Engineered heart tissue: Design considerations and the state of the art

Abstract

Originally developed more than 20 years ago, engineered heart tissue (EHT) has become an important tool in cardiovascular research for applications such as disease modeling and drug screening. Innovations in biomaterials, stem cell biology, and bioengineering, among other fields, have enabled EHT technologies to recapitulate many aspects of cardiac physiology and pathophysiology. While initial EHT designs were inspired by the isolated-trabecula culture system, current designs encompass a variety of formats, each of which have unique strengths and limitations. In this review, we describe the most common EHT formats, and then systematically evaluate each aspect of their design, emphasizing the rational selection of components for each application.

I. INTRODUCTION AND MOTIVATION

Since their initial development more than two decades ago, engineered heart tissues (EHTs) have become nearly ubiquitous in cardiovascular research. EHTs are commonly used for disease modeling and drug testing and have shown promise in therapeutic applications. A common goal for EHT platforms is to recapitulate cardiac physiology and pathophysiology while addressing shortcomings of cell culture and animal models. However, EHT platforms differ in their design and application. To motivate our comparison of different kinds of EHTs, we view the heart as a complex system comprised of various biological host factors, including a genetic background and a physical structure dictated by specialized cell types, performing mechanical work under tight neurohormonal and electrical regulation. Within this framework, we survey various EHT systems, sorting them by substrate shape and matrix composition. We then address specific aspects of their design which permit unique applications. We define engineered heart tissues to be biological tissue constructs composed of multiple layers of cardiomyocytes (CMs) and/or supporting cells comprising a three-dimensional (3D) structure with functional characteristics that mimic the behavior of in vivo cardiac tissue.

Early EHTs were inspired by isolated trabecula culture systems. The trabecular culture system, initially developed by Jansen and colleagues, enabled multi-day culture and functional characterization of isolated ventricular trabeculae, addressing shortcomings of primary cardiomyocyte culture models.^1,2 In this system, a trabecula (about 3.5 mm long, 0.3 mm wide, and 0.2 mm thick when excised from a rat right ventricle) is mounted between a force transducer and length controller within an organ bath. Experimental interventions like electrical stimulation, mechanical loading, and pharmacologic treatment can easily be achieved, unlike in an intact animal. Many seminal observations about cardiac physiology, such as the mechanism of length-dependent activation (LDA), the presence of calcium sparks, the role of protein phosphorylation in heart failure, and the effect of mechanical loading on contractility, have been made in trabecular systems.^3–6 Moreover, the ability to perform pharmacological studies and genetic manipulation while longitudinally measuring function was an attractive prospect and something not easily achievable in vivo or in neonatal cardiomyocyte cell lines.

For isolated trabecula culture systems to work successfully, there are several crucial design parameters: a sterile incubation chamber with continuous flow, appropriate cell culture media, nondestructive techniques of muscle preparation and attachment, and force-measurement equipment. By optimizing these parameters, medium-term (48 h) trabecular culture was achieved with physiologic force production and appropriate inotropic response without rundown. However, because trabeculae had to be surgically extracted from an experimental animal or human subject, throughput was very low and samples were exquisitely sensitive to physical handling. Damage to the preparation was difficult to avoid, particularly when attaching the trabecula to the force transducer with sutures or microtweezers.² Given these limitations, there was a need to create robust in vitro muscle preparations that could be produced systematically at higher throughput, while maintaining the physiological properties of intact muscle. The development of new techniques in the field of biomaterials, coinciding with basic discoveries about embryonic stem cell differentiation, ushered in the development of engineered heart tissue that could achieve much of what trabecula-culture systems achieved and at higher throughput.

Since the first report showing the feasibility of EHTs in 1997,⁷ the field has grown and diversified, but the goal of achieving a model system faithful to adult myocardium has persisted. To this end, EHTs should exhibit several important structural and functional features, including aligned and anisotropic muscle fibers with a high density of myofibrils and mitochondria; gene and protein expression profiles that are consistent with adult myocardium; a functional response to inotropic stimulation; demonstration of length-dependent activation, an analog of the Frank–Starling relationship; and mature calcium handling, as evidenced by a positive force-frequency and appropriate post-rest potentiation.

In this review, we survey different categories of EHTs, focusing on linear constructs that are directly inspired by trabecular culture, then address specific considerations regarding substrate composition, form factor (i.e., EHT geometry, dimensions, and substrate), cell population, media selection, mechano-electrical loading, and functional measurement technique. A schematic for how these design choices may be made is provided in Fig. 1. We appreciate that new biological techniques, such as organoid culture, and new engineering innovations, such as 3D bio-printing, have led to the development of EHTs that have different structures than the linear constructs we focus on here. We refer readers interested in these alternative approaches to other excellent articles and reviews.^8–14 We note that single-layer cardiomyocyte constructs deposited directly onto a substrate do not constitute an engineered heart tissue by our definition; however, these models, which include micropatterned constructs, have yielded important insights.^15,16 We also recognize that EHT technology has made personalized, patient-specific disease modeling possible, but we limit our discussion of disease modeling to cases in which the EHT design philosophy directly reveals the pathophysiological phenotype. Finally, we mention several aspects of EHT design that lend themselves to therapeutic application, but for sake of brevity, refer the reader to other reviews that discuss clinical translation in greater detail.^17,18

FIG. 1.

Flowchart showing principles of EHT design that can be followed in the creation of a new engineered heart tissue system. There may be other considerations not mentioned here, especially in the case of patient-derived induced pluripotent stem cell (iPSC) lines. Figure uses Biorender icons.

II. PHYSICAL STRUCTURE OF ENGINEERED HEART TISSUE

In this section, we explore the physical design of different EHT platforms, grouping the EHTs by substrate type and geometry. Throughout this paper, we often use the term “form factor” to describe the combination of physical size, shape, and matrix composition of a given EHT.

A. Fibrin hydrogel EHTs

The first EHTs were constructed using hydrogel matrices and addressed many of the unphysiological characteristics of cardiomyocytes grown on 2D surfaces. Many formulations of hydrogel-based EHTs exist, differing slightly in substrate composition, throughput, and functional analysis.

A common form factor for 3D hydrogel-based EHTs, developed by Eschenhagen and colleagues, is a polymerized fibrin substrate held between two flexible posts, into which cardiomyocytes and stromal cells are seeded.¹⁹ These tissues have dimensions of approximately 12 × 1.5 × 1.5 mm³ after tissue remodeling. Spontaneously beating EHTs are observed 2–5 D after seeding and last up to 100 D in culture, with contractile kinetics that closely resemble that of adult myocardium.²⁰ Moreover, the silicone posts used to hold either end of the tissue provide diastolic tension which promotes cardiomyocyte alignment; their deflection can also be used to measure contractile function via video analysis. The auxotonic nature of loading with silicone posts is more realistic than 2D culture, but isometric culture in this system is more difficult to achieve, requiring insertion of a rigid rod into the silicone post. Absolute measurement of force production necessitates transfer to a separate testing bath, usually under nonsterile conditions, involving manual handling and careful attachment of the end of the tissue to the force transducer. For this reason, fibrin-post EHTs are more amenable to video analysis in culture.²¹ These EHTs are medium-throughput, relatively cost-effective, and suitable for longitudinal and indirect measurement of contractility under auxotonic conditions. This format of EHT is also suitable for drug screening, recently being used to evaluate 10 iPSC (induced pluripotent stem cell)-derived cardiomyocyte lines with seven inotropic compounds.²²

The major limitation of silicone-post, hydrogel EHTs is that mechanical loading is auxotonic. As a result, when video analysis is used to functionally characterize their contractility, isometric twitch forces cannot be obtained. Moreover, since the tissue length changes during the contraction, it is not possible to characterize the length-dependent activation (LDA) of these EHTs during culture, although a recent publication reports an isometric-testing apparatus to enable LDA.²³ While other functional measurements, such as measurement of calcium transients and action potentials can be made, the effects of strain dependance on these parameters are more difficult to ascertain. Nevertheless, the tuning of post-stiffness to increase afterload was shown to enhance aspects of EHT maturation, as measured by increased sarcomere length and calcium handling. On the other hand, maximum afterload resulted in pro-fibrotic changes consistent with pathological hypertrophy.²⁴ This robust design has enabled fine-tuning of afterload across two orders of magnitude, improving physiological properties at one regime, while uncovering a pathophysiological signal in another.

An alternative design approach for fibrin hydrogel EHTs involves casting the cell-hydrogel mixture between two polymer strands (“Biowire II”).²⁵ The production of these EHTs can be accomplished without using PDMS (polydimethylsiloxane), which sequesters hydrophobic compounds. Typical Biowire II tissues have dimensions of 5 × 1 × 0.3 mm³. Since the physical properties of the polymer wire are known ahead of time, force measurements can be made by monitoring the deflection of the wire. This system, which exhibits positive force-frequency relationship and post-rest potentiation, has been used to uncover chamber-specific responses to drugs,²⁵ model the effects of iPSCs derived from hypertensive patients,²⁵ characterize mutations in sodium channel that lead to dilated cardiomyopathy,²⁶ model cardiac fibrosis,²⁷ and recapitulate angiotensin-induced cardiac remodeling.²⁸ However, because the polymer strands on both ends of the tissue shorten, it remains to be seen whether the Biowire II paradigm could be modified to allow for non-auxotonic loading.

An advantage of hydrogel EHTs is the relative versatility of incorporating hardware that enables real-time functional measurements. For example, the Biowire II platform, described above, has been refunctionalized with additional layers that permit measurement of tissue contraction with elastic microwires and electrical pacing with integrated carbon electrodes.²⁹ Similarly, in a reformulation of the classic silicone post-based EHT, magnets have been incorporated into the tips of posts so that post-bending can be electromagnetically measured as a proxy for auxotonic twitch force.^30,31 However, the increased complexity of such designs necessitates technical expertise across several engineering fields. To address this, platforms such as the milliPillar have been developed to allow for easier dissemination of EHT technology among labs, especially those that may be new to cardiac tissue engineering. Importantly, the milliPillar platform maintains all the mature functional characteristics of its original, custom-built counterpart.³² Typical milliPillar EHTs are cylindrical with a length of approximately 4.5 mm and a diameter of approximately 0.5 mm after 21 days in culture.

Silicone post-hydrogel EHTs have been optimized in various ways, including advanced maturation with an intensity-training protocol and incorporation of EHTs within multi-organ chips linked by vasculature.^33,34 The intensity-training regime, which involves ramp pacing in culture, helps EHTs achieve hallmarks of mature myocardium, such as a positive force-frequency relationship, high density of mitochondria, robust T-tubulation, and functional calcium handling that responds appropriately to inotropic stimulation. Such tissues are 6 mm long with a diameter of 1.8 mm. This intensity-training protocol can be applied to other types of EHT as well. Other design factors that have improved EHT maturity include the deposition of decellularized porcine extracellular matrix (dECM) within a traditional hydrogel EHT, which has been shown to enhance contractile function, calcium handling, and electrophysiological properties.³⁵ dECM can also be incorporated into bioinks to enable 3D printing of EHTs.³⁶

Importantly, the design of fibrin-based EHTs can easily be modified to different sizes and form factors, either smaller for high-throughput applications, or larger for in vivo use as remuscularization patches. For example, GMP (good manufacturing practices)-compliant human EHT patches of a larger size (5 × 7 cm, 450 × 10⁶ cells), but not smaller, improve contractile function in a guinea pig model of cryoinjury as well as in a healthy porcine heart.³⁷ Importantly, changing the form factor from a strip-based EHT to a mesh-like patch improved cell distribution and oxygen diffusion, with resultant improvements in conduction velocity, but with slightly lower force production per cell number than other constructs.³⁷

Similar work performed by the Bursac lab, using a “frame-hydrogel” methodology has also been used to create implantable muscular patches of different architectures with remarkably mature performance (specific force, 20 mN/mm²; conduction velocity, 25 cm/s; patches ranging in surface area from 7 × 7 to 36 × 36 mm²).³⁸ Because the frame-hydrogel method relies on CNC (computer numerical control) production of Teflon molds, a variety of constructs can be created: isotropic tissues to promote random myocyte alignment, patches with elliptical holes to promote anisotropy, epicardial-mimetic patches that replicate realistic fiber orientation as determined by cardiac MRI, and highly aligned cylindrical bundles.³⁸ As with fibrin patches developed by the Eschenhagen group, “frame-hydrogel” patches were successfully implanted in mice and rats with transmission of action potentials and calcium transients, though they failed to achieve electrical coupling with host tissue.^39,40 One difficulty regarding application of hydrogel-based EHT patches for therapeutic purpose is the resultant immune response, which is beyond the scope of this review but discussed at length here.⁴¹ Fibrin-based patches tend to elicit CD3+ immune cell infiltrates and require transplantation in immunodeficient hosts to ensure long-term survival [e.g., up to 110 D in an immunodeficient rat, but only 14 D in a WT (wild type) rat].⁴¹ As a result, although therapeutic application of EHTs remains an attractive goal, barriers such as immune response and a lack of robust electrical coupling to host tissue usually result in graft failure.

B. Collagen-based EHTs

Collagen-based EHTs are conceptually similar to fibrin-based EHTs but arguably provide a more native ECM substrate, since collagen is the most naturally abundant extracellular protein in the body and the heart. Cardiomyocytes avidly bind collagen through integrins, and the collagen network supports the heart's tensile strength, fiber orientation, and nutrient supply.^42–44 Early studies using collagen casting of neonatal rat cardiomyocytes required the presence of Matrigel; however, optimization of experimental parameters enabled this non-chemically defined matrix component to be eliminated. Collagen-based EHTs can also exhibit significant batch-to-batch variability due to differences in the gelling properties of each batch, but this issue can be addressed by producing acid-solubilized collagen in house from standard protocols.⁴⁵

Many collagen-based EHTs are produced in a ring-shaped mold to permit collagen condensation over several days, and the dimensions of the mold can be altered depending on the application. If auxotonic loading is desired, analogous to EHTs held between flexible posts, ring-shaped collagen EHTs can be placed on stretchers with silicone holders, while static stretchers, which promote isometric loading, can be made from stainless steel. As with fibrin-based EHTs, the dimensions and features of the collagen mold can easily be modified to suit the application, which can range from disease modeling and drug screening to regenerative medicine.

A recent area of investigation is the combination of collagen and fibrin to make a composite scaffold that maintains the biologically relevant cues provided by collagen with the rapid polymerization properties of fibrin.⁴⁶ Given the almost infinite combinations of ECM constituents, a design-of-experiments approach, such as one used by Coulombe and colleagues to compare three iPSC-derived cell lines, will help guide the optimal parameters required for a specific application.⁴⁷

Recent improvements in the physiological maturity of collagen-based EHTs have made it possible to perform chamber-specific drug testing in low- and medium-throughput formats, model the hallmarks of heart disease, and perform myocardial repair.^48–51 For example, a ring-shaped construct (120 μm wall thickness, 300 μm inner diameter), used by Goldfracht et al., enables arrhythmia modeling;⁵¹ a mesh-shaped construct (circular patch with diameter of 15 mm) enables vascularization of engineered heart muscle in a model of transmural heart wall replacement;⁵⁰ and a miniature format EHT (6.3 mm length, 3.4 mm width) enables creation of up to 48 constructs in one plate for drug screening.⁴⁸ Auxotonic contraction in ring-shaped constructs combined with electrical pacing improves tissue maturation as revealed by positive force-frequency relationship, enhanced T-tubulation, and improved post-rest potentiation.⁵²

Finally, recent work has upscaled the size of collagen-based EHTs to the so-called “mega-scale,” consisting of a 65 × 75 mm², collagen mesh scaffold supporting the growth of 1 × 10⁹ cardiomyocytes. This mega-scale EHT was successfully used in a porcine model of myocardial ischemia, with no arrhythmogenesis seen up to 4 weeks after implantation.⁵³

C. Native ECM EHTs

In contrast to hydrogel-based EHTs, in which cardiomyocytes are reconstituted within a homogeneous matrix, EHTs can also be created from decellularized native extracellular matrix (ECM) scaffolds which are then seeded with cardiomyocytes. Because the matrix is prepared prior to cell seeding, cells are placed on top of the substrate, rather than cast together with the substrate as in hydrogel-based EHTs; however, as mentioned previously in Sec. II A, ECM can also be denatured and reconstituted within hydrogel substrates. A benefit of using a native substrate is that cardiomyocytes are subject to physiologic cues from components of the extracellular matrix such as collagen, fibrillin, heparin sulfate, and laminin.^54–56 Optimizing the decellularization process is an important consideration because treatment with different detergents results in preferential depletion of certain ECM proteins. For example, a Triton-based decellularization preserves non-collagen proteins at the cost of decreased collagen retention.⁵⁷ As a result, the decellularization process must be optimized for each application. An additional complexity is that ECM composition changes during development, affecting the chemical and physical cues felt by cardiomyocytes, with recent work demonstrating that neonatal-stage ECM best promotes upregulation of transcription factors for cardiomyocyte development.⁵⁸ Finally, the source of the ECM can be crucial to the contractile performance of the EHT, with diseased ECM from a porcine HCM (hypertrophic cardiomyopathy) model causing impaired diastolic function in an iPSC-derived cardiomyocyte model system.⁵⁹

An additional consideration for decellularized ECM EHTs is how tissues are anchored to substrates in a way that permits functional measurements. Tung and colleagues addressed this by affixing vibratome-sectioned, decellularized porcine ECM to a coverslip, creating engineered heart slices (300 μm thick, 5–16 mm diameter disks).^60,61 Under these conditions, properly aligned collagen I, collagen III, and laminin are maintained, which supports cardiomyocyte attachment and survival with a small amount of penetration into the ECM. The slice model enables myocytes to mature into bundles with organized sarcomeres and is excellent for electrophysiologic interrogation and drug testing. However, engineered heart slices display small fractional shortening (<2%, as compared to 30% in the adult heart⁶²), a negative force-frequency relationship, and relatively slow conduction velocities. Nevertheless, this platform enables long-term culture and is an attractive platform for cardiotoxicity testing.⁶⁰

Campbell and colleagues prepared decellularized slices in a similar manner but use a Teflon cassette to immobilize the tissue on both ends, creating a rectangular, bundle-shaped EHT (typically 6–10 mm long with cross-sectional areas around 0.1 mm²).⁶³ This allows for isometric culture and facile manipulation of EHTs into other equipment, such as organ baths for direct force measurements or optical mapping. Under optimized pacing and media conditions, these tissues develop a positive force-frequency relationship, respond physiologically to inotropic stimulation, and recapitulate many genetic cardiomyopathies.^64–69 One limitation of this type of EHT is that tissues are maintained in isometric culture, although recently developed bioreactor systems mitigate this limitation by allowing for dynamic length control in culture.^70,71

D. Muscular thin films (MTFs), microtissues, and other platforms

The muscular thin film (MTF) is a unique approach to measuring the functional parameters of engineered heart tissue.⁷² In an MTF, a layer of PDMS is cured atop a rectangular pattern of PIPAAm [poly(N-isopropylacrylamide)] polymers, onto which cardiomyocytes are seeded. Typical MTFs are 1.2–3 mm long × 0.3–2 mm wide. A layer of fibronectin, previously applied by microcontact printing, enables cell attachment. MTFs are cultured within a simple, one-chamber microfluidic device, which allows for high-throughput measurements of contractility. MTFs are different from other linear EHTs in their thinness and in only being anchored to their substrate by one end. However, the advantages of this design include uniformity and reproducibility of the tissues, rapid mass production, and a simple readout using optical measurement of muscle-film bending. A newer iteration of the technology has incorporated micro-cracked titanium-gold thin film sensors which act as flexible strain gauges. Endothelial barrier inserts isolate each well in a 24-well thin-film platform.⁷³

The adaptable design of MTF devices has enabled application in drug testing and disease modeling. Examples of applications include modeling mitochondrial myopathies, identifying mechanisms of angiotensin-based cardiac injury, and determining the cellular pathogenesis of catecholaminergic polymorphic ventricular tachycardia.^74–76 Additionally, by incorporating a multi-channel design into the thin-film apparatus, McCain and colleagues were able to model the hypoxic border zone of an infarct and differentiate the pathological signatures of uniform hypoxia vs a gradient of hypoxia.⁷⁷

Several other EHT platforms that deviate from the linear trabecular form factor exist, some of which include novel applications of biomaterials. For instance, cardiac microtissues can be created from PDMS-based molds into which iPSC-derived cells can be seeded. Healy and colleagues first developed this system in 2015 to perform drug screening, and then used metabolic conditioning to improve their electrophysiological phenotype.^78,79 Finally, in a unique design by Black and colleagues, an aligned silk-based scaffold is functionalized with cardiac-tissue derived ECM, which results in finely tunable architecture, degradation, and mechanical properties. In an in vivo application, the silk-ECM construct developed robust vascularization, and in vitro, was shown to promote a more functional phenotype in HL-1 atrial myocytes.⁸⁰

III. DESIGN PRINCIPLES OF EHT FABRICATION

We next discuss various design principles to consider when building an EHT system. These factors include cell composition (e.g., iPSC-derived cardiomyocytes vs primary cells and presence of non-myocytes); media selection (including maturation media); type of functional measurement (e.g., direct force measurement vs indirect readout); mechanical and electrical loading (including chronic mechanical stimulus and electrical pacing); and throughput. Some of these considerations, such as mechanical loading and functional measurements, are determined in large part by the form factor of the EHT. Others, such as media and cell composition, can apply to any kind of EHT. A summary of these design choices is provided in Fig. 2.

FIG. 2.

Table summarizing characteristics of different EHT platforms with regard to key design decisions, including substrate structure and ECM, measurement technique, feasibility of electro-mechanical conditioning, throughput, cell composition, and media. Boxes with gray background represent common design choices that can apply to any format of EHT. Figure uses Biorender icons.

A. Cell composition

Early EHTs constructed with primary cells, such as rat neonatal ventricular myocytes, recapitulated many aspects of adult cardiac physiology. However, iPSC-derived cardiomyocytes have now become the preferred cell source for EHT constructs because they are of human origin and because differentiation protocols are robust, yielding very large numbers of highly pure, chamber-specific human cells under chemically defined conditions.^81,82 Techniques like cryopreservation and cell expansion enable order-of-magnitude increases in iPSC-derived cardiomyocyte quantity.⁸³ Moreover, many studies have used patient-derived iPSCs, with creation of isogenic control lines using CRISPR (clustered regularly interspaced palindromic repeats)-Cas9 gene editing, to personalize disease modeling and treatment, as summarized in this review.⁸⁴ A wide variety of donor cells can be used to make iPSCs, including dermal fibroblasts and lymphocytes, but epigenetic mechanisms may make certain cell types more suitable than others.^85,86 Direct-reprogramming techniques, previously only successful in mouse fibroblasts, have even made cardiomyocyte transdifferentiation from human fibroblasts possible.⁸⁷ In this section, we review the pathways that underlie cardiomyocyte development, revisit the cellular composition of the adult heart, and address current approaches to recapitulate native cellular populations in EHTs.

Advances in developmental biology have made efficient cardiomyocyte differentiation possible. The most common paradigm for cardiac differentiation is induction of mesoderm by activation of Wnt/β-catenin signaling, followed by timed Wnt antagonism, cardiac progenitor development, and myocyte formation.⁸⁸ Such differentiations can be performed in either monolayer or embryoid-body format. Though a multitude of differentiation techniques are used, common features include basal media [usually RPMI (Roswell Park Memorial Institute), DMEM (Dulbecco's Modified Eagle Medium), CDM3 (chemically-defined media, 3 components)], strict temporal modulation of activin/BMP (bone morphogenic protein) and/or Wnt signaling, and metabolic purification. Subtype-specific differentiations can also be performed. For example, addition of retinoic acid early in differentiation promotes an atrial-like phenotype through modulation of HOX genes that define the second heart field.⁸⁸ These preparations have been used to investigate atrial fibrillation.²⁵ For all subtypes, a crucial consideration is cardiomyocyte purity which can be verified using markers such as cTnT, α-actinin, and NKX2.5.⁸⁹ Additionally, ventricular cardiomyocytes should predominantly express β-myosin heavy chain (MHC) and ventricular light chain (MLC-1v), while atrial-like cardiomyocytes should express the faster α-MHC isoform and atrial-light chain (ALC-1).⁹⁰ The biological mechanisms that underpin these transitions are beyond the scope of this review but are well-summarized here.^88,91,92 As with other design choices made in the development of an EHT system, the source, identity, and purity of constituent cardiomyocytes is just as important as its physical structure.

Although cardiomyocytes account for 70%–85% of the volume of the human heart, they comprise only 25%–35% of the total cell number.⁹³ The selection of cell type proportion is another design consideration to be optimized, made particularly difficult by the wide variability among estimates of non-myocyte populations. Older work suggests that fibroblasts form the majority of cells (64% fibroblasts in rat heart and 58% “mesenchymal” cells in human heart by flow cytometry), but newer studies identify endothelial cells as the predominant non-myocyte cell type.^94,95 Because of the small size and relative immaturity of iPSC-derived cardiomyocytes, precise recapitulation of these ratios is not a primary goal in EHT design. Rather, EHT cell composition decisions should emphasize a reductionistic approach which prioritizes the specific research question at hand. For example, a study that investigates cardiac fibrosis should optimize the ratio of cardiomyocytes to fibroblasts to achieve a faithful disease model;⁹⁶ but, an experiment that investigates a non-fibrotic disease may not have to alter this variable.

Across the spectrum of EHT form factors, inclusion of fibroblasts improves the electrical and mechanical function of tissues, enhancing contraction rate, promoting better organization of F-actin fibers, increasing the amplitude of contraction with a trend toward positive force-frequency response, and improving structural maturity, without necessarily increasing pro-arrhythmic effects.^97–100 The source and age of fibroblasts also matter, with some studies showing improved tissue maturity in the presence of fetal-like, as opposed to adult fibroblasts.¹⁰¹ Moreover, recent work implicating the importance of second heart field progenitors has given rise to iPSC-derived fibroblasts, which improve the electrophysiological properties of cardiomyocytes in 2D co-culture systems.¹⁰² Finally, epicardial cells, which can form fibroblast grafts in infarcted rat models, have also been shown to improve calcium handling and contractile function in EHTs.¹⁰³

In addition to fibroblasts, endothelial cells have also been shown to promote enhanced phenotypes in EHTs. Early studies using neonatal cardiomyocytes within hydrogel matrices showed that endothelial cells promote more synchronized EHT contraction by upregulating Cx43 expression.¹⁰⁴ Endothelial-fibroblast co-culture in iPSC-CM microtissues can improve T-tubulation, contractility, and mitochondrial respiration.¹⁰⁵ iPSC-cardiomyocyte-endothelial-amniotic mesenchymal stem cell co-culture, meanwhile, improves contractility and troponin T expression and upregulates L-type calcium channels.¹⁰⁶ Simultaneous differentiation of cardiomyocytes and endothelial from the same sample of iPSC cells can be achieved by co-stimulation with VEGF (vascular endothelial growth factor) and a Wnt inhibitor after initial cardiac mesoderm formation.¹⁰⁵

Increased attention to the role of inflammation in the heart has led to the introduction of immune cells, such as macrophages, to EHT systems. Cardiac macrophages are a particularly heterogeneous cell type with multiple origins and manifold functions, including phagocytosis after cardiac injury, regulation of arterial tone within the coronary vasculature, endocardial remodeling, and angiogenesis.^107–110 They can have both anti-inflammatory and pro-inflammatory effects. Macrophages have been shown to regulate electrical coupling in the heart by expressing Cx43 and forming junctions with cardiomyocytes in vivo.¹¹¹ Similarly, co-culture of iPSC-CMs with macrophages was used to model myocardial inflammation such as that seen after myocardial infarction.¹¹² For more in-depth discussions of cardiac macrophage biology, we refer readers to the following reviews.^113–115

The cellular composition of engineered heart tissue can be made as simple or complex as necessary to model specific physiological processes. Advantages of including multiple cell types include enhanced tissue function and the ability to better recapitulate disease models, especially when cell-to-cell communication or paracrine signaling is important to pathogenesis. However, complications of incorporating more cell types include technical complexity of multiple iPSC differentiations, which always result in cellular heterogeneity; the differing media requirements of each cell type, which can somewhat be addressed using custom formulations or volumetric mixtures of different media; and the genetic mosaicism that can result when commercial cell lines are mixed with patient-derived iPSC cell lines. Moreover, because the mitotic capacities of fibroblasts and endothelial cells differ and are greater than that of cardiomyocytes, it is possible that the cell composition could change in unpredictable ways over long-term culture. Certain physiological and pathophysiological models compel the use of non-cardiomyocytes. Models of post-infarct fibrosis, for example, could use greater proportions of fibroblasts, while models of cardiac inflammation would be well-served by addition of macrophages or innate immune cells.

B. Media composition

Selection of media is a crucial aspect of EHT culture and can be considered an aspect of the EHT design process itself. Most cell culture media contain high concentrations of glucose, which is inconsistent with the metabolism of the adult heart, which largely consumes fatty acids.¹¹⁶ Moreover, the chemical composition of blood changes over time, from a state of relative hypoxia in the fetal circulation, to a normoxic state in the adult circulation. Cell culture media typically does not reflect the biochemical composition of adult plasma or accommodate the changes in composition seen during fetal development, adulthood, and disease onset. However, reductionist studies with a variety of cardiomyocyte and EHT systems have shown several important principles that can serve as a starting point for media optimization.

To begin, the addition of hormones to media, such as thyroid hormone and glucocorticoids, can promote cardiomyocyte maturity. Specifically, addition of T3 (triiodothyronine) together with dexamethasone, though not independently, promotes T-tubulation and enhanced Ca²⁺ transients, though these studies were done on Matrigel mattresses rather than true 3D EHTs.¹¹⁷ Additionally, T3 promotes longer sarcomere lengths and enhanced mitochondrial respiration.¹¹⁸ However, the addition of T3 promotes a myosin isoform shift from β- to α-myosin heavy chain (MHC), away from the ratio seen in the native adult ventricle.¹¹⁹ As a result, EHT maturation protocols which seek to enhance β-MHC production may be adversely affected by addition of thyroid hormone.

Fatty acid-containing media can promote the metabolic maturation of cardiomyocytes or EHTs.¹²⁰ Traditional media formulations, such as those containing B27, a supplement originally designed for hippocampal culture, do not contain physiologic concentrations of fatty acids. So far, the most comprehensive maturation media contains glucose-free DMEM, supplemented with vitamins found in RPMI but not DMEM, additional amino acids, such as glutamine and proline, lactate, lipid-rich BSA (bovine serum albumin), and carnitine, among other metabolic constituents.¹²¹ The evaluation of this media combination includes extensive drug response testing, among other physiological read-outs. Disease-modeling experiments show that recapitulating a phenotype sometimes requires experiments to be done in maturation media. Importantly, the molecular mechanism of metabolic maturation of hiPSC-CMs depends on aberrant upregulation of the HIF-1α – lactate dehydrogenase A system, and inhibition of these two enzymes is sufficient to promote switching from glycolysis to oxidative phosphorylation.⁸⁰ However, because many other aspects influence EHT maturation, not every media constituent needs to be added for each experiment. For example, one study noted that switching glucose for galactose was sufficient to promote adult-like cardiomyocyte phenotypes in the presence of fatty acid-containing media.¹²² Other work has demonstrated that substituting DMEM, which has physiological levels of calcium, for RPMI can increase contractile force and promote positive force-frequency relationship when combined with ramp pacing.⁶⁴

Similar to the complexity of cell-type selection for EHT production, the choice of media is a crucial design consideration in which the benefits of custom media are weighed with the complexity of developing the appropriate formulation. Previously published maturation media recipes are an ideal place for researchers to start; however, developing new formulations with many components that have to be optimized in a combinatorial fashion is difficult, especially if the functional readout is low-throughput and does not permit easy interrogation of several media variables at once. Without a priori hypotheses about the most successful modifications, extensive high-throughput testing of media recipes is necessary. An important consideration that should apply to all custom media formulations is the elimination of animal-derived sera, such as fetal bovine serum, because these products are subject to fluctuations in quality and composition and are not chemically defined. Use of knock-out serum replacement is a viable alternative.

C. Mechanical and electrical stimulation

EHT culture systems which provide some of the mechanical and/or electrical aspects of in vivo cardiac tissue can promote the development of mature, functionally useful constructs. In addition, such systems often enable direct interrogation of the role of mechanical loading and/or electrical stimulation in cardiac physiology and disease by making these factors an independent variable. In this section, we discuss various approaches taken to implement mechanical and electrical loading in EHT systems.

1. Mechanical loading

Mechanical loading schemes can be grouped into two categories: dynamic and static. Dynamic systems rely on either the endogenous compliant properties of materials which comprise the EHT or on direct control of EHT stress or strain using external motorized systems. Static systems instead load tissues mechanically at a fixed length without reference to their contractile behavior. That is, they apply a static mechanical load to EHTs without allowing for EHT shortening during contraction.

Many EHT systems incorporate compliant mechanisms in order to measure EHT contraction force using material deformation as a proxy for EHT force generation, as described in Sec. III D.^30,72 In this section, we will focus on approaches that explicitly incorporate loading as a parameter of the experiment, rather than ones which include mechanical loading as a by-product of force measurement or other features.

a. Dynamic mechanical systems

Dynamic mechanical systems may be placed into two categories: passive systems, which use the natural elasticity of integrated materials to approximate the shortening undergone by in vivo cardiac tissue, and active systems, which use external actuators to directly control EHT length during a contraction cycle. The most common passive system design approach involves the use of flexible posts onto which EHTs are cultured or attached, as previously discussed. These posts, often made of silicone³¹ or PDMS,^32,34 provide an elastic load which allows some EHT shortening during contraction and then returns the tissue to its original length following relaxation. This spring-like loading scenario, in which the load experienced by the contracting tissue increases linearly with the distance shortened, is referred to as auxotonic loading. Posts, microposts, or pillars that provide auxotonic loading are relatively easy to design and manufacture and offer a robust option for tuning EHT shortening. They can also be created by microfabrication and soft-lithography techniques.¹²³ Moreover, imaging of post-deflection can be used to estimate EHT force, as discussed in detail in Sec. III D. In a slightly different approach, in which “dyn-EHTs” are created within a bent PDMS strip, modulation of strip dimensions allows for auxotonic contraction and fine tuning of preload; notably, dynamic culture was required to uncover a disease phenotype in this model of arrhythmogenic cardiomyopathy.¹²⁴ dyn-EHTs are typically 6 mm long with a cross-sectional area around 0.08 mm². Finally, modulation of mechanical loading can also occur through a spring-based system, in which storage of torsional potential energy between two posts provides auxotonic loading to an EHT.⁵²

Active systems, though more complex to design and implement, provide the opportunity to directly control EHT length in coordination with electrical stimulation and other culture parameters. For example, development of a voice-coil driven EHT bioreactor, in which either end of a tissue is held by a claw, enabled chronic dynamic culture under a physiologic shortening regime; the use of position feedback control also allowed for the measurement of work loops.¹²⁵ Direct control of EHT length also enables more complex mechanical stimuli which can modulate in character over the course of EHT culture (e.g., to simulate evolving mechanical conditions during in vivo cardiac development, bouts of exercise, or in progressive maturation schemes). Older work with isolated trabeculae achieved force clamping by implementing feed-forward control, iteratively adjusting the length control profile applied to the muscle until achieving a desired work loop or load profile during each contraction–relaxation cycle.¹²⁶ A similar approach was successfully implemented for acute mechanical loading of a single EHT by Sewanan et al.¹²⁷ Future work remains to make active systems tenable for different EHT form-factors, to increase the throughput of active systems, and to enable real-time force measurement and other functional and/or biochemical readout during culture.

b. Static mechanical loading

Static mechanical loading systems focus on modulating EHT strain during isometric culture, rather than attempting to re-create the auxotonic conditions of in vivo cardiac tissue. While this approach does not mimic in vivo conditions, it does allow for reliable comparisons between groups as they respond to a uniform mechanical stimulus. Static systems come in more forms than active systems. Microposts with increased rigidity can be used to culture EHTs at varied static lengths. Alternatively, isometric culture can be promoted by forming collagen-based EHTs around fixed nylon tabs at either end of the tissue; a “no-stress” group is formed by cutting the tissue on one end.¹²⁸ Static mechanical loading systems can also be implemented using microfluidic techniques, an approach which benefits from scalability and throughput.⁷⁹ Finally, EHTs that are made within rigid frames can be made to undergo isometric loading by stretching the matrix across the frame at a set length, as discussed in earlier sections describing the frame-hydrogel and decellularized-matrix approaches.^40,63,129

2. Electrical stimulation

Electrical potentials propagate through the myocardium during each cardiac cycle in precise ways. Normal electrical activity is important to cardiac growth and function and aberrant electrical activity leads to disease states. Efforts have been made to implement electrical pacing systems to both improve EHT maturity and to model aspects of electrical dysfunction during heart disease.

In this section, we briefly review systems for electrically pacing EHTs, focusing on design considerations and key experimental parameters. A typical electrical pacing system consists of a stimulus generator connected to two electrodes which protrude into either EHT culture media or make direct contact with EHTs themselves. Electrical pacing systems are primarily differentiated by their electrode material and the voltage profiles applied. In addition, the mode of stimulation, either near-field stimulation (NFS) or point stimulation,⁶¹ is a significant distinction, but as most systems discussed in this review utilize NFS, we will restrict our discussion to NFS-based systems. Point stimulation, however, is necessary to measure the conduction velocity of EHT preparations.

Stimulus electrodes are commonly made of carbon, platinum, or stainless steel. Ultimately, the best choice for a given application may depend on form-factor and ease-of-use considerations. Platinum electrodes have been used effectively,¹²⁹ although there is some evidence that chronic stimulation with platinum electrodes can lead to cytotoxic effects in non-cardiac cell culture models.¹³⁰ Carbon electrodes are easily acquired, available in different form factors (rod-shaped,³⁴ flat,¹²⁵ etc.), and are highly biocompatible, exhibiting minimal release of toxic compounds during chronic culture.^{25,27,29,32,64,128,131} Stainless steel electrodes are a readily available, inexpensive option and come in a variety of sizes, shapes, and alloys.¹³² They are more robust and less brittle than carbon electrodes, although corrosive buildup can occur during chronic culture. In one study in which electrode materials (carbon, stainless steel, titanium, and titanium nitride) were compared head-to-head, carbon electrodes were shown to have the highest charge-injection capacity, the lowest excitation threshold, and the highest maximum capture rate.¹³³

The stimulus voltage profile is another important consideration in EHT pacing systems. Typically, square wave pulses ranging from 2 to 5 ms¹²⁸ in duration with an amplitude of 2.5–5 V/cm are used, depending on proximity to EHTs and the conductivity of the fluid media and organ bath.^34,64,133 Baseline pacing frequency is often 1 Hz, although protocols involving ramp pacing up to 6 Hz can be used to promote EHT maturation.^34,64 Another key factor is the use of monophasic or biphasic stimulation (i.e., alternating the positive electrode in each pacing cycle). Biphasic stimulation¹³⁴ minimizes electrochemical buildup and the generation of cytotoxic reactive oxygen species,^32,64 while monophasic stimulation better approximates the unidirectional nature of cardiac depolarization.³⁴ Tissues can be paced with the two electrodes aligned perpendicularly on either side of the EHT or on the longitudinal ends.

D. Functional measurement technique

An important requirement for EHT systems is the ability to functionally characterize physiological parameters such as contractility. A variety of measurement techniques, both direct and indirect, can be used to measure force. A direct force measurement (i.e., using a calibrated force transducer) is considered the gold-standard mechanical readout. However, there are situations in which indirect measurements (e.g., optical readouts or electrical sensors) are preferred. In this section, we review the technical considerations of each type of functional force measurement, highlighting the benefits and drawbacks of each system. To show that both direct and indirect force measurements can yield qualitatively similar data, we display representative twitches from six EHT designs in the literature (Fig. 3). Because this review focuses on linear EHT constructs, rather than chamber-mimetic constructs or spherical organoids, we limit our discussion of functional characterization to one-dimensional force assessment.

FIG. 3.

Comparison of six representative engineered heart tissue designs, with corresponding twitch-force measurements, intended to facilitate qualitative comparisons between methodologies. Twitches were digitized from respective publications from the figure listed. Each twitch was measured under 1 Hz stimulation, except for the ring-shaped EHT which was stimulated at 2 Hz and the Dyn-EHT which was allowed to contract spontaneously. Time-axis scale bar is 500 ms for each twitch. The units of the force measurement were kept consistent with original publications, though are analogous. In each table, a brief description of the ECM substrate, the number and type of cells seeded, the measurement technique, and a pertinent fact about the index publication are mentioned. (a) Reproduced with permission from Zhao et al., Cell 176(4), 913–927 (2019). Copyright 2019 Elsevier.²⁵ (b) Reproduced with permission from Lind et al., Lab Chip 17(21), 3692–3703 (2017). Copyright 2017 Royal Society of Chemistry.⁷³ (c) Reproduced with permission from Goldfracht et al., Nat. Commun. 11(1), 75 (2020). Copyright 2020 Author(s), licensed under a Creative Commons Attribution (CC BY) License.⁵¹ (d) Reproduced with permission from Schwan et al., Sci. Rep. 6(1), 32068 (2016). Copyright 2016 Author(s), licensed under a Creative Commons Attribution (CC BY) License.⁶³ (e) Reproduced with permission from Bliley et al., Sci. Transl. Med. 13(603), eabd1817 (2021). Copyright 2021 AAAS.¹²⁴ (f) Reproduced with permission from Ronaldson-Bouchard et al., Nature 556(7700), 239–243 (2018). Copyright 2018 Springer Nature.³⁴

Direct force measurement is an excellent way to functionally characterize engineered heart tissue. However, it is usually a low-throughput approach that requires expensive equipment such as force transducers. Because of concerns about sterility and tissue handling, direct force measurements are usually an end point result that require removal of EHTs from the culture environment and placement into an organ testing bath. These testing baths must incorporate temperature and pH control, flow control (to avoid buildup of metabolic byproducts and permit acute interrogation of pharmacological responses), electrical pacing circuitry to stimulate the tissue, and a force transducer which can be calibrated. Optional components of a testing setup include length control to test length-dependent activation and microscopy to permit measurement of action potentials or calcium transients with voltage-sensitive dyes. These custom setups offer comprehensive interrogation of EHT physiology.

Several aspects of direct force measurements of EHTs require careful consideration. First, tissues must be handled delicately while being moved from the culture environment to the testing bath, often requiring the use of micromanipulators. Second, the way that a tissue construct is attached to the force transducer is a non-trivial consideration that requires optimization. For example, in the original studies that developed the trabecula culture system, muscle fibers were attached to the force transducer with a basket-like extension and pinned on the opposing end using a hook. Other mounting preparations, such as one using microtweezers and one using silk sutures, were reported to cause declines in specific force.² Today, some groups mount hydrogel EHTs within commercially available organ bath systems.³⁴ Others use custom-built systems that mount EHTs using pins or claw-shaped protrusions.^38,135 In cases of ring-shaped constructs mounted on posts, one post can contain a force transducer.⁵¹ Care must be taken during the mounting procedure that excess compliance or slack does not develop while tissue length is being controlled. Once the challenges of mounting and sample handling are overcome, purely isometric, direct force measurements can easily be made, and further manipulation of tissue length (to determine length-dependent activation) becomes possible. Drawbacks of direct-force measurement techniques include low throughput, the need for expensive equipment that is often custom-built, and the difficulty in returning tissues to culture after measurement because of contamination concerns.

Video readouts of tissue contractility have been widely adopted as an alternative to direct force measurement. The physical design of many types of EHTs, such as hydrogel-post EHTs or microtissues lend themselves well to this measurement modality. Because contraction of a hydrogel-post EHT is auxotonic, with quasi-linear deformations of posts relative to EHT length, contraction can be tracked by recording the position of the posts. Video analysis is less useful in isometrically loaded preparations because in a purely isometric setting, the only obvious deformations are perpendicular to the axis of force production. Although feature-tracking algorithms can be used to extract descriptions of motion, interpreting their physiological significance is challenging. Notably, the use of video tracking is not limited to hydrogel-post EHTs but can also be applied to muscular thin films, in which the bending of cantilevers by cardiomyocytes is measured.^72,78 Many research groups have developed their own pipelines for video analysis of EHTs, and several commercial technologies exist.^{32,78,123,136–138}

There are some limitations to video-based force measurements. This is an indirect approach and therefore relies on accurate characterization of the material properties of EHT substrates as well as theoretical models of material deformation. Video processing is time-consuming and may lower throughput or prevent real-time readouts of force during experiments which would otherwise allow for the monitoring of drug responses. Another important consideration is that in such systems, measurement sensitivity and mechanical loading are linked: increased post-stiffness will increase load but decrease force resolution owing to smaller displacements during contraction.

In addition to direct force measurements and indirect video measurements of EHT contractility, there are several unique EHT designs in which the force-sensing apparatus is integrated within the EHT. For example, incorporation of micro gold-cracked strain gauges within muscular thin film EHTs offers an electrical readout of contractility that is continuous, not subject to the post-processing requirements of video analysis and permits high-throughput measurement of contractility.⁷³ A similar approach that can be scaled up to 96-well format relies on light-based 3D printing of micro force-gauge arrays.¹³⁹ Meanwhile, a modification of the common hydrogel-post EHT to include a magnet in one cantilever provides a way to measure a contractility signal using a Hall effect sensor placed under the tissue.³⁰

Many aspects of physiology and pathophysiology can be gleaned by carefully optimizing the functional testing protocol. Isolated force measurements under a single preload/afterload combination represent only a small portion of possible measurements. Other functional measurements may be important depending on the specific physiologic question at hand and can be addressed through careful selection of EHT design parameters. Length-dependent activation is a crucial aspect of cardiac physiology that can be altered in disease.¹⁴⁰ Force-frequency relationship, post-rest potentiation, and diastolic excess fraction, meanwhile, characterize the maturity of the calcium-handling apparatus.^141–143 The response to an inotrope represents functional maturation of the adrenergic signaling pathway.¹⁴⁴ Functional measurements can also include calcium transient analysis, which is central to mechanical and electrical phenotypes, and optical mapping, which is useful for EHT models of arrhythmias. Given the possible complexity of these readouts, optimization of the protocol is itself an important design consideration, encompassing not only considerations about force measurement, but also choices such as the solubility, cytotoxicity, and excitation/emission characteristics of fluorescent calcium and voltage sensors.

E. Throughput

Throughput is an essential consideration for EHT design and influences other choices discussed above. We define low-throughput EHTs as those in which each tissue requires significant and dexterous labor from the researcher, usually resulting in the production of tens of tissues at a time. Medium-throughput EHTs are defined as those that can be made in 24–48 well plates, while high-throughput EHTs permit 96-well fabrication. Generally, medium- and high-throughput EHT models consist of smaller tissues and are more likely to use video analysis or another indirect force measurement. Conversely, larger EHTs are more easily mounted onto force transducers, but more difficult to produce in a high-throughput manner. Medium- and high-throughput EHT systems are ideal for applications like drug screening and may help optimize experimental conditions such as media composition, which can then be confirmed in low-throughput, direct force-measurement systems. Although many advances have been made in developing (relatively) high-throughput EHT systems, even the highest throughput systems are still at least an order of magnitude below 2D monolayers of cardiomyocytes, which are routinely used in 96- and 384-well format for drug and genetic screening.

EHTs of different form factors, including ring-shaped constructs and microtissues, can be implemented in high-throughput format. Use of hydrogel-based ECM substrates easily supports high-throughput designs because preparations of iPSC-CMs and ECM proteins can be easily pipetted and distributed among many wells. However, designs that use native ECM are typically harder to implement at high throughput because of the difficulty of manipulating and decellularizing tissue at scale. High-throughput EHT systems sometimes use microfluidic chambers, with one design faithfully recapitulating aspects of volume overload within a microtissue of only 5000 cells.¹⁴⁵ In general, higher-throughput systems necessitate complex and automated fabrication processes. For example, modification of the Biowire II platform to a 96-well format required complex 3D-printed composite microwires atop a polystyrene plate pre-embedded with carbon electrodes; however, this yielded a tenfold improvement in throughput and more mature microtissues.¹⁴⁶ The Cardiac MicroRings platform, which miniaturized ring-format EHTs to 96-well format, used a robotic liquid-handling system to automate EHT seeding and measured function using an optical readout.¹⁴⁷ Meanwhile, the InVADE platform, which can be used to model organs other than the heart, relies on 3D printing of a novel biopolymer to enable vascularization and endothelization of the tissue.¹⁴⁸ Muscular thin films, which are constructed from a sheet of elastomeric cantilevers which support myocyte growth, permit high-throughput screening by virtue of their size and the fact that dozens of cantilevers can be etched on a single substrate.⁷² Hydrogel-post EHTs have been commercialized into 48-well arrays,⁴⁸ and the recent development of a mosaic imaging system now enables 96-well recordings of these EHTs with a nominal sensitivity of 0.2 μN.¹⁴⁹

Though there is still a trade-off between the robustness of functional readouts and throughput, the above technical developments are shrinking the gap. Nevertheless, many high-throughput EHT systems require complex fabrication of molds, masters, and special 96-well plates. Meanwhile, it is difficult to miniaturize force transducers to a 96-well format because of size and cost limitations, so many high-throughput systems cannot produce direct force readouts. Even an optical readout requires careful optimization of design constraints, such as the working distance between camera and tissues, the temporal and spatial resolution of the imaging system, the presence of lens arrays to improve focus, and the storage and filtering of data. For a detailed discussion of how these design constraints are addressed, we refer readers to papers such as that by the Chen and Bifano groups, in which resolution is optimized and validated (temporal, 60 Hz; spatial, 1–10 μm) using a system of bi-telecentric and doublet lenses.¹⁴⁹ The creation of complex high-throughput EHT platforms requires not only the biological expertise needed to create EHTs but also expertise in optical and electrical engineering. Commercial systems address these issues but their use is presently limited by cost and a lack of customizability. The proliferation of systems and novel designs by many different academic and industry groups may address these drawbacks over time.

IV. CONCLUSION

More than two decades after Eschenhagen's landmark demonstration of feasibility, the use of engineered cardiac muscle constructs remains a robust and active field. Innovations continue to emerge, driven by compelling applications that range from basic cardiac biology to the development of cutting-edge therapies. In fact, as of publication, at least 10 clinical trials using EHT technologies are enrolling subjects, investigating diseases as varied as heritable arrhythmias, ischemic cardiomyopathy, heart failure, and cancer chemotoxicity. At least one case report has demonstrated the safety of a decellularized ECM product in aneurysm repair¹⁵⁰ (though not without limitations¹⁵¹); the first-in-patient application of cardiac ECM hydrogels has demonstrated safety;¹⁵² and embryonic stem cell-fibrin patches produced symptomatic improvement in a cohort of patients with severe LV (left ventricular) dysfunction.¹⁵³ Though many of these studies have only shown equivocally positive results, we believe this signifies a promising field still in its infancy. Going forward, efforts to improve maturity and throughput of EHT systems will enable significant advances in therapeutics and more granular approaches to disease modeling. New challenges that must be overcome include vascularizing EHTs to permit larger and thicker constructs; promoting successful engraftment of large EHTs for in vivo use; and more faithfully modeling cardiovascular anatomy using newer techniques such as 3D bioprinting. These wide-ranging applications, along with design constraints common in all engineering pursuits, have led to a truly diverse set of approaches to creating, manipulating, and extracting information from EHTs. This fertile intellectual landscape should make it increasingly possible for researchers to rationally select design components for EHTs that are optimal for their specific application.

AUTHOR DECLARATIONS

Conflict of Interest

S.G.C. has equity ownership in Propria, LLC, which develops engineered heart tissue technology.

Author Contributions

Ilhan Gokhan: Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Thomas S. Blum: Writing – original draft (supporting); Writing – review & editing (supporting). Stuart G. Campbell: Conceptualization (lead); Supervision (lead); Writing – review & editing (supporting).

DATA AVAILABILITY

Data sharing is not applicable to this article as no new data were created or analyzed in this review.