dehydrogenase 2
RNA-seq RNA sequencing
ROCKi Rho kinase inhibitor
SA sinoatrial
SNAI1 snail family zinc finger 1
SHF second heart field
SMPX small muscle protein, X-linked
SOX2 SRY-box 2
T brachyury
TBX5 T-box 5
TE extraembryonic trophectoderm
TGFβ transforming growth factor-beta
TNNT2 cardiac troponin T2
TTN titin
VEGF vascular endothelial growth factor
Wnt wingless integrated
ZFP42 ZFP42 zinc finger protein
ZSCAN10 zinc finger and SCAN domain
containing 10
α-ACT alpha actinin
α-SMA alpha smooth muscle actin
1
Stem cells have been regarded, since a long time ago, as undifferentiated cells that are capable
of self-renewal and production of specialized progeny through differentiation (Blau et al., 2001;
Weissman, 2000). To accomplish that, stem cells undergo asymmetric cell division, by which the cell
divides to generate one cell with a stem-cell fate (self-renewal) and one cell that commit differentiation,
maintaining homeostasis (Blanpain and Simons, 2013; Knoblich, 2008; Morrison and Kimble, 2006).
In both plant and animal kingdoms, the multicellularity of highly regulated and postembryonic
tissues is dependent of the generation of new cells for growth and repair. Therefore, biological systems
are driven by a balance between cell death and cell proliferation, preserving form and function in tissues.
From this point of view, stem cells are the units of such attributes: development, regeneration and
evolution (Sánchez Alvarado and Yamanaka, 2014; Weissman, 2000).
From conception to death, as cells develop, derived from embryonic tissue, they become
progressively restricted in their developmental potency reaching the point when each cell can only
differentiate into a single specific cell type. In the beginning, the earliest cells in ontogeny are totipotent,
giving rise to all embryonic and extra-embryonic tissues, i.e., only a totipotent cell can originate an entire
organism (De Los Angeles et al., 2015; Sánchez Alvarado and Yamanaka, 2014). Through
embryogenesis, when the pluripotent state is reached, a pluripotent stem cell (PSC) can originate all the
cells from all the tissues of the body, though the contributions to the extra-embryonic membranes or
placenta is limited (De Los Angeles et al., 2015). On the other hand, a multipotent stem cell is restricted
to the generation of the mature cell type of its tissue of origin and finally, a unipotent stem cell display
limited developmental potential, giving rise to only a single cell type (De Los Angeles et al., 2015). In an
adult organism, stem cells can be found in most tissues throughout the body, even within relatively
dormant tissues. These stem cells experience low or no division in normal homeostasis, remaining
quiescent for extended periods of time. However, these cells can respond efficiently to stimuli upon
initiation of homeostasis or injury (Hsu and Fuchs, 2012).
Regardless of whether stem cells are pluripotent or multipotent, stem cell fate is regulated by a
combination of intrinsic and extrinsic mechanisms. For instance, a local tissue microenvironment that
hosts and influences the behaviour or characteristics of stem cells is known as niche and the signals
provided by it are seen as extrinsic mechanisms. These signals can include growth factors, the extracellular
matrix (ECM) and contact with other cells, coordinating stem cell fate with the behaviour of their neighbour
cells. Thus, stem cell communication through tissue structure is a function of their position within a tissue,
the architecture of that tissue and the signals that are transmitted (Watt and Huck, 2013; Xin et al., 2016).
2
Pluripotency can be defined as a transient property of cells within the early embryo, where PSCs
have the capacity to form tissues of all three germ layers of the developing embryo and, later, of the adult
organism - ectoderm, mesoderm and endoderm, and still the germ lineage. As previously mentioned,
PSCs typically provide little or no contribution to the trophoblast layers of placenta (Li and Belmonte,
2017; De Los Angeles et al., 2015).
The first PSCs to be isolated and investigated in culture were derived from mouse
teratocarcinomas - a tumour of germ cell origin that maintain a wide variety of diversely differentiated
tissues – known as embryonal carcinoma (EC) cells (De Los Angeles et al., 2015; Solter and Solter, 2006).
Nevertheless, PSCs can be isolated from several sources through the development (Wu et al., 2016), as
murine (Evans and Kaufman, 1981; Martin, 1981) and human blastocyst (Thomson, 1998) or even from
the post-implantation epiblast (Brons et al., 2007; Tesar et al., 2007) or germ line (Matsui et al., 1992;
Shamblott et al., 1998). Also, pluripotency can be recapitulated in vitro by reprogramming somatic cells
to become induced pluripotent stem cells (Takahashi and Yamanaka, 2006; Takahashi et al., 2007; Yu
et al., 2007).
There are some molecular mechanisms that characterize PSCs where an interconnected
pluripotency gene regulatory network (PGRN), functionally anchored by a selected set of core transcription
factors, is essential to establish pluripotency. As part of the core pluripotency transcription factors encoding
genes are octamer-binding transcription factor 4 (OCT4), SRY-box 2 (SOX2) and NANOG. In certain
circumstances, the loss of SOX2 or NANOG or their substitution can be tolerated (Li and Belmonte, 2017;
De Los Angeles et al., 2015). Despite that, PSCs can be classified into different states of pluripotency
based on molecular signatures, with the terms “naive" and “primed” being introduced to describe early
and late phases of ontogeny, respectively (Weinberger et al., 2016).
Pluripotency can be suggested by such molecular signatures, but only functional assays can reveal
the developmental potential of a cell. Functional assays to assess pluripotency include: differentiation into
three germ layers in vitro, teratoma formation in vivo, chimaera formation, germline transmission through
blastocyst injection, tetraploid complementation and single-cell chimaera formation (De Los Angeles et
al., 2015). For human PSCs, teratoma formation remains the gold standard of functional assays.
Mammalian embryogenesis starts with a single totipotent cell, the zygote. After the first cell
division, the two-cell embryo is composed by two equal blastomeres. In the earlier stages, including
two‑cell and four‑cell embryos, cells are still considered totipotent. Later, in what is called blastocyst, it is
possible to distinguish the extraembryonic trophectoderm (TE) on the outside and the inner cell mass (ICM)
(Wu and Izpisua Belmonte, 2016; Wu et al., 2016). It is in the ICM that pluripotent cells first arise. The
ICM cells, cultured in conditions that allow indefinite self-renewal and maintenance of the pluripotent
state, are known as embryonic stem cells (ESCs) and were first derived from mouse (mESCs) in 1981 by
Martin Evans (Evans and Kaufman, 1981) and Gail Martin (Martin, 1981). These cultures proved to have
3
all the properties previously established for EC cell cultures, as well as a completely normal karyotype
(Evans, 2011). Only by the year of 1998, Thomson derived the first ESC lines from human blastocists
(Thomson, 1998), the so called human embryonic stem cells (hESCs).
Throughout the normal development, the amount of ESCs is limited and their existence is
constrained in the time course of development being present for only a short time. In contrast, tissue
culture allows the generation and maintenance of millions of ESCs indefinitely, preserving their pluripotent
state (Evans, 2011).
The concept that molecular mechanisms “by which the genes of the genotype bring about
phenotypic effects” – epigenetics – was captured by Conrad Waddington in the iconic image of the
epigenetic landscape that influences cellular fate during development, analogously to the movement of a
marble (Stricker et al., 2017). Since then, the possibility that cells can change their identity has fascinated
scientists. This notion was first suggested by John Gurdon, establishing that in vivo plasticity of the
differentiated state can be induced artificially by directly manipulating cells and their environment (Merrell
and Stanger, 2016). It was demonstrated that the marble can be rolled back to the top of the hill, i.e.,
cells can be reprogrammed back to a wider developmental potential.
As the possibility to reprogram cells, not by transplanting their nuclei, but by introducing
pluripotency factors into cells became a reality, cells with a gene expression profile and developmental
potential similar to ESCs were generated in 2006. This accomplishment was obtained using mouse
somatic cells using a cocktail of four transcription factors (Takahashi and Yamanaka, 2006). These cells
were termed induced pluripotent stem cells (iPSCs) and were generated after retrovirally introducing genes
encoding the four transcription factors, OCT4, SOX2, Kruppel like factor 4 (KLF4) and the MYC proto-
oncogene, bHLH transcription factor (MYC) – the “Yamanaka factors”. After successfully generating
mouse iPSCs, in 2007 iPSCs were generated from human fibroblasts, using the same four factors and
alternatively NANOG and Lin-28 homolog A (LIN28) instead of KLF4 and MYC (Takahashi et al., 2007;
Yu et al., 2007). It was the establishment of human induced pluripotent stem cells (hiPSCs). Since then,
iPSCs technology of reprogramming became a robust method to convert differentiated cells to a
pluripotent state (Shi et al., 2016; Takahashi and Yamanaka, 2016).
Non-integrating methods have been developed and include reprogramming using episomal
DNAs, adenovirus, Sendai virus, PiggyBac transposons, minicircles, recombinant proteins, synthetically
modified mRNAs, microRNAs and, more recently, small molecules (Shi et al., 2016). These new
techniques avoid insertional mutagenesis and transgene reactivation leading to a lower variability
between the cell lines.
The derivation of iPSCs is the apogee of cellular de-differentiation, a category of cellular
conversion that defines the reversion of a committed or differentiated cell into a cell with greater
developmental potential (Merrell and Stanger, 2016).
4
Since the isolation of ESCs from human embryos, a research interest in using pluripotent stem
cells as a potential tool for medicine has been growing. Besides that, after finding that somatic cells can
revert all the way back to the embryonic state through transcription factors, manipulation of signalling
pathways aiming for cell differentiation has been studied contributing to iPSCs application in biomedicine
(Figure 1). Accordingly, several protocols have been described for direct differentiation in vitro to neurons,
haematopoietic cells, hepatocytes, smooth muscle cells and cardiomyocytes, among other cell types
(Tabar and Studer, 2014).
An obvious application of PSCs in medicine is in cell therapy. The potential of regenerative
medicine based on the use of stem cells to promote endogenous regeneration or to replace damaged
tissues after cellular transplantation has been shown to successfully induce functional recoveries (Shi et
al., 2016). Indeed, iPSCs enable the generation of autologous cells, i.e. patient-specific cells, supressing
the risk of rejection and infection. The first clinical study using hiPSC products was performed in 2014 by
transplanting retinoic pigment epithelium sheets (Kimbrel and Lanza, 2015). However, the acquisition of
chromosomal aberrations, due to the reprograming process and subsequent culture, present some of the
disadvantages of these cells (Lamm et al., 2016). Due to PSC tumorigenicity, it is critical to ensure that
the transplanted product does not contain undifferentiated cells with the potential to generate teratomas
(Shi et al., 2016).
Another equally important biomedical application of PSCs is in disease modelling. It is expected
that in vitro PSC-based disease models help to identify the pathological mechanisms underlying human
diseases, a key to novel therapeutic strategies for their prevention and treatment. Both human ESCs and
iPSCs have been used for modelling human genetic diseases, establishing isogenic cell lines with novel
gene editing tools (e.g. CRISPR-Cas9) to induce disease-causing mutations or to silence mutation carried
by patient-specific cells (Avior et al., 2016; Sterneckert et al., 2014).
Modelling of human diseases is the motivation to develop therapeutic agents allowing the
diseases to be treated, alleviated or cured. Therefore, drug screening is also considered as a potential
application of PSCs. Animal models have been used in drug screening but differences from the actual
human setting lead these experiments to fail due to inaccurate forecasting of their effects. Moreover,
animal models are not suitable for high-throughput screening of small-molecule libraries (Avior et al.,
2016; Sayed et al., 2016). Until now, many drug screens have been conducted and potential drug
candidates have been identified. Moreover, it is not only important to assess efficacy but also toxicity,
predicting the likelihood of candidate drugs to cause serious side effects (Shi et al., 2016). A specific
patient has a specific genetic background and this fact implies different responses to medications for each
individual. Accordingly, hiPSC-based drug screening is the key for a personalized therapy, an emerging
approach known as precision medicine (Sayed et al., 2016).
5
Figure 1 | Clinical application of induced pluripotent stem cells and their derivatives.
After iPSC differentiation, the originated cells have several applications: cell therapy, (A) disease modelling, (B) drug
screening and (C) toxicity tests (Adapted from Bellin et al., 2012).
The growing knowledge about niches, stem cell maintenance and differentiation triggered the
development of new three-dimensional in vitro culture technologies to which the name organoids was
given (Fatehullah et al., 2016). Organoids recapitulate biological parameters like the spatial organization
of tissue-specific cells, cell-cell interactions, cell-matrix interactions trying to be representative of the in
vivo physiology of the organism (Yin et al., 2016). Therefore, this is a promising technology in the sense
that embryonic development, lineage specification, tissue homeostasis, as well as disease modelling can
be brought to light. In fact, this technology has proven to be efficient to study brain disorders and as an
example of that, microcephaly was successfully modelled using RNA interference and patient-specific
iPSCs (Lancaster et al., 2013).
Also, a new generation of reprograming, in vivo cellular reprogramming, seems to bring new
opportunities to the table. Using lineage-restricted transcription factors and microRNAs, the directly
reprogramming of one somatic cell into another desired cell type can help damaged organs to regenerate
lost tissue (Srivastava and DeWitt, 2016).
Just as new technologies are being developed, the greater will be the potential applicability of
iPSCs cells in the emerging fields of regenerative medicine, disease modelling, and drug screening.
From a developmental standpoint, iPSCs allow for the recapitulation of early stages of
embryogenesis through in vitro differentiation. Exposure to developmental morphogens can direct the in
vitro differentiation into the three germ layers (Müller and Lengerke, 2009). This suggests that the use of
a developmental biology approach will end in the creation of protocols that enable the derivation of
highly specialized cell populations.
6
In the course of embryonic development, pluripotent stem cells give rise to precursor cells that
transit through a differentiation pathway increasingly restricting their lineage potential (Blanpain and
Simons, 2013). During embryogenesis, more precisely during the process of gastrulation, there is the
formation of a transient structure in the region of the epiblast, the primitive streak. Throughout this
process, uncommitted cells migrate through the primitive streak ending in the generation of the three
germ layers: mesoderm, endoderm and ectoderm.
For years, mesoderm has been identified as the germ layer associated with a broad spectrum of
families of extracellular signalling molecules like wingless integrated (Wnt) proteins, fibroblast growth
factors (FGFs) and many members of the transforming growth factor-beta (TGFβ) superfamily, including
bone morphogenetic proteins (BMPs), Nodal and Activin, which are responsible for mammalian
cardiogenesis (Loh et al., 2016; Noseda et al., 2011). Thus, cardiac cells specification depends on a
combination of signalling factors and downstream transcriptional events, leading to the expression of
cardiac-specific factors (Später et al., 2014).
Much of the knowledge in this area is based on mouse embryogenesis (Figure 2). The development
path of the heart has its origin prior to or shortly after the beginning of gastrulation, when high levels of
Nodal lead to mesoderm induction (Burridge et al., 2015). Nodal acts to maintain the expression of key
patterning genes such as Bmp4 and consequently of Wnt3 in the epiblast. Then, Nodal and Wnt signalling
are restricted to the posterior region since their antagonists, left right determination factor 1 (Lefty1),
cerberus 1 (Cer1) and dickkopf Wnt signalling pathway inhibitor 1 (Dkk1) start to be expressed anteriorly.
Later, WNT3 leads to the expression of mesendodermal transcription factor genes namely,
brachyury (T ) and eomesodermin (Eomes) (Burridge et al., 2015).
T, as well as EOMES, are responsible for the activation of mesoderm posterior 1 (Mesp1) gene
expression, which encodes a key transcription factor for the regulation of the epithelial-to-mesenchymal
transition (EMT). This process consists in the ability of MESP1 to induce snail family zinc finger 1 (Snai1),
which in turn downregulates E-cadherin (E-cad also known as Cdh1), allowing these cells to migrate away
from the primitive streak and to expand, forming the lateral plate mesoderm (Burridge et al., 2015; Paige
et al., 2015). Many believed that MESP1 was the specific marker of cardiac committed cells, but its
presence has been demonstrated in other mesoderm lineages as hematopoietic and skeletal muscle ones
(Chan et al., 2013).
It is then from the newly formed lateral plate mesoderm, that a cell population form what is called
the first heart field (FHF), giving rise to a crescent-shaped structure, the cardiac crescent. From another
cell population, it is formed the second heart field (SHF). Hyperpolarization-activated cyclic nucleotide-
gated K+4 (Hcn4) and LIM homeodomain transcription factor Islet1 (Isl1) has been proposed to encode
markers for FHF and SHF, respectively, while myocardial regulatory genes as GATA binding protein 4
(Gata4) and NK2 homeobox 5 (Nkx2-5) are present in both the FHF and SHF (Burridge et al., 2015). The
cells of the FHF differentiate and proliferate forming the linear heart tube that eventually will begin to
contract pumping blood and nutrients throughout the embryo. At later stages, each heart field has
7
Figure 2 | Cardiac development in the mouse embryo: specification and progression of the cardiac cell lineages.
Schematic representation demonstrating the main relationships between cardiac cell lineages, sublineages and marker
expression of each cell type. Abbreviations: PS, primitive streak; CM, cardiogenic mesoderm; FHF, first heart field;
SHF, second heart field; ML, midline; OFT, outflow tract; PHT, primitive heart tube; PM, pharyngeal mesoderm; VP, venous
poles; LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle; SVC and IVC, superior and inferior vena cava;
PEO, proepicardial organ; CNCCs, cardiac neural crest cells; EPC, epicardium; EMT, epithelial-to-mesenchymal transition;
HSC, hematopoietic stem cell. (Adapted from Später et al., 2014; Burridge et al., 2015).
different contributions to the several regions and chambers of the heart. FHF-derived heart tube seems to
contribute to the left ventricle and portions of the right and left atrium, while the SHF will form the right
ventricle, outflow tract, both atria, and inflow myocardium (Paige et al., 2015). Alongside with these two
major cell sources, FHF and SHF, there is also a contribution from the proepicardial organ (PEO) and
cardiac neural crest cells (CNCCs) to the heart development. The PEO is a transient structure that starts
to eventually give rise to the epicardium, which by turn can contribute to several lineages within the heart
(Mordwinkin et al., 2013; Später et al., 2014).
Transcription factor encoding genes such as T-box 5 (Tbx5) and T-box 20 (Tbx20), in combination
with Nkx2-5 and Gata4, have been shown to play a regulatory role in cardiac chamber morphogenesis.
8
This process includes induction of natriuretic peptide A (Nppa), genes encoding gap junction proteins like
connexin 40 (CX40) and connexin 43 (CX43), as well as genes encoding cytoskeletal proteins like the
small muscle protein, X-linked (SMPX) (Paige et al., 2015).
Therefore, synchronized series of interactions and consequent stem cell differentiation result in
predicable structural and functional changes, leading to the adult phenotype of cardiac tissue.
The heart is often referred to as a single entity, but as shown, this hides a great deal of complexity.
When initially formed, the heart appears to be a single tube, but after complex remodelling, it ends in a
four-chambered heart: two upper atria and two lower ventricles. Overall, the heart, as it is known, is the
sum of several layers like the so-called myocardium that form the muscular walls of the heart, the
epicardium which is the outer epithelial lining of the heart and the endocardium that form the interior
lining of blood vessels and cardiac valves (Xin et al., 2013). All these layers have as main cellular
constituents cardiomyocytes (~25-35% of all cells), smooth muscle cells and cardiac non-myocytes, like
endothelial cells (>60% of all non-myocytes) and cardiac fibroblasts (<20% of all non-myocytes) (Pinto
et al., 2016). Together, these cells contribute to the structural, biochemical, mechanical and electrical
properties of the heart resulting in an organ that becomes functional, as blood is pumped throughout the
body.
The FHF and SHF are the major sources of cardiomyocytes in the heart, with a minor contribution
from the PEO. It is now know that is under the control of myogenic transcription factors like SRF, MEF2C
and NKX2-5, that clusters of microRNAs, such as the miR-1/miR-133, are responsible for the
differentiation of early cardiomyocyte progenitors into fetal cardiomyocytes via negative feedback loop
with MYOCD (Burridge et al., 2015).
Following cardiomyocyte lineage specification, these cells must undergo proliferation and
terminal differentiation in a specific cardiomyocyte subtype in order to establish fully functional cells, such
ventricular and atrial cardiomyocytes and cardiac conduction cells (nodal cells and Purkinje fibers). In this
process, paracrine signals from the epicardium and the endocardium are involved (Später et al., 2014).
These morphological events are reflected by the expression of specific cardiac markers, in particular
structural sarcomeric proteins that can even allow the distinction of chamber specific and/or left–right
specific cardiomyocytes. These proteins compose the fundamental unit of muscle contraction, the
sarcomere, a repeating micrometre-sized element. In humans, these specific markers are encoded by
genes such as the cardiac troponin T2 (TNNT2) expressed in the thin filaments and actinin alpha 1
(ACTN1) in the Z-discs. Also, myosin heavy chains (MYH6 and MYH7, atrial and ventricular, respectively),
myosin light chains (MYL2, MYL3, MYL4 and MYL7, with MYL2 being ventricular and MYL7 being present
in both atrium and ventricle), Myosin-binding protein C (MYBPC3) and Titin (TTN) are specific markers
whose genes are expressed in cardiac thick filaments (Hwang and Sykes, 2015; De Sousa Lopes et al.,
2006).
9
Mammalian life as it is known, from embryogenesis to adulthood, depends on the second to second
uninterrupted function of the heart. Given its important role, congenital as well as acquired diseases of
the heart can have catastrophic consequences. Actually, cardiovascular diseases (CVDs) represent the
number one human cause of death and it was estimated that in 2015 17.7 million people died from
CVDs, representing 31% of all global death (WHO, 2017). Only in the United States, the annual direct
and indirect cost of CVDs was estimated to be 320.1 billion dollars in 2011 and projections estimate a
total cost of 1208 billion dollars by the year of 2030 (Mozaffarian et al., 2015).
CVDs include numerous insults like coronary artery diseases (e.g. angina and myocardial
infarction, also known as heart attack), hypertension, genetic mutations and heart failure, which are
associated with loss or dysfunction of cardiac muscle cells, diminished pump function, arrhythmias
(irregular beating rate) and eventual death (Xin et al., 2013). Accordingly, cardiovascular diseases are
one of the most serious challenges for modern medicine.
Organ regeneration has long been recognized and can be seen widely across the animal kingdom.
With regard to mending broken hearts, the first definitive report of heart regeneration concerns zebrafish
(Poss, 2002). Since then, it is known that in both fish and amphibians, functional cardiomyocytes
repopulate the injury site (Vivien et al., 2016). Only after almost ten years, the first report of mammalian
heart regeneration was described in neonatal mouse (Porrello et al., 2011), demonstrating that heart
regenerative repair is not entirely lost to mammals in evolution.
However, there has been a longstanding dogma that adult humans fail to regenerate their hearts.
After cardiac injury, there is no regeneration of the majority of the lost cardiomyocytes, in fact necrotic
muscle is usually replaced with scar tissue. Thus, the low intrinsic cardiomyocyte turnover rate compromise
the contractility of the remaining myocardium, leading to heart failure and even death if the extent of
injury is severe (Bergmann et al., 2015; Uygur and Lee, 2016).
In fact, some recent studies support the concept of age-dependent regeneration in the human
heart. A case of regeneration with full functional recovery was reported in a human new-born heart after
myocardial infarction (Haubner et al., 2016). After birth, human cardiomyocytes become binucleated
and/or polyploid. Only a small population of mononuclear diploid cardiomyocytes persists and it was
proposed that these can be the primary cell type for cardiomyocyte renewal in the adult human heart
(Patterson et al., 2017).
A broad range of approaches envisioning heart regeneration have been proposed and they include
cell therapy, biomaterials, tissue engineering, in vivo reprogramming and modulation of endogenous
repair (Cahill et al., 2017).
10
For years, clinicians have seen cell-based regenerative therapies to repair damaged cardiac tissue
as a solution, regenerating the heart by reconstitution of the cardiomyocyte substrate. Previous studies in
animal models and humans have demonstrated that this approach can have modest beneficial effects on
cardiac function (Table 1), with stem cells and other cell types being injected directly into the injured heart
or into the coronary circulation (Xin et al., 2013).
Most of the initial cardiac cell-based regeneration strategies were done by injection of autologous
skeletal myoblasts into ischaemic myocardium. Myoblasts have the potential to differentiate into myotube;
however they do not beat in synchrony with the surrounding myocardium. Also, bone marrow-derived
mononuclear stem cells (BMMNCs) were of the first adult stem cells transplanted into infarcted hearts
(Behfar et al., 2014). Therapies based on purified cell population like human mesenchymal stem cells
(hMSCs) have been implemented as well. Yet, the benefits of BMMNCs and hMSCs seen in some clinical
studies are increasingly attributed to paracrine effects (Cahill et al., 2017; Choudry et al., 2016;
Karantalis et al., 2014).
Similarly, a controversial topic relating cardiac regeneration relies on the existence of endogenous
stem cells, the cardiac progenitor cells (CPCs) found in the adult heart. However, it is now recognized that
when transplanted these cells have more prosurvival or pro-angiogenic effects than cardiomyogenic
potential (Tzahor and Poss, 2017).
The capacity of PSCs for multilineage differentiation is, therefore, an important advantage that
makes them promising candidates to promote adequate myocardium regeneration. Protocols for directed
differentiation of ESCs and iPSCs, using growth factors or small molecules, have allowed the production
of de novo cardiomyocytes (Tzahor and Poss, 2017). In 2014, grafts of human ESC‑derived
cardiomyocytes (hESC-CMs) were successfully perfused by the host, in a non-human primate heart
(Chong et al., 2014). Later, human ESC-derived cardiac progenitor cells were transplanted for the first
time in humans with severe heart failure, resulting in the patient recovery with new-onset contractility
observed in the patched region and no complications (Menasché et al., 2015). However, the efficacy
remains speculative. Besides all the progress, ethical concerns regarding the use of blastocycts and the
need for immunosuppression to prevent rejection of the cells is a reality that hESC-CMs have to face as a
regenerative therapy (Cahill et al., 2017).
The possibility of generate patient- and stage-specific cardiac progenitors from iPSCs, allows the
stem cells to match the specific environment of a certain disease, breaking the limitations imposed by
other approaches (Sayed et al., 2016). In the last years, several pre-clinical studies have studied the effects
of intramyocardial injection of iPSCs–derived cardiomyocytes into murine and porcine models after
cardiac injury (Lalit et al., 2014). Indeed, the use of cynomolgus monkey iPSC-derived cardiomyocytes in
a non-human primate model showed functional improvement very recently, nevertheless a significant
ventricular arrhythmia rate after cell transplant was present (Shiba et al., 2016).
11
Table 1 | List of clinical trials using cell therapy approach for heart regeneration in acute myocardial infarction and
heart failure (Adapted from Cahill et al., 2017).
Name Design
Patient
group and
number of
patients
Cell type, dose and
delivery route
Primary end
point Outcome Comment
Menasche et
al. (2001)
Case report • Ischaemic
heart failure
undergoing
CABG
• n = 1
• Skeletal myoblasts
• 800 × 106 cells
• Intramyocardial
injection during
CABG
NA NA • First‑in‑man report
of skeletal myoblast
injection
• Improved wall
motion and perfusion
on PET
Strauer et al.
(2002)
• Non-randomized • Open-label
• Acute
myocardial
infarction
• n = 20
• Autologous bone
marrow-derived cells
• 2.8 × 107 cells
• Intracoronary
delivery
Not specified Reduced infarct
size in cell therapy
arm
First study of bone
marrow cells in acute
myocardial infarction
Perin et al.
(2003)
• Non-
randomized
• Open-label
• Ischaemic
heart failure
• n = 21
• Autologous bone
marrow-derived cells
• 25.5 × 106 cells
• Transendocardial
injection
Safety • Improved left
ventricular function
• Reduced
reversible
perfusion defect
First study of bone
marrow cells in heart
failure
BOOST
(2004)
• Randomized
• Non-placebo
Controlled
• Acute
myocardial
infarction
• n = 60
• Autologous bone
marrow-derived cells
• 24.6 × 108
nucleated cells
• Intracoronary
delivery
Change in
LVEF
Improved global
left ventricular
function
First randomized
study of bone marrow
cells
ASTAMI
(2006)
• Randomized
• Non-placebo
Controlled
• Acute
myocardial
infarction
• n = 100
• Autologous bone
marrow-derived cells
• 68 × 106 cells
• Intracoronary
delivery
Change in
LVEF
No change in left
ventricular function
end-diastolic
volume or infarct
size at 6 months
Negative trial
concurrent with
REPAIR-AMI
REPAIR-AMI
(2006)
• Randomized
• Double-blind
• Placebo
Controlled
• Acute
myocardial
infarction
• n = 204
• Autologous bone
marrow-derived cells
• 236 × 106 cells
• Intracoronary
delivery
Change in
LVEF
Significant
improvement in
global left
ventricular function
at 4 months
Largest trial of bone
marrow cells. Showed
reduction in clinical
end point of death,
recurrent myocardial
infarction and
revascularization
Janssens et
al. (2006)
• Randomized
• Double-blind
• Placebo
Controlled
• Acute
myocardial
infarction
• n = 67
• Autologous bone
marrow-derived cells
• 304 × 106
nucleated cells
• Intracoronary
delivery
Change in
LVEF
Negative for
primary end point
Reduction in infarct
volume
MAGIC
(2008)
• Randomized
• Double-blind
• Placebo
Controlled
• Heart
failure and
previous
myocardial
infarction
undergoing
CABG
• n = 97
• Skeletal myoblasts
• 400 × 106 cells (low
dose) – 800 × 106
cells (high dose)
• Surgical injection
during CABG
Change in
regional
and global
left
ventricular
function
Negative for
primary efficacy
end points
• High-cell-dose
arm had reduced
left ventricular
remodelling with
decreased left
ventricular volumes
• Increased
arrhythmias in the
cell therapy arms
SCIPIO*
(2011)
• Randomized
• Open-label
• Non-placebo
Controlled
• Ischaemic
heart failure
• n = 23
• Autologous KIT+
cardiac stem cells
• 1 × 106 cells
• Intracoronary
delivery
Safety No adverse events
reported
Increase in LVEF and
decrease in infarct
size reported in cell
therapy recipients
(secondary end
points)
CADUCEUS
(2012)
• Randomized
• Non-placebo
Controlled
• Acute
myocardial
infarction
• n = 25
• Cardiosphere-
derived cells
• 12.5–25 × 106 cells
• Intracoronary
delivery
Safety:
unexpected or
arrhythmia-
related death,
myocardial
infarction,
tumour
formation or
MACE
Met safety end
point
Reduction in scar
size and mass in cell
therapy arm
12
Table 1 (cont.) | List of clinical trials using cell therapy approach for heart regeneration in acute myocardial
infarction and heart failure.
Name Design
Patient
group and
number of
patients
Cell type, dose and
delivery route
Primary end
point Outcome Comment
FOCUS-CCTRN
(2012)
• Randomized • Double-blind • Placebo controlled
• Ischaemic
heart failure
• n = 92
• Autologous bone
marrow-derived cells
• 100 × 106 cells
• Transendocardial
injection
• Change in
left ventricular
end-systolic
volume
• Maximal O2
consumption,
• Reversibility
on SPECT
Negative for
primary end
points
-
SWISS-AMI
(2013)
• Randomized
• Open-label
• Non-placebo
controlled
• Acute
myocardial
infarction
(either early
(5–7 days)
or late
(3–4 weeks))
• n = 200
• Autologous bone
marrow-derived cells
• 140–160 × 106
nucleated cells
• Intracoronary
delivery
Change in
LVEF
Negative for
primary end
point
-
PROMETHEUS
(2014)
• Non-
randomized
• Non-placebo
controlled
• Ischaemic
heart failure
undergoing
CABG
• n = 6
• Mesenchymal stem
cells
• 2 × 107 cells (low
dose) – 2 × 108 cells
(high dose)
• Intramyocardial
injection during
CABG
NA NA • Increased ejection
fraction
• Decreased scar
mass
Menasche et al.
(2015)
Case report • Ischaemic
heart failure
undergoing
CABG
• n = 1
• Human
ESC‑derived
cardiac progenitor
cells on a fibrin
scaffold
• Surgical patch
implantation
NA NA • First-in-man study
• New-onset
contractility
observed in the
patched region
REGENERATEAMI
(2016)
• Randomized
• Double-blind
• Placebo
controlled
• Acute
myocardial
infarction
• n = 100
• Autologous bone
marrow‑derived cells
• 59.8 × 106 cells
• Intracoronary
delivery
Change in
LVEF
Negative for
primary end
point
Large,
double-blinded study
that failed to meet
primary efficacy end
point
Abbreviations: CABG, coronary artery bypass grafting; ESC, embryonic stem cell; LVEF, left ventricular ejection fraction;
MACE, major adverse cardiovascular events; NA, not applicable; PET, positron emission tomography, SPECT, single-photon
emission-computed tomography. *Subject to expression of concern. ‡See individual studies for precise characteristics of
bone marrow-derived cells used.
However, the use of PSCs for cardiovascular regeneration still faces some challenges. One of
them is the survival of PSC-derivatives to the inflammatory environment of infarcted myocardium leading
to the failure of a fully engraftment into the host. This fact provides the need for a sufficiently large cell
graft, requiring time and great cost. Also, the misalignment of the engrafted cardiac cells into the host
myocardium can cause ventricular arrhythmias (Sayed et al., 2016). Bioengineering tools like tissue
scaffolds can be the answer to assist engraftment of transplanted cardiomyocytes. Among them are
collagen, fibrin, gelatin, hyaluronic acid, chitosan, alginate, or decellularized tissues (Tzahor and Poss,
2017). Taking advantage of these knowledge, cell-sheet systems composed of multiple hiPSC-derived
cardiovascular cell lineages along with biomaterials, have been tested as a solution for easily produce a
transplantable tissue-like structure (Masumoto et al., 2014, 2016).
The regeneration of the adult human heart represents one of the major challenges in science
given the magnitude of heart disease. Although initial studies are promising, clinical treatments need to
be more efficient until hiPSCs-derived cardiomyocytes (hiPSC-CMs) can be translated into human benefit.
13
Cardiac regeneration will probably require more than simply injecting the right type of cells in the right
place. The foundation of knowledge of cardiac developmental mechanisms, both in normal development
and after injury, seems to provide a strong platform to achieve this goal.
The use of animals as model systems might lead the scientists to fail to recapitulate the human
disease phenotypes, creating the need for disease modelling using human samples. The same problem
is inherent to drug screening. For cardiovascular diseases, it is difficult to compare data between other
mammalians and humans given some variation in their heart rates and electrophysiological properties
(Sayed et al., 2016). The solution for such limitations can depend on hiPSCs.
The use of hiPSCs for disease modelling, but also as a platform for drug screening, requires the
identification of disease-specific phenotypes. For cardiovascular diseases, physiological phenotypes
(which are the most challenging to assess) are particularly relevant, making it essential to evaluate features
like ion exchange, channel activity and contractility. This evaluation is typically performed using multi-
electrode arrays or by patch-clamp analysis of the cardiac muscles (Avior et al., 2016).
However, for diseases that do not affect ion channels encoding genes, phenotypic characterization
has relied on cellular characteristics such as the cardiomyocyte size and redistribution of cardiac proteins
from the cytoplasm to the nucleus. At the molecular level, some assays have been equally considered,
including the analysis of mitochondrial activity, glucose and fatty acid metabolism and still the expression
of stress-associated proteins, apoptosis and disease markers. The platform of hiPSCs differentiation into
cardiomyocytes can also be important since three-dimensional cultures can be particularly interesting to
evaluate structural phenotypes like spatial orientation (Avior et al., 2016).
Among the different diseases studied with hiPSCs, there are several heart diseases that have been
a target of interest by the scientific community. An illustration of that are long QT syndromes (LQTS) that
are caused by mutations in ion channels, such as the potassium, sodium and calcium channel encoding
genes. hiPSCs from patients affected by LQTS allowed to know the characteristic electrophysiological
features of the disease such as prolonged action potential duration and arrhythmia, demonstrating that
hiPSCs could model the features of the mutation. hiPSC-CMs were also differentiated from patients with
LEOPARD syndrome, a syndrome that leads to mutations in a gene that interferes in the RAS-MAPK
signalling pathway that regulates several aspects of cellular function. These cardiomyocytes could present
features of hypertrophy, a typical trait of this multi-systemic disease (Bellin et al., 2012).
At the same time, hiPSC technology also contributes to the screening of several drugs in
cardiovascular diseases. These cells may be advantageous over immortalized cell lines since they have
functional characteristics similar to those of primary human cardiomyocytes (Navarrete et al., 2013). One
common cause of drug development failure rely on drug-induced arrhythmia and hiPSC-CMs have shown
to be an efficient platform for arrhythmia screening using microelectrode arrays (MEA) (Navarrete et al.,
2013). Drug-induced cardiotoxicity profiles were also recapitulated for healthy subjects, LQTS,
hypertrophic cardiomyopathy, and dilated cardiomyopathy patients using de novo cardiomyocytes for the
14
first time (Liang et al., 2013). Propanolol and nadolol are two β‑adrenergic blocking agents that could
attenuate arrhythmia in hiPSC-CMs from patients with LQTS, constituting an effective therapy that is
already in clinical use (Bellin et al., 2012).
hiPSC-CMs can provide an effective platform for drug screening and toxicology studies, still they
present several limitations related with cardiomyocyte maturation state that often resemble fetal
cardiomyocytes at many levels (Sayed et al., 2016). Some effort has been made to mature hiPSC-CMs
and this topic will be approached latter on.
Virtually, it is possible to differentiate PSCs in any desired cell type, but in order to do that, it is
essential to control cell differentiation and to direct PSCs along specific pathways. In fact, most of the
differentiation protocols are based on recapitulating differentiation events that occur in the natural
development of the embryo. Stepwise progression aiming for a certain state of differentiation can be
possible thanks to signalling pathway agonists and antagonists in specific concentrations and sequences
(Bellin et al., 2012).
Considerable progress has been made in the past few years for the development of defined
protocols for cardiac derivation of functional cardiomyocytes. Differentiation into cardiomyocytes is very
delicate and there must be an optimization of each individual component that takes part in the process.
Between the most successful strategies to generate cardiac cells at high efficiencies are the ones that
manipulate the main regulatory signalling processes and respective signalling molecules involved in
cardiogenesis in vivo. Until this moment, no evidence can be found that can deny the resemblance
between the gene regulatory networks that control PSC differentiation and those that control human
development (Burridge et al., 2015). Among the recurring players are Wnt proteins, FGFs and many
members of the TGFβ superfamily, including BMPs, Nodal and Activin (Burridge et al., 2012).
Also, two basic systems for in vitro PSC cardiomyocyte differentiation are used: culture as three-
dimensional (3D) aggregates or as a two-dimensional (2D) adherent monolayer. In both cases, sequential
inductive environments using growth factors and/or small molecules are used (Burridge et al., 2012).
Protocol efficiency is usually assessed by specific markers of cardiomyocytes, such as cardiac troponin T
(cTnT) and/or alpha actinin (α-ACT), encoded by the TNNT2 and ACTN1 genes, respectively.
The first attempts for in vitro differentiation of PSCs to cardiomyocytes arise with EC cells and
mESCs cultures. When these cells are dissociated to single cells, they form suspended colonies that
develop into three-dimensional aggregates with a spheroid shape. The similarity between these
aggregates and early post-implantation embryos, led to the term embryoid bodies (EBs) (Stevens, 1960).
15
These EBs spontaneously recapitulate early steps of embryogenesis in an uncontrolled fashion, ending in
an highly variable structure with differentiated cells of all three germ layers, where cardiomyocytes can
be present (Mummery et al., 2012). The initial efforts to differentiate hESCs into cardiomyocytes using this
technique were performed with fetal bovine serum (FBS) and had differentiation efficiencies of ~5-10%
(Kehat et al., 2001), proving this method to be inefficient. Even so, differentiation protocols from hiPSC
lines were also tested using this protocol (Zhang et al., 2009; Zwi et al., 2009).
Soon, morphogens as Activin A and FGF2 were used for mesoderm induction in hESCs EBs
(Burridge et al., 2007) (Figure 3A), raising the efficacy to ~25%. Additional growth factors, like BMP4
and WNT3A, also proved to play an important role in the optimization of EB mesoderm induction for both
hESC and hiPSC lines (Burridge et al., 2011; Tran et al., 2009; Yang et al., 2008). Lastly, the mesoderm
induction via growth factors was replaced by a small molecule, the glycogen synthase kinase 3 beta
(GSK3β) inhibitor CHIR99021 (Gonzalez et al., 2011).
For some growth factors such as Wnts that promote mesoderm induction at early stages, it may be
important to inhibit signalling later (Ueno et al., 2007). Accordingly, the addition of the Wnt inhibitor
DKK1 represented a major breakthrough in cardiac specification (Yang et al., 2008). Similar to what
happened for the induction of mesoderm, some small molecules also operate in cardiac specification,
like SB431542 and dorsomorphin (Kattman et al., 2011) and the inhibitor of Wnt response 1 (IWR-1)
(Ren et al., 2011; Willems et al., 2011). Besides the cardiac specification, some additional factors can
improve the efficiency of cardiomyocyte differentiation from EBs as vascular endothelial growth factor
(VEGF) and lower concentrations of insulin (Mummery et al., 2012).
To overcome EB size heterogeneity some of these experiments were based on strategy characterized
by “forced aggregation” using U- or V-shaped wells, leading to 3D aggregates with identical size in each
well (Burridge et al., 2007, 2011; Elliott et al., 2011) (Figure 3B). This feature can be controlled by varying
the starting number of cells in each well. A high-throughput approach to produce thousands of
aggregates per area is commercially available as AggreWells™ plates (by STEMCELL Technologies)
(Burridge et al., 2012; Mummery et al., 2012). In addition, aggregates cultured in microwells exhibited
higher levels of Wnt signalling than EBs and they showed upregulation of genes associated with
cardiogenesis (Azarin et al., 2012).
An alternative approach to obtain PSC-CMs is characterized by the expansion and differentiation
of cells as a 2D adherent monolayer (Figure 3C). The first attempt of this method cultured hPSCs to a
high density using activin A followed by BMP4, resulting in a differentiation efficiency of ~30% (Laflamme
et al., 2007). No factors were added to cardiac specification, but not changing the medium for some
days allows the accumulation of potential secreted factors as DKK1. Similarly to 3D methods, the addition
of WNT3A and DKK1 were only tested in subsequent studies (Paige et al., 2010). Naturally, small
molecules like the IWR-1 or others, like the inhibitor of Wnt production 4 (IWP-4), have also proven
effective in 2D protocols (Hudson et al., 2011).
16
Figure 3 | Methods for cardiomyocyte differentiation of pluripotent stem cells.
(A) 3D suspension EBs, (B) 3D aggregates and (C) 2D adherent monolayer as the three methods for cardiomyocyte
differentiation (Adapted from Burridge et al., 2012).
An efficient and growth factor-free protocol was developed by Lian and colleagues, which relay on
the temporal modulation of the canonical Wnt signalling pathway by firstly activating Wnt signalling for
mesendoderm induction, using the GSK3B inhibitor CHIR99021 followed by a step of Wnt inhibition for
cardiac lineage specification using IWP-2 or IWP-4 (Lian et al., 2012). Media change was performed
every 2 or 3 days until day 15, with beating clusters observed about day 10. iPSCs differentiation was
induced in the chemically defined RPMI 1640 medium supplemented with B27 (a complex mixture of 21
components, some of animal origin). This protocol, which uses a GSK3B inhibitor followed by a Wnt
inhibitor (termed GiWi protocol), can generate up to 98% cTnT+ cardiomyocytes (Lian et al., 2012,
2013a). An important fact related with this protocol lies on the use of B27 without insulin in the first 7
days of differentiation and with insulin on the following days, since insulin inhibits cardiac mesoderm
induction (Lian et al., 2013b). A similar protocol but with a different Wnt inhibitor, KY02111, was also
published (Minami et al., 2012).
Despite the advance in cardiac differentiation protocols, there are some obstacles like the
complexity of the used medium and of the involved secretome that have to be overcome. Accordingly, the
2D monolayer system was tested more recently under chemically defined conditions on synthetic matrices
(Burridge et al., 2014). In this protocol the unnecessary components of B27 were removed ending in a
medium formulation containing only 3 components: the basal medium RPMI 1640, L-ascorbic acid
2-phospate and rice derived recombinant human albumin. The termed CDM3 medium along with
17
biphasic Wnt induction with CHIR99021 and Wnt-C59 led to the production of cardiomyocytes at >85%
purity. According to the same line of thought, the authors of the GiWi protocol optimized it as a an
albumin-free cardiomyocyte differentiation platform, the GiWi2 (Lian et al., 2015), that produces
88–98% cTnT+ cells.
Theoretically, the 2D system can present some advantages over the 3D system, since a uniform
adherent monolayer of cells is free of complex diffusional barriers present in EBs and aggregates. Despite
that, all these protocols suggest the same message: in vitro differentiation protocols are highly dependent
on the same signalling and gene expression pathways required in vivo. Both 3D and 2D methodologies
play with the same major signalling pathways such as BMP, TGFβ/Activin/Nodal, Wnt, and FGF, all of
them with highly specific temporal windows for effectiveness. Indeed, the same markers that are present
during cardiogenesis in vivo has proven to be a constant between methodologies throughout
cardiomyocyte differentiation (Burridge et al., 2012, 2015).
Regenerative therapies, like the ones developed for myocardial infarction, can require up to 108 to
109
cardiomyocytes per patient. Consequently, clinical therapies and other industrial applications of
hPSC-CMs, will require large-scale productions with large cell quantities of cells being generated under
highly robust, well-defined, and economically viable conditions (Kempf et al., 2014; Mummery et al.,
2012).
To achieve this goal, the expansion of hPSCs has to be as scalable as the differentiation
methodology. Indeed, the scalability of hPSCs cultures has seen some improvements in the last years
using defined culture media and defined reagents in a broad spectrum of bioreactors. Among the many
advantages of this technology, it can be highlighted that some bioreactors allow continuous perfusion as
well as automated control of pH, oxygen tension and temperature (Kropp et al., 2017).
The first description of large-scale production of PSC-CMs was performed with mESCs as a
suspension culture of EBs (Zandstra et al., 2003). It is important to this kind of culture to be dynamic,
preventing the settling of the spheroids and ensuring an adequate exchange of metabolites, nutrients,
and important molecules for cardiac differentiation. Likewise, the dynamic nature of the culture should
have as a priority the maintenance of homeostasis of physicochemical properties like pH and oxygen
levels (Mummery et al., 2012).
Several scalable cardiac differentiation culture systems have arisen and some particular cases can
be emphasized such as the encapsulation of EBs in hydrogels to prevent agglomeration (Bauwens et al.,
2005); single-cell suspensions in bioreactors with optimized conditions and cell densities to generate
aggregates of a specific size (Schroeder et al., 2005) and forced aggregation to produce size-specified
aggregates followed by loading into the bioreactor (Niebruegge et al., 2009). The use of microcarriers,
allowing to culture adherent cells in suspension conditions, has been studied as a platform for long-term
PSCs expansion. This technology was also reported for cardiac differentiation but with low efficiencies,
with a maximum of 20% of α-ACTN+ cells (Lecina et al., 2010).
18
It should be noted that the development of scale-up platforms go hand in hand with cardiac
differentiation protocols. Example of that is the application of the GiWi protocol developed by Lian and
his group to a scalable suspension culture resulting in the generation of 4x107 to 5x10
7 hPSC-CMs per
differentiation batch at >80% purity (Kempf et al., 2014, 2015). The same protocol was also applied in
the production of hPSC-CMs in suspended microcarrier cultures, this time with a higher differentiation
efficiency with 65% of cTnT+ cells (Ting et al., 2014).
Novel methods for large-scale cardiomyocyte differentiation also include the 2D culture system, in
which the hiPSCs are cultured in multilayer culture plates with active gas ventilation, resulting in
productions up to 1.5x109 differentiated cells with a differentiation efficiency of 66-87% (Tohyama et al.,
2017).
An alternative approach to the described methods for production of hPSC-CMs is the generation
of multipotent cardiovascular progenitor cells (CVPCs). These cells can differentiate into cardiomyocytes,
smooth muscle cells, and endothelial cells and were first reported as cKit+ cells isolated from rat hearts
(Beltrami et al., 2003) and later they were also reported in human hearts (Messina et al., 2004). These
cells can proliferate for extended periods of time without loss of potential. However their great weakness
is that none of the described cell surface markers until now are exclusively expressed on cardiac cell
lineages (Mummery et al., 2012). Meanwhile, a chemically defined system for robust generation of hPSC-
derived CVPCs were described by simultaneously inhibiting GSK3β, BMP, and TGFβ/Nodal/Activin
signalling (Cao et al., 2013). This method could be a promising alternative as a simplified route for
generating large numbers of cardiovascular cells.
Another method for generation of de novo cardiomyocytes is transdifferentiation, the conversion
of a cell from one specialized state to another (Merrell and Stanger, 2016). Direct reprogramming of
cardiac fibroblasts into cardiomyocytes has been reported using cardiac-specific transcription factors such
as GATA4, MEF2C, and TBX5 (GMT factors), though from the generated cardiomyocytes only less than
10% were able to spontaneously contract (Ieda et al., 2010). Several improvements for this technology
were published recently with different cocktails of transcription factors, but in contrast to this conventional
reprogramming, a chemical approach with small molecules that modulate endogenous factors in the
starting cell type was also demonstrated (Cao et al., 2016). However, transdifferentiation has some
limitations as the low amount of starting cellular material resulting in a small number of cardiomyocytes
produced (Burridge et al., 2015). Integrating retro/lentiviruses used in transdifferentiation may hinder the
use of this technology for in vivo translational purposes but the novel chemical models can overcome this
problem.
The generation of de novo cardiomyocytes frequently results in a mixed cardiovascular population
as protocols produce predominantly ventricular cells, with ∼15–20% atrial cells and ∼5% nodal cells
(Burridge et al., 2015). The use of mixed populations can be problematic for some applications, indeed
19
the production and/or isolation of specific cardiomyocyte subtypes can be useful to model and treat
diseases that affect specific regions of the heart (Lee et al., 2017). For instance, ischemic damage of the
ventricular myocardium after myocardial infarction would require hPSC-derived ventricular
cardiomyocytes, likewise, the screening of atrial specific drugs would require hPSC-derived atrial cells and
cardiac rhythm malfunctions would require hPSC-derived nodal cells.
Protocols for production of specific cardiomyocyte subpopulations have been developed based in
established protocols. The addition of the small molecule AG1478 that inhibits neuregulin signalling,
BMS-189453 that inhibits retinoic acid signalling or the addition of retinoic acid itself, can increase the
nodal population up to ∼50% (Zhu et al., 2010), the ventricular population to 83%, and the atrial
population to 94% (Zhang et al., 2011), respectively.
Besides that, as previously approached, FHF and SHF have different contributions concerning the
different subtypes of cardiomyocytes. Very recently, it was demonstrated that the distinct mesoderm
populations that give rise to human ventricular and atrial cardiomyocytes are specified with different
concentrations of BMP4 and Activin A and can be identified based on the expression of glycophorin A
(GYPA) and retinaldehyde dehydrogenase 2 (RALDH2) genes, respectively (Lee et al., 2017). This finding
can be a platform to develop protocols with efficient generation of specific cardiomyocyte subtypes.
In humans, the different subtypes of cardiomyocytes have some specific markers as the already
mentioned MYL7 that is expressed throughout the atria, ventricles, and outflow tract and MYL2 that is
specific for ventricular cells. Beside these makers, there are other ones such as iroquois homeobox 4
(IRX4) (ventricular), HCN4 and TBX18 (nodal), and NPPA and NPPB (atrial and ventricular) (Burridge et
al., 2015).
It is the characterization of de novo cardiomyocytes that defines the success of differentiation.
Indeed, the first sign of contractile foci is probably the first clue of a successful cardiac differentiation.
However, more stringent tests have been applied to evaluate the efficiency of current differentiation
protocols, in which the fraction of cardiomyocytes is estimated by flow cytometry for a cardiac-specific
protein like cTnT. Quantitative real-time polymerase chain reaction (qRT-PCR) has been used as well for
PSC-CMs characterization purposes assessing the expression of several cardiac genes, whether they are
genes encoding transcription factors or myofilament proteins that are known to be present in
cardiomyocytes in vivo. These proteins can also be immunolabeled with antibodies (Mummery et al.,
2012). A more high-throughput approach includes, for example, the study of the transcriptome, consisting
in the complete set of RNA transcripts that are produced by the genome for a specific developmental
stage. RNA sequencing (RNA-seq) is one of the approaches that allow transcriptomic analysis using deep-
sequencing technologies (Wang et al., 2009). In fact, transcriptomics analysis can be particularly
interesting for cardiomyocyte production since it can quantify the changing expression levels of each
transcript during cell differentiation.
20
One of the main proprieties of the myocardium is its contraction and relaxation, coordinated on a
beat-to-beat basis by a cycling of calcium from the cytoplasm, sarcoplasmic reticulum, and the sarcomere.
It is necessary the generation of action potentials (APs) for propagation of the electric signal for triggering
intracellular calcium release and subsequent contraction (Li et al., 2013). Similar machinery seems to be
present in PSC-CMs and accordingly, functional assays have particular importance in the characterization
of cardiomyocytes. Cardiac APs have been demonstrated by use of microelectrode arrays and their
properties can provide insights on the type and maturity of the derived cardiomyocytes. Besides that, ionic
currents responsible for cardiac APs can also be evaluated by single-cell patch-clamp and the
electrophysiology of multicellular preparations can also be monitored by multielectrode arrays (Mummery
et al., 2012). During the calcium cycling, the rise and subsequent decay of calcium, known as the calcium
transient, modulates both the contractile force and frequency contraction of cardiomyocytes. This
phenomenon can also be recorded with fluorescent calcium indicators in beating clusters or single cells
(Li et al., 2013).
To take full advantage of the potential applications of PSC-CMs, these cells must recapitulate the
properties of adult cardiomyocytes in vivo. However, using the abovementioned methodologies, it has
been reported that PSC-CMs do not exhibit the morphological and functional characteristics of adult
cardiomyocytes (Table 2). PSC-CMs usually display an immature phenotype resembling that of fetal
cardiomyocytes, in terms of automaticity (spontaneous contraction), proliferation, metabolism, cellular
structure, calcium handling, and electrophysiological properties (Yang et al., 2014).
Morphologically, a human adult cardiomyocyte is elongated and cylindrical (~150 x 10 µm for
ventricular cells) whereas a fetal cardiomyocyte is smaller. Comparatively, hPSC-CMs are smaller as well
and circular (~5-10 µm in diameter), but with a more developed shape in late hPSC-CMs (~30 x 10
µm). In addition, adult cardiomyocytes are usually bi- or multinucleated, while immature hPSC-CMs are
mono-nuclear (Robertson et al., 2013). Over time, the sarcomeric structure becomes more organized and
it increases in length, with each sarcomere reaching ~2.2 µm in relaxed adult cardiomyocytes and ~1.65
µm in immature hPSC-CMs. During maturation, myofibrillar protein isoforms undergo switching, having
impact in the contractility, and these isoforms can also be studied. Also, mitochondria is a critical
component of cardiomyocyte maturation occupying ~20% to 40% of the adult cardiomyocyte volume
while in immature hPSC-CMs it occupies a small fraction of cell volume (Yang et al., 2014).
Another parameter to assess cardiomyocyte maturation is related with cell junctions. In adult
cardiomyocytes, the gap junction protein CX43 and the adherens junction protein N-cadherin (N-CAD,
also known as CDH2), are concentrated into intercalated disks at the ends of the cells, but in immature
hPSC-CMs they are circumferentially distributed. This redistribution has impact in the conduction velocity
of the cell (Yang et al., 2014).
In what concerns to functional parameters, the study of APs is particularly relevant, allowing the
distinction of the different subtypes of cardiomyocytes. Based on the AP it has been reported that
hPSC-CMs consist of a heterogeneous population with atrial -, nodal-, and ventricular-like
21
Table 2 | Summary of differential properties of cardiomyocytes dependent on the maturation status.
Different features of immature and mature cardiomyocytes. (Adapted from Sayed et al., 2016).
Parameters Immature Mature
Morphology
Shape Round or polygonal Rod and elongated
Size 20–30 pF 150 pF
Nuclei per cell Mononucleated ∼25% Multinucleated
Multicellular organization Disorganized Polarized
Sarcomere appearance Disorganized Organized
Sarcomere length Shorter (∼1.6 μm) Longer (∼2.2 μm)
Sarcomeric proteins
• MHC β > α β >> α
• Titin N2BA N2B
• Troponin I ssTnI cTnI
Sarcomere units
• Z-discs and I-bands Formed Formed
• H-zones and A-bands Formed (prolonged
differentiation) Formed
• M-bands and T-tubules Absent Present
Electrophysiology
Action potential properties
• Resting membrane potential ∼−60 mV −90 mV
• Upstroke velocity ∼50 V/s ∼250 V/s
• Amplitude Small Large
• Spontaneous automaticity Exhibited Absent
Ion currents
• Hyperpolarization-activated pacemaker (If) Present Absent
• Sodium (INa) Low High
• Inward rectifier potassium (IK1) Low or absent High
• Transient outward potassium current (Ito) Inactivated Activated
• ATP-sensitive K+ current (IK,ATP) Not reported Present
• L- and T-type calcium (ICa,L and ICa,T), rapid
and slow rectifier potassium currents
(IKr and IKs), Na+-Ca
+ exchange current
(INCX) and acetylcholine-activated K+(IK,ACh)
Similar to adult CMs
Conduction velocity Propagation of signal Slower (∼0.1 m/s) Faster (0.3–1.0 m/s)
Gap junctions Distribution Circumferential Polarized to intercalated
disks
Calcium handing
Ca2+
transient Inefficient Efficient
Amplitudes of Ca2 +
transient Small and decrease with
pacing Increase with pacing
Excitation-contraction coupling Slow Fast
Contractile force ∼nN range/cell ∼μN range/cell
Ca2+
handling proteins
• CASQ2, RyR2, and PLN Low or absent Normal
Force-frequency relationship Positive Negative
Mitochondrial
bioenergetics
Mitochondrial number Low High
Mitochondrial volume Low High
Mitochondrial structure Irregular distribution,
perinuclear
Regular distribution,
aligned
Mitochondrial proteins
DRP-1 and OPA1 Low High
Metabolic substrate Glycolysis (glucose) Oxidative (fatty acid)
Responses to β-
adrenergic stimulation
Response Lack of inotropic reaction Inotropic reaction
Cardiac alpha-adrenergic receptor ADRA1A Absent Present
Abbreviations: ATP, adenosine triphosphate; CM, cardiomyocyte; MHC, major histocompatibility complex.
22
cardiomyocytes. Overall, the reported AP properties are typical for cardiomyocytes less mature than adult
cardiomyocytes (Robertson et al., 2013).
Calcium handling is another functional parameter that has been widely reported with hPSC-CMs
exhibiting Ca2+
transients. Differential Ca2+
handling properties are also explained by variability in the
maturation status of hPSC-CMs. In adult cardiomyocytes, the sarcoplasmic reticulum is responsible for
almost 70% of calcium release, whereas in immature hPSC-CMs it seems that sarcoplasmic reticulum has
little contribution to calcium transients that truly occurs through the cell membrane, being smaller and
slower. After the beginning of the differentiation process, spontaneous contractions could be seen as early
as day 5, with different basal rhythms ranging from 21 to 52 beats per minute (BPM) (Robertson et al.,
2013). Additionally, β-adrenergic stimulation has been studied in hPSC-CMs and it was found that the β-
adrenergic agonist, isoproterenol, is able to induce a dose-dependent increase of the frequency of
spontaneous beating (Yang et al., 2014).
It takes years for human cardiomyocytes to reach adult form. For hPSC-CMs there are a plenty of
approaches that can be applied envisioning the maturation of de novo cardiomyocytes. These cells can
mature over time in culture, but besides that, another methodology for improving cardiomyocyte
maturation includes the addition of certain molecules. Triiodothyronine is an example of that and it is
able to upregulate genes like MYH7 along with numerous ion channels. Insulin-like growth factor 1
(IGF-1) is an additional example of these molecules. Other techniques comprise adrenergic stimulation,
electrical stimulation, stretch stimulation, micropatterning and maturation of metabolic profiles by
switching metabolism from glycolysis to oxidative phosphorylation (Burridge et al., 2015; Yang et al.,
2014). An example of that is the platform Biowire, that combines a 3D culture system with electrical
stimulation to mature cardiomyocytes, improving myofibril ultrastructural organization, conduction
velocity and both electrophysiological and calcium handling properties (Nunes et al., 2013).
A multiplicity of shapes can be found throughout the biological kingdom creating the complex
morphology of the different tissues. Physical properties of the ECM and mechanical forces are essential
to the morphogenetic processes present in embryonic development, defining tissue architecture and
driving specific cell differentiation programs. Force and geometry, physical aspects of the cellular
environment, can be sensed by mechanotransduction systems that can then transduce the physical signals
into biochemical responses. This transduction possibly activates many signalling pathways producing
controlled functional responses. Therefore, sensing of geometry is crucial to the cellular sensing of 2D
versus 3D matrices (Downing et al., 2013; Dupont et al., 2011; Vogel and Sheetz, 2006). Indeed, the 3D
fibrous meshwork of the natural ECM is very different from an uniformly coated 2D surface, providing
different biochemical and physical cues.
In the heart, in a 3D oriented configuration, each cardiomyocyte establishes cell–cell interactions
with approximately other eleven cardiomyocytes. In the 2D culture system, this extensive network of
cell-cell and cell-ECM interactions does not exist. Therefore, cardiac tissues obtained from 3D culture
23
system exhibit more physiologic electrical and contractile function than the constructs obtained in 2D
(Atmanli and Domian, 2017). Matrix elasticity can also be highlighted as a major difference between
both culture systems, as ECM has impact in how the cardiac tissue contract. 3D constructs are capable of
contract against the ECM, while in 2D, there is only a layer of ECM allowing the cells to sense the stiffness
of the underlying surface and to contract against it. This fact can has a negative impact on cardiomyocyte
differentiation since cells are viable on stiff substrates and gene expression associated to cardiac
development is downregulated (Atmanli and Domian, 2017).
“Because life is 3D”, the 2D culture interaction between cells and surface are seen as an artificial
condition. In contrast, it is believed that 3D culture better mimic the physiological conditions, making it
an emerging trend as building blocks for tissue engineering (Laschke and Menger, 2017).
Although extensive studies have been performed on the roles of growth factors and chemical
compounds in cell differentiation, the role of biophysical cues in this process is still to be clearly
understood. Overall, it has been established that gene expression for proteins of the cytoskeleton,
proliferation, transcription, translation, ECM production and inter- and intracellular signalling pathways,
is shape-dependent (Vogel and Sheetz, 2006).
24
25
As previously mentioned, CVDs represent the number one human cause of death worldwide, and
it was estimated that in 2015, 17.7 million died from CVDs, representing 31% of all global death (WHO,
2017). Accordingly, cardiomyocytes derived from hiPSCs represent a promising cell source for heart
regeneration, disease modelling and drug screening. In the literature, current approaches are based on
the cardiomyocyte differentiation of 2D adherent monolayer and 3D aggregates. However, similarities
and differences between these two different culture platforms are still poorly explored.
At the first place, this work aims to characterize an optimized 3D platform for derivation of hiPSCs
into cardiomyocytes, establishing a comparison with a previously reported 2D culture system. The present
3D culture is expected to overcome some limitations described on distinct 3D approaches as the variability
in cardiac differentiation efficiencies, lack of reproducibility and uncontrolled aggregate size.
This study also aims to bring novel insights on the gene expression patterns throughout sequential
stages of cardiomyocyte differentiation for the 3D culture system. To do so, a transcriptomic analysis
based on RNA-seq data, supported by other techniques as well, was here performed.
As a remarkable goal, this work aims to unveil how the particular microenvironment of both 3D
and 2D culture systems can be translated into distinct molecular signatures that may exist between
platforms. In case of a different profile between the two different culture modes, its origin must be further
analysed.
At last, but not least, it is expected to explore the maturation status and functional properties of the
generated cardiomyocytes, given the importance of such attributes to a full benefit of hiPSC-CMs
potential. Simultaneously, the results should aid to understand how close de novo cardiomyocytes
resemble adult phenotypes.
26
27
In this work, the majority of the experiments were performed using the hiPSC line iPS-DF6-9-9T.B,
provided by WiCell Bank (Wisconsin, USA). This cell line is vector free and was reprogramed from foreskin
fibroblasts with a karyotype 46, XY that were collected from healthy donors using defined factors in the
Laboratory of Dr. James Thomson, at University of Wisconsin. When referred, other two cell lines were
used: the hiPSC line WT-F002.1A.13 provided by TCLab (Tecnologias Celulares para Aplicação Médica,
Unipessoal, Lda.) and Gibco® Episomal hiPSC line provided by Thermo Fisher Scientific.
As adhesion substrate for hiPSCs, Matrigel® matrix (Corning®, #354277) was used. Matrigel®
was aliquoted and stored at -20ºC. Aliquots were thawed on ice and diluted 1:100 (v/v) in cold Dulbecco’s
Modified Eagle Medium: Nutrient Mixture F-12 (DMEM/F12) (Gibco™/ Thermo Fisher Scientific,
#32500-035). This formulation was used to coat 6-well tissue culture plates (Corning®) that were left at
room temperature least at for two hours, or stored at 4ºC for later use up to 2 weeks.
For cell culture, mTeSR™1 medium (STEMCELL Technologies™, #85850) was used. mTeSR™1
5x supplement was thawed at room temperature or at 4ºC overnight and mixed with basal medium. The
final formulation was aliquoted and stored at -20ºC. Aliquots were thawed overnight at 4ºC and
supplemented 1:200 (v/v) with penicillin/streptomycin (PenStrep) (Gibco™/ Thermo Fisher Scientific,
#15140122). mTeSR™1 was pre-warmed at room temperature before use.
For each experiment, cells cryopreserved in liquid nitrogen at -196ºC were thawed at 37ºC for
30 s. The cell content of a cryovial was then gently resuspended in culture medium in a centrifuge tube
and centrifuged at 200 xg for 3 min. Cell pellet was resuspended in culture medium and the resulting cell
suspension was seeded onto Matrigel® coated culture plate with a final volume of 1.5 mL. For
maintenance of hiPSCs, cells were kept in a CO2 incubator at 37°C, 5% CO2 and 20% O2 and culture
medium was changed daily.
28
When hiPSCs cultures attained a confluence of 60-70%, enzyme-free cell passaging was
performed using EDTA solution (Invitrogen™/ Thermo Fisher Scientific, #15575-038) diluted in
phosphate-buffered saline (PBS) (Sigma-Aldrich®/ Merck, #D8662) at a concentration of 0.5 mM. After
culture medium removal, cells were washed twice and then incubated with EDTA at room temperature for
5 min. After EDTA removal, cells were flushed with culture medium and collected to a centrifuge tube
followed by seeding onto new Matrigel® coated culture plates. Splits 1:3 were performed.
For hiPSCs storage, after culture medium removal, cells were washed twice and then incubated
with EDTA at room temperature for 5 min. After EDTA removal, cells were flushed with culture medium,
collected to a centrifuge tube and centrifuged at 200 xg for 3 min. Cell pellet was resuspended in
KnockOut™ serum replacement (KO-SR) (Gibco™/ Thermo Fisher Scientific, #10828028) with 1:10 (v/v)
of dimethylsulfoxide (DMSO) (Sigma-Aldrich®/ Merck, #276855) at a cell density of about 1.5x106
cells/vial for a final volume of 250 µL/vial. The cell suspension was loaded into a cryovial and stored in
a freezing container at -80ºC for 24h followed by storage in liquid nitrogen at -196ºC.
Whenever it was necessary to assess cell density, cell counting was performed. Cell suspension
sample was diluted 1:1 in Trypan Blue solution (Gibco™/ Thermo Fisher Scientific, #15250-061) and
after homogenization, cells were counted using a haemocytometer under an inverted optical microscope.
Cells were seeded onto Matrigel® coated 12-well tissue culture plates (Corning®) at a cell
density of 4x105 cells/well for a final volume of 1 mL/well (1x10
5 cells/cm
2). Culture medium was changed
daily until a confluence of 90%-95% was achieved.
Prior to cell aggregation, cells were incubated with culture medium supplemented 1:1000 (v/v)
with Rho kinase inhibitor (ROCKi, Y-27632) (STEMCELL Technologies™, #72304) at 37ºC for 1 h. After
culture medium removal, cells were washed with PBS and then incubated with Accutase® solution
(Sigma-Aldrich®/ Merck, #A6964) at 37ºC for 7 min for single-cell formation. Cells were flushed with
Accutase® solution and collected to a centrifuge tube with culture medium, for inactivation of enzymatic
digestion, and centrifuged at 200 xg for 3 min. Cell pellet was resuspended in culture medium
29
supplemented with ROCKi followed by seeding onto AggreWell™800 plate (STEMCELL Technologies™)
with a cell density of 9x105 cells/well (3x10
3 cells/microwell) for a final volume of 1.5 mL/well. Before
seeding AggreWell™800 plate was centrifuged at 2700 xg for 5 min and after seeding at 300 xg for 3
min. Cells settle by gravitation and self-aggregate to form 3D aggregates. Cells were kept in a CO2
incubator at 37°C, 5% CO2 and 20% O2 and after the first 24h culture medium was changed daily without
ROCKi until an aggregate diameter of 300 µm was achieved.
During 3D aggregate expansion their size was monitored using an optical microscope DMI
3000B (Leica) with a digital camera DXM 1200F (Nikon). Aggregates were measured using ImageJ
Software. For diameter calculation, diameter of a circle of equal projection area (EQPC) was calculated
for aggregates during expansion (Equation 1). Average of maximum and minimum Feret diameter
(distance between two parallel tangents of the aggregate at a given angle) was calculated for aggregates
during differentiation due to a decrease in circularity and roundness.
𝑑 = 2 × √𝐴 𝜋⁄
Equation 1
For hiPSCs differentiation into cardiomyocytes, the GiWi protocol was performed (Lian et al.,
2013a). Roswell Park Memorial Institute (RPMI) 1640 medium (Gibco™/ Thermo Fisher Scientific,
#11875-093) was used as basal medium. From day 0 to day 6, cells were cultured in RPMI supplemented
with B-27™ minus insulin (Gibco™/ Thermo Fisher Scientific, #A1895601), the first basal medium, and
from day 7 until the end of differentiation, cells were cultured in B-27™ (Gibco™/ Thermo Fisher Scientific,
#17504001), the second basal medium. Basal medium was always supplemented 1:200 (v/v) with
PenStrep. At day 0 of differentiation, after mTeSR™1 culture medium removal, cells were cultured in the
first basal medium supplemented with the GSK3 inhibitor CHIR99021 (Stemgent, #04-0004) at a final
concentration of 6 µM for 2D and 11 µM for 3D. After 24 hours, medium was changed to the first basal
medium. At day 3, cells were cultured in the first basal medium supplemented with the Wnt inhibitor
IWP-4 (Stemgent, #04-0036) at a final concentration of 5 µM. At day 5, medium was changed to the first
basal medium. At day 7, medium was changed to the second basal medium and in the case of 3D
aggregates, they were flushed from the AggreWell™800 plate and transferred to a 6 Well Ultra-Low
Attachment plate (Corning®). Thereafter, medium was changed every 3 days until cell harvest.
30
For flow cytometry, cell samples of the selected time points were collected. After culture medium
removal, cells were washed with PBS and then incubated with Accutase® solution, in the case of
pluripotent cells, or with 0.25% (v/v) trypsin-EDTA (Gibco™/ Thermo Fisher Scientific, #R001100) in PBS,
in the case of differentiated cells, at 37ºC for 7 min. Then, culture medium was added for inactivation of
enzymatic digestion and after homogenization, cells were centrifuged at 200 xg for 3 min. Cell pellet was
washed with PBS and centrifuged again at 200 xg for 3 min. After PSB removal, cell pellet was fixed by
resuspension in 2% (v/v) paraformaldehyde (PFA) reagent (Sigma-Aldrich®/ Merck) in PBS and stored at
4ºC until flow cytometry.
Fixed cells were centrifuged at 200 xg for 3 min. Cell pellet was resuspended and incubated with
90% (v/v) cold methanol in Milli-Q® water at 4ºC for 15 min. Cells were then washed 3 times with flow
cytometry buffer 1 (FB1), constituted by 0.5% (v/v) bovine serum albumin (BSA) solution (Sigma-Aldrich®/
Merck, #A8327) in PBS, and centrifuged at 200 xg for 3 min, each time. Cell pellet was resuspended
and incubated in the primary antibody mouse IgG anti-cTnT (Invitrogen™/ Thermo Fisher Scientific,
#MA5-12960) diluted 1:800 (v/v) in FB2, constituted by 0.1% (v/v) Triton X-100 (Sigma-Aldrich®/ Merck,
#T9284) in FB1, at room temperature for 1 h. After incubation, cells were washed with FB2 and cell pellet
was resuspended and incubated in the secondary antibody Alexa Fluor™ 488 goat anti-mouse IgG
(Invitrogen™/ Thermo Fisher Scientific, #A11001) diluted 1:1000 (v/v) in FB2, at room temperature for
30 min in the dark. Cells were washed twice with FB2 and centrifuged at 200 xg for 3 min, each time.
Cell pellet were resuspended in FB1 for a final volume of 300 μL/FACS tube. Flow cytometry was
performed using a FACSCalibur™ flow cytometer (Becton Dickinson) and data analysis using Flowing
Software 2.0.
Fixed cells were centrifuged at 200 xg for 3 min. Cell pellet was washed twice with of 1% (v/v)
normal goat serum (NGS) solution (Sigma-Aldrich®/ Merck, #G9023) in PBS and centrifuged at 200 xg
for 3 min, each time. Cell pellet was resuspended in 3% (v/v) NGS solution in PBS, cell suspension
collected to a microtube and centrifuged at 200 xg for 3 min. Cell pellet was resuspended and incubated
in 1:1 of 3% NGS and 1% (v/v) saponin (Sigma-Aldrich®/ Merck, #47036-506-F) in PBS, at room
temperature for 15 min. After incubation, cells were centrifuged at 200 xg for 3 min and cell pellet was
resuspended and incubated in 3% NGS at room temperature for 15 min. After incubation, cells were
centrifuged at 200 xg for 3 min and cell pellet was resuspended and incubated in the primary antibody
mouse IgG anti-OCT4 (Merck Millipore/ Merck, #MAB4419) diluted 1:250 (v/v) in 3% NGS, at room
temperature for 90 min. After incubation, cells were washed twice with 1% NGS, and centrifuged at 200
31
xg for 3 min, each time. Cell pellet was resuspended and incubated in the secondary antibody Alexa
Fluor™ 488 goat anti-mouse IgG diluted 1:500 (v/v) in 1% NGS, at room temperature for 45 min in the
dark. After incubation, cells were washed twice with 1% NGS and centrifuged at 200 xg for 3 min, each
time. Cell pellet was resuspended in PBS for a final volume of 300 μL/FACS tube. Flow cytometry was
performed using a FACSCalibur™ flow cytometer and data analysis using Flowing Software 2.0.
Histological analysis with hematoxylin and eosin was performed by the Histology and
Comparative Pathology Laboratory at Instituto de Medicina Molecular (Lisbon, Portugal).
For immunocytochemistry, cell samples of the selected time points were stored. In the case of
single-cell, samples were collected after incubation with 0.25% (v/v) trypsin-EDTA in PBS and replated
onto Matrigel® coated Nunc™ Lab-Tek™ II chambered coverglass 8 well (Thermo Scientific™/ Thermo
Fisher Scientific). For all samples, after culture medium removal, cells were washed with PBS and then
fixed in 4% (v/v) PFA in PBS at 4ºC for 20 min. After PFA removal, cells were stored in PBS at 4ºC until
analysis. At the time of the analysis, in the case of 3D aggregates, cells were incubated with 15% (w/v)
sucrose in PBS, at 4ºC overnight; embedded in Tissue-Tek® optimum cutting temperature (O.C.T.)
(Sakura, #4583); and then frozen in isopenthane at -80ºC. Aggregate sections were cut on a cryostat-
microtome and collected on microscope slides by the Histology and Comparative Pathology Laboratory
at Instituto de Medicina Molecular (Lisbon, Portugal). For all samples, cells were incubated in 0.1 M
glycine (Merck Millipore/ Merck) for 10 min at room temperature to remove PFA residues, permeabilized
with 0.1% (v/v) Triton X-100 (Sigma-Aldrich®/ Merck) in PBS, at room temperature for 10 min and
blocked with 10% (v/v) fetal goat serum (FGS) (Gibco™/ Thermo Fisher Scientific) in TBST, constituted by
20 mM Tris-HCl pH 8.0, 150 mM NaCl and 0.05% (v/v) Tween-20 (Sigma-Aldrich®/ Merck) in PBS, at
room temperature for 30 min. Cells were incubated with the primary antibody (Table 3) diluted in blocking
solution at 4ºC overnight followed by incubation with secondary antibody at room for 30 min. Cells were
then incubated with secondary antibody (Table 3) at room temperature for 30 min and nuclear
counterstaining was performed using 0.15% (v/v) 4’,6-diamidino-2-phenylindole (DAPI) dye (Sigma-
Aldrich®/ Merck) in PBS, at room temperature for 5 min. In the case of 3D aggregates, after brief drying,
sections were mounted in Mowiol® (Sigma-Aldrich®/ Merck, #81381). Finally, immunofluorescence
staining images were acquired with a LSM 710 Confocal Laser Point-Scanning Microscope (Zeiss) for 3D
aggregate sections and single-cell samples, and with a fluorescence microscope DMI 3000B (Leica) with
digital camera DXM 1200F (Nikon), for 2D adherent monolayer; data analysis was performed using ZEN
Imaging Software (Zeiss) and ImageJ Software, respectively.
32
Table 3 | List of primary and secondary antibodies used for immunocytochemistry.
Primary Antibody Brand/ Company, #Cat. No. Dilution (v/v)
CD31 Mouse IgG Dako/ Agilent Technologies, #M0823 1:50
cTnT Mouse IgG Invitrogen™/ Thermo Fisher Scientific, #MA5-12960 1:200
CX43 Rabbit IgG Sigma-Aldrich®/ Merck, #C6219 1:400
E-CAD Rabbit IgG Cell Signaling Technology, #3195S 1:200
NANOG Mouse IgG BD Biosciences/ Becton Dickinson, #560259 1:200
N-CAD Mouse IgG BD Biosciences/ Becton Dickinson, #610920 1:500
NKX2-5 Rabbit IgG Santa Cruz Biotechnology, #SC-14033 1:500
α-ACT Mouse IgG Sigma-Aldrich®/ Merck, #A7811 1:200
α-SMA Mouse IgG Dako/ Agilent Technologies, #M0851 1:200
Secondary Antibody Brand/ Company Dilution (v/v)
Alexa Fluor™ 546 Mouse Anti-Mouse IgG Invitrogen™/ Thermo Fisher Scientific, #A11010 1:400
Alexa Fluor™ 488 Mouse Anti-Rabbit IgG Invitrogen™/ Thermo Fisher Scientific, #A11008 1:400
Prior RNA isolation, cell samples from sequential stages of cardiomyocyte differentiation were
collected. After culture medium removal, cells were washed with PBS and then incubated with
Accutase® solution at 37ºC for 7 min. Then, culture medium was added for inactivation of enzymatic
digestion and after homogenization, cells were centrifuged at 200 xg for 3 min. Cell pellet was washed
with PBS and centrifuged again at 200 xg for 3 min. After PSB removal, cell pellet was stored at -80ºC
until RNA extraction. Total RNA was extracted using High Pure RNA Isolation Kit (Roche) following the
provided instructions. RNA quality control was performed by the Gene Expression Unit at Instituto
Gulbenkian de Ciência (Lisbon, Portugal) and RNA-seq was performed using QuantSeq 3' mRNA-Seq
Library Prep Kit FWD for Illumina (Lexogen) in a HiSeq™ Sequencing System (Illumina) by the Genomics
Unit at Centre for Genomic Regulation (Barcelona, Spain). RNA-seq data analysis was performed using
QuantSeq FWD analysis pipeline available on Bluebee® genomics analysis platform (Bluebee Holding
BV). For each sample individually, the pipeline includes reads trimming using BBDuk, alignment against
genomic sequence using STAR aligner, gene read counting using HTSeq and quality control using RSeQC.
Based on these data, QuantSeq was able to perform a transcriptomic analysis using DESeq2 and
reporting a list of differentially expressed genes. Transcriptomic results present a p-value according to
Wald statistical test. p-values were corrected using the Benjamin-Hochberg method. Genes were
considered up or downregulated when a two-fold or greater change was exhibited and whose expression
values have statistically significance associated (p-value <0.05).
Total RNA from cell samples of sequential stages of cardiomyocyte differentiation was extracted
using High Pure RNA Isolation Kit following the provided instructions. RNA was quantified using a
nanodrop and 1 µg of RNA was converted into cDNA with High Capacity cDNA Reverse Transcription Kit
33
(Applied Biosystems™/ Thermo Fisher Scientific) also following the provided instructions. PCR reactions
were run using 12.5 ng of cDNA and 250 µM of Taqman™ Gene Expression Assays (Applied
Biosystems™/ Thermo Fisher Scientific) (Table 4), along with Taqman™ Gene Expression Master Mix
(Applied Biosystems™/ Thermo Fisher Scientific). Reactions were run in triplicate using ViiA™ 7
Real-Time PCR Systems (Applied Biosystems™/ Thermo Fisher Scientific) and data were analysed using
QuantStudio™ Real-Time PCR Software (Applied Biosystems™/ Thermo Fisher Scientific). The analysis was
performed using the ΔΔ
Ct method and values were normalized against the expression of the housekeeping
gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
Table 4 | Genes (and Assay ID) used for quantitative real-time polymerase chain reaction.
Gene Assay ID Gene Assay ID
GAPDH Hs02758991_g1 MYH6 Hs01101425_m1
OCT4 Hs00999634_gH MYH7 Hs01110634_m1
T Hs00610080_m1 MYL2 Hs00166405_m1
ISL1 Hs01099687_m1 MYL7 Hs01085598_g1
NKX2-5 Hs00231763_m1 E-CAD Hs01023894_m1
TNNT2 Hs00165960_m1 - -
Gene ontology analysis was performed on the selected genes using PANTHER Classification
System available on Gene Ontology Consortium (http://www.geneontology.org). Results were sorted by
fold enrichment that represents the number of provided genes divided by the expected number (based on
the reference list). If fold enrichment is greater than 1, it indicates that the category is overrepresented in
the experiment. Genes were grouped into categories with a p-value < 0.05 for biological processes,
molecular functions and cellular components. Protein-protein interaction networks were constructed using
STRING database (https://string-db.org/), which provides a prediction pipeline based on in vivo and
experimental assays (co-occurrence in genomes/ phylogeny, co-expression, fusion events, genomic
neighbourhood/ synteny) (Szklarczyk et al., 2015). Cluster positions were determined by an algorithm
based on global confidence binding score (minimum required interaction score of 0.4, i.e. medium
confidence). Based on this score, proteins were clustered applying the Markov Cluster Algorithm (MLC)
with an inflation parameter = 4. Expression heatmaps and principal component analysis of selected
genes were generated using ClustVis (http://biit.cs.ut.ee/clustvis).
Calcium-handling analysis was performed by imaging of spontaneous calcium transients in 3D
aggregates. Aggregates were transferred to Nunc™ Lab-Tek™ II chambered coverglass 8 well and after
culture medium removal, aggregates were resuspended and incubated in pre-warmed fura-2
acetoxymethyl ester (fura-2 AM) dye in Fluo-4 Direct™ Calcium Assay Buffer (Invitrogen™/ Thermo Fisher
Scientific) for a final concentration of 5 µM, at 37ºC for 45 min. Aggregates were then washed and
34
resuspend in pre-warmed Fluo-4 Direct™ Calcium Assay Buffer and image pairs obtained by exciting the
cells at 340 nm and 380 nm using an inverted microscope with epifluorescent optics every 0.7 s at 37ºC.
Excitation wavelength were changed with the high speed multiple excitation fluorimetric system Lambda
DG4 (Sutter Instrument) and fluorescence was recorded at 510 nm with a CDD camera (Photometrics).
To evaluate cardiomyocyte response to adrenergic stimuli, cells were stimulated with pre-warmed
isoproterenol hydrochloride (Sigma-Aldrich®/ Merck) for a final concentration of 1 µM. Data analysis
was performed using MetaFluor software (Universal Imaging).
35
For 2D adherent monolayer cultures, the temporal modulation of Wnt signalling with small
molecules has been described as an efficient protocol for differentiation of hiPSCs into cardiomyocytes
(Lian et al., 2012). As a starting point, the hiPSC line DF6-9-9T.B (henceforth referred as DF6) was
screened for its cardiogenic potential using the mentioned approach, with the GSK3 inhibitor,
CHIR99021, followed by the Wnt inhibitor, IWP-4 (Figure 4A). For this task, some parameters of the stated
protocol were pre-optimized by the laboratory. The presence of the specific marker of cardiomyocytes
cTnT was analysed by flow cytometry (60% of cTnT+ cells) and, therefore, DF6 was selected for further
investigation (Supplementary Figure 1).
The same protocol was already tested for 3D aggregate cultures in dynamic (Fonoudi et al., 2015;
Kempf et al., 2014) and static conditions (Dahlmann et al., 2013), but the variability in cardiac
differentiation efficiencies represents one of the main challenges of these studies. In suspension cultures,
the reason for heterogeneous differentiations could be related with uncontrolled size leading to diffusional
gradients (Bauwens et al., 2011). Even with controlled size by forced aggregation, the lack of an
exhaustive optimization could explain the inefficiency of such protocols. Accordingly, a detailed
optimization was performed in the laboratory for a robust cardiomyocyte differentiation of size-controlled
3D aggregates by forced aggregation for the DF6 cell line. This optimization was accomplished by a face-
centred factorial design with focus on the aggregate size in combination with the concentration of
CHIR99021 (Supplementary Figure 2). The differentiation protocol followed in this work for 3D
aggregates is based on these results.
Before expansion, a flow cytometry characterization was performed to confirm the expression of
pluripotent stem cell markers, whose analysis revealed them to be 98% OCT4+ (Figure 4B), suggesting
their pluripotency for following 2D and 3D approaches. For the 2D culture, expansion started with a cell
density of 0.4x105 cells per well (0.4x10
5 cells/mL) and they were expanded until 90% confluence for the
beginning of differentiation (~1.5x106 cell/mL). For the 3D culture, single cells were seeded at an
inoculation density of 3000 cells per microwell (6x105 cells/mL). After cells settling in the microwells by
gravitation, they self-organized in 3D aggregates and their size was monitored for periods of 24h (Figure
4C). After the first 24h, aggregates presented an average size of 240.6 ± 5.3 µm, a size that is relatively
close to the one optimized by factorial design (224 µm). Frequently by day 3 of aggregates expansion,
they reached the optimal size for the beginning of differentiation with an average of 296.1 ± 2 µm
(~4.2x105 cells/mL). As represented in Figure 4D, 89% of the aggregates presented a diameter within
36
Figure 4 | Characterization of hiPSC cardiomyocyte differentiation in 2D and 3D culture systems.
(A) Schematic outlining of the procedures used to differentiate hiPSCs into cardiomyocytes in 2D adherent monolayer and
in 3D aggregate cultures. (B) Quantification of pluripotency by flow-cytometry analysis for OCT4 on hiPSCs from DF6 cell
line. n=1. (C) Aggregate size during expansion in microwells by diameter measurement. Mean ± SEM; n=5 experiments *
37
the range of 280-310 µm. On day 0 of differentiation, 96.6% of the cells of 2D culture and 96.2% of the
cells of 3D culture were OCT4+ (Supplementary Figure 3A,B), confirming that over the period of
monolayer expansion and during short-term aggregation of hiPSCs, the pluripotent state of the cells, that
soon will undergo cardiac differentiation, it is not affected.
After the beginning of differentiation in 2D culture, cell beating was evident during the experiment,
the first typical indicator of a successful differentiation into cardiomyocytes. Spontaneous contractions
were firstly observed between day 11 and 12 of differentiation (n=3). Other studies using the same
protocol registered the first beating cells between day 8 and day 10, variability that was attributed to cell
line variability (Lian et al., 2012). In general, the number of beating clusters gradually increased in
number over time. Figure 4E represents the normal morphology of cardiac tissue in a 2D adherent
monolayer culture after 15 days of differentiation. By this day, large regions of the monolayer culture
were vigorously contracting and the beating did not cease until the cells were harvested for further
analysis.
In 3D culture, the aggregates increased their size reaching an average of 637.9 ± 9.7 µm on day
12 of differentiation (Figure 4F). This increment in size could be observed mainly after aggregates being
transferred to low attachment cell culture plates, where they remained from day 7 until day 15 of
differentiation in static conditions (Figure 4G). One of the novelties of this protocol relies on the extended
period of forced aggregation assuring, as can be seen, the homogeneity in size and avoiding the different
aggregates to fuse between each other during differentiation. Indeed, aggregates only tend to fuse when
in suspension, mostly by day 15 (for that reason diameter on day 15 is not presented), but at this point
of differentiation it is expected that cells already constitute cardiomyocytes (Lian et al., 2013a).
Spontaneous cell beating was perceived between day 8 and 10 (n=6), starting always after passaging to
suspension conditions and always some days in advance of 2D culture system. This is at some extent in
agreement with recent reports of similar approaches of forced aggregation, where first beating cells were
detected by day 6 (Dahlmann et al., 2013). Also, in the present work, the number of beating aggregates
as well as the strength of contraction gradually increased over time.
Histological analysis with hematoxylin and eosin (H&E) indicated that at day 15, 3D aggregates
consisted of polygonal and fusiform cells densely packed, associated with rare acidophil ECM, probably
collagen (Figure 4H). Occasionally, some areas of apoptosis and necrosis were visible in the core of the
______________________________
*for each time point, measuring 23 aggregates for each experiment. (D) Aggregate size distribution on day 0 of
differentiation. n=5 experiments, measuring 23 aggregates for each experiment. (E) Bright-field image of cardiomyocytes
generated on 2D adherent monolayer on day 15 of differentiation. Scale bar represents 100 µm. (F) Aggregate size during
differentiation by ferret diameter measurement. Mean ± SEM; n=2-4 experiments for each time point with replicates
ranging from 5-59. (G) Bright-field images of time-course cardiomyocyte differentiation on 3D aggregate culture.
Aggregates in microwells from day 0 to 6 and in suspension from day 7 to 15. Scale bars represent 100 µm. (H) Histological
analysis by hematoxylin and eosin of a cryosection of a day 15 aggregate. Scale bar represents 50 µm. (I) Differentiation
Efficiency by flow-cytometry analysis for cTnT on cells differentiated in 2D adherent monolayer culture. n=3. (J)
Differentiation Efficiency by flow-cytometry analysis for cTnT on cells differentiated in 3D aggregate culture. n=6. (L), (M)
Immunofluorescent staining for cTnT, α-ACT, α-SMA and CD31 in red of differentiated cells in 2D adherent monolayer and
of differentiated cells in cryosections of 3D aggregates on day 15 of differentiation. Nuclei in blue were stained with DAPI.
Scale bars represent 50 µm.
38
aggregates (Supplementary Figure 4), most probably due to hypoxic conditions given the size of the
multicellular 3D structure.
In order to evaluate the efficiency of the differentiation protocol, the presence of cTnT was accessed
by flow cytometry. Samples of differentiated tissue from 2D culture presented an average of 60.3 ± 3.6%
cTnT+ cells (Figure 4I) and an average of 61.2 ± 7.6% cTnT
+ cells for 3D culture (Figure 4J). These results
show that in terms of differentiation efficiency there were no differences between the experiments
performed in 2D and 3D systems (Figure 4K), suggesting that the format of the platform of differentiation
has no impact on the percentage of cTnT+ cardiomyocytes. However, for the DF6 cell line, the original
protocol claims that 89.5 ± 2.1% of the cells were cardiomyocytes in a 2D culture system (Lian et al.,
2013a). Taking this into account, the 2D culture system could actually give rise to higher differentiation
efficiencies than the reported ones in this work, but in fact, it proves to be a very sensitive platform while
the 3D culture proves to be a very robust system, that after aggregate size optimization always leaded to
successful differentiations. During this work, it was clearly perceptible that for differentiation in 2D culture
the initial inoculation density and the time of expansion until day 0 were very critical for a successful
differentiation and this may be in the origin of lower efficiency rates than those reported on literature. In
contrast, for 3D culture, the optimal size proved to be a more stringent variable. Also, the long-term
expansion of the cell line in question could have some impact on its potential to differentiate into
cardiomyocytes. As the protocol for 2D culture differentiation was performed very closely to the directors
of Lian et al., a new protocol optimization may be necessary.
To test the reproducibility of the optimized 3D procedure, the protocol was also tested for other two
cell lines, WT- F002.1A.13 (henceforth referred as TCLab) and Gibco® Episomal hiPSC line (henceforth
referred as Gibco®) and then the efficiency of differentiation was assessed at day 15 (Supplementary
Figure 5A,4B). Gibco® presented an average of 62.1 ± 3.2% cTnT+ cells showing very similar results to
the ones obtained with DF6. Interestingly, TCLab presented an average of 94.7 ± 1.5% cTnT+ cells. Not
all pluripotent cell lines are equal in their capacity to differentiate into desired cell types in vitro and this
fact can explain differences in differentiation efficiency between different hiPSCs lines. Genetic and
epigenetic variations contribute to functional variability between cell lines and heterogeneity within clones,
altering the signalling response of developmental pathways (Cahan and Daley, 2013). For this reason,
more replicates of this experiment should be performed to obtained significant conclusions. Nevertheless,
the 3D aggregate protocol proved to be efficient with other cell lines beyond DF6.
Besides that, immunofluorescence staining of the adherent monolayer and of aggregates
cryosections was performed for further characterization of culture cell composition. (Figure 4L, Figure
4M). Along with cell composition, cryosections of 3D aggregates allow to know the internal organization
and architecture of the differentiated structure. For both culture systems, staining with the cardiomyocyte
specific markers cTnT and α-ACT proves the efficient differentiation of hiPSCs into cardiomyocytes. In
addition, other cardiac cells types that are naturally present in vivo were found in the differentiated tissue.
It was the case of smooth muscle cells, detected by positive staining of alpha smooth muscle actin
(α-SMA, also known as ACTA2) and endothelial cells, with positive staining for cluster of differentiation
31 (CD31, also known as PECAM-1). It is possible to see by the staining patterns that the majority of the
39
cells that constitute the aggregates are cardiomyocytes with cTnT and α-ACT staining with the second
major population being the smooth muscle cells. Endothelial cells seem to be, for both 2D and 3D
cultures, the cell population represented in minority with little microvascular networks. All these markers,
cTnT, α-ACT, α-SMA and CD31, demonstrated positive staining for TCLab and Gibco® cell lines as well
(Supplementary Figure 5C, 4D).
It is true that the protocol was designed for cardiomyocyte production, but given the autonomous
nature of in vitro differentiation, particularly in this case where all the protocol lies in the modulation of a
single signalling pathway solely by two small molecules, it is expected that the line between
cardiomyocytes and other cardiovascular cell lineages would be tenuous. Highly correlated with the
parameters optimized before this work, the bulk cell density (cell number per culture volume) and the
concentration of the GSK3 inhibitor CHIR99021 have been identified to regulate mesendodermal
patterning of hiPSCs, in which the same starting small molecule can originate a broad spectrum of cell
lineages by differential secretomes (Kempf et al., 2016). In fact, it was demonstrated very recently by the
group which developed the GiWi protocol, that slightly changes after GSK3 inhibitor and Wnt inhibitor
addition could end in the generation of another cardiovascular cell lines such as smooth muscle cells or
even others like cardiac fibroblasts and epicardial cells (Bao et al., 2017). A similar manipulation of the
Wnt signalling with GSK3 inhibitor also leads to endothelial progenitor differentiation (Lian et al., 2014).
These findings highly support the presence of α-SMA+ smooth muscle cells and CD31
+ endothelial cells
in the present differentiation. However, it it has been reported that, the use of engineered constructs
including multiple reporters for different cell lineages are more likely to replicate the in vivo dynamic
organization and function of cardiac tissue, particularly for endothelial cells and also smooth muscle cells
that can contribute to increased structural maturation (Masumoto et al., 2016).
As already postulated, hiPSCs recapitulate the early stages of embryogenesis through in vitro
differentiation. Accordingly, cardiomyocyte specification depends on a combination of signalling factors
and downstream transcriptional events, leading to the expression of cardiac-specific factors. In this work,
to better understand the gene expression patterns in 3D culture system and to further study cardiac-
specific enriched genes, a transcriptomic analysis throughout sequential stages of cardiomyocyte
differentiation was performed using RNA-seq data.
At first, as a fairly preliminary analysis, two extreme time points of cardiomyocyte differentiation
were compared, day 0 and day 20. With these time points, it was possible to compare the gene expression
profile of a starting pluripotent cell population with a final cardiomyocyte-like population. Globally,
28617 genes were represented in the analysis. For further investigation were considered as regulated
genes the ones that exhibited a two-fold or greater change and those whose expression values have
statistically significance associated (p-value <0.05). According to these criteria, from the total genes,
1776 (6.2%) were significantly upregulated and 1122 (3.9%) were downregulated. 8884 (31%) had a too
low read count to be analysed. In this study, an expression heatmap was built for the top 10 upregulated
40
Figure 5 | Transcriptomic analysis of cardiomyocyte differentiation in 3D culture and protein-protein interaction.
(A) Expression heatmap for the top 10 upregulated and top 10 downregulated genes assessed by transcriptomic analysis
when comparing day 0 and day 20 of differentiation in 3D culture (fold change > 2, p-value < 0.05). Genes are displayed
on the vertical axis from the most upregulated (MYH6) to the most downregulated (ESRG) gene, with the different time
points on the horizontal axis. High expression levels are in dark red and low expression levels are in light red. n=3
experiments for each time point. (B), (C) Protein-protein interaction networks based on protein encoding genes of the top
10 up and downregulated genes assessed by transcriptomic analysis when comparing day 0 and day 20. The interaction
networks were constructed by STRING that predicts associations based on in vivo and in vitro experimental assays
(Szklarczyk et al., 2015). Interactions with a minimum required interaction score of 0.4 are visualized, i.e. medium
confidence. Lines represent protein-protein associations (more thickness, more confidence).
and top 10 downregulated genes (Figure 5A, Supplementary Table 1). Among the downregulated
genes, there are known markers responsible for hiPSCs pluripotency whose expression were basically
undetectable at day 20. They were LINE1 type transposase domain containing 1 (L1TD1), NANOG,
ZFP42 zinc finger protein (ZFP42), zinc finger and SCAN domain containing 10 (ZSCAN10) and POU5F1
also known as OCT4. To understand how these genes and their encoded proteins could be associated,
an analysis using STRING database was performed (Figure 5B). STRING predicts protein-protein
associations based on experimental assays. As expected, the analysis found high predicted interaction
between OCT4 and NANOG that are known to be part of the PGRN (De Los Angeles et al., 2015), and
accordingly, it is expected that both genes are downregulated at day 20 of differentiation. ZFP42 and
ZSCAN10 belong to the class of zinc finger transcription factors. ZFP42 is a known marker for
undifferentiated ESCs and it is a direct target of NANOG and augmented by OCT4 (Shi et al., 2006).
41
ZSCAN10 is required to maintain ESCs at a pluripotent state, indeed, it works together with OCT4 and
SOX2 as a transcriptional regulatory network (Yu et al., 2009), hence it showed high predicted interaction
with OCT4 and NANOG. In addition, L1TD1 also has an important role in stemness regulation with its
depletion leading to OCT4 and NANOG downregulation (Närvä et al., 2012). It was indeed the most
downregulated gene codifying for a protein between day 0 and 20 of differentiation. Based on this
information, the downregulation of such genes is expectable at this stage of differentiation, suggesting
that the cells were no logger pluripotent.
Some of the downregulated genes do not codify for a protein, their transcripts are non-coding RNA
molecules such as long intergenic non-coding RNA (lincRNA)1, sense intronic RNA
2 and antisense RNA
(asRNA)3. In this particular case the sense intronic RNA embryonic stem cell related (ESRG) and the long
intergenic non-coding RNA 678 (LINC00678) were found between the top downregulated genes. ESRG
is a novel gene only expressed in human ICM and PSCs and its absence leads to expression decrease of
many pluripotency genes (Wang et al., 2014). About LINC00678 little is known.
Among the upregulated genes in 3D cardiomyocyte differentiation, there are known markers of
cardiomyocytes like MYH6, MYBPC3, TNNT2, MYL4 and MYL7, all genes that codify key structural
proteins of the sarcomere in cardiomyocytes (Hwang and Sykes, 2015). In fact, the presence of this
structural genes seems to be an early event in cardiomyocyte maturation (Yang et al., 2014). Surprisingly,
the structural protein TNNI1 also raised as one of the top upregulated genes at day 20 comparing with
day 0. TNNI1 is a specific protein of other type of striated muscle than cardiac, more properly skeletal
muscle (Hwang and Sykes, 2015). As previously debated, other mesoderm derivatives in addition to
cardiomyocytes were present in the 3D culture such as smooth muscle and endothelial cells, and skeletal
muscle cells appear to be no exception. Actually, transcriptomic analysis confirmed their presence
identifying smooth muscle and endothelial markers (data not shown).
An interactomic analysis using STRING was performed using the protein codifying genes between
the top 10 upregulated ones (Figure 5C). A strong interaction was predicted within all the structural
sarcomeric proteins, but no association was found for the proteins codified by heat shock protein family
B (small) member 7 (HSPB7), SMPX, and heart and neural crest derivatives expressed 2 (HAND2). Though,
it is known that HAND2 is a regulator of heart development and along with other transcription factors
can give rise to cardiomyocytes by reprogramming of fibroblasts (Anderson et al., 2016). Interestingly, a
gene with a regulatory role over HAND2, HAND2-AS1 (that codifies for an asRNA), was one of the top
10 upregulated genes. HAND2 has already been reported as an enriched transcript in other
cardiomyocyte differentiations (Hartogh et al., 2016).
In brief, this preliminary analysis of the top 10 upregulated and top 10 downregulated genes from
day 0 versus day 20 of differentiation in 3D culture, confirm the enrichment of previously identified genes
with regulatory and functional properties in cardiac development and also the downregulation of genes
associated with pluripotency.
1 Transcript that does not overlap within a coding gene.
2 Transcript within an intron of a coding or non-coding gene.
3 Transcript that overlaps a protein coding gene on the opposite strand (in exonic or intronic sequences).
42
Table 5 | Gene ontology analysis of enriched transcripts at day 20 of differentiation. Of the 28617 genes represented in the comparison analysis between day 0 and 20, the 1776 upregulated genes (fold change > 2, p-value < 0.05, n=3) were grouped into Gene Ontology (GO) categories (p-value < 0.05). Only 1453 genes were mapped. This table represents the top 20 results (based on fold enrichment) for biological processes, molecular functions and cellular components.
Gene Ontology Category H. sapiens
Genes
Genes in
Category
Fold
Enrichment p-value
Bio
logic
al Process
Cardiac muscle fibre development 9 8 12.29 3.80E-03
SA node cell to atrial cardiac muscle cell signalling 8 7 12.1 2.21E-02
SA node cell action potential 8 7 12.1 2.21E-02
SA node cell to atrial cardiac muscle cell communication 9 7 10.75 4.74E-02
Cardiac myofibril assembly 16 12 10.37 3.45E-05
Cardiac chamber formation 11 8 10.06 1.67E-02
Cell migration involved in heart development 14 10 9.88 1.05E-03
Positive regulation of stem cell differentiation 14 9 8.89 1.05E-02
Vascular endothelial growth factor signalling pathway 14 9 8.89 1.05E-02
Ventricular cardiac muscle cell action potential 19 12 8.73 2.22E-04
Semi-lunar valve development 16 10 8.64 3.50E-03
Regulation of cardiac muscle contraction by regulation of the release of
sequestered calcium ion 16 10 8.64 3.50E-03
Regulation of release of sequestered calcium ion into cytosol by
sarcoplasmic reticulum 24 15 8.64 4.82E-06
Regulation of cardiac muscle contraction by calcium ion signalling 21 13 8.56 7.45E-05
Outflow tract septum morphogenesis 21 13 8.56 7.45E-05
Atrial cardiac muscle cell to AV node cell communication 15 9 8.3 1.83E-02
Cell-cell signalling involved in cardiac conduction 27 16 8.19 2.77E-06
Sarcomere organization 39 23 8.15 4.50E-10
Cardiac muscle cell development 48 28 8.07 1.12E-12
Myofibril assembly 60 35 8.07 1.86E-16
Mole
cula
r Function
Titin binding 11 8 10.06 6.05E-03
Structural molecule activity conferring elasticity 17 9 7.32 1.81E-02
Structural constituent of muscle 46 20 6.01 1.38E-06
Actinin binding 42 17 5.6 7.59E-05
Alpha-actinin binding 33 13 5.45 4.36E-03
Extracellular matrix structural constituent 80 22 3.8 6.17E-04
Transmembrane receptor protein kinase activity 82 22 3.71 9.29E-04
Growth factor binding 130 32 3.4 1.61E-05
Cytokine binding 98 23 3.24 4.78E-03
Heparin binding 161 37 3.18 6.46E-06
Transcription factor activity, RNA polymerase II distal enhancer
sequence-specific binding
100 22 3.04 2.17E-02
Actin filament binding 153 33 2.98 2.00E-04
Ion channel binding 117 25 2.95 8.81E-03
Glycosaminoglycan binding 213 45 2.92 1.66E-06
Growth factor activity 163 33 2.8 8.17E-04
Ligand-gated channel activity 139 27 2.69 2.00E-02
Ligand-gated ion channel activity 139 27 2.69 2.00E-02
Sulfur compound binding 232 45 2.68 2.11E-05
Actin binding 413 78 2.61 2.53E-10
Ras guanyl-nucleotide exchange factor activity 248 41 2.29 5.55E-03
Cellula
r C
om
ponent
Cardiac myofibril 7 6 11.85 2.01E-02
Myofilament 25 15 8.3 1.29E-06
Striated muscle thin filament 22 13 8.17 1.98E-05
Fibrillar collagen trimer 12 7 8.07 4.56E-02
Banded collagen fibril 12 7 8.07 4.56E-02
Complex of collagen trimmers 20 10 6.91 3.90E-03
Z disc 124 56 6.24 3.08E-23
I band 140 62 6.12 1.80E-25
Sarcomere 191 81 5.86 1.43E-32
Contractile fibre part 208 88 5.85 1.73E-35
M band 29 12 5.72 2.86E-03
Myofibril 212 86 5.61 2.54E-33
Contractile fibre 223 90 5.58 8.39E-35
Sarcoplasmic reticulum membrane 40 16 5.53 9.68E-05
A band 43 17 5.47 4.52E-05
T-tubule 44 16 5.03 3.41E-04
Sarcolemma 120 43 4.95 7.19E-14
Sarcoplasm 68 23 4.68 3.48E-06
Sarcoplasmic reticulum 61 20 4.53 6.06E-05
Extracellular matrix component 122 36 4.08 6.01E-09
43
To assess the potential function of the enriched transcripts on day 20 of differentiation, they were
grouped into Gene Ontology (GO) categories. The 1776 upregulated genes were classified in the
biological processes, molecular functions and cellular components categories (Table 5) with a statistically
significant value (p-value <0.05). Concerning the involved biological processes, the analysis identified
several GO terms related with cardiac development processes including muscle fibre and cell
development from its formation to the mature structure, cardiac myofibril assembly, cardiac chamber
formation, cell migration and stem cell differentiation and still sarcomere organization. Also,
heart-specific regulatory and functional processes were identified. With high fold enrichment values, some
processes involving sinoatrial (SA) node cells can be highlighted. These SA node is known as the heart
pacemaker and its cells generate spontaneous action potentials controlling the rate of cardiac contraction
(Später et al., 2014). Consequently, regulation processes related with cardiac muscle contraction and
calcium ion signalling were also emphasised by the analysis.
Molecular functions were also evaluated using GO analysis that in majority demonstrated that
structural binding of muscle components was between the main enriched functions in differentiated 3D
aggregates. Namely, titin and actin binding functions, but also the binding of ions and molecules like
growth factors and cytokines were assessed. Besides that, transport activity seems to be another molecular
function highly represented. Titin biding GO term presented the highest fold enrichment and this protein
in particular is a giant protein that runs from the Z-disk to the M-band in sarcomeres with several biding
partners being required for sarcomere assembly (Hwang and Sykes, 2015). Accordingly, there must have
been a lot of genes related with this assembly between the upregulated transcripts.
Finally, the cellular components identified by GO analysis were mainly related with the
cytoskeleton, including the major components that give rise to the known sarcomere architecture such as
myofibril and their filaments (myofilaments), Z discs, I, M and A band. Interestingly, T-tubule and
sarcoplasmic reticulum are between the enriched cellular components on day 20. Other studies in 2D
culture have reported few or no T-tubule in the cardiomyocytes and consequent poor sarcoplasmic
reticulum calcium release (Yang et al., 2014), but in this work it is clear that genes involved with this
cellular components were being upregulated. This fact constitutes another prove that cardiomyocyte
maturation was occurring within the 3D aggregates at this stage.
Overall, this transcriptomic analysis demonstrated that at day 20 of differentiation the gene
expression patterns became specified towards cardiac differentiation with downregulation of pluripotency
genes. At this point, cardiac developmental processes take place and structural and functional cardiac
genes are among the enriched transcripts. These generic data are in agreement with previous reports of
transcriptomics analysis at sequential stages of 2D cardiomyocyte differentiation (from hESCs using BMP4,
Activin A and CHIR99021) (Hartogh et al., 2016). In the mentioned study, a clear shift between early
cardiac progenitor stages (day 5 and 7) and cardiac-lineage-committed progenitors and cardiomyocytes
(day 10 and 14) was identified. GO terms more related with broad developmental processes latter
included heart-specific regulatory and functional GO terms. Indeed, another published work comparing
2D and 3D culture systems at the transcriptome level (Zhang et al., 2015), using a similar protocol to the
present one, defends that there is a high degree of similarity between 2D and 3D culture systems.
44
Nevertheless, as a difference, this previous comparative study highlights a fraction of non-cardiac
and pluripotency genes between the enriched transcripts for the 3D culture at day 14 of differentiation,
such as LIN28, CDH1, SOX2, ZIC2, and OTX2 (pluripotent and neural-specific), and still APOA2 and TF
(liver‑specific). In the present work, using the optimized 3D protocol, part of these genes is in fact between
the significant downregulated genes and none of them are between the upregulated ones from the
comparative analysis between day 0 and day 20. However, a few upregulated genes in the present work
also fit in some GO categories of non-cardiac cells namely kidney development processes.
To analyse the progression of gene expression throughout the sequential stages of differentiation
in 3D culture, an expression heatmap was generated up to day 20 (Figure 6A). Therefore, key regulatory
genes which exhibit temporal expression were selected to this analysis.
In the beginning of differentiation, after induction with CHIR99021, gene expression analysis
demonstrated that the pluripotency gene OCT4 and NANOG started to be downregulated. By that time,
cells start to narrow their potency. Simultaneously with this event of loss of pluripotency, the relative peak
in mesendodermal gene expression, assessed by expression of T, was clearly perceptible at day 1 of
differentiation. The same is true for EOMES. This peak in day 1 for T and EOMES expression is in
agreement with another observations for 3D culture with the same protocol and forced aggregation
(Zhang et al., 2015). At this stage, cells can activate MESP1 gene expression, that at day 3 is visibly
upregulated. This gene indirectly regulates EMT, leading to mesoderm commitment, the germ layer that
is responsible to give rise to the cardiac lineage (Burridge et al., 2015). After exposure to the Wnt
signalling inhibitor IWP-4 at day 3, cardiac progenitor genes started to be upregulated. ISL1 had its
Figure 6 | Gene expression profile of regulatory and structural genes at sequential stages in 3D culture.
Expression heatmaps for normalized counts from RNA-seq of (A) regulatory and (B) structural genes at sequential stages
of hiPSC cardiomyocyte differentiation when cells were cultured as 3D aggregates. Genes are displayed on the vertical axis
with the different time points on the horizontal axis. High expression levels are in dark red and low expression levels are in
light red. Heatmap shows average values from n=3 experiments for each time point.
45
maximum expression at day 7, while NKX2-5 and GATA4 were still being upregulated with latter
maximum expression. This is also consistent with previous studies (Zhang et al., 2015).
Over time, terminal cardiac genes with structural function have their relative expression increased
and to study their progression some sarcomeric protein encoding genes were selected: TNNT2, MYH6,
MYH7, MYL2, MYL3, MYL4, MYL7, MYBPC3, TTN and ACTC1. For this analysis, an expression heatmap
was created demonstrating that, from day 5, cardiomyocytes were emerging within the 3D aggregates
and at day 18 and 20 the maximum expression of these structural genes was achieved (Figure 6B). Since
some of them are restricted to different cardiomyocyte subtypes (Hwang and Sykes, 2015), this also
suggests a degree of variability in the aggregates, including ventricular- and atrial-like cells.
Looking to the normalized values used to generate the heatmap (Supplementary Figure 6), it can
be easily seen that TTN is one of the most expressed genes codifying for sarcomeric proteins in the
cardiomyocytes of the 3D aggregate. It is also demonstrated that MYH6, a gene encoding specific filament
protein of atrial cardiomyocytes, reached high levels of expression until day 20. On the other hand,
MYH7, MYL3 and MYL2, genes encoding for proteins of ventricular cardiomyocytes, seemed to be much
less expressed in this period of time. This may suggest that ventricular cardiomyocytes arise in the 3D
culture later than the atrial ones. Similar data has been reported in transcriptomic analysis for 2D culture
systems, particularly for MYL2 (Hartogh et al., 2016).
Overall, this RNA-seq analysis, including genes of different developmental stages, suggest a faster
progression from the pluripotent cell state to mesoderm, followed by a cardiac progenitor stage and lastly
a cardiomyocyte-like fate that is relatively stable in the short term.
To confirm these results, a qRT-PCR analysis was performed for some of the key genes previously
selected and for the same crucial time points (Figure 7A). Their expression is always represented as relative
to day 0 and normalized against the expression of the housekeeping gene GAPDH.
Globally, the results obtained by qRT-PCR were in good agreement with the RNA-seq data. In this
analysis was particular perceptibly that after day 1, T relative expression started diminishing, a pattern
that continues until the end of the differentiation protocol, and simultaneously, NKX2-5 reached a plateau
whereas ISL1 relative expression appeared to decrease in the last days. Similarly to NKX2-5, it is possible
to see that after day 9 the majority of the terminal cardiac genes also reached a plateau. It was between
the starting point of their expression and this stable state that the first spontaneous beating cells were
detected in culture. The qRT-PCR analysis was also performed for TCLab cell line (Supplementary Figure
5E) and the trends were very similar to the ones obtained for DF6, demonstrating that using the optimized
3D protocol for different cell lines leads to identical gene expression patterns regarding cardiac
development.
3D culture was also analysed by immunofluorescence staining using antibodies for representative
markers of pluripotency (NANOG), cardiac progenitors (NKX2-5) and cardiomyocytes (cTnT) (Figure 7B).
For this task, cryosections of the 3D aggregates were used as before. Staining for NANOG was evident
within all the aggregate on day 0 but it faded away on day 1 and 3, confirming the loss of pluripotency.
At day 7 is poorly perceptible, but at day 12, it is possible to see the positive staining for NKX2-5,
suggesting that cardiac progenitors are present in the culture. Lastly, a clear progression for the
46
Figure 7 | Relative expression profiles and immunofluorescent staining for pluripotency, mesendoderm, cardiac
progenitor and cardiomyocyte markers during hiPSCs differentiation.
(A) Gene expression was assessed by qRT-PCR at sequential stages for key genes that unveil the progression throughout †
47
cardiomyocyte specific marker cTnT is depicted in the images. A strong positive staining at day 12
demonstrates that at this stage 3D aggregates are fully enriched in cardiomyocytes. The entire section
proved to have the target cells of this work, showing that diffusional barriers are not a problem of the
optimized 3D differentiation protocol.
To reveal common features but also potential differences between the sequential stages of
cardiomyocyte differentiation in 2D and 3D cultures, a comparative analysis for gene expression was
performed using qRT-PCR. The same time points and key genes with temporal and cell-type specific
expression were selected to this analysis.
The variance in relative expression levels in 2D and 3D culture systems for the several time points
and all the genes were evaluated using a principal component analysis (PCA) (Figure 8A). This statistical
method reduces the dimensionality of the gene expression data retaining most of its variation, identifying
directions in which the variance in the data is maximal, named as principal components (Ringnér and
Ringner, 2008). In this work, the maximal variance in relative expression between the samples was
explained by PC1 (46.2%), followed by variation in PC2 (21.2%). As can be seen in the PCA, there are
some considerable differences between 2D and 3D cultures, they seem to be, in fact, very similar for the
initial stages of differentiation for the selected genes, but at latter differentiation stages, there is a
substantial difference for the 3D culture. From day 9 to day 20, the 3D culture is negatively correlated
with PC1. This may be a hint that from day 9 on, 3D culture has a different dynamic.
To analyse the progression of gene expression between 2D and 3D cultures throughout the
sequential stages of differentiation, an expression heatmap was generated up to day 20 for the same
selection of genes (Figure 8B). The analysis emphasised the peak of T expression in 3D Culture and
highlighted the high levels of expression of MYH6 in the same system with later peak for MYH7. This
representation corroborates the previous analysis for 3D culture. Globally comparing to the 2D adherent
monolayer, the expression levels are similar to the ones observed in the previous system. Interestingly,
MYH6 also revealed a high level of expression as one of the most upregulated genes in 2D culture.
Even though the expression trends have been very similar, the fold change 3D/2D for each gene
and time point was quantified for a more detailed analysis and to better understand the differences
between these culture systems. In the beginning of differentiation, the pluripotency gene OCT4 was
downregulated slightly faster in 3D culture and this can have induced the differences identified later at
day 1. Even though the peak of T relative expression had been maximum, for both systems, 24h after the
______________________________
† cardiomyocyte differentiation with OCT4 (pluripotency), T (mesendoderm), ISL1 and NKX2-5 (cardiac progenitors) and
TNNT2, MYH6, MYH7, MYL2 and MYL7 (cardiomyocytes). Analysis was performed using the ΔΔCt method with values being
presented as relative expression taking day 0 as reference. The values were normalized against the expression of the
housekeeping gene GAPDH. The Y axis has a log10 scale. Mean ± SEM; n=3 experiments for each time point. (B)
Immunofluorescent staining for NANOG in red (pluripotency), NKX2-5 in green (cardiac progenitors) and cTnT, in red
(cardiomyocytes) in cryosections of 3D aggregates on day 0, 1, 3, 7 and 12 of differentiation. Nuclei in blue were stained
with DAPI. Scale bars represent 50 µm.
48
Figure 8 | Gene expression analysis between 2D and 3D culture systems during cardiomyocyte differentiation.
(A) PCA for relative gene expression of both culture systems. Each blue circle, 3D, or red square, 2D, represents an individual
time point. X and Y axis show PC1 and PC2 that explain 46.2% and 21.2% of the total variance, respectively. Prediction
ellipses are such that with probability 0.95, a new observation from the same group will fall inside the ellipse. For 3D, n=3.
or 2D, n=1. (B) Expression heatmap for relative expression of both culture systems. Genes are displayed on the vertical axis,
with the different time points on the horizontal axis. High expression levels are in dark red and low expression levels are in
light red. Data are no available from day 9 on for OCT4 and T and from day 0 to 5 for TNNT2, MYH6, MYH7, MYL2 and
MYL7. For 3D, heatmap shows average values from n=3 experiments for each time point. For 2D, n=1. (C) Gene expression
was assessed by qRT-PCR for key genes throughout cardiomyocyte differentiation. Analysis was performed using the ΔΔCt
method taking day 0 as reference. The values were normalized against GAPDH gene. The Y axis has a log10 scale. Mean
fold change 3D/2D ± SEM; n=3.
beginning of differentiation (as predicted in the original protocol for 2D culture (Lian et al., 2013a)), it
was substantially higher in 3D culture with a fold change of 14, continuing on day 3 with a fold change
of 29. This suggests that in 3D culture, pluripotency balance is broken sooner with consequent
49
upregulation of mesendordermal transcription factors leading to earlier EMT. Actually, it was reported
that forced aggregation in microwells favour mesendoderm lineages by itself, in which 3D cultured hESCs
showed a strong peak in expression of T and no such peak in 2D cultured hESCs (Hsiao et al., 2014).
Therefore, these results demonstrate that the 3D culture can has a critical role on EMT in the first steps of
differentiation.
Concerning the cardiac progenitor genes, for ISL1, both culture systems had its peak by day 7 with
low fold differences. Still, at day 3 of differentiation, NKX2-5 relative expression presented a fold change
of 15. Similar to the other transcription factors, this earlier expression profile comparing to 2D culture
supports the idea of a faster progression in cardiac development for aggregate cells. The same is
particular evident for TNNT2, the gene that encodes a well-known cardiomyocyte structural protein, cTnT.
The fold change was about 182 at day 5 and at day 20 it was still about 12. Between the structural genes,
the exception was probably MYL7 that presented a low fold increase in 3D culture comparing to 2D.
Although a previous work comparing 2D and 3D culture systems suggested a synchronized
differentiation of the cells (it is not clear if the experiment was performed with hESCs or hiPSCs) (Zhang
et al., 2015), in this work it is clear that there were differences concerning the progression of
cardiomyocyte differentiation at the gene expression level. As the molecular mechanism underlying
cardiac differentiation become better understood and biophysical cues from the cardiac
microenvironment are defined these differences can be clarified. For now, one can only speculate that
cell-cell signalling and cell-ECM interactions in 3D culture can be in the origin of such differences in
hiPSC-CMs development and somehow, this intercellular communication have a positive effect on the
development of cardiac tissue.
Since cell-cell and cell-ECM interactions play an important role on cardiomyocyte differentiation
and since the 3D culture system create a particular extracellular microenvironment (Atmanli and Domian,
2017), this can be a starting point to unveil differences across 2D and 3D culture systems. Moreover, one
of the main differences found between both culture systems in gene expression analysis was related with
high fold change of the mesendodermal transcription factor T. To address this issue, a characterization
of some surface markers was performed with focus on some molecules with an important role in EMT
and subsequent cardiomyocyte differentiation. Thus, the presence of the cell adhesion molecules E-CAD
and N-CAD and also of the gap junction protein CX43, were assessed by immunofluorescence staining
(Figure 9A). Once again, cryosection of the 3D aggregates was used for this purpose and cTnT was used
as an indicative of successful cardiomyocyte differentiation.
At day 0 of differentiation, it is visible that E-CAD had its maximum cell staining. After this point
until the end of the experiment, its presence clearly diminished among the aggregate section. At day 1, it
is already hard to distinguish positive staining for this marker, and at day 3 no staining is visible. This
process recapitulates the in vivo EMT, when pluripotent epithelial cells transit to form the mesoderm germ
layer (Lamouille et al., 2014). This finding is in agreement with obtained RNA-seq results for the gene
50
Figure 9 | Epithelial-mesenchymal transition and cell-cell interaction during cardiomyocyte differentiation. ‡
51
expression profile of T, EOMES, MESP1 and SNAI1, which have a critical role in EMT, reaching their peaks
from day 1 to 3. Additionally, differences concerning E-CAD expression in 3D comparing to 2D culture
were found by qRT-PCR (Figure 9B). This gene presented higher expression profile in 3D aggregates (1.8x
higher) and its decrease appeared to be faster.
Based upon previous work and in the obtained results, here it is hypothesized that a positive
feedback mechanism can be responsible for the higher fold change in T expression and for the subsequent
fold increase in regulatory and structural genes in 3D culture comparing to 2D. The following model is
proposed (Figure 9C): pluripotent 3D aggregates generated by forced aggregation in microwells present
higher levels of E-CAD than in 2D adherent monolayer, possibly by increased cell-cell contact with
neighbouring cells, as already reported (Azarin et al., 2012; Hsiao et al., 2014). Along with the high
levels of E-CAD, a downregulation of Wnt signalling can be observed since both compete for β-catenin,
that acts as a physical linkage between the cell adhesion molecule and the cytoskeleton in the first case,
and as a transcription co-factor in the second (Azarin et al., 2012). After GSK3 inhibition by CHIR99021
at day 0 of differentiation, nuclear β-catenin starts Wnt signalling induction (Lian et al., 2012), which
results is T expression and, later, the downregulation of E-CAD by SNAI1 (from day 0 to 3). The
disassemble of adherens junctions due to decrease in E-CAD trough EMT, go together with the release of
high amounts of β-catenin. The cytoplasmic β-catenin can now be translocated to the nucleus to act as a
transcription co-factor in Wnt signalling, increasing even more T expression and following pathway. In
fact, the existence of a pathway that physically feeds Wnt signalling from a β-catenin pool upon adherens
junctions dissociation was described (Kam and Quaranta, 2009). This process would act as a positive
feedback until total EMT is achieved. As E-CAD is present in lower amounts in 2D adherent monolayer,
the release of β-catenin during EMT will be smaller. Therefore, a weaker Wnt signalling induction leads
to lower expression profile of mesendodermal transcription factors in the 2D culture system. Additionally,
there is work that defends that β-catenin is indeed cleaved by calpain in hPSCs 3D aggregates (Konze et
al., 2014), but this cleavage can actually increase the stability of β-catenin as transcription co-factor (Abe
and Takeichi, 2007). It should also be borne in mind that aggregates with different sizes proved to be
different in E-CAD composition (Azarin et al., 2012). Thus, in addition to the contribution of this parameter
to the bulk cell density, the correlation size/E-CAD can be one of the several reasons that make this
______________________________
‡ (A) Immunofluorescent staining for E-CAD, N-CAD and CX43 in green for cryosections of 3D aggregates from day 0 to 9
of differentiation. E-CAD and CX43 staining were assessed in addition to cTnT in red, used as a control of successful
cardiomyocyte differentiation. Nuclei in blue were stained with DAPI. Scale bars represent 50 µm. (B) E-CAD gene
expression was assessed by qRT-PCR at the first days of differentiation. Analysis was performed using the ΔΔCt method
taking 2D as reference. The values were normalized against the expression of the housekeeping gene GAPDH. The Y axis
has a log2 scale. Fold change 3D/2D; n=1. (C) Schematic representation of the proposed model in which 3D aggregates
can generate higher pools of β-catenin upon adherens junction dissociation in the first steps of cardiac differentiation,
leading to a stronger Wnt signalling induction comparing to 2D adherent monolayer. (D) Protein-protein interaction
networks based on upregulated genes assessed by transcriptomic analysis when comparing day 0 and day 3 from GO
categories related to epithelial-mesenchymal transition. The interaction networks were constructed by STRING. 5 more
proteins were predicted for the network. Interactions with a minimum required interaction score of 0.4 are visualized, i.e.
medium confidence. Lines represent protein-protein associations; more thickness, more confidence. Proteins are clustered
using the MCL algorithm and they presented in different colours - inflation parameter = 4. Dashed lines represent inter-
cluster association.
52
parameter so important in 3D cardiac differentiation. In general, this is a hypothesis that requires further
experiments to be validated.
Following E-CAD decline during EMT it is well established that there is a raise of N-CAD (Lamouille
et al., 2014). Such increment was corroborated by the results obtained in this work, in which an obvious
increase on N-CAD positive staining is perceptible since day 3 of differentiation. It has been known for a
long time that N-CAD play an important role in anchoring myofibrils in cell-cell contact in cardiomyocytes
(Zuppinger et al., 2000) and the immunofluorescence images obtained support that.
To better characterize the changes in cell communication during EMT, it was possible to observe
that the gap junction protein CX43 also change its levels during this event. This connexin was present
within the aggregate during pluripotency state at day 0 and it was detected until day 3. After EMT, this
protein was reduced to small areas that still do not express the cardiomyocyte marker cTnT. However, at
day 9 of differentiation little dots of positive staining were present between cTnT+ cardiomyocytes. Even
though CX43 has long been considered the most abundant connexin in the heart (Fishman et al., 1991),
this oscillatory behaviour had already been described, in which following EMT, CX43 is downregulated
by SNAI1 and then upregulated again with the maturation of cardiomyocytes (De Boer et al., 2007). Not
only was this the verified pattern but also the assessed SNAI1 expression profile is coordinated with it,
since SNAI1 maximum expression was detected at day 3 and after that, no CX43 is detected between de
novo cardiomyocytes. Comparing CX43 and N-CAD, the first seemed to resurge just after the
establishment of the other, what is supported by data suggesting that N-CAD is required to maintain
CX43 (Li et al., 2005).
Furthermore, using the transcriptomic analysis data, GO categories as “epithelial to mesenchymal
transition”, “regulation of epithelial to mesenchymal transition” and “positive regulation of epithelial to
mesenchymal transition” were found between the enriched categories when comparing day 0 and day 3.
These categories included a total of 24 genes and the association of their corresponding proteins was
assessed by STRING (Figure 9D). The platform was able to predict interactions of the proteins under study
with other 5 proteins, and interestingly, β-catenin is among them, reinforcing the importance of this
protein on EMT.
These results give some insights of how cell-cell interaction is involved in EMT and consequently in
the development of cardiomyocytes. Some clues were uncovered as some of the differences found
between the 2D and 3D systems can be explained based on the intercellular communication characteristic
of each culture.
It was already stated that the generated cardiomyocytes must recapitulate the properties of adult
cardiomyocytes in vivo, so full advantage can be taken of their potential. In this work, during the
differentiation of hiPSCs to cardiomyocytes it was possible to see that the generated cells presented
spontaneous beating and upregulation of genes encoding for several structural proteins and for proteins
53
associated with muscle contraction and calcium ion signalling. For further insights in cardiomyocyte
maturation state, morphological and functional features were selected for analysis.
Firstly, regarding morphologic features, immunofluorescence staining of single cell cardiomyocytes
from 2D and 3D cultures, at day 15, was compared using cell shape as a parameter. Cells differentiated
as an adherent monolayer were in general round (Figure 10A), a shape that is characteristic of immature
cardiomyocytes. On the other hand, differentiated cells from aggregates, although round in part, also
presented several elongated cardiomyocytes (Figure 10B) as would be expected from a mature cardiac
tissue. Sometimes, the images even suggest that binucleated cardiomyocytes were present in the 3D
culture, with multinucleation being a characteristic of late hiPSC-CMs. These results suggest that
cardiomyocytes produced in 3D aggregates, generated by forced aggregation, present a more adult-like
morphology.
Since long time ago, CX43 has also been related to cardiomyocyte maturation, and now it is known
that this connexin participates in the functional electromechanical coupling between cardiomyocytes,
promoting propagation properties during heart development (Pervolaraki et al., 2017). Above, some
perceptions about CX43 presence during the first steps of differentiation in 3D culture were considered,
but one may reasonably ask if at late stages, have the differentiated cells a high content in CX43?
Recurring again to immunofluorescence staining of aggregate cryosections, it was demonstrated that a
lot of cardiomyocytes already present CX43 as a gap junction protein by day 20 (Figure 10C).
Besides the maturation evidences provided by morphological and structural analysis,
transcriptional regulators of cardiomyocyte maturation were sought among the upregulated genes
delineated in the transcriptomic analysis between day 0 and 20 of differentiation. This search was based
on transcriptional regulators previously identified as critical for cardiomyocyte maturation in vivo (Uosaki
et al., 2015). Between them, genes from the peroxisome proliferator-activated receptor (PPAR) pathway
were found among the identified upregulated genes in this work. They were retinoid X receptor alpha
(RXRA) (fold change <2), PPARD (fold change <2), PPARA and PPARG coactivator 1 alpha (PPARGC1A).
These genes are master regulators of fatty acid metabolism that prevail in mature cardiomyocytes to the
detriment of glycolysis. Actually, in cardiomyocytes differentiated in 2D culture (from mouse PSCs with
Activin A, BMP4, VEGF, bFGF and FGF10) these regulators were found inactive in contrast with in vivo
samples (Uosaki et al., 2015). A recent study, also using 2D culture differentiation (from hPSCs with
CHIR99021, Activin A, Ascorbic acid and IWR-1), reported that shifting from glucose-containing to
galactose- and fatty acid-containing medium is a viable procedure to promote cardiomyocyte maturation
(Correia et al., 2017). Here, similarly to adult cardiomyocytes, hiPSC-CMs generated in 3D aggregates
seemed to have already their fatty acid metabolism pathway increased, but further investigation on how
distinct carbon sources in synergy with 3D culture affect cell maturation would be interesting.
To understand if these results correlate with physiological features, the hiPSC-CM calcium-handling
in 3D aggregates was analysed using calcium imaging. Intracellular calcium concentration in
cardiomyocytes was measured using the calcium-sensitive fluorescent dye fura-2 acetoxymethyl ester
(Fura-2 AM). The results are depicted as a Fura-2 AM ratio between bound (380 nm) and unbound (340
nm) to calcium (Figure 10D). In general, the recording of intracellular calcium transients displayed strange
54
Figure 10 | Morphological and structural features in cardiomyocyte maturation and calcium handling analysis.
Single-cell cardiomyocytes from (A) 2D and (B) 3D culture system at day 15 of differentiation, immunostained for cTnT in
red. In (B) the arrows in the left image highlight elongated cardiomyocytes and in the right image, a possible binucleation.
(C) Immunofluorescent staining for cTnT in red and CX43 in green for cryosections of 3D aggregates at day 20 of
differentiation. Nuclei in blue were stained with DAPI. Scale bars represent 20 µm. (D) Intracellular calcium imaging for 3D
aggregates at day 20 of differentiation. Ratio images are depicted. H for high intracellular calcium and L for low intracellular
calcium. (E) Calcium transients profile for 3D aggregates at day 20 of differentiation, before and after addition of the β-
agonist isoproterenol. The sets of calcium transients are highlighted in red.
patterns of cyclic calcium handling. Spontaneous beating frequency was regular in time but with sets of
calcium transients spaced by long intervals of time (Figure 10E). The reason for this atypical behaviour
may lie in the protocol for image recording itself and not in the generated cardiomyocytes, since they
have a regular beat rate when in culture. For that reason, the optimization of this protocol should be
performed in a near future. However, for the sets of calcium transients obtained, beating frequency was
possible to measure. At day 20, hiPSC-CMs presented a frequency of 23 BPM. The obtained results are
between the reported beating frequencies for hPSC-CMs that range from 21 to 52 BPM (Robertson et al.,
cTnT/CX43/DAPI
55
2013). Regardless of whether spontaneous beats did not have a normal pattern as a whole, when calcium
transients were recorded they appeared to have typical hiPSC-CMs physiological features.
Setbacks aside, an important parameter of cardiomyocyte physiology is the response to
neurohormonal regulation. Accordingly, the ability of de novo cardiomyocytes to respond to adrenergic
stimuli were assessed evaluating the positive chronotropic effect (higher beating rate) induced by the
-agonist isoproterenol. After isoproterenol addition, calcium transients were recorded ceaseless without
long intervals of time between them, but the increase in beating frequency was not verified (20 BPM).
Even if an appropriate pharmacological response was not detected, isoproterenol stimulation suggested
the existence of functional β-adrenergic signalling. Given that basal calcium handling may have been
affected by the performed protocol, the presented pharmacological response may not correspond to the
actual response of the generated cardiomyocytes.
Overall, several hints lead one to believe that hiPSC-CMs generated by the optimized 3D protocol
presented a certain degree of similarity with adult cardiomyocytes. However, optimization of the protocol
for calcium handling analysis should be performed for a proper examination of physiological features.
56
57
Cardiomyocyte differentiation is a landscape made up of complex and challenging mechanisms.
In the past years, scientific community started to explore this landscape with a much sharper resolution,
as key fundamental concepts have emerged from a collective and longstanding effort to understand the
biology of the heart from a developmental point of view. This knowledge must be expanded into novel
experimental systems and methodologies leading one to unprecedented levels of in vivo resemblance in
tissue engineering. The present interrogation of cardiomyocyte differentiation from hiPSCs in both 2D and
3D culture systems lead to an evaluation of in vitro parameters aiming to frame the understanding in
cardiac tissue generation.
The present work, unveiled similarities and different features of both culture systems. At a glance,
the platform of differentiation had no impact on the differentiation efficiency analysed by flow cytometry,
however 3D proved to be a more robust platform with more chances to lead to successful differentiations
while 2D was characterized by some variability. One of the major problems of 3D culture, related with
size heterogeneity, was even surpassed with extended periods of forced aggregation in the optimized
protocol.
From a gene expression standpoint, transcriptomic analysis revealed that 3D aggregates of hiPSCs
could recapitulate in vivo cardiogenesis at sequential stages from the pluripotent state to mesoderm,
passing through cardiac progenitors ending at a cardiomyocyte-like state. The same was true to 2D
adherent monolayer. Despite of the simultaneity of relative expression peak for mesendodermal genes in
both 2D and 3D systems, the fold change revealed a relative expression substantially higher for these
genes in 3D culture. Consequently, earlier expression patterns for several transcription factor and
structural protein encoding genes in 3D culture, comparing to 2D, support a faster progression in cardiac
development for 3D aggregates.
The discovery of such differences opened the door for investigation of the role of cell-cell
interactions on cardiomyocyte differentiation with particular focus on EMT. Briefly, it is proposed that
based on larger amounts of E-CAD on 3D aggregates, these cells can generate higher pools of β-catenin
upon adherens junction dissociation in the first steps of cardiac differentiation, leading to a stronger Wnt
signalling induction comparing to 2D adherent monolayer. This fact supports the maturation status of the
cardiomyocytes found within differentiated 3D aggregates, which is closer to adult cardiomyocytes mainly
at the morphological, structural and metabolic level.
The present data must not restrict the understanding of cardiac tissue engineering to a set of
attributes but to encourage future work on the field. Transcriptomic analysis of 2D adherent monolayer
seems to be the logical next step of this study allowing a more detailed comparison between the different
culture systems. Comparative studies upon cardiomyocyte maturation also have some space for
58
improvement. In the main signalling pathways responsible for cardiogenesis, there are some important
players whose gene expression profile could also be assessed for a more clarifying analysis. Since this
study was based on a hiPSC-derived heterogeneous population to represent cardiomyocyte
differentiation, containing non-cardiomyocytes, a single-cell approach would be interesting to be
performed. Lastly, comparison of the current in vitro experiment with in vivo samples or even
computational tools that evaluate how the results resemble in vivo properties could also be explored
aiming for novel solutions to the production of fully functional mature cardiomyocytes.
To a large extent, the resulting findings reflect the molecular signatures of 3D culture system during
cardiomyocyte differentiation elevating its considerable potential. Also, this work aids in the basic
understanding of mechanisms governing either pluripotent spheroids or monolayer during mesoderm
commitment, demonstrating that cardiomyocyte differentiation is indeed culture shape-dependent.
Beyond cardiac differentiation protocols refining, this work further broadens the potential applicability of
hiPSC-CMs to regenerative medicine, disease modelling, and drug screening. Therefore, one must look
forward to a future in which simple cues can be applied to take full advantage of the potential of stem
cells. And by doing so, being able to enhance human condition, all the while aiming, literally, for a
change of heart.
59
Anderson, K.M., Anderson, D.M., McAnally, J.R., Shelton, J.M., Bassel-Duby, R., and Olson, E.N. (2016).
Transcription of the non-coding RNA upperhand controls Hand2 expression and heart development.
Nature 539, 433–436.
Atmanli, A., and Domian, I.J. (2017). Recreating the Cardiac Microenvironment in Pluripotent Stem Cell
Models of Human Physiology and Disease. Trends Cell Biol. 27, 352–364.
Avior, Y., Sagi, I., and Benvenisty, N. (2016). Pluripotent stem cells in disease modelling and drug
discovery. Nat. Rev. Mol. Cell Biol. 17, 170–182.
Azarin, S.M., Lian, X., Larson, E.A., Popelka, H.M., de Pablo, J.J., and Palecek, S.P. (2012). Modulation
of Wnt/β-catenin signaling in human embryonic stem cells using a 3-D microwell array. Biomaterials 33,
2041–2049.
Bao, X., Lian, X., Qian, T., Bhute, V.J., Han, T., and Palecek, S.P. (2017). Directed differentiation and
long-term maintenance of epicardial cells derived from human pluripotent stem cells under fully defined
conditions. Nat. Protoc. 12, 1890–1900.
Bauwens, C., Yin, T., Dang, S., Peerani, R., and Zandstra, P.W. (2005). Development of a perfusion fed
bioreactor for embryonic stem cell-derived cardiomyocyte generation: Oxygen-mediated enhancement
of cardiomyocyte output. Biotechnol. Bioeng. 90, 452–461.
Bauwens, C.L., Song, H., Thavandiran, N., Ungrin, M., Massé, S., Nanthakumar, K., Seguin, C., and
Zandstra, P.W. (2011). Geometric control of cardiomyogenic induction in human pluripotent stem cells.
Tissue Eng. 17, 1901–1909.
Behfar, A., Crespo-Diaz, R., Terzic, A., and Gersh, B.J. (2014). Cell therapy for cardiac repair—lessons
from clinical trials. Nat. Rev. Cardiol. 11, 232–246.
Bellin, M., Marchetto, M.C., Gage, F.H., and Mummery, C.L. (2012). Induced pluripotent stem cells: the
new patient? Nat. Rev. Mol. Cell Biol. 13, 713–726.
Beltrami, A.P., Barlucchi, L., Torella, D., Baker, M., Limana, F., Chimenti, S., Kasahara, H., Rota, M.,
Musso, E., Urbanek, K., et al. (2003). Adult Cardiac Stem Cells Are Multipotent and Support Myocardial
Regeneration we have documented the existence of cycling ventricu- lar myocytes in the normal and
pathologic adult mam. Cell 114, 763–776.
Bergmann, O., Zdunek, S., Felker, A., Salehpour, M., Alkass, K., Bernard, S., Sjostrom, S.L.,
Szewczykowska, M., Jackowska, T., Dos Remedios, C., et al. (2015). Dynamics of Cell Generation and
Turnover in the Human Heart. Cell 161, 1566–1575.
Blanpain, C., and Simons, B.D. (2013). Unravelling stem cell dynamics by lineage tracing. Nat. Rev. Mol.
Cell Biol. 14, 489–502.
60
Blau, H.M., Brazelton, T.R., and Weimann, J.M. (2001). The evolving concept of a stem cell: Entity or
function? Cell 105, 829–841.
De Boer, T.P., Van Veen, T.A.B., Bierhuizen, M.F.A., Kok, B., Rook, M.B., Boonen, K.J.M., Vos, M.A.,
Doevendans, P.A., De Bakker, J.M.T., and Van Der Heyden, M.A.G. (2007). Connexin43 repression
following epithelium-to-mesenchyme transition in embryonal carcinoma cells requires Snail1 transcription
factor. Differentiation 75, 208–218.
Brons, I.G.M., Smithers, L.E., Trotter, M.W.B., Rugg-Gunn, P., Sun, B., Chuva de Sousa Lopes, S.M.,
Howlett, S.K., Clarkson, A., Ahrlund-Richter, L., Pedersen, R.A., et al. (2007). Derivation of pluripotent
epiblast stem cells from mammalian embryos. Nature 448, 191–195.
Burridge, P.W., Anderson, D., Priddle, H., Barbadillo Muñoz, 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., 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.
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.
Burridge, P.W., Sharma, A., and Wu, J.C. (2015). Genetic and Epigenetic Regulation of Human Cardiac
Reprogramming and Differentiation in Regenerative Medicine. Annu. Rev. Genet. 49, 461–484.
Cahill, T.J., Choudhury, R.P., and Riley, P.R. (2017). Heart regeneration and repair after myocardial
infarction: translational opportunities for novel therapeutics. Nat. Rev. Drug Discov. 16, 699–717.
Cao, N., Liang, H., Huang, J., Wang, J., Chen, Y., Chen, Z., and Yang, H. (2013). Highly efficient
induction and long-term maintenance of multipotent cardiovascular progenitors from human pluripotent
stem cells under defined conditions. Nat. Publ. Gr. 23, 1119–1132.
Cao, N., Huang, Y., Zheng, J., Spencer, C.I., Zhang, Y., Fu, J.-D., Nie, B., Xie, M., Zhang, M., Wang,
H., et al. (2016). Conversion of human fibroblasts into functional cardiomyocytes by small molecules.
Science (80-. ). 352, 1216–1220.
Chan, S.S.K., Shi, X., Toyama, A., Arpke, R.W., Dandapat, A., Iacovino, M., Kang, J., Le, G., Hagen,
H.R., Garry, D.J., et al. (2013). Mesp1 patterns mesoderm into cardiac, hematopoietic, or skeletal
myogenic progenitors in a context-dependent manner. Cell Stem Cell 12, 587–601.
Chong, J.J.H., Yang, X., Don, C.W., Minami, E., Liu, Y.-W., Weyers, J.J., Mahoney, W.M., Van Biber, B.,
Cook, S.M., Palpant, N.J., et al. (2014). Human embryonic-stem-cell-derived cardiomyocytes regenerate
non-human primate hearts. Nature 510, 273–277.
61
Choudry, F., Hamshere, S., Saunders, N., Veerapen, J., Bavnbek, K., Knight, C., Pellerin, D., Locca, D.,
Westwood, M., Rakhit, R., et al. (2016). A randomized double-blind control study of early intra-coronary
autologous bone marrow cell infusion in acute myocardial infarction: The REGENERATE-AMI clinical trial.
Eur. Heart J. 37, 256–263.
Correia, C., Koshkin, A., Duarte, P., Hu, D., Teixeira, A., Domian, I., Serra, M., and Alves, P.M. (2017).
Distinct carbon sources affect structural and functional maturation of cardiomyocytes derived from human
pluripotent stem cells. Sci. Rep. 7, 8590.
Dahlmann, J., Kensah, G., Kempf, H., Skvorc, D., Gawol, A., Elliott, D.A., Dräger, G., Zweigerdt, R.,
Martin, U., and Gruh, I. (2013). The use of agarose microwells for scalable embryoid body formation
and cardiac differentiation of human and murine pluripotent stem cells. Biomaterials 34, 2463–2471.
Downing, T.L., Soto, J., Morez, C., Houssin, T., Fritz, A., Yuan, F., Chu, J., Patel, S., Schaffer, D. V., and
Li, S. (2013). Biophysical regulation of epigenetic state and cell reprogramming. Nat. Mater. 12, 1154–
1162.
Dupont, S., Morsut, L., Aragona, M., Enzo, E., Giulitti, S., Cordenonsi, M., Zanconato, F., Le Digabel, J.,
Forcato, M., Bicciato, S., et al. (2011). Role of YAP/TAZ in mechanotransduction. Nature 474, 179–183.
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. Nat. Methods 8, 1037–1040.
Evans, M. (2011). Discovering pluripotency: 30 years of mouse embryonic stem cells. Nat. Rev. Mol. Cell
Biol. 12, 680–686.
Evans, M.J., and Kaufman, M.H. (1981). Establishment in culture of pluripotential cells from mouse
embryos. Nature 292, 154–156.
Fatehullah, A., Tan, S.H., and Barker, N. (2016). Organoids as an in vitro model of human development
and disease. Nat. Cell Biol. 18, 246–254.
Fishman, G.I., Hertzberg, E.L., Spray, D.C., and Leinwand, L.A. (1991). Expression of connexin43 in the
developing rat heart. Circ. Res. 68, 782–787.
Fonoudi, H., Ansari, H., Abbasalizadeh, S., Larijani, M.R., Kiani, S., Hashemizadeh, S., Zarchi, A.S.,
Bosman, A., Blue, G.M., Pahlavan, S., et al. (2015). A Universal and Robust Integrated Platform for the
Scalable Production of Human Cardiomyocytes From Pluripotent Stem Cells. Stem Cells Transl. Med. 4,
1482–1494.
Gonzalez, R., Lee, J.W., and Schultz, P.G. (2011). Stepwise chemically induced cardiomyocyte
specification of human embryonic stem cells. Angew. Chemie - Int. Ed. 50, 11181–11185.
Hartogh, S.C. Den, Wolstencroft, K., Mummery, C.L., and Passier, R. (2016). A comprehensive gene
expression analysis at sequential stages of in vitro cardiac differentiation from isolated MESP1-expressing-
mesoderm progenitors. Nat. Publ. Gr. 1–16.
Haubner, B.J., Schneider, J., Schweigmann, U., Schuetz, T., Dichtl, W., Velik-Salchner, C., Stein, J.I., and
Penninger, J.M. (2016). Functional Recovery of a Human Neonatal Heart after Severe Myocardial
62
Infarction. Circ. Res. 118, 216–221.
Hsiao, C., Tomai, M., Glynn, J., and Palecek, S.P. (2014). Effects of 3-D microwell culture on initial fate
specification in human embryonic stem cells. AIChE J. 60, 1225–1235.
Hsu, Y.-C., and Fuchs, E. (2012). A family business: stem cell progeny join the niche to regulate
homeostasis. Nat. Rev. Mol. Cell Biol. 13, 103–114.
Hudson, J., Titmarsh, D., Hidalgo, A., Wolvetang, E., and Cooper-White, J. (2011). Primitive Cardiac
Cells from Human Embryonic Stem Cells. Stem Cells Dev. 21, 1513–1523.
Hwang, P.M., and Sykes, B.D. (2015). Targeting the sarcomere to correct muscle function. Nat. Rev. Drug
Discov. 14, 313–328.
Ieda, M., Fu, J.D., Delgado-Olguin, P., Vedantham, V., Hayashi, Y., Bruneau, B.G., and Srivastava, D.
(2010). Direct reprogramming of fibroblasts into functional cardiomyocytes by defined factors. Cell 142,
375–386.
Kam, Y., and Quaranta, V. (2009). Cadherin-Bound β-Catenin Feeds into the Wnt Pathway upon
Adherens Junctions Dissociation: Evidence for an Intersection between β-Catenin Pools. PLoS One 4,
e4580.
Karantalis, V., Difede, D.L., Gerstenblith, G., Pham, S., Symes, J., Zambrano, J.P., Fishman, J., Pattany,
P., McNiece, I., Conte, J., et al. (2014). Autologous mesenchymal stem cells produce concordant
improvements in regional function, tissue perfusion, and fibrotic burden when administered to patients
undergoing coronary artery bypass grafting: The prospective randomized study of mesenchymal stem ce.
Circ. Res. 114, 1302–1310.
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., 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.
Kempf, H., Olmer, R., Kropp, C., R??ckert, M., Jara-Avaca, M., Robles-Diaz, D., Franke, A., Elliott, D.A.,
Wojciechowski, D., Fischer, M., et al. (2014). Controlling expansion and cardiomyogenic differentiation
of human pluripotent stem cells in scalable suspension culture. Stem Cell Reports 3, 1132–1146.
Kempf, H., Kropp, C., Olmer, R., Martin, U., and Zweigerdt, R. (2015). Cardiac differentiation of human
pluripotent stem cells in scalable suspension culture. Nat. Protoc. 10, 1345–1361.
Kempf, H., Olmer, R., Haase, A., Franke, A., Bolesani, E., Schwanke, K., Robles-Diaz, D., Coffee, M.,
Göhring, G., Dräger, G., et al. (2016). Bulk cell density and Wnt/TGFbeta signalling regulate
mesendodermal patterning of human pluripotent stem cells. Nat. Commun. 7, 13602.
Kimbrel, E. a., and Lanza, R. (2015). Current status of pluripotent stem cells: moving the first therapies to
the clinic. Nat. Rev. Drug Discov. 14, 681–692.
Knoblich, J.A. (2008). Mechanisms of Asymmetric Stem Cell Division. Cell 132, 583–597.
63
Kropp, C., Massai, D., and Zweigerdt, R. (2017). Progress and challenges in large-scale expansion of
human pluripotent stem cells. Process Biochem. 59, 244–254.
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.
Lalit, P.A., Hei, D.J., Raval, A.N., and Kamp, T.J. (2014). Induced pluripotent stem cells for post-
myocardial infarction repair: Remarkable opportunities and challenges. Circ. Res. 114, 1328–1345.
Lamm, N., Ben-David, U., Golan-Lev, T., Storchová, Z., Benvenisty, N., and Kerem, B. (2016). Genomic
Instability in Human Pluripotent Stem Cells Arises from Replicative Stress and Chromosome Condensation
Defects. Cell Stem Cell 18, 253–261.
Lamouille, S., Xu, J., and Derynck, R. (2014). Molecular mechanisms of epithelial–mesenchymal
transition. Nat. Rev. Mol. Cell Biol. 15, 178–196.
Lancaster, M.A., Renner, M., Martin, C.-A., Wenzel, D., Bicknell, L.S., Hurles, M.E., Homfray, T.,
Penninger, J.M., Jackson, A.P., and Knoblich, J.A. (2013). Cerebral organoids model human brain
development and microcephaly. Nature 501, 373–379.
Laschke, M.W., and Menger, M.D. (2017). Life is 3D: Boosting Spheroid Function for Tissue Engineering.
Trends Biotechnol. 35, 133–144.
Lecina, M., Ting, S., Choo, A., Reuveny, S., and Oh, S. (2010). Scalable Platform for Human Embryonic
Stem Cell Differentiation to Cardiomyocytes in Suspended Microcarrier Cultures. Tissue Eng. Part C
Methods 16, 1609–1619.
Lee, J.H., Protze, S.I., Laksman, Z., Backx, P.H., and Keller, G.M. (2017). Human Pluripotent Stem Cell-
Derived Atrial and Ventricular Cardiomyocytes Develop from Distinct Mesoderm Populations. Cell Stem
Cell 21, 179–194.e4.
Li, M., and Belmonte, J.C.I. (2017). Ground rules of the pluripotency gene regulatory network. Nat. Rev.
Genet. 18, 180–191.
Li, J., Patel, V. V., Kostetskii, I., Xiong, Y., Chu, A.F., Jacobson, J.T., Yu, C., Morley, G.E., Molkentin, J.D.,
and Radice, G.L. (2005). Cardiac-specific loss of N-cadherin leads to alteration in connexins with
conduction slowing and arrhythmogenesis. Circ. Res. 97, 474–481.
Li, S., Chen, G., and Li, R.A. (2013). Calcium signalling of human pluripotent stem cell-derived
cardiomyocytes. J. Physiol. 591, 5279–5290.
Lian, X., Hsiao, C., Wilson, G., Zhu, K., Hazeltine, L.B., Azarin, S.M., Raval, K.K., Zhang, J., Kamp, T.J.,
and Palecek, S.P. (2012). Robust cardiomyocyte differentiation from human pluripotent stem cells via
temporal modulation of canonical Wnt signaling. Proc. Natl. Acad. Sci. 109, E1848–E1857.
Lian, X., Zhang, J., Azarin, S.M., Zhu, K., Hazeltine, L.B., Bao, X., Hsiao, C., Kamp, T.J., and Palecek,
S.P. (2013a). Directed cardiomyocyte differentiation from human pluripotent stem cells by modulating
Wnt/β-catenin signaling under fully defined conditions. Nat. Protoc. 8, 162–175.
Lian, X., Zhang, J., Zhu, K., Kamp, T.J., and Palecek, S.P. (2013b). Insulin inhibits cardiac mesoderm,
64
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.
Lian, X., Bao, X., Al-Ahmad, A., Liu, J., Wu, Y., Dong, W., Dunn, K.K., Shusta, E. V., and Palecek, S.P.
(2014). Efficient differentiation of human pluripotent stem cells to endothelial progenitors via small-
molecule activation of WNT signaling. Stem Cell Reports 3, 804–816.
Lian, X., Bao, X., Zilberter, M., Westman, M., Fisahn, A., Hsiao, C., Hazeltine, L.B., Dunn, K.K., Kamp,
T.J., and Palecek, S.P. (2015). Chemically defined, albumin-free human cardiomyocyte generation. Nat.
Methods 12, 595–596.
Liang, P., Lan, F., Lee, A.S., Gong, T., Sanchez-Freire, V., Wang, Y., Diecke, S., Sallam, K., Knowles,
J.W., Wang, P.J., et al. (2013). Drug screening using a library of human induced pluripotent stem cell-
derived cardiomyocytes reveals disease-specific patterns of cardiotoxicity. Circulation 127, 1677–1691.
Loh, K.M., Chen, A., Koh, P.W., Deng, T.Z., Sinha, R., Tsai, J.M., Barkal, A.A., Shen, K.Y., Jain, R.,
Morganti, R.M., et al. (2016). Mapping the Pairwise Choices Leading from Pluripotency to Human Bone,
Heart, and Other Mesoderm Cell Types. Cell 166, 451–468.
De Los Angeles, A., Ferrari, F., Xi, R., Fujiwara, Y., Benvenisty, N., Deng, H., Hochedlinger, K., Jaenisch,
R., Lee, S., Leitch, H.G., et al. (2015). Hallmarks of pluripotency. Nature 525, 469–478.
Martin, G.R. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium
conditioned by teratocarcinoma stem cells. Proc. Natl. Acad. Sci. U. S. A. 78, 7634–7638.
Masumoto, H., Ikuno, T., Takeda, M., Fukushima, H., Marui, A., Katayama, S., Shimizu, T., Ikeda, T.,
Okano, T., Sakata, R., et al. (2014). Human iPS cell-engineered cardiac tissue sheets with cardiomyocytes
and vascular cells for cardiac regeneration. Sci. Rep. 4, 1–7.
Masumoto, H., Nakane, T., Tinney, J.P., Yuan, F., Ye, F., Kowalski, W.J., Minakata, K., Sakata, R.,
Yamashita, J.K., and Keller, B.B. (2016). The myocardial regenerative potential of three-dimensional
engineered cardiac tissues composed of multiple human iPS cell-derived cardiovascular cell lineages. Sci.
Rep. 6, 1–10.
Matsui, Y., Zsebo, K., and Hogan, B.L.M. (1992). Derivation of pluripotential embryonic stem cells from
murine primordial germ cells in culture. Cell 70, 841–847.
Menasché, P., Vanneaux, V., Hagège, A., Bel, A., Cholley, B., Cacciapuoti, I., Parouchev, A.,
Benhamouda, N., Tachdjian, G., Tosca, L., et al. (2015). Human embryonic stem cell-derived cardiac
progenitors for severe heart failure treatment: First clinical case report. Eur. Heart J. 36, 2011–2017.
Merrell, A.J., and Stanger, B.Z. (2016). Adult cell plasticity in vivo: de-differentiation and
transdifferentiation are back in style. Nat. Rev. Mol. Cell Biol. 17, 413–425.
Messina, E., De Angelis, L., Frati, G., Morrone, S., Chimenti, S., Fiordaliso, F., Salio, M., Battaglia, M.,
Latronico, M.V.G., Coletta, M., et al. (2004). Isolation and expansion of adult cardiac stem cells from
human and murine heart. Circ. Res. 95, 911–921.
Minami, I., Yamada, K., Otsuji, T.G., Yamamoto, T., Shen, Y., Otsuka, S., Kadota, S., Morone, N., Barve,
M., Asai, Y., et al. (2012). A Small Molecule that Promotes Cardiac Differentiation of Human Pluripotent
65
Stem Cells under Defined, Cytokine- and Xeno-free Conditions. Cell Rep. 2, 1448–1460.
Mordwinkin, N.M., Burridge, P.W., and Wu, J.C. (2013). A review of human pluripotent stem cell-derived
cardiomyocytes for high-throughput drug discovery, cardiotoxicity screening, and publication standards.
J. Cardiovasc. Transl. Res. 6, 22–30.
Morrison, S.J., and Kimble, J. (2006). Asymmetric and symmetric stem-cell divisions in development and
cancer. Nature 441, 1068–1074.
Mozaffarian, D., Benjamin, E.J., Go, A.S., Arnett, D.K., Blaha, M.J., Cushman, M., De Ferranti, S.,
Després, J.P., Fullerton, H.J., Howard, V.J., et al. (2015). Heart disease and stroke statistics-2015 update :
A report from the American Heart Association. Circulation 131, e29–e39.
Müller, R., and Lengerke, C. (2009). Patient-specific pluripotent stem cells: promises and challenges. Nat.
Rev. Endocrinol. 5, 195–203.
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. Circ. Res. 111, 344–358.
Närvä, E., Rahkonen, N., Emani, M.R., Lund, R., Pursiheimo, J.P., Nästi, J., Autio, R., Rasool, O.,
Denessiouk, K., Lähdesmäki, H., et al. (2012). RNA-binding protein L1TD1 interacts with LIN28 via RNA
and is required for human embryonic stem cell self-renewal and cancer cell proliferation. Stem Cells 30,
452–460.
Navarrete, E.G., Liang, P., Lan, F., Sanchez-Freire, V., Simmons, C., Gong, T., Sharma, A., Burridge,
P.W., Patlolla, B., Lee, A.S., et al. (2013). Screening drug-induced arrhythmia events using human
induced pluripotent stem cell-derived cardiomyocytes and low-impedance microelectrode arrays.
Circulation 128.
Niebruegge, S., Bauwens, C.L., Peerani, R., Thavandiran, N., Masse, S., Sevaptisidis, E., Nanthakumar,
K., Woodhouse, K., Husain, M., Kumacheva, E., et al. (2009). Generation of human embryonic stem cell-
derived mesoderm and cardiac cells using size-specified aggregates in an oxygen-controlled bioreactor.
Biotechnol. Bioeng. 102, 493–507.
Noseda, M., Peterkin, T., Simões, F.C., Patient, R., and Schneider, M.D. (2011). Cardiopoietic factors
extracellular signals for cardiac lineage commitment. Circ. Res. 108, 129–152.
Nunes, S.S., Miklas, J.W., Liu, J., Aschar-Sobbi, R., Xiao, Y., Zhang, B., Jiang, J., Massé, 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.
Paige, S.L., Osugi, T., Afanasiev, O.K., Pabon, L., Reinecke, H., and Murry, C.E. (2010). Endogenous
Wnt/β-Catenin signaling is required for cardiac differentiation in human embryonic stem cells. PLoS One
5.
Paige, S.L., Plonowska, K., Xu, A., and Wu, S.M. (2015). Molecular regulation of cardiomyocyte
differentiation. Circ. Res. 116, 341–353.
Patterson, M., Barske, L., Van Handel, B., Rau, C.D., Gan, P., Sharma, A., Parikh, S., Denholtz, M.,
66
Huang, Y., Yamaguchi, Y., et al. (2017). Frequency of mononuclear diploid cardiomyocytes underlies
natural variation in heart regeneration. Nat. Genet. 49, 1346–1353.
Pervolaraki, E., Dachtler, J., Anderson, R.A., and Holden, A. V. (2017). Ventricular myocardium
development and the role of connexins in the human fetal heart. Sci. Rep. 7, 12272.
Pinto, A.R., Ilinykh, A., Ivey, M.J., Kuwabara, J.T., D’antoni, M.L., Debuque, R., Chandran, A., Wang, L.,
Arora, K., Rosenthal, N.A., et al. (2016). Revisiting cardiac cellular composition. Circ. Res. 118, 400–
409.
Porrello, E.R., Mahmoud, A.I., Simpson, E., Hill, J.A., Richardson, J.A., Olson, E.N., and Sadek, H.A.
(2011). Transient Regenerative Potential of the Neonatal Mouse Heart. Science (80-. ). 331, 1078–1080.
Poss, K.D. (2002). Heart Regeneration in Zebrafish. Science (80-. ). 298, 2188–2190.
Ren, Y., Lee, M.Y., Schliffke, S., Paavola, J., Amos, P.J., Ge, X., Ye, M., Zhu, S., Senyei, G., Lum, L., et al.
(2011). Small molecule Wnt inhibitors enhance the efficiency of BMP-4-directed cardiac differentiation of
human pluripotent stem cells. J. Mol. Cell. Cardiol. 51, 280–287.
Ringnér, M., and Ringner, M. (2008). What is principal component analysis? Nat Biotechnol 26, 303–
304.
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.
Sánchez Alvarado, A., and Yamanaka, S. (2014). Rethinking differentiation: Stem cells, regeneration,
and plasticity. Cell 157, 110–119.
Sayed, N., Liu, C., and Wu, J.C. (2016). Translation of Human-Induced Pluripotent Stem Cells from
Clinical Trial in a Dish to Precision Medicine. J. Am. Coll. Cardiol. 67, 2161–2176.
Schroeder, M., Niebruegge, S., Werner, A., Willbold, E., Burg, M., Ruediger, M., Field, L.J., Lehmann, J.,
and Zweigerdt, R. (2005). Differentiation and lineage selection of mouse embryonic stem cells in a stirred
bench scale bioreactor with automated process control. Biotechnol. Bioeng. 92, 920–933.
Shamblott, M.J., Axelman, J., Wang, S., Bugg, E.M., Littlefield, J.W., Donovan, P.J., Blumenthal, P.D.,
Huggins, G.R., and Gearhart, J.D. (1998). Derivation of pluripotent stem cells from cultured human
primordial germ cells. Proc. Natl. Acad. Sci. U. S. A. 95, 13726–13731.
Shi, W., Wang, H., Pan, G., Geng, Y., Guo, Y., and Pei, D. (2006). Regulation of the Pluripotency Marker
Rex-1 by Nanog and Sox2. J. Biol. Chem. 281, 23319–23325.
Shi, Y., Inoue, H., Wu, J.C., and Yamanaka, S. (2016). Induced pluripotent stem cell technology: a decade
of progress. Nat. Rev. Drug Discov. 16, 115–130.
Shiba, Y., Gomibuchi, T., Seto, T., Wada, Y., Ichimura, H., Tanaka, Y., Ogasawara, T., Okada, K., Shiba,
N., Sakamoto, K., et al. (2016). Allogeneic transplantation of iPS cell-derived cardiomyocytes regenerates
primate hearts. Nature 538, 388–391.
Solter, D., and Solter, D. (2006). From teratocarcinomas to embryonic stem cells and beyond: a history
of embryonic stem cell research. Nat. Rev. Genet. 7, 319–327.
De Sousa Lopes, S.M.C., Hassink, R.J., Feijen, A., Van Rooijen, M.A., Doevendans, P.A., Tertoolen, L.,
67
De La Rivière, A.B., and Mummery, C.L. (2006). Patterning the heart, a template for human cardiomyocyte
development. Dev. Dyn. 235, 1994–2002.
Später, D., Hansson, E.M., Zangi, L., and Chien, K.R. (2014). How to make a cardiomyocyte.
Development 141, 4418–4431.
Srivastava, D., and DeWitt, N. (2016). In??Vivo Cellular Reprogramming: The Next Generation. Cell 166,
1386–1396.
Sterneckert, J.L., Reinhardt, P., and Scholer, H.R. (2014). Investigating human disease using stem cell
models. Nat Rev Genet 15, 625–639.
Stevens, L.C. (1960). Embryonic potency of embryoid bodies derived from a transplantable testicular
teratoma of the mouse. Dev. Biol. 2, 285–297.
Stricker, S.H., Köferle, A., and Beck, S. (2017). From profiles to function in epigenomics. Nat. Rev. Genet.
18, 51–66.
Szklarczyk, D., Franceschini, A., Wyder, S., Forslund, K., Heller, D., Huerta-Cepas, J., Simonovic, M.,
Roth, A., Santos, A., Tsafou, K.P., et al. (2015). STRING v10: Protein-protein interaction networks,
integrated over the tree of life. Nucleic Acids Res. 43, D447–D452.
Tabar, V., and Studer, L. (2014). Pluripotent stem cells in regenerative medicine: challenges and recent
progress. Nat Rev Genet 15, 82–92.
Takahashi, K., and Yamanaka, S. (2006). Induction of Pluripotent Stem Cells from Mouse Embryonic and
Adult Fibroblast Cultures by Defined Factors. Cell 126, 663–676.
Takahashi, K., and Yamanaka, S. (2016). A decade of transcription factor-mediated reprogramming to
pluripotency. Nat. Rev. Mol. Cell Biol. 17, 183–193.
Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S. (2007).
Induction of Pluripotent Stem Cells from Adult Human Fibroblasts by Defined Factors. Cell 131, 861–
872.
Tesar, P.J., Chenoweth, J.G., Brook, F. a, Davies, T.J., Evans, E.P., Mack, D.L., Gardner, R.L., and McKay,
R.D. (2007). New cell lines from mouse epiblast share defining features with human embryonic stem cells.
Nature 448, 196–199.
Thomson, J.A. (1998). Embryonic Stem Cell Lines Derived from Human Blastocysts. Science (80-. ). 282,
1145–1147.
Ting, S., Chen, A., Reuveny, S., and Oh, S. (2014). An intermittent rocking platform for integrated
expansion and differentiation of human pluripotent stem cells to cardiomyocytes in suspended
microcarrier cultures. Stem Cell Res. 13, 202–213.
Tohyama, S., Fujita, J., Fujita, C., Yamaguchi, M., Kanaami, S., Ohno, R., Sakamoto, K., Kodama, M.,
Kurokawa, J., Kanazawa, H., et al. (2017). Stem Cell Reports Efficient Large-Scale 2D Culture System for
Human Induced Pluripotent Stem Cells and Differentiated Cardiomyocytes. Stem Cell Reports 9, 1–9.
Tran, T.H., Wang, X., Browne, C., Zhang, Y., Schinke, M., Izumo, S., and Burcin, M. (2009). Wnt3a-
induced mesoderm formation and cardiomyogenesis in human embryonic stem cells. Stem Cells 27,
68
1869–1878.
Tzahor, E., and Poss, K.D. (2017). Cardiac regeneration strategies: Staying young at heart. Science (80-
. ). 356, 1035 LP-1039.
Ueno, S., Weidinger, G., Osugi, T., Kohn, A.D., Golob, J.L., Pabon, L., Reinecke, H., Moon, R.T., and
Murry, C.E. (2007). Biphasic role for Wnt/beta-catenin signaling in cardiac specification in zebrafish and
embryonic stem cells. Proc. Natl. Acad. Sci. U. S. A. 104, 9685–9690.
Uosaki, H., Cahan, P., Lee, D.I., Wang, S., Miyamoto, M., Fernandez, L., Kass, D.A., and Kwon, C.
(2015). Transcriptional Landscape of Cardiomyocyte Maturation. Cell Rep. 13, 1705–1716.
Uygur, A., and Lee, R.T. (2016). Mechanisms of Cardiac Regeneration. Dev. Cell 36, 362–374.
Vivien, C.J., Hudson, J.E., and Porrello, E.R. (2016). Evolution, comparative biology and ontogeny of
vertebrate heart regeneration. NPJ Regen. Med. 1, 16012.
Vogel, V., and Sheetz, M. (2006). Local force and geometry sensing regulate cell functions. Nat. Rev. Mol.
Cell Biol. 7, 265–275.
Wang, J., Xie, G., Singh, M., Ghanbarian, A.T., Raskó, T., Szvetnik, A., Cai, H., Besser, D., Prigione, A.,
Fuchs, N. V., et al. (2014). Primate-specific endogenous retrovirus-driven transcription defines naive-like
stem cells. Nature 516, 405–409.
Wang, Z., Gerstein, M., and Snyder, M. (2009). RNA-Seq: a revolutionary tool for transcriptomics. Nat.
Rev. Genet. 10, 57–63.
Weinberger, L., Ayyash, M., Novershtern, N., and Hanna, J.H. (2016). Dynamic stem cell states: naive to
primed pluripotency in rodents and humans. Nat. Rev. Mol. Cell Biol. 17, 155–169.
Weissman, I.L. (2000). Stem cells: units of development, units of regeneration, and units in evolution. Cell
100, 157–168.
Willems, E., Spiering, S., Davidovics, H., Lanier, M., Xia, Z., Dawson, M., Cashman, J., and Mercola, M.
(2011). Small-molecule inhibitors of the Wnt pathway potently promote cardiomyocytes from human
embryonic stem cell-derived mesoderm. Circ. Res. 109, 360–364.
Wu, J., and Izpisua Belmonte, J.C. (2016). Stem Cells: A Renaissance in Human Biology Research. Cell
165, 1572–1585.
Wu, J., Yamauchi, T., and Izpisua Belmonte, J.C. (2016). An overview of mammalian pluripotency.
Development 143, 1644–1648.
Xin, M., Olson, E.N., and Bassel-Duby, R. (2013). Mending broken hearts: cardiac development as a
basis for adult heart regeneration and repair. Nat. Rev. Mol. Cell Biol. 14, 529–541.
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.
Yang, X., Pabon, L., and Murry, C.E. (2014). Engineering adolescence: Maturation of human pluripotent
stem cell-derived cardiomyocytes. Circ. Res. 114, 511–523.
Yin, X., Mead, B.E., Safaee, H., Langer, R., Karp, J.M., and Levy, O. (2016). Engineering Stem Cell
69
Organoids. Cell Stem Cell 18, 25–38.
Yu, H.B., Kunarso, G., Hong, F.H., and Stanton, L.W. (2009). Zfp206, Oct4, and Sox2 are integrated
components of a transcriptional regulatory network in embryonic stem cells. J. Biol. Chem. 284, 31327–
31335.
Yu, J., Vodyanik, M., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J., Tian, S., Nie, J., Jonsdottir, G.,
Ruotti, V., Stewart, R., et al. (2007). Induced Pluripotent Stem Cell Lines Derived from Human Somatic
Cells. Science (80-. ). 318, 1917–1920.
Zandstra, P.W., Bauwens, C., Yin, T., Liu, Q., Schiller, H., Zweigerdt, R., Pasumarthi, K.B.S., and Field,
L.J. (2003). Scalable Production of Embryonic Stem Cell-Derived Cardiomyocytes. Tissue Eng. 9, 767–
778.
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. Circ. Res.
104.
Zhang, M., Schulte, J.S., Heinick, A., Piccini, I., Rao, J., Quaranta, R., Zeuschner, D., Malan, D., Kim,
K.P., Röpke, A., et al. (2015). Universal cardiac induction of human pluripotent stem cells in two and
three-dimensional formats: Implications for in vitro maturation. Stem Cells 33, 1456–1469.
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.
Zhu, W.Z., Xie, Y., Moyes, K.W., Gold, J.D., Askari, B., and Laflamme, M.A. (2010). Neuregulin/ErbB
signaling regulates cardiac subtype specification in differentiating human embryonic stem cells. Circ. Res.
107, 776–786.
Zuppinger, C., Eppenberger-Eberhardt, M., and Eppenberger, H.M. (2000). N-Cadherin: Structure,
function and importance in the formation of new intercalated disc-like cell contacts in cardiomyocytes.
Heart Fail. Rev. 5, 251–257.
Zwi, L., Caspi, O., Arbel, G., Huber, I., Gepstein, A., Park, I.H., and Gepstein, L. (2009). Cardiomyocyte
differentiation of human induced pluripotent stem cells. Circulation 120, 1513–1523.
70
71
Supplementary Figure 1 | Cardiomyocyte differentiation efficiency for DF6 cell line assessed by flow cytometry.
Supplementary Figure 2 | Optimization of hiPSC cardiomyocyte differentiation in a 3D culture system by a face-
centred composite design.
Quadratic model relating initial aggregate diameter and small molecule CHIR concentration with cTnT percentage at day
15 of differentiation. (A) 3D representation and (B) 2D heatmap of the obtained results. (C) Percentage of cTnT at day 15
of differentiation related with aggregate diameter and (D) with CHIR concentration. Discontinuous line represents the
relation predicted by the model and the solid lines represent the standard deviation at +/-95% of confidence. Blue dots
represent the experimental results obtained for the model. (E) Significance contribution of each factor to the 3D
cardiomyocyte differentiation model (Diam – Aggregate diameter; CHIR – CHIR concentration). Contribution of each factor
was considered statistically significant for p-values<0.05. Mariana Branco’s unpublished results.
72
Supplementary Figure 3 | Pluripotency at day 0 of differentiation for 2D and 3D culture systems.
Quantification of pluripotency by flow-cytometry analysis for OCT4 on hiPSCs at day 0 of differentiation for (A) 2D adherent
monolayer and (B) 3D aggregates. n=1.
Supplementary Figure 4 | Hematoxylin and eosin analysis of 3D aggregates.
Histological analysis by H&E of cryosection of a day 15 aggregate. The arrow highlights an area of apoptosis and necrosis
within the aggregate. Scale bar represents 50 µm.
73
Supplementary Figure 5 | Characterization of hiPSC cardiomyocyte differentiation in 3D culture systems for TCLab
and Gibco® cell lines. ǁ
74
Supplementary Table 1 | Transcriptomic analysis of cardiomyocyte differentiation in 3D culture.
Top 10 upregulated and top 10 downregulated genes assessed by transcriptomic analysis when comparing day 0 and day
20 of differentiation in 3D culture (fold change > 2, p-value < 0.05).
Gene Gene Description Gene Type Fold Change (log2) p value
MYH6 myosin heavy chain 6 Protein coding 11.74 7.05E-226
TNNI1 troponin I1, slow skeletal type Protein coding 11.32 5.68E-73
HSPB7 heat shock protein family B
(small) member 7 Protein coding 11.31 4.51E-33
MYBPC3 myosin binding protein C, cardiac Protein coding 11.04 6.35E-114
TNNT2 troponin T2, cardiac type Protein coding 10.78 3.05E-124
SMPX small muscle protein, X-linked Protein coding 10.66 6.65E-48
MYL4 myosin light chain 4 Protein coding 10.60 3.93E-34
HAND2 heart and neural crest derivatives
expressed 2 Protein coding 10.52 6.33E-49
MYL7 myosin light chain 7 Protein coding 10.46 1.25E-135
HAND2-AS1 HAND2 antisense RNA 1 (head to
head) Antisense RNA 10.42 2.46E-29
POU5F1 POU class 5 homeobox 1 Protein coding -8.53 1.23E-24
AC022140.2 - TEC4 -8.84 3.72E-19
MT1G metallothionein 1G Protein coding -8.87 5.26E-19
ZSCAN10 zinc finger and SCAN domain
containing 10 Protein coding -8.94 6.52E-20
AC106864.1 - Antisense RNA -9.23 2.47E-65
ZFP42 ZFP42 zinc finger protein Protein coding -9.35 9.46E-31
AC009446.1 - lincRNA -9.49 5.48E-32
NANOG Nanog homeobox Protein coding -9.73 9.25E-25
LINC00678 long intergenic non-protein
coding RNA 678 lincRNA -9.83 3.04E-34
L1TD1 LINE1 type transposase domain
containing 1 Protein coding -10.00 5.99E-132
ESRG embryonic stem cell related Sense intronic -11.29 8.61E-60
ǁ A) TCLab and (B) Gibco® differentiation efficiency by flow-cytometry analysis for cTnT on cells differentiated in 3D culture.
n=3. (C) TCLab and (D) Gibco® immunofluorescent staining for cTnT, α-ACT, α-SMA and CD31 in red of differentiated cells
in cryosections of 3D aggregates on day 15 of differentiation. Nuclei in blue were stained with DAPI. Scale bars represent
50 µm. (E) Gene expression comparison for DF6 (•) and TCLab (∎) cell lines assessed by qRT-PCR at sequential stages for
key genes that unveil the progression throughout cardiomyocyte differentiation with OCT4 (pluripotency), T
(mesendoderm), ISL1 and NKX2-5 (cardiac progenitors) and TNNT2, MYH6, MYH7, MYL2 and MYL7 (cardiomyocytes).
Analysis and data normalization was performed as for DF6 cell line. The Y axis has a log10 scale. Mean ± SEM; n=1
experiments for each time point.
4 To be Experimentally Confirmed.
75
Supplementary Figure 6 | Genes expression normalized values from RNA-seq.
Normalized counts from RNA-seq of (A) regulatory and (B) structural genes at sequential stages of cardiomyocyte
differentiation when cells were cultured as 3D aggregates.
76
77