Human Induced Pluripotent Stem-Cell-Derived Cardiomyocytes as Models for Genetic Cardiomyopathies

In the last few decades, many pathogenic or likely pathogenic genetic mutations in over hundred different genes have been described for non-ischemic, genetic cardiomyopathies. However, the functional knowledge about most of these mutations is still limited because the generation of adequate animal models is time-consuming and challenging. Therefore, human induced pluripotent stem cells (iPSCs) carrying specific cardiomyopathy-associated mutations are a promising alternative. Since the original discovery that pluripotency can be artificially induced by the expression of different transcription factors, various patient-specific-induced pluripotent stem cell lines have been generated to model non-ischemic, genetic cardiomyopathies in vitro. In this review, we describe the genetic landscape of non-ischemic, genetic cardiomyopathies and give an overview about different human iPSC lines, which have been developed for the disease modeling of inherited cardiomyopathies. We summarize different methods and protocols for the general differentiation of human iPSCs into cardiomyocytes. In addition, we describe methods and technologies to investigate functionally human iPSC-derived cardiomyocytes. Furthermore, we summarize novel genome editing approaches for the genetic manipulation of human iPSCs. This review provides an overview about the genetic landscape of inherited cardiomyopathies with a focus on iPSC technology, which might be of interest for clinicians and basic scientists interested in genetic cardiomyopathies.


Introduction
At the beginning of this century, the human genome project was finished [1]. The development of next generation sequencing (NGS) technologies significantly reduced the price and time, allowing for efficient genome and exome analyses, even in clinical routine procedures. However, even 20 years later, the clinical interpretation of genetic sequence variants (GSVs) is still challenging because the functional and structural impact of many variants is unknown. Therefore, multi-disciplinary approaches are often necessary for the interpretation and functional analysis of novel GSVs [2]. At present, in clinical routine procedures, the pathological impact of GSVs is classified due to standards and guidelines of the American College of Medical Genetics and Genomics (ACMG) [3].
Cardiomyopathies are diseases that affect the heart muscle, leading to functional and structural abnormalities [4], and are the main indication for heart transplantation (HTx) [5]. Beside environmental factors, like myocarditis or cardiotoxicity of cancer drugs, non-ischemic cardiomyopathies often have a genetic etiology with dominant inheritance. However, because pathogenic mutations in more than 100 different genes are associated with non-ischemic cardiomyopathies, the interpretation of novel GSVs is still challenging [6]. Moreover, little is currently known on digenic, or even polygenic, etiologies of cardiomyopathies [7]. Incomplete penetrance, different expressivity, and pleiotropy make the clinical interpretation even more challenging.
Functional analyses using adequate cell and animal models can lead to a more sophisticated interpretation of GSVs, which might be not only relevant for genetic counseling but also for the development of personalized therapies. According to the ACMG guidelines, in vitro or/and in vivo functional analyses provide strong criteria (PS3) for the classification of GSVs [3,8]. However, the generation of animal models is still time consuming and expensive. Moreover, in some cases, human cardiomyopathies cannot be modeled using animal models because of species differences. For example, TMEM43-p.S358L is a mutation with full penetrance in several families with arrhythmogenic cardiomyopathy (ACM) [9][10][11]. In contrast, the Tmem43 knock-out, as well as the knock-in mice carrying this specific mutation, do not develop an ACM phenotype [12]. Because of these limitations, human iPSC-derived cardiomyocytes are unprecedented research tools to model and investigate genetic cardiomyopathies.
Here, we provide an overview about the genetic landscape of inherited cardiomyopathies and summarize the development of important human iPSC lines for modelling human cardiomyopathies in vitro. In addition, we review the differentiation into cardiomyocytes and discuss relevant methods used for the cellular and molecular characterization of human iPSC-derived cardiomyocytes.

Clinical Background
In clinical cardiology, cardiomyopathies are classified into five major structural subtypes ( Figure 1). Dilated cardiomyopathy (DCM, MIM #604145) is mainly characterized by left-ventricular dilation in combination with a decrease of the wall diameter [13]. These structural changes decrease the cardiac ejection fraction. Hypertrophic cardiomyopathy (HCM, MIM #160760) is characterized by the hypertrophy of the ventricular walls and/or the septum [14], leading to a reduced cardiac output. Restrictive cardiomyopathy (RCM, MIM #115210) is caused by an increase in ventricular stiffness, leading to dilated atria and diastolic dysfunction [15]. Hyper-trabeculation of the left ventricular wall is a hallmark for (left-ventricular) non-compaction cardiomyopathy (NCCM, MIM #604169) [16]. It mainly affects the left ventricle, but isolated right ventricular or biventricular forms of NCCM have been reported [17]. Ventricular arrhythmias and predominant right or biventricular dilation are the main clinical symptoms of ACM (MIM #609040) [18]. The fibro fatty replacement of the myocardial tissue is a pathognomonic feature characteristic of ACM [19]. However, at the early stage of the disease, structural changes may be absent or subtle [20]. Because ACM is a progressive disease, left ventricular involvement develops frequently at a later stage [21].

Genetic Basis of Inherited Cardiomyopathies
Thirty years ago, Seidmans' group discovered the first pathogenic mutation in MYH7, encoding for β-myosin heavy chain, in a four-generation family, in which several members developed HCM [22]. At present, genetic variants have been described in more than 100 different genes associated with non-ischemic cardiomyopathies or syndromes with cardiac involvement such as Marfan or Leopard syndrome (for an overview, see Table 1). Of note, the spectrum of affected genes and mutations partially overlaps between the different non-ischemic cardiomyopathies ( Figure 1) [32][33][34]. However, the molecular reasons why mutations in the same gene can cause different cardiac phenotypes are largely unknown.
From a genetic point of view, non-ischemic cardiomyopathies are quite heterogeneous [35][36][37]. However, the different non-ischemic cardiomyopathies are characterized by an accumulation of mutations in a distinct set of genes encoding for proteins that are essential for cardiomyocyte function. For example, HCM is mainly caused by mutations in genes encoding sarcomeric proteins

Genetic Basis of Inherited Cardiomyopathies
Thirty years ago, Seidmans' group discovered the first pathogenic mutation in MYH7, encoding for β-myosin heavy chain, in a four-generation family, in which several members developed HCM [22]. At present, genetic variants have been described in more than 100 different genes associated with non-ischemic cardiomyopathies or syndromes with cardiac involvement such as Marfan or Leopard syndrome (for an overview, see Table 1). Of note, the spectrum of affected genes and mutations partially overlaps between the different non-ischemic cardiomyopathies ( Figure 1 , are also known to cause HCM (see Figure 1).
NCCM is the third most frequent non-ischemic cardiomyopathy [54,55] and can occur as a primary cardiomyopathy or can be part of a syndromic disease like the Barth syndrome (MIM, #302060) [56]. Mutations in over 20 different genes having a significant overlap with HCM-or DCM-associated genes have been described in NCCM patients so far (see Figure 1 and Table 1 [69,70]. The majority of affected RCM genes, which partially overlap with the group of HCM-associated genes, encode for sarcomere or cytoskeleton proteins (see Figure 1 and Table 1). The first RCM-associated mutation was identified in TNNI3, encoding cardiac troponin I [71]. More recently, there is growing evidence that FLNC mutations, encoding the cytolinker protein filamin-C, are frequently associated with RCM [72][73][74][75][76].
In summary, a relevant amount of all non-ischemic cardiomyopathies have a genetic etiology. Although in most cases, cardiomyopathies are inherited monogenetically, the underlying genetic landscape is complex, diverse, and currently only partially known.

Generation of Patient-Specific-Induced Pluripotent Stem Cells Via Reprogramming
In the 1960s, Gurdon et al. cloned Xenopus laevis for the first time [274,275]. Consequently, Gurdon was awarded the Nobel Prize in medicine in 2012, together with Yamanaka [276]. The cloning of mammals by nuclear transfer from somatic cells into enucleated unfertilized mammalian eggs over twenty years ago demonstrated that the cellular differentiation can be artificially turned back into a pluripotent state [277]. The next breakthrough was the identification of essential reprogramming factors by the Yamanaka group [278,279]. Initially, reprogramming was performed with 24 candidate transcription factor genes. Out of these, four critical genes were identified to be crucial for iPSC generation: Sox2, Oct4, Klf4, and c-Myc [278]. Depending on the donor cell type, the set of reprogramming factors can vary since specific cell types might endogenously express some of the necessary factors. For example, c-Myc is not required for the reprogramming of fibroblasts [280].
Different delivery methods were developed for reprogramming of somatic cell types like fibroblasts, lymphocytes, keratinocytes, urine-derived, or intestinal cells into iPSCs (see Figure 2). Initially, iPSCs were generated using retroviral transduction [278,279,281]. The Moloney-based retroviral vector system used by the Yamanaka lab has the advantage of undergoing silencing in the iPSCs state but is restricted to dividing cell types. Therefore, lentiviruses were used to improve the transduction efficiency of dividing and non-dividing cell types. However, after lentiviral transduction, the expression of the reprogramming factors are poorly silenced [282,283], leading to difficult differentiation of these iPSCs [284]. Therefore, inducible systems were used, allowing for the silencing of the Yamanaka factors in iPSCs [284,285].

Generation of Patient-Specific-Induced Pluripotent Stem Cells Via Reprogramming
In the 1960s, Gurdon et al. cloned Xenopus laevis for the first time [274,275]. Consequently, Gurdon was awarded the Nobel Prize in medicine in 2012, together with Yamanaka [276]. The cloning of mammals by nuclear transfer from somatic cells into enucleated unfertilized mammalian eggs over twenty years ago demonstrated that the cellular differentiation can be artificially turned back into a pluripotent state [277]. The next breakthrough was the identification of essential reprogramming factors by the Yamanaka group [278,279]. Initially, reprogramming was performed with 24 candidate transcription factor genes. Out of these, four critical genes were identified to be crucial for iPSC generation: Sox2, Oct4, Klf4, and c-Myc [278]. Depending on the donor cell type, the set of reprogramming factors can vary since specific cell types might endogenously express some of the necessary factors. For example, c-Myc is not required for the reprogramming of fibroblasts [280].
Different delivery methods were developed for reprogramming of somatic cell types like fibroblasts, lymphocytes, keratinocytes, urine-derived, or intestinal cells into iPSCs (see Figure 2). Initially, iPSCs were generated using retroviral transduction [278,279,281]. The Moloney-based retroviral vector system used by the Yamanaka lab has the advantage of undergoing silencing in the iPSCs state but is restricted to dividing cell types. Therefore, lentiviruses were used to improve the transduction efficiency of dividing and non-dividing cell types. However, after lentiviral transduction, the expression of the reprogramming factors are poorly silenced [282,283], leading to difficult differentiation of these iPSCs [284]. Therefore, inducible systems were used, allowing for the silencing of the Yamanaka factors in iPSCs [284,285]. However, usage of integrating viral systems enhances the risk for insertional mutagenesis, limiting their application [286]. Furthermore, the transgene reactivation of c-Myc showed increased tumorigenicity in chimeric mice [280], limiting the usage of iPSCs for clinical approaches. To overcome these limitations, non-integrating delivery methods have been developed. Transient transfection of the PiggyBac transposon with a Cre-mediated excisable system was one of the first non-integrating methods ( Figure 2). Minimized genome modification, in combination with silencing of the reprogramming factor expression in the iPSC state, are the main advantages of this system [287]. Another approach is the adenoviral transduction leading to an overexpression of the reprogramming factors in the host cells without genomic integration [288]. Transient transfection or However, usage of integrating viral systems enhances the risk for insertional mutagenesis, limiting their application [286]. Furthermore, the transgene reactivation of c-Myc showed increased tumorigenicity in chimeric mice [280], limiting the usage of iPSCs for clinical approaches. To overcome these limitations, non-integrating delivery methods have been developed. Transient transfection of the PiggyBac transposon with a Cre-mediated excisable system was one of the first non-integrating methods ( Figure 2). Minimized genome modification, in combination with silencing of the reprogramming factor expression in the iPSC state, are the main advantages of this system [287]. Another approach is the adenoviral transduction leading to an overexpression of the reprogramming factors in the host cells without genomic integration [288]. Transient transfection or electroporation with episomal plasmids encoding the reprogramming factors is an alternative method to produce virus-free iPSCs [289] ( Figure 2). However, the efficiency of this delivery method is quite low [290]. More promising non-inserting delivery methods include the use of Sendai viruses [291], which are RNA viruses that do not enter the nucleus, thereby decreasing the risk of genomic insertion.
Reprogramming using miRNAs that are specifically expressed in embryonic pluripotent stem cells (ESCs) can enhance the reprogramming efficiency [292]. For example, the miR302/367 cluster is highly expressed in pluripotent cells, but not in differentiated cells, and its promoter is transcriptionally regulated by the reprogramming factors Oct4 and Sox2 [293]. This cluster is functionally involved in regulation of the cell cycle and maintenance of pluripotency. Overexpression of the miRNA cluster miR302/367 can promote the reprogramming of somatic cells [294]. In combination with the reprogramming factors, a higher efficiency can be achieved [292]. Although RNA-based reprogramming methods show higher efficiency compared to Sendai virus and episomal methods, the reliability is significantly lower [295]. Non-integrating delivery methods provide iPSCs that are more applicable for clinical disease modeling. Besides the integrating and non-integrating delivery systems, DNA-free approaches with transgene free reprogramming have been established. Small compounds or recombinant reprogramming factors were used ( Figure 2) [296,297]. For example, the histone deacetylase inhibitor valproic acid improves the reprogramming efficiency [298,299]. The efficient synthesis of large amounts of purified native recombinant proteins and the permeabilization of the plasma membranes are crucial for this reprogramming method [300]. More recently, the CRISPR-dCas9-based synergistic activation mediator (SAM) system has been developed and applied for reprogramming [301,302]. This system is based on a fusion protein of the enzymatic inactive form of Cas9 (dCas9) and a transcription activator domain forming an artificial transcription factor which, in combination with specific guide RNAs, is able to activate the transcription of endogenous genes with minimal off-target activity. Weltner et al. successfully used this system for the expression of different reprogramming factors to generate iPSCs [302].
In summary, different integrating and non-integrating approaches have been developed for reprogramming different cell types into iPSCs to improve the efficiency and to reduce the risk of further genomic alterations (see Figure 2).

Genetic Modification of Induced Pluripotent Stem Cell Lines
Besides the generation of human iPSCs from the primary cells of mutation carriers by direct reprogramming [278,281], specific genetic mutations can also be inserted using genome editing techniques like clustered regularly interspaced short palindromic repeats associated protein 9 (CRISPR-Cas9) [303], CPF1 [304], or transcription activator-like effector nucleases (TALENs) [305,306]. In addition to genome editing approaches, iPSCs or the differentiated cardiomyocytes can be genetically modified by overexpressing specific mutant proteins [307,308] or by decreasing the expression of specific mutant proteins, e.g., by RNA interference [309].
Using patient-derived iPSCs, it is sometimes challenging to correlate directly functional effects in vitro with the specific genetic variants because the genetic and epigenetic background of the cells is largely unknown [310]. In contrast to patient-derived iPSCs, which carry the sum of all genetic sequence variants of the affected patients, genome edited iPSC lines carry specifically inserted mutations. Therefore, the effects of particular mutations can be directly compared with their corresponding isogenic wild-type controls in genome-edited iPSCs.
Genome editing techniques like CRISPR-Cas9 are based on endonuclease activity, which insert double-strand breaks (DSBs) into the DNA double helix at specific sites. Different endogenous cellular repair mechanisms like non-homologous end joining (NHEJ) or homology directed repair (HDR) are used for the repair of these DSBs. However, NHEJ is an imprecise process, which might lead to the insertion, deletion, or substitution of nucleotides [311]. Indel variants frequently cause frameshifts, and consequently, premature termination codons (PTCs). PTCs are recognized by nonsense mediated RNA decay (NMD) degrading the mutant mRNA. Therefore, DSBs can be efficiently used to generate knock-out models. In contrast, HDR uses DNA template molecules for the specific repair of the DSBs. In combination with suitable donor molecules, e.g., single-stranded oligonucleotides or double-stranded DNA templates like PCR products or plasmids, HDR can be used to insert specific point mutations [312], small peptide-encoding tags [313], or even larger fluorescence proteins at specific positions [314][315][316]. Unfortunately, the ratio of HDR to NHEJ is low, limiting the efficiency of knock-in strategies [317]. Therefore, different approaches for inhibiting NHEJ or promoting HDR have been developed (for reviews, see References [317][318][319]). The delivery of donor template molecules in close proximity to the DSBs by coupling Cas9 with the donor molecule might be a promising strategy [320][321][322]. An alternative are dCas9-related base pair editors [323][324][325], which can be used to exchange relevant nucleotides at specific positions.

Differentiation of Human Induced Pluripotent Stem Cells into Cardiomyocytes
The human adult heart is a post-mitotic organ with a very limited capacity for regeneration [326]. Beside the murine, atrial cardiomyocytes-related HL-1 cell line [327], no further contracting human cardiomyocytes cell lines are therefore currently available. Because of ethical and technical issues, the isolation of primary human cardiomyocytes from human surgical material and their long-time culture is in most cases impossible. Primary cardiomyocytes isolated from rodent hearts have characteristic differences like a different electrophysiology in comparison to the human ones. Therefore, cardiomyocytes derived from human ESCs or iPSCs are the predominant human cell resource [328,329].
Originally, Zhang et al. described the differentiation of cardiomyocytes from human iPSCs [330]. Comparable to ESCs, human iPSCs form embryonic bodies in suspension that can be further differentiated into cardiomyocytes [330][331][332][333][334]. However, the efficiency of this process was limited. In addition, monolayers of iPSC-derived cardiomyocytes can be generated [335,336]. In vivo, cardiogenesis is a complex cellular and molecular process where different transcription factors, growth factors, and miRNAs are time dependently expressed and regulated [337][338][339][340][341]. Driven by discoveries from development biology, it has been recognized that different recombinant growth factors, e.g., BMP4, can also be used to increase the efficiency of in vitro differentiation into cardiomyocytes [342][343][344]. In addition, modulation of the Wnt pathway by small molecules, e.g., CHIR99021 and IWP2, efficiently increases the differentiation into cardiomyocytes about 90% [344,345]. Furthermore, metabolic selection by glucose depletion, in combination with lactate supplementation, can be applied for further accumulation of cardiomyocytes [346,347]. Recently, Zhao et al. developed a method for the differentiation and generation of heteropolar cardiac tissue with atrial and ventricular ends [348]. Talkhabi et al. has previously reviewed the differentiation of iPSCs into cardiomyocytes in detail [349].

Methods for the Functional Analysis of Cardiomyocytes Derived from Induced Pluripotent Stem Cells
Besides general histochemical or molecular methods, e.g., RNA-Seq or proteomics, specific techniques for the functional in vitro analysis of the electrophysiological and contractile properties of iPSC-derived cardiomyocytes are frequently used. Patch clamping and multiple electron arrays (MEAs) are frequently used for the electrophysiological analysis of iPSC-derived cardiomyocyte monolayers [350,351]. The application of Ca 2+ specific fluorescence dyes, e.g., Indo1 or Fura-2, allows for the microscopic analysis of Ca 2+ transients [352][353][354]. Additionally, voltage-sensitive fluorescence dyes like di-4-ANEPPS can be used for the analysis of the electrophysiological properties [355]. For the analysis of the contractile properties of iPSC-derived cardiomyocytes, microscopic techniques like traction force measurements have also been used [356]. Atomic force microscopy can also be applied for measuring the contraction forces of iPSC-derived cardiomyocytes [357,358]. Feaster and coworkers developed a method to culture iPSC-derived cardiomyocytes on Matrigel mattresses, allowing for the contractility measurement by cell shortening [359].

Overview about Existing iPSC Lines Carrying Cardiomyopathy Associated Mutations
In 2010, Carvajal-Vergara and co-workers published a landmark paper about the generation of an iPSC line carrying the heterozygous mutation PTPN11-p.T468M [360]. Mutations in PTPN11 cause the Leopard syndrome [361,362], which is frequently associated with severe HCM [363]. Interestingly, these iPSC-derived cardiomyocytes were larger and presented an abnormal, nuclear localization of NFATc4 [360]. Members of the NFAT family are involved in the calcineurin-NFAT signaling regulating hypertrophy [364]. Since this original report, about 70 different iPSC lines carrying cardiomyopathy-associated mutations in several different genes have been generated ( Table 2). The majority of these mutant iPSC lines have been used for phenotypic modeling of genetic cardiomyopathies using electrophysiological and/or contraction measurements ( Table 2). Besides modeling genetic cardiomyopathies, iPSC-derived cardiomyocytes were also used for the modeling of non-genetic causes of cardiomyopathies, e.g., doxorubicin cardiotoxicity [263,365], hypoxia [366], peripartum [367], or diabetic cardiomyopathy [368][369][370][371], or even infection with Trypanosoma cruzi [372] or with coxsackievirus B3 [373].
In the beginning, iPSC lines generated from healthy probands were frequently used as controls for experiments. However, because different iPSC lines have a variable genetic background, this approach has limitations. Since the development of efficient genome editing technologies like CRISPR-Cas9 or TALENs [303], it is common to generate isogenic control lines [374]. Interestingly, the reverse approach by inserting specific mutations in iPSCs from healthy control persons is also sometimes used [375]. In some cases, the rationale of these studies is the functional characterization of specific cardiomyopathy-associated mutations, which might contribute to a pathogenicity classification. In addition, iPSC-derived cardiomyocytes were used for the development of therapeutic strategies, e.g., genome editing. An interesting application of iPSC-derived cardiomyocytes is the testing of specific gene therapeutic concepts [376]. For example, Gramlich et al. applied antisense-mediated exon skipping in iPSC-derived cardiomyocytes with a truncating TTN (TTNtv) mutation for restoring the expression of titin [377]. However, at present, it appears that some of the TTNtv do not lead to premature translation termination in failing human hearts [378]. Thus, iPSCs might therefore be useful in future to check and modulate possible read-throughs of TTNtv mutations as well. Similarly, Kyrychenko et al. used CRISPR-Cas9 to delete whole exons within the DMD gene to correct the reading frame [379]. Of note, this strategy restores contractility in the iPSC-derived cardiomyocytes [379]. Hopefully, the combination of iPSC-derived cardiomyocytes with adequate modern genetic engineering tools will contribute in future to the development of therapeutic options in the context of personalized medicine.

Limitations of Human Induced Pluripotent Stem-Cell-Derived Cardiomyocytes
Besides cardiomyocytes, the human adult heart consists of several different cell types like fibroblasts, endothelial cells, leukocytes, pericytes, and smooth muscle cells. It has been estimated that the proportion of cardiomyocytes in myocardial tissue is around 25-35%, indicating that the majority of the cardiac cells are non-cardiomyocytes [460]. However, the molecular and cellular interactions and interferences between the different cardiac cell types are poorly understood. In particular, under pathological conditions like inflammation or fibrosis, the cellular composition of the heart of cardiomyopathy patients can vary and might change over time. Therefore, it is in general challenging to model the complex cellular and molecular networks using iPSC-derived cardiomyocytes in vitro, although the artificial generation of cardiac tissue has been impressively improved during the last few years [461][462][463][464][465]. Besides these general limitations, iPSCs and iPSC-derived cardiomyocytes have some specific limitations, which are outlined in the following paragraphs.

Genomic Instability
Genomic instability of iPSCs can be a fundamental problem limiting the clinical application of iPSC-derived cells because of safety concerns [466]. Mayshar et al. showed that a significant portion of iPSC and ESC lines carry full or partial chromosomal aberrations [467]. However, even for in vitro analysis, genomic instability could be an important issue, especially in the context of modeling genetic diseases like cardiomyopathies. Therefore, novel iPSC lines should be genetically characterized in general. Karyotype analysis using Giemsa staining or comparative genomic hybridization arrays can be used to detect larger chromosomal abnormalities, while next generation sequencing assays can be applied for genetic analysis at the single nucleotide level.
Three different mechanisms contribute to the mutagenesis in iPSCs: besides the existence of genetic variants in the parental somatic donor cells, mutations can be introduced during reprogramming procedure or during the long-time culture of iPSCs [468]. Of note, mutations might accumulate in iPSCs over the culturing time [469]. Therefore, it is advisable to use early passages and to repeat analyses for genetic stability from time to time.

Heterogeneity of iPSC-Derived Cardiomyocytes
Although cardiac differentiation protocols for iPSCs have been improved significantly over recent years [345,470], it should be kept in mind that iPSC-derived cardiomyocytes are still a heterogeneous cell population. Especially for bulk down-stream applications like proteomics, genomics, or metabolomics, this might have a significant impact.

Cellular, Molecular, and Functional Differences of Adult Ventricular Cardiomyocytes and iPSC-Derived Cardiomyocytes
Even though human iPSC-derived cardiomyocytes are contractile cell types, there are important cellular, molecular, and functional differences compared to adult cardiomyocytes. The most obvious differences are the size and shape of iPSC-derived cardiomyocytes. Adult ventricular cardiomyocytes have a typical rod-like shape and are relatively large cells with lengths of about 100 µm and diameters of 10-25 µm [471]. In contrast, iPSC-derived cardiomyocytes are much smaller [472] and are morphologically heterogeneous. The geometry of iPSC-derived cardiomyocytes ranges from round to rectangular or polygonal shapes [473,474]. In adult ventricular cardiomyocytes, the sarcomeric structure is highly organized and the Z-bands are in parallel with the intercalated disc. On the contrary, iPSC-derived cardiomyocytes have a more irregular and amorphous sarcomeric organization with diverse orientations [462,475]. In human myocardial tissue, the closed-ends of the plasma membranes connect the cardiomyocytes longitudinally and these ends of the cardiomyocytes "cylinders" are called intercalated discs. Multi-protein complexes mediate the cell-cell interactions at the intercalated discs and are subdivided into desmosomes, adherens, and gap junctions [476]. Although desmosomes and adherens junctions are also formed in iPSC-derived cardiomyocytes [472,477], the cellular distribution of these cell-cell junctions are not conserved [478,479]. Another important difference is the number of nuclei. Whereas a significant number of the human cardiomyocytes in vivo are binuclear cells [480], iPSC-derived cardiomyocytes are mononuclear cells [481]. In addition, there are significant differences in contraction and electrical properties of iPSC-derived cardiomyocytes in comparison to adult ones [474]. In summary, the structural and functional properties of iPSC-derived cardiomyocytes are more similar to fetal cardiomyocytes than to adult cardiomyocytes [482]. To overcome these limitations, different natural engineering approaches were established to drive cardiomyocytes maturation. One method is to stimulate the cardiomyocytes with electrical or mechanical impulses [483]. The composition of the extracellular matrix can also affect the interaction of the CMs, therefore influencing the cellular behavior [484,485]. Another promising approach is the co-culture of iPSC-derived cardiomyocytes with non-cardiomyocytes, enabling a more likely cardiac environment with different cellular interactions [486]. Physical, chemical, electrical, and genetic factors are being tested as stimuli for further maturation [487]. However, maturation of iPSC-derived cardiomyocytes is incompletely understood at the molecular level and more studies are needed in future.

Testing of Gene Therapies Using iPSC-Derived Cardiomyocytes as in Vitro Models
An interesting research topic is the development of personalized therapeutic strategies for genetic cardiomyopathies in vitro. Beyond the opportunities that reprogramming technologies offer for therapeutic myocardial regeneration, iPSC-derived cardiomyocytes are a promising platform to develop and test different gene therapies for genetic, non-ischemic cardiomyopathies. In general, the pathomechanisms of inherited cardiomyopathies can be classified into loss of function (LOF) or gain of function (GOF) mechanisms. LOF can be caused by (haplo)insufficiency or by the expression of non-functional proteins. For example, several HCM-associated MYBPC3 mutations cause haploinsufficiency [415,488]. GOF is caused by mutant and toxic proteins such as those shown for several DES missense mutations [489,490].
Genome editing using CRISPR-Cas9 or TALENs has been applied to repair different mutations in iPSC-derived cardiomyocytes. After the insertion of DSBs, iPSCs repair these DSBs using NHEJ or HDR. Template molecules like oligonucleotides, plasmids, PCR products, or even the second chromosome might be used for HDR. Recently, Ma et al. even applied CRISPR-Cas9 for the repair of a pathogenic MYBPC3 mutation in human pre-implanted embryos [491]. However, because the efficiency of HDR is low, the direct repair of mutations in iPSCs via genome editing is challenging. Therefore, single iPSC clones were frequently generated in vitro and the direct translational transfer of this method is consequently limited. A second therapeutic strategy is exon skipping [492]. Exon skipping corrects the open reading frame (ORF) of an affected gene via skipping of the mutant or multiple exons and restores the expression of the truncated, but still functional, protein. For this approach, specific antisense oligonucleotides binding to the mutant exons can be used [493]. Besides its application in iPSC-derived cardiomyocytes carrying mutations in DMD [494] or TTN [377], antisense-mediated exon skipping was also directly applied in human patients with Duchenne's muscular dystrophy [495]. Recently, Eric Olson's group applied CRISPR-CPF1 or -Cas9-mediated genome editing for exon skipping in iPSC-derived cardiomyocytes [379,496,497]. Prondzynski et al. applied trans-splicing and total gene replacement for the artificial increased expression of MYBPC3 in iPSC-derived cardiomyocytes carrying a heterozygous frameshift mutation in MYBPC3 [419]. The authors used adeno-associated viruses (serotype 2/9, AAV2/9) for the transduction of iPSC-derived cardiomyocytes with 5 -and 3 -pre-trans-splicing molecules and the total cDNA of MYBPC3. However, the efficiency of the trans-splicing approach was low. In contrast, the total gene replacement strategy increased the MYBPC3 expression to over 80% in comparison with wild-type controls and was able to prevent cellular hypertrophy [419].
The combination of the iPSC-derived cardiomyocytes platform with gene therapy tools is a promising therapeutic approach enabling pre-clinical demonstration of proof-of-principle for inherited cardiomyopathies.

Summary
Human iPSC-derived cardiomyocytes represent the only available human cellular model for the direct functional analysis of specific genetic cardiomyopathies and might therefore overcome the limitation of species differences. Impressive progress in the reprogramming and differentiation procedure during the last decade allows, in combination with novel genome editing techniques like CRISPR-Cas9, for the development of defined/patient specific cardiomyocyte models including generation of their isogenic control lines. In summary, iPSC-derived cardiomyocytes have been used for: (a) the characterization of genetic variants of unknown significance, which might be helpful for genetic counseling [375]; (b) analyses of the molecular pathomechanisms [415]; and (c) the development of specific therapies [377,497].
However, the cellular and molecular crosstalk between inflammatory cells, fibroblasts, myoblasts, and cardiomyocytes is difficult to model using iPSC-derived cardiomyocytes. Therefore, in our opinion, iPSC-derived cardiomyocytes should also be combined with animal models or with ex vivo investigations of explanted human myocardial tissue whenever possible to overcome the specific limitations of iPSC-derived cardiomyocytes.
Interestingly, for some genes like DMD, PKP2, MYBPC3, or MYH7, several different iPSC lines have been generated. In contrast, for rare cardiomyopathy genes, e.g., TMEM43, no iPSC lines have been developed yet. The genetic analysis in the past few decades has revealed a high heterogeneity of inherited, non-ischemic cardiomyopathies. In our view, it is therefore important to generate further novel iPSC lines also carrying mutations in rare cardiomyopathy genes to compare the molecular differences and commonalities leading to non-ischemic cardiomyopathies. Hopefully, iPSC-derived cardiomyocytes will contribute to unravelling the pathomechanisms of genetic cardiomyopathies and will help in efficient drug development in future.

Conflicts of Interest:
The authors declare no conflict of interest.