Modelling sarcomeric cardiomyopathies with human cardiomyocytes derived from induced pluripotent stem cells

Abstract

Cardiomyocytes derived from human induced pluripotent stem cells (iPSCs) provide a unique opportunity to understand the pathophysiological effects of genetic cardiomyopathy mutations. In particular, these cells hold the potential to unmask the effects of mutations on contractile behaviour in vitro, providing new insights into genotype–phenotype relationships. With this goal in mind, several groups have established iPSC lines that contain sarcomeric gene mutations linked to cardiomyopathy in patient populations. Their studies have employed diverse systems and methods for performing mechanical measurements of contractility, ranging from single cell techniques to multicellular tissue-like constructs. Here, we review published results to date within the growing field of iPSC-based sarcomeric cardiomyopathy disease models. We devote special attention to the methods of mechanical characterization selected in each case, and how these relate to the paradigms of classical muscle mechanics. An appreciation of these somewhat subtle paradigms can inform efforts to compare the results of different studies and possibly reconcile discrepancies. Although more work remains to be done to improve and possibly standardize methods for producing, maturing, and mechanically interrogating iPSC-derived cardiomyocytes, the initial results indicate that this approach to modelling cardiomyopathies will continue to provide critical insights into these devastating diseases.

Graphical Abstract

In this review, current progress in the field of modelling sarcomeric cardiomyopathies using induced pluripotent stem cell-derived cardiomyocytes is described. In many cases, these models have been developed for the purpose of characterizing the contractile phenotype that underlies disease. The various methods that have been employed to measure contractile behaviour of iPSC-derived cardiomyocytes are discussed and an overview of the phenotypic data that are accumulating from these studies is provided.

Introduction

Familial hypertrophic cardiomyopathy (HCM) is a lethal inherited autosomal dominant disorder that manifests as abnormal thickening of the heart wall, cardiomyocyte remodelling, diastolic dysfunction, fibre disarray, and myocardial fibrosis (Tardiff, 2005; Braunwald & Maron, 2012; Marian & Braunwald, 2017). Affecting between 1 in 250 and 1 in 500 people in the general population of the USA, HCM remains the most common monogenic cardiovascular disease (Maron et al. 2012). Since the discovery of the first missense mutation clearly linked to HCM in the β-cardiac myosin heavy chain (MYH7) approximately 17 years ago by the Seidman team (Geisterfer-Lowrance et al. 1990), a relentless tide of genetic studies has established genetic HCM as primarily a disease of the cardiac sarcomere.

Of the 1500-plus variants linked to HCM, over 90% occur in genes encoding thick and thin filament proteins, such as MYH7, cardiac myosin binding protein C (MYBPC3), cardiac troponin T (TNNT2), and α-tropomyosin (TPM1) (Tardiff, 2005; Marian & Braunwald, 2017). The advent of next generation sequencing and whole exome sequencing has led to the identification of new candidate HCM variants in sarcomeric and sarcomere-related genes such as Z-disc elements (Alfares et al. 2015; Walsh et al. 2017), reaffirming the conclusion that HCM is primarily a monogenic disease of the sarcomere. Despite significant advances in cardiac pathophysiology and treatment of disorders such as arrhythmia and heart failure, current medical therapy for HCM has been confined thus far to targeting symptomatology rather than underlying causes (Marian & Braunwald, 2017; Czepluch et al. 2018). The development of precise, effective therapies for HCM is likely to be delayed until a clear understanding of HCM pathomechanisms emerges, linking the proximal cause to intermediate and distal consequences.

At the organ level, the various HCM-related sarcomeric mutations converge on a constellation of remodelling events. These include asymmetric left ventricular hypertrophy, interstitial fibrosis, and myocardial fibre disarray, concomitant with abnormalities of both relaxation and contraction of the left ventricle (Marian & Braunwald, 2017). Yet, a definitive consensus as to the pathological effects of HCM-associated missense mutations at the level of the sarcomere has not completely emerged (Van Der Velden et al. 2015). Animal models of HCM have led to some critical insights (Maass & Leinwand, 2000; Sheikh & Chen, 2007). For instance, the first mouse model of HCM revealed that α-myosin (MYH6) R403G heterozygotes experienced altered cardiac function that preceded histopathological remodelling (Geisterfer-Lowrance et al. 1996). This reinforced the intuitive hypothesis that mutated sarcomeric proteins directly alter contractile function to drive long-term tissue pathologies.

Since then, numerous sarcomeric protein mutations have been described and investigated in mice, protein expression systems, and biopsies of patient myocardium. The prevailing but not-yet-universal view is that HCM-linked sarcomeric mutations tend to increase myofilament calcium sensitivity, increase the crossbridge duty cycle, and increase energy cost of tension generation (Redwood et al. 1999; Ashrafian et al. 2003; Crilley et al. 2003; Ferrantini et al. 2009; Van Dijk et al. 2011; Fraysse et al. 2012; Bai et al. 2013; Sequeira et al. 2013; Van Der Velden et al. 2013; Coppini et al. 2014; Lopes & Elliott, 2014; Witjas-Paalberends et al. 2014; Spudich et al. 2016; Cohn et al. 2018; Ujfalusi et al. 2018).

iPSC-derived cardiomyocytes

A central motivation for pursuing human induced pluripotent stem cell (iPSC) models of sarcomeric HCM is the opportunity to confirm and expand upon studies made with patient myectomy samples, which represent advanced states of disease. Experiments that use iPSC-derived cardiomyocytes (iPSC-CMs) approximate the very earliest stages of cardiac development and should afford glimpses of contractile alterations that appear first in the disease, unobscured by systemic, decades-old compensatory responses (Buikema & Wu, 2017). Peering into this early window may also reveal how these alterations lead to later manifestations of HCM. The ability to perform these studies in human cells is a significant motivation for the use of iPSC-CMs.

iPSC technology itself, as a new avenue for modelling human disease and associated genetic insults, was made possible by the discovery of the Yamanaka factors (OCT4, SOX2, LIN28, NANOG) that lead human adult somatic cells such as fibroblasts or T-cells to revert towards a pluripotent state (Takahashi et al. 2007). Within 4 years of this discovery, iPSC-CMs could be reproducibly differentiated (Mummery et al. 2012). These methods currently include inhibition of pathways such as WNT in iPSC monolayers (Lian et al. 2013) and the activation of BMP in embryoid bodies (EB) (Zhang et al. 2009), both of which drive iPSCs to mimic development of the heart by forming definitive mesoderm and later cardiac progenitor cells (CPCs). CPCs develop over time into a mixture of atrial and ventricular cardiomyocytes. While these methods can still be considered in their infancy in terms of reliably pure cardiomyocyte differentiation, consistent cardiomyocyte differentiation can be achieved through careful protocol optimization, cell line selection, and the use of cardiomyocyte purification methods.

The characteristics of iPSC-CMs, particularly relative to adult human cardiomyocytes, must be considered. On the whole, they appear to be similar to isolated fetal cardiomyocytes (Bedada et al. 2016), typically having a circular shape, immature electrophysiology (e.g. lacking T-tubules and having a higher resting membrane potential), glycolytic rather than oxidative metabolism, and disorganized sarcomeres. Their sarcomeres contain functional sarcomeric proteins but not necessarily their full complement (e.g. myomesin) or a fully adult isoform expression profile (e.g. cTnI rather than ssTnI). Such immature characteristics result in crucial functional differences from adult cardiomyocytes, notably differences in conduction velocity, adrenergic response, inotropic response, and excitation–contraction coupling (Schwan & Campbell, 2015; Jeziorowska et al. 2017).

In spite of these potential limitations, iPSC-CMs have already been leveraged in the study of inherited cardiomyopathies. Lan et al. (2013) were the first to use iPSC technology to model HCM. In this landmark study, they generated iPSC-CMs from a family of 10 that included genotype-negative and genotype-positive individuals with the HCM-associated mutation R663H in the β-myosin heavy chain gene (MYH7). Comprehensive clinical testing established that the mutant gene segregated with clinical symptoms such as left ventricular hypertrophy, premature ventricular contractions, dyspnoea and exertional angina. Even teenage carriers showed hyperdynamic function via echocardiography. iPSCs were generated using the standard EB protocols common at the time, yielding spontaneously contracting cardiomyocytes of about 60% purity. iPSC-CMs with the MYH7 R663H mutation recapitulated a surprising array of HCM-like phenotypes at 40 days post-differentiation. Mutant iPSC-CMs were larger, more frequently multinucleated, had increased myofibrillar content, and had increased sarcomere disorganization compared with control cardiomyocytes. HCM iPSC-CMs also displayed upregulation of GATA4, TNNT2, MYL2 and MYH7 when compared with controls. Electrophysiological studies indicated that after day 30, HCM cardiomyocytes exhibited behaviours and beat patterns that were highly reminiscent of arrhythmic waveforms. Lan and co-workers also analysed light microscopy videos of contracting HCM iPSC-CMs, finding that mutant cells beat at irregular frequencies and may have been hypercontractile in comparison to control cells. Cellular calcium cycling was also abnormal, with mutant iPSC-CMs having an increased level of diastolic calcium along with decreased release of SR calcium, a finding which preceded cellular hypertrophy and remodelling. Significantly, administration of calcium channel blockers verapamil and diltiazem seemed to prevent HCM phenotypes of hypertrophy, abnormal calcium handling, and arrhythmias.

Altogether, Dr Wu’s team successfully harnessed iPSC-CM technology to create a plausible explanation for the pathogenicity of MYH7 R663H by connecting clinically relevant diagnostic findings with autonomous cellular events, and even suggested putative patient treatments. By our own survey of the literature (Table 1), this ground-breaking paper has been followed by 24 additional studies from a number of groups using iPSC-CMs to model sarcomeric disease. The studied sarcomeric mutations are associated with both HCM and dilated cardiomyopathy (DCM), encompassing variants in MYH7, MYBPC3, TNNT2, TTN, MYL3 and TPM1. Mutations in non-sarcomeric cardiomyopathy genes, such those involved with ion channels, metabolic function, and growth pathways, are outside the scope of this review, but these too have been profitably studied in iPSC systems (Moretti et al. 2010; Cashman et al. 2016; Hinson et al. 2016; Joanne et al. 2016; Josowitz et al. 2016; Judge et al. 2017; Lee et al. 2017; Liu et al. 2017; Lu et al. 2017; Ben Jehuda et al. 2018; Hallas et al. 2018; Moreau et al. 2018; Zhan et al. 2018; Li et al. 2018b).

Table 1.

Survey of iPSC-CM studies of HCM and DCM sarcomeric mutations

Reference	Disease	Sarcomeric mutation (gene/protein)	Cell line type	Age* (days)	Contractility characterization method	Contractile phenotype
Mosqueira et al. (2018)	HCM	MYH7 c.C9123T/MHC-β R453C	Isogenic	24–30	Fibrin EHT; auxotonic optical measurement	Reduced force; increased time to peak activation; no change in relaxation time
Chang et al. (2018)	DCM HCM	MYBPC3/cMYBPC R493x; MYH7/MHC-β R723C; MYH7/MHC-β R403Q; MYH7/MHC-β R663H; TNNT2/cTnT R173W; TTN cN22577fs+/-/TTN;TTN c.67745delT/TTN;	Patient	30	NA	NA
Ma et al. (2018)	HCM (VUS)	MYL c.170C>A/MYL3 A57D	Patient and isogenic	45–50	CM; optical recording	No change in contraction or relaxation velocity
Prajapati et al. (2018)	HCM	MYBPC3/MYPBC Gln1061X; TPM1/TPM-α D175N	Patient	48–85	NA	NA
Chopra et al. (2018)	DCM	TTN/TTNtv A band	Patient	30	CM; TFM	Decreased diastolic stress
Wang et al. (2018)	HCM	TNNT2 c.236T>A/cTnT I79N	Isogenic	35	CM in Matrigel mattress; edge detection	Increased cell shortening; no change in time to peak shortening; increase in time to 90% relaxation
Viswanathan et al. (2018)	HCM	MYBPC3 del25bp/cMYBPC MYBPC3/cMYPBC D389V	Patient	NR	NA	NA
Wang et al. (2017)	HCM	TNNT2 c.236T>A/cTnT I79N	Isogenic	NR	NA	NA
Streckfuss-Bömeke et al. (2017)	DCM	RBM20/RBM20 S635A	Patient	60	Circular EHMs; direct isometric measurement	Decreased active and passive force
Karakikes et al. (2017)	DCM	TNNT2 c.517C>T/cTnT R173W	Patient and isogenic	30	NA	NA
Prondzynski et al. (2017)	HCM	MYBPC3 c.1358-1359insC/cMYPBC	Patient	21	NA	NA
Wyles et al. (2016b)	DCM	RBM20/RBM20 R636S	Patient	20	NA	NA
Davis et al. (2016)	HCM	MYBPC3/cMYBPC R943ter	Patient	30	CM; TFM	Increased tension-time index; increased peak stress
Broughton et al. (2016)	DCM	TNNT2/cTnT R173W	Patient	5–14	CM; DIC Line scan	Decreased cell shortening
Pioner et al. (2016)	FCM	MYH7/MHC-β E848G	Patient	80–100	Isolated myofibril measurement	Increased calcium sensitivity; decreased maximum isometric tension; increased maximal kACT; faster submaximal slow kREL
Ojala et al. (2016)	HCM	MYBPC3/MYPBC Gln1061X TPM1/TPM-α D175N	Patient	22–72	NA	NA
Wyles et al. (2016a)	DCM	RBM20/RBM20 R636S	Patient	20	NA	NA
Birket et al. (2015)	HCM	MYBPC3 c.2372dupG/cMYBPC	Patient	21–25	CM; TFM	Decreased traction stress
Wu et al. (2015)	DCM	TNNT2/cTnT R173W	Patient	12–60	CM; TFM	Lower peak tangential stress; lower maximum contraction rate
Hinson et al. (2015)	DCM	Various TTN/TTN A/I band mutations	Isogenic	30–40	Cardiac microtissue; optical auxotonic measurement	Decreased peak twitch force
Gramlich et al. (2015)	DCM	TTN/TTN Ser14450fsX4	Patient	20–30	NA	NA
Tanaka et al. (2014)	HCM	MYBPC3/cMYBPC Gly999-Glb1004del	Patient	30–90	CM; motion vector analysis	Increased variability in contractile direction
Han et al. (2014)	HCM	MYH7/MHC-β Arg442Gly	Patient	22–30	NA	NA
Lan et al. (2013)	HCM	MYH7/MHC-β Arg663His	Patient	20–40	CM; video analysis of contractile pixel concentration per cell	Increased normalized contractile motion
Sun et al. (2012)	DCM	TNNT2/cTnT R173W	Patient	18–48	CM; AFM	Decreased contraction force

^*Age in days is counted from the start of differentiation protocol; where reported, this was directly taken in methods or results sections. In some cases, the age was calculated by a combination of differentiation protocol days and then days post-differentiation until cardiomyocytes were assayed. In some cases, various assays and measurements were conducted at different days, so this was accounted for as well.

cMYBPC, cardiac myosin binding protein C; cTnT, cardiac muscle troponin T; MHC, myosin heavy chain; NA, not assayed; NR, not reported; TPM, tropomyosin; TTN, titin; TTNtv, titin-truncating variants.

Although initial studies used patient-derived cell lines and single-cell functional assays, work in this area has made use of increasingly sophisticated techniques to enhance the power of the iPSC-CM disease modelling approach. A growing number of investigators have taken advantage of developments in genetic engineering techniques, most notably CRISPR/Cas9, to introduce cardiomyopathy mutations into otherwise healthy cell lines or to correct cell lines with mutations. This has the advantage of providing isogenic cell lines that accentuate mutation-mediated phenotypes. Assays of cardiomyocyte function have also expanded in variety and scope, with investigators using iPSC-CMs to look at metabolic behaviour, electrophysiology, calcium transients and contractility. Several of these studies have evaluated gene expression in sarcomeric cardiomyopathy iPSC-CMs, often via qPCR and occasionally via whole transcriptome sequencing (Han et al. 2014; Tanaka et al. 2014; Gramlich et al. 2015; Hinson et al. 2015; Wu et al. 2015; Prondzynski et al. 2017; Streckfuss-Bömeke et al. 2017; Mosqueira et al. 2018; Li et al. 2018b), reviewed recently (Eschenhagen & Carrier, 2018). In spite of gains in the areas of both genetic editing and assay development, much work remains to be done. For instance, it remains unclear whether the available mutants constitute a full representation of possible HCM disease pathways. Although building iPSC disease models of every clinically identified HCM mutation may not be feasible, a range of cell lines, carefully selected, should ultimately enable clear identification of the most common disease mechanisms. On the functional side, the field has not yet converged on an approach for maturing iPSC-CMs and measuring their contractile function that is sufficiently standardized to enable satisfactory comparisons between research groups and across mutation types.

Mechanics of iPSC-CMs

A central aspect of cardiomyocyte physiology is contractile function. It is undeniably important in the study of sarcomeric cardiomyopathies, as the affected genes are integrally associated with the production of contractile force. Herein lies an important challenge for the use of iPSC-CMs as disease models for cardiomyopathy.

Traditional cardiac muscle mechanics experiments typically make use of linear muscle preparations such as trabeculae or papillary muscles to investigate force production and dynamics (Sonnenblick, 1962; Janssen, 2010). In this form, cardiac muscle can be subjected to precisely defined loads and accurate force measurements. These preparations can give a rich variety of phenotypic information including isometric twitch dynamics (kinetics), force–length relationships, force–frequency relationships, load–power/force–velocity relationships, and passive mechanical properties.

iPSC-CMs do not naturally resemble native linear muscle preparations because they are produced in either 2D monolayers or embryoid bodies. The field has therefore been confronted with two basic options for pursuing contractile measurements in iPSC-CMs: (i) grow the cells on a static substrate and measure contractility under a single, defined loading condition or (ii) engineer cells into a construct that resembles a linear muscle with the possibility of more sophisticated mechanical characterization. Both approaches are evident in the array of methods developed so far for assessing iPSC-CM contractility (Fig. 1). Several of these techniques have already been used for the study of sarcomeric cardiomyopathies.

Figure 1. A survey of mechanical methods used to measure iPSC-CM contractile function from single cell to 3D engineered heart tissue.

Multiple iPSC-CM studies have made use of single-cell techniques for assaying contractility. These include atomic force microscopy (AFM; Chang et al. 2013; Pesl et al. 2016), traction force microscopy (TFM; Hazeltine et al. 2012), impedance-based measurement (Peters et al. 2015) and analysis of light microscopy videos (Liu & Padfield, 2012; Ahola et al. 2014; Huebsch et al. 2015). These approaches have proven adequate for comparisons of contractile function between mutant and non-mutant iPSC-CMs and have revealed significant differences. Such readouts are meaningful in the sense that they support the hypothesis that sarcomeric mutations assert their pathogenicity by way of altered contractile behaviour. At the same time, mechanistic interpretations (i.e. exactly how a mutation alters contraction) are difficult because they cannot be unambiguously related to the traditional muscle mechanics experiments that were designed to illuminate complex relationships between molecular function and macroscopic contractile behaviour. Nevertheless, technological advances are improving the quality and information density of single cell measurements. A particularly innovative approach has been developed by Ribeiro et al., who grew iPSC-CMs on appropriately stiff substrates with micropatterned templates that produced rod-like CMs (Ribeiro et al. 2015, 2017). These maturely shaped cells were then mechanically characterized by TFM, light microscopy, and even live sarcomere length measurements, which are quite similar to the unloaded shortening experiments traditionally made on isolated adult cardiomyocytes.

It is interesting to consider the loading experienced by iPSC-CMs in the various single-cell contraction assays. After analysing several methods, we estimate cardiomyocytes shorten by 5–10% as they contract against the load imposed by the substrate. These contractions may best be described as auxotonic (shortening against an elastic load) (Huntsman et al. 1979). In reality this is a simplification; a cell attached to a flat substrate is subject to heterogeneous loading, with the top surface of the cell being practically unrestrained while the bottom experiences a much larger resistive load. The ultimate impact of subcellular loading gradients on contractile measurements is not clear.

The main alternative to the single-cell approach is that of tissue engineered myocardium. Engineered heart tissues (EHTs) originate from the early work in embryonic chick (Eschenhagen et al. 1997) and neonatal rat (Zimmermann et al. 2000) models that were readily adapted for pluripotent stem cell-derived cardiomyocyte technology to make intact 3D muscle structures whose force and contractility can be measured through various direct or indirect methods (Schaaf et al. 2011; Tulloch et al. 2011). These include muscular thin films (Lind et al. 2017), hydrogel-based tissues (Schaaf et al. 2011; Boudou et al. 2012; Rupert & Coulombe, 2017; Tiburcy et al. 2017), re-seeded decellularized myocardial strips (Schwan et al. 2016) and even cardiac mini-chamber organoids (MacQueen et al. 2018; Li et al. 2018a) (Fig. 1). Some of these 3D methods provide measurements of isometric force, auxotonic force, and velocity of loaded shortening that are more relatable to the intact muscle preparations of traditional muscle physiology.

Regardless of whether iPSC-CM contractility is measured in a 2D or 3D format, single cell or multicellular tissue, the interpretation of measurements should be done carefully in the context of applied loads. These include the specific preload condition (e.g. stretch and sarcomere length), afterload condition (e.g. isometric, isotonic, auxotonic, unloaded), and substrate properties (e.g. substrate stiffness, particularly with regards to physiological stiffness). Each of these factors can profoundly impact both the intrinsic and apparent contractile properties of the cardiomyocyte.

Mechanical characterization of iPSC-CM models of sarcomeric cardiomyopathies

Given the diversity in preload, afterload and substrate characteristics represented by the various contractile measurement techniques (Fig. 1), looking for an emergent phenotypic pattern among HCM and DCM cell lines can only be done in an approximate way. Based on currently available data, iPSC-CM studies of sarcomeric cardiomyopathies do not fully confirm the results obtained in earlier model systems. A survey of the 25 studies involving HCM and DCM mutant iPSC cell lines to date is only partially consistent with the idea that excess contractility leads to pathological hypertrophy and insufficient contractility results in dilated remodelling as captured by the tension-time index (TTI) (Davis et al. 2016).

The contractility of iPSC-CMs expressing DCM mutations in TTN, TNNT2, and RBM20 has been measured and suggests a general finding of hypocontractility, albeit through different assays. Sun et al. (2012) looked at one of the first iPSC-CM models of DCM, a patient-derived cardiac troponin T (cTnT) R173W cell line. Using AFM, they found that the DCM mutant iPSC-CMs were severely hypocontractile compared to controls. Hinson et al. (2015) elegantly compared a variety of carefully picked titin A/I band mutants through a series of isogenic iPSC lines engineered into microtissues. These tissues grow around the ends of mildly flexible cantilevers, such that during contraction the cantilever ends are displaced by a small but measurable amount (~2.5% of total tissue length for wild-type tissue by our own analysis of literature data). This displacement is optically measured and analysed to obtain auxotonic contraction force. They determined that peak active twitch force was consistently decreased in DCM-associated TTN mutations.

The TNNT2 R173W mutation in patient-derived cell lines in particular has been examined by two different groups, using two different types of measurements. Broughton et al. (2016) used DIC line scanning to show that these diseased iPSC-CMs had decreased cell shortening, whereas Wu et al. (2015) characterized cardiomyocyte contractility using TFM and found a decreased peak tangential stress compared to controls. In this case, both groups found similar phenotypes for this mutation despite using assays with starkly different loading conditions.

Streckfuss-Bömeke et al. (2017) studied the RBM20 mutation S635A, which is not a direct sarcomere mutation but causes DCM by affecting the splicing of TTN. Engineered heart muscle rings measured directly under isometric conditions exhibited both decreased active and passive force. This is particularly interesting since normal cardiac function encompasses physiological stress generation and regulation in systole and diastole, but both have often not been measured together.

Pioner et al. (2016) looked at MYH7 E848G in patient-derived iPSC-CMs and were able to isolate myofibrils from these cardiomyocytes to perform a classic muscle physiology experiment in which solutions with varying calcium can be used to study tension generation and rate of tension generation and relaxation in these microscopic preparations. The disease diagnosis of this MYH7 E848G cohort was said to be familial cardiomyopathy without a clear claim of HCM or DCM, and the clinical phenotype encompassed systolic dysfunction, diastolic dysfunction and arrythmias, with no clear signs of hypertrophy. Given the systolic dysfunction, this FCM might be more suggestive of a DCM phenotype. Using myofibrils from these iPSC-CMs, they found that maximum isometric tension was decreased, along with an increase in maximal activation time constant and an increase in submaximal slow relaxation time constant. Although TTI was not directly measured in any of these studies, all DCM mutants showed changes in one or more contractile parameters that are consistent with decreased contractility and presumably decreased TTI.

iPSC-CM models of HCM are more numerous and have involved more extensive characterizations of contractility, though it should be pointed out that only seven of the 14 HCM reports we identified included some form of contractile characterization (Table 1). Studies have utilized a range of experimental systems, from single cells on various substrates to EHTs, covering mutations in the genes MYH7, MYL3, TNNT2, and MYBPC3. Looking at these results in aggregate, there seems to be an equal number of investigations reporting hypocontractility and hypercontractility.

Hypercontractility has been reported in some studies of mutations in MYH7, MYBPC3, and TNNT2. As mentioned above, Lan et al. (2013) found that patient-derived single iPSC-CMs with the MYH7 R663H mutation had increased normalized contractile pixel concentration per cell. Similarly, Davis et al. (2016) in their study looked at MYBPC3 R943ter iPSC-CMs using TFM, finding that these diseased iPSC-CMS had massively increased TTI through increased peak stress, increased time to peak contraction, increased time for relaxation. As the only investigation published characterizing contractility of a thin filament mutation in iPSC-CMs, Wang et al. studied the TNNT2 I79N mutant in single CMs seeded on the simple but effective Matrigel mattress format, which results in iPSC-CM structural maturation into rod-like cells (Wang et al. 2018). In this isogenic comparison with a wild-type cell line using edge detection, they found that the diseased iPSC-CMs exhibited increased cell shortening and increase in late relaxation time but no change in time to peak shortening.

Hypocontractility has been reported in cell lines with HCM mutations in MYH7 and MYBPC3. Recently, Mosqueira et al. (2018) built a series of isogenic MYH7 R453C cell lines and measured their contractile properties by optical tracking of auxotonic contractions (~10% shortening for a wild-type tissue by our own analysis of published data) against a cantilever post of known stiffness in fibrin EHTs. Strikingly, they found that their mutation-carrying EHTs had reduced force and longer time to peak contraction but no defects in relaxation. Birket et al. (2015) used TFM in a patient-derived MYBC3c.2372dupG IPSC-CM cell line to characterize contractility of iPSC-CMs grown under a cocktail of maturing factors (dexamethasone, insulin-like growth factor 1 (IGF1), and triiodothyronine (T3)); they found that these HCM mutants had decreased active traction stress generation.

Based on the studies detailed above, the HCM phenotype is not clearly hypercontractile, and indeed, there is conflicting evidence for even similar mutations in MYH7 and MYBPC3. At the sarcomere level itself, perhaps different mechanisms are at play for different mutations. For instance, haploinsufficiency may be a key player in truncating MYBPC3 mutants, whereas increased calcium sensitivity might be a key player in thin filament missense mutants.

Reconciling findings of hypocontractility and hypercontractility in HCM mutants, particularly in similar mutations in the same gene, might be possible by considering how different measurement and cellular formats affect mechanical behaviour. In particular, measurements where substantial shortening is allowed, including unloaded and auxotonic preparations, would vary significantly from isometric conditions (Janssen & de Tombe, 1997). For instance, the auxotonic twitches reported by Mosqueira et al. could look different under isometric conditions. Isometric conditions are well suited to revealing the numbers and speed of myosin binding, whereas auxotonic contractions include information about how myosin functions to generate shortening velocity against a given load.

In other cases, conflicting results may be attributed to differences in culture media and age. For instance, the studies by Birket et al. and Davis et al. present conflicting results for different truncating MYBPC3 mutations using TFM, but Davis et al. grew their iPSC-CMs for 30 days under standard (RPMI + B27 + insulin) medium conditions (Davis et al. 2016) whereas Birket et al. grew their iPSC-CMs for 21 days with 5 days of enriched medium (RPMI + B27 + insulin + T3 + IGF + dexamethasone) (Birket et al. 2015). If we assume that truncating MYBPC3 mutations should have a similar effect on cellular contractility and work through a similar putative mechanism such as haploinsufficiency, then it might be possible that the age and culture conditions played a central role in this discrepancy.

Conclusions

The field of cardiomyopathy modelling with iPSC-derived cardiomyocytes is still in its infancy. Challenges and opportunities abound in equal measure. Foremost among the challenges is the need to continue to improve cardiomyocyte maturation. It is not clear how far maturation must progress before iPSC-CMs can be accepted as a reasonable platform for investigating sarcomeric cardiomyopathy. We would argue that correct isoform expression of major sarcomeric constituents and realistic twitch dynamics (Chung et al. 2018) are sufficient goals (Schwan & Campbell, 2015; Bedada et al. 2016). Adult matrix substrate (Fong et al. 2016; Schwan et al. 2016) and electromechanical conditioning (Ronaldson-Bouchard et al. 2018) among other approaches are offering steady improvements in these aspects of maturation.

The ~13 studies reporting HCM and DCM iPSC-CM mechanical phenotypes do not form a clear consensus, but it would be premature to suppose that iPSC-CMs as a model system cannot provide decisive information. The picture emerging from the handful of contractile measurements in DCM cell lines at least is remarkably consistent. For HCM mutations, the diverging results may reflect a diversity of measurement approaches and experimental conditions. Accordingly, we anticipate that the picture will clear substantially as cell and tissue maturation are improved, culture conditions and duration are standardized, and as more comprehensive mechanical characterizations (including both isometric and isotonic-type loading) are performed. Characterization of the same mutant iPSC-CM cells using several of the available techniques could also provide important insights. If these steps fail to produce a consensus contractile phenotype, it may be necessary to explore other mechanical measurements that might reveal a common disease mechanism, such as metabolic efficiency or muscle power output.

Biography

Lorenzo Rakesh Sewanan is an MD-PhD candidate in the Department of Biomedical Engineering at Yale University. His current research interests lie at the intersection of cardiac tissue engineering, biomechanics, and mechanobiology, with particular application to understanding mechanisms underlying cardiomyopathy and heart failure. He hopes to pursue a career as an academic cardiologist and a physician–scientist with a continued emphasis on the role of mechanics and mechanobiology in cardiac health and disease.

Stuart Campbell is an Associate Professor of Biomedical Engineering at Yale University. His laboratory is focused on studying the biophysical mechanisms underlying inherited cardiomyopathies. His approach combines new computational models of muscle function and engineered heart tissue constructs made from human induced pluripotent stem cells. He holds degrees in bioengineering from Washington State University (BS, 2004) and the University of California San Diego (PhD, 2010).