Biomaterializing the Promise of Cardiac Tissue Engineering

Abstract

During an average individual's lifespan, the human heart pumps nearly 200 million liters of blood delivered by approximately 3 billion heartbeats. Therefore, it is not surprising that native myocardium under this incredible demand is extraordinarily complex, both structurally and functionally. As a result, successful engineering of adult-mimetic functional cardiac tissues is likely to require utilization of highly specialized biomaterials representative of the native extracellular microenvironment. There is currently no single biomaterial that fully recapitulates the architecture or the biochemical and biomechanical properties of adult myocardium. However, significant effort has gone toward designing highly functional materials and tissue constructs that may one day provide a ready source of cardiac tissue grafts to address the overwhelming burden of cardiomyopathic disease. In the near term, biomaterial-based scaffolds are helping to generate in vitro systems for querying the mechanisms underlying human heart homeostasis and disease and discovering new, patient-specific therapeutics. When combined with advances in minimally-invasive cardiac delivery, ongoing efforts will likely lead to scalable cell and biomaterial technologies for use in clinical practice. In this review, we describe recent progress in the field of cardiac tissue engineering with particular emphasis on use of biomaterials for therapeutic tissue design and delivery.

1. Introduction:

Cardiovascular disease (CVD) continues to yield profound morbidity and mortality across the globe despite more than a half-century of advances in preventive medicine. According to the most recent global statistics published in 2017, there were greater than 400 million CVD cases annually with nearly 18 million deaths (Roth et al., 2017). Regenerative medicine strategies offer promise for secondary management of CVD, especially in the setting of myocardial infarction (MI) and congestive heart failure (CHF) which continue to have an increasing prevalence worldwide (Sanchis-Gomar et al., 2016; Savarese and Lund, 2017). Unlike organ systems such as lung (Zacharias et al., 2018), liver (Wang et al., 2015), intestine (Beumer and Clevers, 2016), skeletal muscle (Relaix and Zammit, 2012) and skin (Blanpain and Fuchs, 2006), which have functional regenerative capacity via tissue-resident stem cells, the mammalian heart is virtually non-regenerative (Bergmann et al., 2015) as it does not possess a robust pool of progenitor cells able to generate new cardiomyocytes (Li, Y. et al., 2018; van Berlo et al., 2014). Growing evidence supports that cardiomyocytes can undergo limited self-renewal through innate processes of dedifferentiation and proliferation (Bergmann et al., 2015; Nakada et al., 2017; Wang, W.E. et al., 2017), with ongoing attempts to potentiate these effects after myocardial injury (Bassat et al., 2017; D'Uva et al., 2015; Leach et al., 2017; Mohamed et al., 2018); however, safety and efficacy of these strategies remain to be proven in large animals and humans. Given its inability to renew functional myocardium after acute, or chronic insults, the mammalian heart instead relies upon tissue remodeling via endogenous cellular components to adapt toward a new, and often maladaptive, physiologic state (Cohn et al., 2000). A potential approach to address this adverse cardiac remodeling is the implantation of regenerative exogenous stem or cardiac progenitor cells into the heart. Still, despite twenty years of clinical research, adult stem cell therapies for ischemic heart disease show only marginal benefits, primarily due to low cell engraftment rates and their inability to generate new functional muscle. Accordingly, there has been a high level of commitment in the research community towards developing functional engineered cardiac tissues (fECTs) for use in cell-based regenerative therapies (Jackman et al., 2015; Shadrin et al., 2017), cardiac disease modeling (Benam et al., 2015; Sadeghi et al., 2017; Turnbull et al., 2018; Wang et al., 2014) and discovery science (Liau et al., 2017; Mills et al., 2017). Recent advances in biomaterials for construction and delivery of fECTs (Capulli et al., 2016; Radisic and Christman, 2013; Reis et al., 2016; Weinberger et al., 2017) as well as for cell-free injection (Ungerleider and Christman, 2014) have become the foundation for the development of new bioengineering therapies for myocardial infarction and heart failure. In this review, we discuss various strategies for in vitro fabrication of fECTs and their applications in cardiac regenerative therapies in vivo, while focusing on biomaterial aspects of tissue design and delivery.

2. Cell Sources and Architectures for Cardiac Tissue Engineering

The main goals of cardiac tissue engineering are to develop high-fidelity, 3-dimensional (3D) cell culture systems to study heart function, dysfunction, and pharmacology in vitro as well as to provide a ready source of functional tissues for therapeutic transplantation. Before achieving these goals, proof-of-concept myocardial tissue engineering studies were needed to demonstrate that cardiomyocytes could be successfully cultured in a functional 3D network in vitro. The first such study, published in 1997 by Eschenhagen et al. demonstrated survival of embryonic chick cardiomyocytes in a collagen-I hydrogel matrix anchored against two Velcro-coated glass cylinders (Eschenhagen et al., 1997). In response to electric shocks, cultured chick cardiomyocytes in this 3D environment maintained force-generating capacity for up to 11 days. Together with early pioneering work by Vandenburgh et al., and the first engineering of fECTs with mammalian cells by Bursac et al., these studies laid the groundwork for the following two decades of dedicated striated muscle tissue engineering research (Bursac et al., 1999; Powell et al., 1999; Vandenburgh and Karlisch, 1989). As visualized in Figure 1, fECTs have diversified to encompass constructs fabricated from various cell sources and possess a multitude of 3D sizes and architectures necessary for specific in vitro and in vivo applications

Figure 1. Cell sources and architectures for engineered cardiac tissues.

A) Cardiomyocytes for engineering of fECTs can be differentiated from human pluripotent stem cells or isolated from neonatal rodent ventricles; B) A variety of fECT architectures have been developed, including cell sheets, spheroids, bundles, patches, and single ventricle analogs.

2.1. Neonatal Rodent Ventricular Myocytes

In cardiovascular research, transgenic mouse models provide a versatile, syngeneic platform for systematic studies of heart development and disease in vivo (Vandamme, 2014) while rat models are mostly used as a therapeutic testbed, and as a source of primary heart cells due to their larger hearts. Adult ventricular myocytes have been dissociated from both rats and mice using the Langendorff perfusion method, and have been used to study single-cell electrophysiology and contractile function (Louch et al., 2011; Wolska and Solaro, 1996). However, these cells exhibit low viability in long-term culture which makes them unsuitable for use in cardiac tissue engineering. Neonatal rat ventricular myocytes (NRVMs), on the other hand, can be isolated with high yield and purity with relatively simple enzymatic dissociation protocols and they exhibit long-term viability in culture making them an ideal cell source for tissue engineering (Figure 1A). Therefore, since the late 1990s NRVMs have been used to generate 3D fECTs in which cells have been shown to undergo progressive electrical and mechanical maturation with time in culture (Bursac et al., 1999; Bursac et al., 2003; Papadaki et al., 2001; Shimizu et al., 2002; Zimmermann et al., 2000; Zimmermann et al., 2002). Methods for engineering of fECTs from neonatal mouse ventricular myocytes (NMVMs) (de Lange et al., 2011) were developed more than ten years after those for NRVMs. While one obvious advantage of mouse over rat-based systems is the availability of transgenic mouse models, difficulties with achieving high purity and yield of dissociated NMVMs have been the main hurdle to wide-spread use of this model in cardiac research. Although transgenic rats are likely to become more ubiquitous in the future, methodologies for their generation have been technically challenging (Pradhan and Majumdar, 2016).

Collective cardiac tissue engineering efforts from multiple groups using rodent-sourced cardiomyocytes, whether primary or mouse pluripotent stem cell-derived, have led to development of 3D culture conditions that permit continued functional maturation of cells akin to postnatal cardiac development in vivo (Bian et al., 2014a; Christoforou et al., 2013; Didie et al., 2013; Kensah et al., 2013; Liau et al., 2011; Miki et al., 2012; Stohr et al., 2013). Recently, Jackman et al. have described small cylindrical fECTs that, after 2 weeks of culture, generate high contractile stresses (~60 mN/mm²) and conduction velocities (52.5 cm/s) approaching those of adult rat myocardium (Jackman et al., 2016). Use of dynamic culture (on a rocking platform) to enhance mass transport drove many of the functional improvements which were mediated via mTORC1 signaling rather than by mechanotransduction. Further structural and functional maturation beyond 2-week culture was achieved by low-rate (0.2 Hz) electrical pacing to maintain tissue contractions and the application of thyroid hormone (T3) (Jackman, C. et al., 2018). After a total of 5 weeks, resulting fECTs showed improved cardiomyocyte size, electromechanical function, expression of contractile proteins, and ultrastructural properties (t-tubules, M-lines, Z-disks, intercalated disks) that quantitatively matched those of similarly-aged native rat ventricular myocytes in vivo. In a large number of proof-of-concept studies starting in 1999, in vitro engineered rodent fECTs have been implanted in mouse and rat models of heart disease showing their ability to reverse pathological cardiac remodeling and improve contractile function, thus paving the way for the translational studies with human cells (Li et al., 1999).

2.2. Human Pluripotent Stem Cell-Derived Cardiomyocytes

Rodent fECTs have been critical for developing methods to improve cardiomyocyte maturation in vitro. Given that they achieve the most advanced functional maturation to date, rodent fECTs remain a valuable platform for in vitro studies of postnatal cardiac development. However, for preclinical research and therapeutic applications, human pluripotent stem cells (hPSCs) represent an ideal, and potentially unlimited source, of bona fide functional human cardiomyocytes (hPSC-CMs)(Figure 1A). Rapid advances in cardiac differentiation of hPSCs from early low-efficiency embryoid body-based protocols (Itskovitz-Eldor et al., 2000) to reporter-based CM selection (Anderson et al., 2007), and finally transgene-free directed differentiation procedures (Kempf et al., 2015; Lian et al., 2012), have unlocked the potential of producing large numbers of disease-specific, autologous cardiac-lineage cells for basic research and therapeutic indications (Jackman et al., 2015; Ogle et al., 2016; Oikonomopoulos et al., 2018). Building on extensive experience with rodent cardiomyocytes, robust human hPSC-CM differentiation protocols have enabled generation of the first human fECTs and evidence that CM maturation in a 3D environment is enhanced compared to that in standard 2D culture (Nunes et al., 2013; Zhang et al., 2013). Further progress has resulted in optimization of culture conditions to advance structural and functional maturation of 3D cultured hPSC-CMs (Jackman et al., 2016; Ronaldson-Bouchard et al., 2018) as well as the scale-up of engineered tissue size to clinically relevant dimensions without loss in magnitude or spatial homogeneity of functional output (Shadrin et al., 2017). Similar to rodent studies, survival and beneficial effects of human fECTs in rodent (Riegler et al., 2015; Shadrin et al., 2017; Weinberger et al., 2016; Wendel et al., 2015), and recently porcine (Gao et al., 2018; Kawamura et al., 2017), myocardial infarction models have been reported. However, the modest beneficial effects are most likely a result of mechanical effects and paracrine action rather than direct functional coupling between the fECT and host myocardium.

2.3. Non-Myocytes

By volume, cardiomyocytes represent the largest fraction of resident myocardial cells. Consensus estimates over the past several decades suggest 70-80% of myocardial volume is comprised of cardiomyocytes (Zhou and Pu, 2016). However, non-myocytes, inclusive of cardiac fibroblasts (CFs), endothelial cells (ECs), vascular smooth muscle cells (VSMCs) and immune system cells comprise the largest fraction of myocardium when measured by cell number. Given the importance of non-myocytes for homeostatic function and disease response in native myocardium, efforts to recapitulate native heterocellular environments in rodent and human fECTs are rapidly expanding. Zhang et al. generated hydrogel-based human fECTs consisting of varying ratios of hPSC-CMs (48-90%) to non-myocytes spontaneously arising from CM differentiation that included cells positive for markers of CFs, VSMCs, and occasionally ECs (Zhang et al., 2013). Interestingly, this group found that the speed of action potential propagation in fECTs linearly increased with CM purity, while generated contractile force showed a biphasic trend with highest force amplitudes occurring at a 70-80% CM fraction. Having a larger than 90% pure CM population was found to impede both tissue formation and contractile function, which likely resulted from the inadequate levels of hydrogel remodeling by insufficient numbers of non-myocytes. These results were further confirmed in a number of follow-up studies that employed either endogenously arising or exogenously added cardiac and dermal fibroblasts or other non-myocyte cell types (Giacomelli et al., 2017b; Li, Y. et al., 2017; Liau et al., 2017; Ronaldson-Bouchard et al., 2018).

2.4. Tissue Architectures

Application of tissue engineering technology for CVD targets is primed for explosive growth over the next decade. Part of the attractiveness of fECT strategies is the ability to fabricate diverse tissue structures that may be exploited for specific engineering and scientific goals (Jackman et al., 2015; Weinberger et al., 2017). These tissue architectures range from stackable cell sheets or cylindrical constructs with relatively small numbers of cells, to large tissue patches of clinically relevant dimensions and rodent-scale single ventricle organoids (Figure 1B). Several major tissue architectures regularly utilized by various groups are detailed below.

2.4.1. Cell Sheets

Traditional two-dimensional (2D) cell culture systems have been extensively used for studies of cardiomyocyte function, metabolism, and drug response. In 2003, the Okano Lab developed thermosensitive poly(N-isopropylacrylamide) (PNIPAAm)-coated cell culture surfaces that, upon cooling to room temperature, could induce the detachment of intact, free-floating 2D cell sheets (Shimizu et al., 2003). Multiple cell sheets can then be stacked into a thicker 3D tissue that can survive for extended times in culture and on infarcted rodent, porcine, canine, and human hearts (Cyranoski, 2018; Kawamura et al., 2017; Shimizu et al., 2009). The main advantage of this scaffold-free approach is that application of sutures during epicardial implantation is not needed since cell sheets spontaneously adhere to the heart surface. Furthermore, the lack of biomaterial scaffold decreases the chance for immune rejection inherent to transplantation of xenobiotic or non-autologous materials. Challenges to using cell-sheet technology primarily lie in mechanical fragility of sheets and the need for a large number of fibroblasts or other matrix secreting non-myocytes to induce successful sheet detachment. Clinical trials with cell sheets made from autologous skeletal myoblasts, and recently allogeneic hPSC-CMs are currently under way in Japan (Imamura et al., 2016; Yui, 2018).

2.4.2. Spheroids

Small, self-aggregating and scaffold-free tissue constructs termed “organoids” are powerful platforms for discovery biology. Contemporary reports detail specific organoids for most human tissue systems (Rossi et al., 2018). As a scientific tool, organoids have extraordinary scalability, given low requirement for cell input (low-thousands) and ease of formation, thus allowing high-throughput screening of heterocellular interactions and drug responses. Like many other cell types used in organoid systems, cardiomyocytes self-aggregate into spheroids and can be supplemented by other cardiac resident cells such as fibroblasts (Kofron et al., 2017), endothelial cells (Giacomelli et al., 2017a), and immune system cells. Recent work with human cardiac spheroids has been aimed at disease modeling (Campbell et al., 2018; Polonchuk et al., 2017). The major drawback of spheroids is the lack of functional architecture, afforded by more structurally advanced systems, that is amenable to direct functional measurement of isometric force generation and electrical conduction. Regardless, we expect to see a growing number of groups publishing chemical and genetic screening studies on spheroid-based fECTs in the near future. Importantly, recent studies with intramyocardial injection of cardiac spheroids in rodents have reported higher cell engraftment rates compared to traditional single cell injection, as well as electrical coupling with host cardiomyocytes, thus warranting future developments towards potential clinical use (Mattapally et al., 2018; Oltolina et al., 2015). Methods to spatially organize and electromechanically integrate spheroids into larger functional structures may further facilitate these efforts (Kim et al., 2018; Ong et al., 2017).

2.4.3. Bundles

Thin, cylindrical cardiac tissue constructs in the form of a cable, ribbon, or ring (here collectively termed “bundles”) remain a mainstay of fECT design with their main appeal being the pseudo-1D geometry that self-directs aligned, anisotropic organization of cells. Structural and functional anisotropy is a hallmark of mature cardiac tissues and is paramount to fast-velocity action potential propagation and strong directional contraction. Cardiac tissue bundles offer the most flexible platform for easy-to-interpret measurements of electrical and mechanical function (Black et al., 2009; Hansen et al., 2010; Hansen et al., 2018; Jackman, C. et al., 2018; Jackman et al., 2016; Nunes et al., 2013; Ronaldson-Bouchard et al., 2018; Schwan et al., 2016). Dependent on dimensions, small bundles have inherent scalability, like spheroids, for medium-throughput screening assays (Mannhardt et al., 2017; Nunes et al., 2013; Ronaldson-Bouchard et al., 2018). Bundles can also be multiplexed to create larger tissue constructs for potential in-vivo regenerative therapy applications, although robust electromechanical communication among stacked individual bundles remains to be demonstrated (Tiburcy et al., 2017; Weinberger et al., 2016; Zimmermann et al., 2006).

2.4.4. Patches

Future regenerative therapies with fECTs will likely require generation of 3D tissues with dimensions sufficient to cover the epicardial surface above an infarct region. Generation of thin multilayer fECTs, or patches, has thus been a major research area in therapeutic cardiac tissue engineering. In their simplest form, patches consist of multiple layers of randomly oriented and electromechanically integrated cardiomyocytes usually formed within hydrogels or porous polymer scaffolds (Bursac et al., 1999; Shadrin et al., 2017). More complex fabrication strategies enable improved patch anisotropy via generation of aligned fibrous scaffolds. Network patches contain microfabricated honeycomb structures that guide local cell alignment through spatially distributed cues while preserving the large surface area of the patch (Bian et al., 2014a; Bian et al., 2014b; Liau et al., 2011). Recently, Shadrin et al. generated hPSC-CM patches with a clinically-relevant surface area of 16cm² and electromechanical properties (conduction velocity of approximately 30 cm/s and specific force of approximately 25 mN/mm²) (Shadrin et al., 2017) approaching metrics of adult human myocardium (Durrer et al., 1970; Hasenfuss et al., 1991; Mulieri et al., 1992). Survival and beneficial effects of NRVM and fECTs upon implantation in rodent and porcine hearts have been demonstrated in multiple studies, but major limitations remain with respect to formation of thicker (greater than 100 μm) tissues and robust electromechanical integration with host myocardium (Gao et al., 2018; Jackman, C.P. et al., 2018; Wendel et al., 2015).

2.4.5. Single Ventricles

In practice, the previously discussed fECT architectures fail to fully recapitulate the complex anatomy of the mammalian heart. Most of the constructs have served to model a subsection of myocardium, or to function as tissue grafts for a section of damaged myocardium, instead of functioning as complete replacement pumps for failing hearts. Several studies have described efforts to generate more realistic heart chamber geometries using hydrogel- and electrospinning-based methods. In an early study, “pouchlike” single ventricle constructs were slipped over the epicardial surface of a whole rodent heart as a ventricular-assist device, although the functional integration with host myocardium was limited (Yildirim et al., 2007). In independent studies, single ventricle constructs were generated by seeding NRVMs in a hydrogel around a modified Foley-catheter balloon (Lee et al., 2008). A similar method was recently used by the same group for generation of ventricle-like cardiac organoid chambers (hvCOC) made from hPSC-CMs with successful demonstration of pressure-volume loop recordings (Li, R.A. et al., 2018). A modified electrospinning method was also used to generate single ventricle human fECTs which the authors cultured in a specialized bioreactor that allowed for continuous hemodynamic monitoring (MacQueen et al., 2018). While interesting methodologically, these single ventricle engineering approaches appear rather difficult to scale up and deliver for human use, and they would also require large numbers of cardiomyocytes if used for higher throughput drug testing applications.

3. Biomaterial Scaffolds for Cardiac Tissue Engineering

Mammalian myocardium is a complex, heterocellular environment built upon, and residing within, a highly specialized extracellular matrix (ECM) (Nielsen et al., 2017; Rienks et al., 2014). The cardiac ECM is composed of structural (e.g. collagen I/III) and basement membrane (e.g. laminin/collagen IV) proteins along with non-structural glucosoaminoglycans and proteoglycans. Human cardiac ECM provides scaffolding for nearly 10 billion cells organized in a compact, heavily vascularized environment (Tirziu et al., 2010). Some estimates place cardiomyocyte density in human adults near 30,000 cells/mm³, all surrounded by a dense vascular network (~3,000 capillaries/mm³) necessary for successful nutrient exchange under high metabolic demand (Bergmann et al., 2015; Hudlicka et al., 1992). At a molecular level, cardiac ECM provides biochemical cues through extracellular protein binding sites along with reservoirs of secreted paracrine factors. From a tissue and organ standpoint, the cardiac ECM possesses advanced biophysical properties that provide both rigidity and elasticity in response to dynamic physiology while also facilitating coordinated transmission of electrical and mechanical signals throughout the myocardium.

Design of biomaterials to recapitulate the native cardiac ECM architecture and composition through various stages of heart development would be an ideal approach to generate high fidelity, biomimetic fECTs (Williams et al., 2014). Unfortunately, no current technology can accurately replicate the complexity of the native ECM environment. Whilst cell sheet and spheroid methods can generate scaffold-free tissues, most studies utilize biomaterial scaffolds to spatially organize cardiomyocytes in a desired 3D geometry while also presenting important cues for their maturation. An ideal 3D scaffold should provide: 1) biomimetic diversity of binding sites to engage functionally relevant integrins in both cardiomyocytes and non-myocytes, 2) sufficient numbers of cell-binding sites to permit physiological tissue density, 3) capacity for significant remodeling of tissue structure to allow rapid cell spreading, alignment, and replacement of original with cell-secreted ECM, 4) appropriate biomechanical properties to enable continuous tissue contractions, 5) minimal immunogenicity when transplanted in vivo, and 6) scalability to allow generation and implantation of tissues of clinically relevant size. As described below, strategies to achieve these requirements have utilized natural and synthetic hydrogels, synthetic polymer scaffolds, decellularized native myocardial ECM, nanomaterial scaffolds, and 3D bioprinting. Table 1 lists many of the most commonly used biomaterials for cardiac tissue engineering while Figure 2 illustrates strategies for deployment of biomaterials in 3D architectures for research and potential therapeutics.

Table 1.

Common biomaterial substrates for functional cardiac tissue engineering.

Scaffold biomaterial	Source	FCTE type supported	Advantages	Disadvantages	Example references
Collagen I/B114Gelatin	Nature-derived	Bundle; Patch	Native cardiac ECM protein; scalable	High Stiffness; slow gelation; batch-to-batch variability	(Eschenhagen et al., 1997; Lee et al., 2017; Nunes et al., 2013; Turnbull et al., 2014)
Fibrin	Nature-derived	Bundle; Patch	Native protein; fast gelation; high biodegradability; high elasticity; scalable	Batch-to-batch variability; rapid biodegradability	(Hirt et al., 2014 Ronaldson-Bouchard et al., 2018; Shadrin et al., 2017; Yuan Ye et al., 2011; Zhang et al., 2013)
Matrigel: Collagen IV + laminin	Murine sarcoma	Bundle; Patch	Basal lamina components; commercial product	Xenogencity; batch-to-batch variability; not a standalone product	(Cashman et al., 2016; Jackman, Li, and Bursac, 2018a; Jackman, Ganapathi, et al., 2018b; Shadrin et al., 2017)
Alginate	Brown algae	Injectable	Native cardiac ECM-like properties; Scalable; 1st clinical trials	Bio-inertness; difficulty to remodel	(Liberski et al., 2016; Mann et al., 2016; Rao et al., 2015)
Polyethylene Glycol (PEG)	Synthetic	Bundle; Injectable	Low-immunogenicity; Readily modifiable; scalable	Bio-inertness; difficulty to remodel	(Jongpaiboonkit et al., 2008; Kraehenbuehl et al., 2009; Rane et al., 2011)
Hyaluronic Acid (HA)	Synthetic	Bundle	Readily modifiable; tunable stiffness; scalable	Bio-inertness; difficulty to remodel	(Young and Engler, 2011)
Poly(DL-lactic-co-glycolic) acid (PLGA)	Synthetic	Bundle	High porosity; clinically approved	High stiffness	(Bursac et al., 2007; Bursac et al., 1999)
Polyglycerol sebacate (PGS)	Synthetic	Bundle	Tunable mechanical properties	Fatigability; low porosity; long degradation time	(Engelmayr Jr. et al., 2008)
Polycaprolactone (PCL)	Synthetic	Single ventricle	Elastomeric; cell alignment	Scalability; 3D porosity; long degradation time	(MacQueen et al., 2018)
PEG-Fibrinogen	Hybrid	Bundle	Early hiPSC to CM differentiation control	Density of cell attachment sites	(Kerscher et al., 2016)
PLGA/PGA-laminin	Hybrid	Electro-spinning	Ease of processing; clinically approved	Scalability, 3D porosity, mechanical properties	(Papadaki et al., 2001; Yu et al., 2014)
Rodent ECM Decellularized	Nature-derived	Bundle; Whole heart	Native cardiac ECM; high biocompatibility	Xenogenicity; scalability; reproducibility	(Lu et al., 2013; Ott et al., 2008)
Porcine ECM Decellularized	Nature-derived	Bundle; Whole heart	Native cardiac ECM; human-scale tissue; 1st in clinical trial	Xenogenicity; scalability; reproducibility	(Freytes et al., 2014; Hodgson et al., 2017; Schwan et al., 2016; Spang and Christman, 2018; Wainwright et al., 2010)
Human ECM Decellularized	Nature-derived	Bundle; Whole heart	Native human cardiac ECM; not	Availability; reproducibility	(Becker et al., 2017; Guyette et al., 2016; Sanchez et al., 2015)
Apple Decellularized	Nature-derived	Patch	Scalable	Xenogenicity; 3D porosity; unclear safety & efficacy	(Modulevsky et al., 2014)
Leaf Decellularized	Nature-derived	Patch	Preset vascular-like network; Scalable	Xenogenicity; 3D porosity; unclear safety & efficacy	(Gershlak et al., 2017)

Becker et al. (2017); Bursac et al. (2007); Bursac et al. (1999); Cashman et al. (2016); Engelmayr Jr. et al. (2008); Eschenhagen et al. (1997); Freytes et al. (2014); Gershlak et al. (2017); Guyette et al. (2016); Hirt et al. (2014); Hodgson et al. (2017); Jackman et al. (2018a); Jackman et al. (2018b); Jongpaiboonkit et al. (2008); Kerscher et al. (2016); Kraehenbuehl et al. (2009); Lee et al. (2017); Liberski et al. (2016); Lu et al. (2013); MacQueen et al. (2018); Mann et al. (2016); Modulevsky et al. (2014); Nunes et al. (2013); Ott et al. (2008); Papadaki et al. (2001); Rane et al. (2011); Rao et al. (2015); Ronaldson-Bouchard et al. (2018); Sanchez et al. (2015); Schwan et al. (2016); Shadrin et al. (2017); Spang and Christman, 2018); Turnbull et al. (2014); Wainwright et al. (2010); Young and Engler (2011); Yu et al. (2014); Yuan Ye et al. (2011); Zhang et al. (2013).

Figure 2. Materials and techniques for engineering cardiac tissues.

Materials and techniques most frequently used for fECT engineering and therapy include: A) hydrogel scaffolds made from either natural or synthetic materials formed with 3D molding or bioprinting; B) fibrous scaffolds generated from synthetic materials including PLGA, PGS and others; C) nanofibrous scaffolds generated through electrospinning or extrusion spinning; D) decellularized tissues either intact or homogenized and crosslinked into a hydrogel; and E) injectable scaffolds to provide bulking and/or beneficial remodeling of the treated heart.

3.1. Hydrogels

Hydrogels are the most widely used biomaterials for cardiac tissue engineering due to their useful biochemical and biophysical properties. This includes the ability to generate numerous geometries in a highly scalable fashion. Successful deployment of hydrogels for fECTs generally requires a casting mold to achieve desired form although 3D bioprinting methods can also serve to provide spatial boundaries. In addition to the mold, anchoring supports within the hydrogel provide mechanical loading necessary for tissue remodeling along with support for cardiomyocyte growth, alignment and maturation. Hydrogels, however, are generally soft scaffolds with stiffnesses less than that of native myocardium. This can create handling difficulties for early engineered tissues which are susceptible to researcher-induced mechanical forces. Nevertheless, as the fECTs mature in culture, the original hydrogel is remodeled through deposition of ECM proteins from seeded cells and by cell-mediated compaction of the original hydrogel substrate. The optimal hydrogel substrate has been the subject of extensive study, but several natural protein products utilized regularly include collagen I, the blood clotting protein fibrin, and the collagen-IV/laminin composite Matrigel. Synthetic and carbohydrate-based hydrogels have also been explored for supporting fECT form and function although they often lack native cell-binding motifs. Accordingly, natural protein-based hydrogels continue to demonstrate the most mature and functional end-products. While conjugating multiple (natural and/or synthetic) hydrogels may further promote biomimetic properties of fECTs, no consensus on the optimal hydrogel milieu exists at this time.

3.1.1. Collagen I

Collagen I and collagen III are the most abundant structural proteins found in the ECM of mammalian myocardium (Takawale et al., 2015). Collagen I fibrils provide the majority of the tensile strength whereas addition of collagen III to collagen I has been found to increase tissue elasticity (Asgari et al., 2017). Thus, collagen I hydrogel has long been utilized for production of fECTs using cardiomyocytes from different species (Eschenhagen et al., 1997; Naito et al., 2006; Ruan et al., 2016; Tiburcy et al., 2017; Turnbull et al., 2014; Zhang et al., 2017). Nevertheless, collagen I scaffolds have several limitations for use in cardiac tissue engineering because 1) they are not easily remodeled to yield physiological cell density and connectivity, 2) they do not stimulate endogenous matrix secretion by cardiac fibroblasts, and 3) are stiffer than other hydrogel scaffolds. As a result, collagen I may not provide ideal conditions for cardiomyocyte maturation and macroscopic contractions. Furthermore, collagen I hydrogels have a long gelling time which can lead to settling of seeded cells during polymerization, yielding nonuniform cell density. A denatured form of collagen I, gelatin, has also been used for coating of 2D culture surfaces, but it requires chemical processing to form a 3D hydrogel that is compatible with cardiomyocyte culture. Specifically, gelatin methacrylol (GelMA) hydrogels have tunable biophysical and biochemical properties and have been increasingly utilized for generation of fECTs (Shin et al., 2016), however, functional properties of cardiomyocytes embedded inside GelMA hydrogels remain far inferior to those in collagen I and fibrin-based tissues.

3.1.2. Fibrin

Fibrin is the product of thrombin mediated cleavage of its pro-protein fibrinogen. Fibrin deposition is crucial to hemostasis in the initial clotting cascade and for the process of early wound healing (Laurens et al., 2006). Advantageous to cardiac tissue engineering, fibrin is extensively remodeled and degraded during wound healing which makes it a good candidate for achieving high cardiomyocyte spreading, density, and interconnectivity in fECTs. Compared to collagen I fibrils, fibrin fibrils have significantly lower tensile stiffness and higher elasticity, comparable to those of native myocardial ECM (Collet et al., 2005; Grassl et al., 2002; Yang et al., 2007). Furthermore, fibrin-based hydrogels stimulate collagen and other ECM secretion from cells within the hydrogel (Ross and Tranquillo, 2003) and their short gelation time can be tuned by both concentration of thrombin and temperature, allowing for uniform seeding of cells in 3D geometries. Given these benefits, along with the fact that it is an FDA approved material, fibrin scaffolds have gained wide adoption in the cardiac tissue engineering field (Hirt et al., 2014; Ronaldson-Bouchard et al., 2018; Schaaf et al., 2011; Shadrin et al., 2017; Yuan Ye et al., 2011; Zhang et al., 2013). Disadvantages to fibrin hydrogels as an fECT scaffold are significant batch-to-batch variability and a fast degradation rate, which can be overcome by inclusion of the anti-fibrinolytic agents aminocaproic acid and aprotinin (Khodabukus and Baar, 2009) in batch-dependent concentrations. Secondary cross-linking of fibrinogen with poly(ethylene glycol), PEG, to generate PEGylated fibrin permits tuning of scaffold porosity and mechanical properties and has been shown to also support the generation of mature fECTs (Kerscher et al., 2016).

3.1.3. Matrigel

While collagen I is the most abundant ECM protein in myocardium, cardiomyocytes directly interact with basal lamina proteins (e.g. laminin, collagen IV). This prompted supplementation of Matrigel (a commercially available basal lamina extract isolated from mouse Engelbreth–Holm–Swarm tumors) to collagen I and fibrin hydrogels at the time of tissue fabrication. Unlike collagen I and fibrin, Matrigel does not require enzymatic or pH modulation to induce gelation since a rise in temperature is sufficient to initiate this process. In addition to use in 3D fECTs (Cashman et al., 2016; Jackman, C. et al., 2018; Jackman et al., 2016; Jackman, C.P. et al., 2018; Zimmermann et al., 2002), Matrigel is widely utilized as a 2D cell-surface coating for culture of hPSCs and cardiomyocytes. A number of studies have shown requisite roles of Matrigel in achieving advanced structural and functional maturation of 3D-cultured cardiomyocytes. Naturally, the major drawback for use of Matrigel-containing fECTs in clinical applications is the origin of Matrigel from murine tumor cells. However, engineering Matrigel-free fECTs necessitates careful optimization of culture media and cell composition (Naito et al., 2006; Tiburcy et al., 2017). Encouragingly, recent studies have identified PEG-based synthetic hydrogels that could serve as an alternative to Matrigel for common cell culture applications (Nguyen et al., 2017).

3.2. Synthetic Polymer Scaffolds

Synthetic polymer scaffolds designed for cardiac tissue engineering are attractive given that there is nearly infinite potential for their modification to suit the biophysical and biochemical needs of fECTs. Synthetic polymers can be designed to take on different 3D geometries, serving as hydrogels in the case of PEG or PNIPAAm, or as fibrous mesh scaffolds if generated from poly(DL-lactic-co-glycolic acid) (PLGA) or poly(glycerol sebacate) (PGS). Successful implementation of synthetic polymers as fECT scaffold replacements or additives necessitates 1) high biocompatibility and porosity, 2) adequate biodegradability and biomechanical characteristics, and 3) low immunogenicity.

PEG is one such candidate given that it is FDA approved, readily modifiable, and has low immunogenicity. The inherently inert and non-remodeling nature of PEG requires incorporation of cell-attachment and/or matrix metalloproteinase (MMP)-sensitive peptide motifs to support cell attachment, remodeling, and formation of physiologically dense tissues. Specifically, conjugating RGD integrin-binding domains onto the PEG backbone can improve cardiomyocyte viability in 3D culture compared to unmodified substrate (Jongpaiboonkit et al., 2008), while biodegradability and stimulation of cardiogenic differentiation of PEG can be improved by adding MMP-sensitivity (Kraehenbuehl et al., 2011). Similarly, use of PEG to crosslink other hydrogels, including collagen, fibrin, hyaluronic acid (HA) (Young and Engler, 2011) and gelatin (Lee et al., 2017) enables additional control of material stiffness to yield pro-maturation benefits on cardiomyocytes.

In addition to hydrogels, synthetic polymer scaffolds with diverse 3D architectures, chemical composition, porosity, degradability, and mechanical properties have long been pursued as biomaterials for cardiac tissue engineering. Specifically, fibrous scaffolds made from PGA or PLGA, two FDA-approved materials, were early candidates for generation of fECTs given their high porosity, tunable biodegradation, and processability (Bursac et al., 2007; Bursac et al., 1999; Papadaki et al., 2001). Unfortunately, the high stiffness of PLGA is not compatible with the need for extensive structural remodeling and macroscopic contractions of fECTs, which can limit the ultimate maturation state of cardiomyocytes. PGS-based 3D scaffolds also generated considerable interest among cardiac tissue engineers given their bioresorbable nature, elastomeric properties, and ability to precisely control their mechanical properties (Engelmayr et al., 2008). When combined with laser etching to generate a honeycomb-like porous architecture, PGS scaffolds exhibited mechanical anisotropy resembling that of ventricular tissue, however, the compartmentalized nature of the scaffold and low porosity created significant impediments to electrical conduction throughout the tissue. Other synthetic elastomeric biomaterials including polyurethane (PU) (Alperin et al., 2005) and poly(octamethylene maleate (anhydride) citrate) (POMaC) (Zhang et al., 2016) have been tested as 3D scaffolds in cardiac tissue engineering, but they are limited by low porosity and low biodegradation as well as material fatigue induced by repeated cycles of stretch.

While tremendous versatility in the design of 3D synthetic scaffolds offers promise for the field of cardiac tissue engineering, achieving biomimetic cardiomyocyte density, anisotropic structural organization, and functional maturation using synthetic substrates comparable to their natural counterparts has not yet been achieved. To address deficiencies of individual scaffolding substrates, a logical leap is to multiplex synthetic and natural biomaterials to synergize advantageous characteristics such as mechanical stability and unique geometries of polymer scaffolds with the high density of bioactive sites and porosity of natural hydrogels. We expect to see a proliferation of studies exploring these combined strategies for successful engineering of robust, therapeutic fECTs.

3.3. Decellularized Scaffolds

De novo generation of 3D scaffolds which are fully mimetic of native cardiac ECM could be a nearly impossible task, regardless of the extensive efforts detailed above. This notion prompted the development of tissue and organ decellularization procedures to extract native myocardial ECM. Decellularized myocardial ECM can be applied directly as a nature-formed scaffold with intact 3D architecture, or as a cardiomyocyte-compatible hydrogel. Methods to decellularize whole hearts and other organs typically use enzymatic and/or detergent-based methods to remove the cells while retaining the most ECM components possible (Gilpin and Yang, 2017). A key quality measure is evaluation of retained DNA as a proxy for level of decellularization and potential for immunogenicity of the scaffold. Ott et al. provided the first proof-of-concept study by reseeding NRVMs onto decellularized whole rat hearts and demonstrating NRVM viability, inotropic response to phenylephrine, and rudimentary pump function after 4 weeks of bioreactor culture (Ott et al., 2008). Currently studied decellularized cardiac ECM scaffolds include those sourced from rodent, pig, and human hearts with the first clinical trial of myocardial injection of a porcine-sourced product currently underway (Spang and Christman, 2018). Noncardiac ECM products are already commercially available for a multitude of human indications (Badylak et al., 2009), which may speed future clinical use of decellularized heart ECM.

3.3.1. Animal

Given their large size and ready availability, porcine hearts are a great source of cardiac ECM. Significant work is underway to optimize porcine heart decellularization with current techniques demonstrating removal of more than 98% of DNA and immunogenic substances (Hodgson et al., 2017; Weymann et al., 2011). Generation of fECTs using porcine myocardial ECM as a solubilized hydrogel matrix or a decellularized 3D tissue scaffold holds promise for therapeutic and experimental applications. Nevertheless, the ability to preserve native ECM component stoichiometry, such as basal lamina versus structural protein fractions, with reproducibility between different donor hearts remains a significant challenge. Further demonstrating compatibility with human cardiomyocytes while maintaining biomechanical stability and remodeling properties is also required before cardiac ECM derivatives compete with state-of-the-art fECTs made of fibrin and collagen I hydrogels. Additionally, recellularization of the whole decellularized hearts with the goal of organ transplantation (Kitahara et al., 2016; Lu et al., 2013; Ott et al., 2008) is faced with a number of daunting challenges including the need for large cell numbers (upwards of 10 billion assuming no cell loss during recellularization), redistribution of specific cell types to their original positions, preservation of original viscoelastic properties, and significant risks of arrhythmias, coronary artery thrombosis, and immune rejection.

3.3.2. Human

Given that future use of animal-derived scaffolds may be limited by their xeno-immunogenicity, decellularized human cardiac ECM may be a more viable alternative for clinical use. While some progress with human ECM-based scaffolds has been made (Guyette et al., 2016; Sanchez et al., 2015), the function of the resulting fECTs has been very low with generated forces and pressure gradients in decellularized scaffolds and whole hearts amounting to only 100uN and 2.5mmHg, respectively (Guyette et al., 2016). Processing smaller human cardiac tissue samples, such as those from surgical biopsies, holds promise for generation of solubilized hydrogels for myocardial injection therapies (Becker et al., 2017). Overall, the limited supply of human-derived ECM components will likely constrain their use to tools for fECT-based discovery science rather than as scaffolds for wide-spread clinical use.

3.3.3. Plant

Supply and immunogenicity challenges for animal- and human-derived scaffolds have prompted research on plant-derived biomaterials for use in human tissue engineering (Fontana et al., 2017). Cellulose scaffolds such as those derived from decellularized apple hypanthium have supported long-term survival of cultured C2C12 cells (Modulevsky et al., 2014) and demonstrated transient immune response upon subcutaneous implantation that gradually disappeared by 8 weeks in vivo (Modulevsky et al., 2016). A similar concept was utilized to produce “pre-vascularized” network scaffolds made from decellularized spinach and parsley leaves (Gershlak et al., 2017). Seeding leaf-based scaffolds with human ECs and CMs generated 2D fECTs that maintained spontaneous contractility up to 21 days in culture. Along with other cellulose-based scaffolds (Entcheva et al., 2004; Wippermann et al., 2009), these novel materials await careful evaluation of their ability to support in vitro engineering and in vivo therapeutic efficacy of 3D fECTs while hopefully limiting immunogenicity.

3.4. Nanofibrous Scaffolds and 3D Bioprinting

While bulk cell seeding of hydrogels, synthetic polymer scaffolds, and decellularized heart ECM represents a more traditional approach to generation of fECTs, recent advances in nanomaterial technology may drive the design of more complex heterocellular microenvironments mimicking that of native myocardium. Specifically, increasingly sophisticated methods for electrospinning and 3D bioprinting offer the ability to use natural and/or synthetic biomaterials while controlling scaffold architecture and cellular composition in a spatially precise fashion. As a result, an increasing number of researchers have been developing new nanomaterial-based strategies to harness fundamental biochemical and biophysical cues from native cardiac ECM that would allow generation of more functional ECTs.

3.4.1. Nanofibrous Scaffolds

Nanofibrous scaffolds that faithfully recapitulate native ECM topography and bioactivity may present unique platforms for generation of functional myocardium as well as other cardiovascular tissues including heart valves, vascular outflow tracts, and blood vessels (Capulli et al., 2016). Electrospinning functions by extrusion of electrically-charged polymers through a small orifice to create nanofibers with precisely controlled dimensions and the ability to generate higher-order 3D geometries based on fiber collection methodologies (Bhardwaj and Kundu, 2010; Senel Ayaz et al., 2014). Most electrospinning studies have utilized synthetic polymers (PLGA, Polycaprolactone (PCL), etc.) and have generated aligned human fECTs with therapeutic potential (Li, J. et al., 2017). Recent methodological advances have permitted increased throughput along with incorporation of native proteins (e.g. gelatin, collagen) into spun nanofibers, thus enabling the formation of aligned patches and single ventricle fECTs made from NRVMs or hPSC-CMs (Badrossamay et al., 2010; Kharaziha et al., 2013; MacQueen et al., 2018). Nevertheless, current nanofiber scaffolding technologies are plagued by several limitations that must be overcome to enable generation of highly functional and therapeutically relevant fECTs including: 1) low porosity that precludes deep penetration of seeded cells, 2) inability to efficiently incorporate various cardiac ECM proteins, and 3) difficulty scaling-up current technologies for human applications.

3.4.2. 3D Bioprinting

Bioprinting is a very promising technology permitting versatile generation of large heterocellular fECTs with complex 3D architecture. Early work by Gaetani et al. utilized RGD-conjugated alginate or HA/alginate matrices with human cardiac progenitor cells (hCPCs) to 3D-print a six-layered, 2cm × 2cm fECT patch which retained high cell viability for 2 weeks in vitro (Gaetani et al., 2012) and 4 weeks implanted on infarcted rat hearts (Gaetani et al., 2015). Arguably the most exciting prospect of soft-tissue 3D bioprinting is the generation of thick perfusable fECTs. Kolesky et al. demonstrated this possibility by generating 1cm-thick, perfused osteogenic tissues printed from a cell-laden dual-crosslinked gelatin-fibrin matrix (Kolesky et al., 2016). Accordingly, a very recent study reported the construction of heterocellular (mouse iPSC-CMs + human ECs), 3D-bioprinted fECTs in a PEG-fibrinogen hydrogel which exhibited improved connectivity to host vasculature upon subcutaneous implantation in NOD-SCID mice (Maiullari et al., 2018). Overall, 3D bioprinting for the purpose of functional cardiac tissue engineering appears to be in early stages of development, awaiting the discovery of optimal bioinks and cell mixtures, but preliminary studies and fast-paced advances suggest enormous potential for creating the next-generation of therapeutic fECTs.

4. Therapeutic Use of Biomaterials without Cells

Intramyocardial or coronary artery injections of exogenous cells in the heart, from initial studies with skeletal myoblasts (Taylor et al., 1998) to more recent transplantations of large numbers of hPSC-CMs into non-human primates (Liu, Y.W. et al., 2018), have demonstrated the potential for cell-based myocardial recovery in humans (Garbern and Lee, 2013). Attempts to improve engraftment of transplanted cells by co-delivery of biomaterials have revealed that injected biomaterials alone, without cells, can also improve compromised cardiac function (Christman et al., 2004). It is now well-understood that injectable biomaterials exert multiple therapeutic benefits to post-MI hearts including: 1) amelioration of maladaptive remodeling induced by altered inflammatory response, 2) reduction of abnormal mechanical loading induced by tissue bulking, and 3) recruitment of endogenous regenerative cells (Reis et al., 2016).

Of the three therapeutic modalities, the most clinically advanced concept is tissue bulking (Table 2 lists recent clinical trial utilizing tissue bulking). The leading biomaterial candidate for this strategy is the inert seaweed-derived compound, alginate (Liberski et al., 2016). In the PRESERVATION I trial, a proprietary sodium alginate and calcium gluconate product was introduced via intracoronary perfusion 2-5 days after coronary artery stenting in patients with ST-elevation myocardial infarction (STEMI) (Rao et al., 2015). At 6 months post injection, alginate-treated and placebo saline-injected patients exhibited similar end-diastolic volumes (a direct measurement of LV chamber dilation) and no difference in mortality events, demonstrating clinical safety but not efficacy of the therapy (Rao et al., 2016). On the other hand, a smaller 12-month trial called AUGMENT-HF that performed intramyocardial injection of a proprietary calcium-alginate solution (Algisyl) vs. a saline control has demonstrated a small, yet significant, improvement in peak VO2 (an oxygen utilization measure) and 6-minute walk tests which are clinically relevant measures of functional capacity in patients with heart failure (Mann et al., 2016). The phase II trial of AUGMENT-HF is already planned with the targeted completion in 2020.

Table 2.

Recent clinical trials of cell-free biomaterials for heart disease.

Trial Name	Clinicaltrials.gov Identifier	Substrate	Route of Administration	Patient numbers	Cardiac Injury	Results	Ref
PRESERVATION I	NCT01226563	Sodium alginate + Calcium Gluconate	Intracoronary	303	Recent STEMI	Complete; No functional improvement; No safety concerns	(Rao et al., 2015; Rao et al., 2016)
AUGMENT HF	NCT01311791	Calcium Alginate	Intramyocardial	58	Congestive Heart Failure; EF < 35%	Complete; Mild improved peak VO2; Mildly improved 6-min walk test	(Mann et al., 2016)
AUGMENT HF II	NCT03082508	Calcium Alginate	Intramyocardial	> 200 planned	Congestive Heart Failure; EF < 35%	Not yet enrolling	N/A
VentriGel	NCT02305602	Porcine ECM	Trans-endocardial	15	Remote STEMI	Closed; no published results	(Self-Naraghi et al., 2013)

Mann et al. (2016); Rao et al. (2015); Rao et al. (2016); Self-Naraghi et al. (2013).

While clinical trials with tissue-bulking agents have shown mixed results, efforts to develop new injectable biomaterials with pleiotropic effects on the diseased heart are ongoing. Specifically, bioactivity appears to be an important biomaterial attribute for the induction of sustained functional benefits. This potentially explains why bio-inert materials, such as alginate or unmodified PEG (Rane et al., 2011), fail clinical efficacy studies. Hence, a soon to be completed Phase I safety study is testing transendocardial injection of a decellularized porcine heart hydrogel product (VentriGel) in 15 patients who have experienced recent STEMI and who have evidence of maladaptive left ventricular remodeling (Seif-Naraghi et al., 2013). As a lack of reproducibility of natural-derived products may hinder wide-spread clinical use, other groups are exploring the therapeutic potential of injectable synthetic hydrogels with built-in bioactive moieties or pro-regenerative gene products (Eckhouse et al., 2014; Purcell et al., 2014; Wang, L.L. et al., 2017). Based on the favorable clinical safety profile, we expect to witness rapid expansion of basic discovery, large animal, and first-in-man studies testing injectable synthetic biomaterials as cardiac therapeutics.

5. The Future of Cardiac Tissue Engineering

The exciting field of cardiac tissue engineering holds potential to greatly improve our understanding of congenital and acquired heart diseases as well as facilitate development of new regenerative therapies for MI and CHF. A simple PubMed search for “cardiac tissue engineering” produces more than 1000 articles published annually over the past three years. Further narrowing this search string by adding the term “biomaterials” produces over 200 annual articles, underlining the importance of technological advances in biomaterial science for the purpose of creating new, or improved, fECTs. Given its tremendous promise, we anticipate that biomaterial research for cardiac tissue engineering and therapeutics will continue to follow its current steep trajectory.

As evidenced by recently active clinical trials discussed in Section 4, the first successful therapies for heart disease based on tissue-engineering principles will likely rely on introduction of cell-free biomaterials as injectable scaffolds. Such scaffolds may serve to provide mechanical stability for acutely injured myocardium after MI, or by acting as a substrate to direct positive myocardial remodeling in the setting of chronic ischemic or infiltrative cardiomyopathies. Modifying bioactivity, mechanical, and degradation properties of injectable biomaterials will allow design of multi-faceted, spatiotemporally controlled therapies to precisely modify immune/inflammatory and regenerative signaling leading to positive clinical outcomes.

The next generation of injectable biomaterials will likely involve incorporation of cell-sourced products to improve their bioactivity. Early examples of this strategy have included use of secreted ECM modulatory proteins (MMPs, TIMP-3) or cardiac-specific miRNAs to improve functional outcomes in animal models of myocardial infarction (Purcell et al., 2014; Wang, L.L. et al., 2017). Moreover, it is evident that the moderate functional improvements seen across a multitude of stem cell-based clinical therapies can be attributed to paracrine action of the transplanted cells (Hodgkinson et al., 2016). Secreted exosomes are currently thought to be likely candidates for such observed therapeutic effects (Sahoo and Losordo, 2014) with the possibility that the exosome-mediated effects could be further augmented via biomaterial delivery through an intramuscular or epicardial route (Liu, B. et al., 2018). Similarly, injection of PLGA-based microparticles containing oxygen-releasing moieties (Fan et al., 2018) or fused with stem cell membrane fragments (Tang et al., 2017) offer additional biomaterial-based approaches to improve heart function after myocardial injury. While the exact therapeutic cargos, reproducibility, and longevity of these strategies remain unknown, their cell-free and likely non-immunogenic nature suggest future developments toward potential clinical applications.

Despite being one of the original goals of cardiac tissue engineering, the clinical use of biomimetic fECTs for epicardial transplantation remains the furthest from completion. While the first clinical trials with human iPSC derivatives for treatment of ischemic heart disease appear to be safe (Menasche et al., 2018), applications of hiPSC-derived fECTs in humans await development of reproducible methods to overcome several technical hurdles including: 1) invasive surgical procedures, 2) immature cardiomyocyte phenotype, 3) small tissue thickness, and 4) lack of electromechanical integration between the grafted fECT and host myocardium (Jackman, C.P. et al., 2018; Shadrin et al., 2017). Recent studies have demonstrated progress towards accelerated structural and functional maturation of hiPSC-CMs in vitro (Ronaldson-Bouchard et al., 2018; Shadrin et al., 2017), which is likely to alleviate arrhythmogenic risks observed in initial cell injection studies in non-human primates (Chong et al., 2014; Liu, Y.W. et al., 2018). However, it is unknown if highly mature cardiomyocytes will be the optimal cell source for achieving thick tissue grafts with the capacity to robustly survive and functionally integrate in vivo. Novel cell engineering and biomaterial strategies employed for fECT construction and delivery will likely play key roles toward improving functional integration as well as promoting long-term viability of engrafted fECTs (Montgomery et al., 2017; Nguyen et al., 2016, 2018; Zhang et al., 2016). True benefit for human disease will be measured in years, rather than weeks-to-months. In the meantime, fECTs will continue to serve as an important in vitro tool for studying human development, biology, and disease including genetic and pharmacological screens to discover and develop new therapeutics. Reproducibility among various research groups, including standardization of cell sources, handling procedures, and animal models of disease would further accelerate progress. As the field moves forward, one can easily envision synergy between biomaterial science and iPSC-based technologies as a powerful, scalable platform for future preclinical and clinical applications in cardiovascular science and medicine.

5. Conclusions

Materializing the promise of cardiac tissue engineering into viable clinical therapies is increasingly becoming a realistic goal. Recent clinical studies with injectable biomaterials are leading ongoing translational efforts, yet synergistic delivery of bioactive cell-based products with biomaterial scaffolds will likely be the next horizon in ameliorating human heart disease. Parallel advances in pluripotent stem cell technology to incorporate adult functionality and patient specificity into engineered cardiac tissues will also drive the field forward. The depth and breadth of efforts toward the goal of functional, human disease modeling and regenerative therapeutics will require coordinated efforts from the diverse pool of talented researchers involved in this exciting field. In this review, we have discussed the present understanding of cell sources, architectures, and biomaterial scaffolds used for creation of functional engineered cardiac tissues (fECTs). Despite significant progress toward generating adult-like fECTs, work remains to demonstrate complete maturation, functional integration, longevity, and human scalability of such constructs. We envision that multiplexing natural and synthetic biomaterials deployed by innovative technologies, such as 3D bioprinting, will further unlock the capacity to create heterocellular, biomimetic fECTs for disease modeling, drug development, and therapeutic applications. Given current progress, we may one day be able to both build and rebuild a heart.

Abbreviations:

(CM)

cardiomyocyte

(CVD)

cardiovascular disease

(CHF)

congestive heart failure

(EC)

endothelial cell

(fECT)

functional engineered cardiac tissue

(ECM)

extracellular matrix

(hPSC)

human pluripotent stem cell

(LV)

left ventricular

(MMP)

matrix metalloproteinase

(MI)

myocardial infarction

(NMVM)

neonatal mouse ventricular myocyte

(NRVM)

neonatal rat ventricular myocyte

(PCL)

polycaprolactone

(PLGA)

poly-DL-lactic-co-glycolic acid

(PEG)

polyethylene glycol

(PGS)

polyglycerol sebacate

(PNIPAAm)

poly(N-isopropylacrylamide)

(POMaC)

poly(octamethylene maleate (anhydride) citrate)

(PU)

polyurethane

(STEMI)

ST-segment elevation myocardial infarction

(VSMC)

vascular smooth muscle cell