Recapturing embryonic potential in the adult epicardium: Prospects for cardiac repair

Abstract Research into potential targets for cardiac repair encompasses recognition of tissue‐resident cells with intrinsic regenerative properties. The adult vertebrate heart is covered by mesothelium, named the epicardium, which becomes active in response to injury and contributes to repair, albeit suboptimally. Motivation to manipulate the epicardium for treatment of myocardial infarction is deeply rooted in its central role in cardiac formation and vasculogenesis during development. Moreover, the epicardium is vital to cardiac muscle regeneration in lower vertebrate and neonatal mammalian‐injured hearts. In this review, we discuss our current understanding of the biology of the mammalian epicardium in development and injury. Considering present challenges in the field, we further contemplate prospects for reinstating full embryonic potential in the adult epicardium to facilitate cardiac regeneration.


| INTRODUCTION
The recognition of the regenerative capacity of lower vertebrate 1 and neonatal mammalian 2 hearts has reinvigorated the search for endogenous reparative pathways that may be translated to regenerate the human heart after injury. Such pathways emulate the complex, coordinated events that occur during embryonic development, upon which our understanding of cardiac formation and composition is founded. 3 The adult mammalian heart lacks regenerative ability, owing to the absence of tractable cardiomyocyte (CM) precursors and the inability of mature CMs to proliferate after injury. [4][5][6] Heart failure is commonly caused by myocardial infarction (MI), with major morbidity and mortality consequences worldwide. 7 Following MI, damaged CMs are replaced by an expansion of tissue-resident cardiac fibroblasts (CFb), which respond by transitioning to myofibroblasts, depositing collagen to the fibrotic scar [8][9][10] to prevent ventricular wall rupture. 9,11 However, excessive fibrosis impairs contractile function and initiates a deleterious cycle of myocardial loss and adverse remodeling. 12 Strategies to limit myocardial damage (eg, restoring adequate perfusion, 13 replacing dying CMs 4,5 and reducing unwarranted fibrosis 14,15 ) are thus vital to prevent heart failure.
A central regulator of the processes outlined above, and therefore an important therapeutic target, is the epicardium.
Commonly described as the outermost layer of the heart in vertebrates, this mesothelial tissue is a source of multipotent progenitors, growth factors, and extracellular matrix (ECM) components during cardiac development and following injury. 16 In this review, we discuss the prospects for reinstating embryonic potential in the adult epicardium to facilitate cardiac regeneration.
We will highlight recent literature and new technologies that are proving invaluable in the quest to harness the epicardium for repair.
2 | RECAPTURING EMBRYONIC POTENTIAL 2.1 | Epicardium: Developmental origin, formation, and function The epicardium originates from a transient structure in the developing embryo called the proepicardial organ (PEO), adjacent to the septum transversum, and proximal to the looping heart tube and sinus venosus (SV). 3,17,18 Lineage tracing studies in the mouse suggest that the PEO derives from Nkx2.5 and Isl1 expressing lateral plate mesoderm, the source of most other cardiac precursor cells. 19 PEO formation is induced by a carefully controlled balance of bone morphogenetic proteins (BMPs) and fibroblast growth factors (FGFs), which determine whether LPM adopts a myocardial or epicardial fate. 17 From embryonic day (E)9.5 in mouse, epicardial progenitors residing in the PEO detach and migrate to envelop the developing myocardium. Epicardial formation from the PEO is widely conserved, having been described in all vertebrate species examined, including zebrafish, 20 Xenopus, 21 chicken, 22 mice, 23 rats, 24 and humans. 25 Due to the scarcity of available tissue to study early human embryology, insights into the formation of the human epicardium are limited. However, examination of carnegie stage (CS) 12 embryonic sections by light microscopy first revealed villous protrusions of mesothelial cells extending from the sinus wall onto the dorsal side of the ventricle and spreading over the heart as a squamous epithelium, 25 supporting a conserved mechanism. Histological analysis of CS11 embryos (4 weeks postcoitum, equivalent to mouse E10) suggested that epicardial formation was already underway at this earlier stage, with "round, progenitor-like cells" described to overlie the compact myocardial layer, albeit these cells were not distinguished by marker analysis and the PEO protrusions were not captured in these samples. 26 Epicardial formation is complete in human embryos by CS15 and is characterized by expression of markers, such as WT1, TCF21, GATA5, TBX18, cytokeratin, and podoplanin, 26 consistent with other species. Unlike the monolayer structure in rodent, chick, and zebrafish hearts, the multilayered human epicardium consists of a mesothelium overlying connective tissue and an expanded subepicardial space containing elastic fibers and blood vessels. 27 This species difference emerges during fetal stages 26,27 and, during adulthood, adipose tissue depots accumulate within the subepicardial space, which can profoundly influence cardiac function. 28 The extent and depth of our understanding of epicardial origin, formation, and role in supporting mammalian cardiac development is due, in large part, to the ease of genetic manipulation of the mouse embryo, for lineage tracing and "knockout" developmental studies.
The epicardium serves three primary functions to support cardiac development.

| Direct cellular contribution
Prior to maturation of the epicardial layer, successive, regionalized waves of epithelial-mesenchymal transition (EMT) between E11.5 and E13.5 mobilize epicardium-derived mesenchymal cells, as precursors for atrioventricular valve mesenchyme, CFb and mural cells (pericytes and vascular smooth muscle cells [vSMCs]). 16,29 A variety of myocardial-derived signals promote EMT, including transforming growth factor-β (TGFβ) and FGFs, although roles specifically in EMT have been difficult to distinguish from roles in migration and fate, as the processes are intricately linked. 30 Progenitors and specialized cells originating from the epicardium are commonly referred to as epicardium-derived cells (EPDCs; Figure 1). Early lineage tracing studies supported further contributions extending to CMs 31,32 and coronary endothelial cells (CECs). 32,33 While constitutive and inducible genetic lineage tracing models in mouse have, in many ways, helped advance our knowledge of the epicardium, they have also contributed toward a muddled narrative and dispute surrounding the extent to which certain fates are adopted. Genetic lineage tracing is predicated on the major assumption that the genetic marker used to drive labeling of the parent progenitor cell is specific and neither independently expressed in its differentiated progeny (derivatives) at later stages, nor in neighboring cell types. 3 Collectively, research over the last two decades has revealed that epicardial markers, taken individually and particularly in the absence of efficient temporal labeling methods, do not meet these criteria. Namely, Tbx18 and Tcf21 are expressed both in epicardial derivatives and in nonepicardially contributed cell types, 18,[34][35][36] whereas Wt1 and Sema3d are expressed in nonepicardial cell types that populate the heart later in development, as summarized in Table 1. 18,37,38 A reappraisal of marker specificity and the utility of fate mapping tools has resulted in a consensus that the epicardium is an unlikely native source for CMs and CECs. 6

| A source of essential paracrine signals
The secretory repertoire of the embryonic epicardium has been elaborated over the years to include an extensive list of growth factors, morphogens, and chemokines that mediate CM proliferation (e.g., IGF2 43 and BMP4 41 ) and coronary vessel growth (e.g., Elabela 44 and CXCL12 45 ). The embryonic epicardial secretome has been reviewed at length elsewhere 46 ; however, a more comprehensive list will soon emerge, with recent advances in "omics" and single-cell resolution technology 47 leading to a surge of data profiling the epicardial transcriptome. Li et al 41 identified Rspo1 expression in the epicardium-an activator of the canonical Wnt signaling pathway-from single-cell RNA sequencing (scRNA-seq) analysis of E10.5 hearts. The authors applied ligand-receptor analysis to identify hundreds of paracrine F I G U R E 1 Epicardium formation and function. The epicardium forms the outermost layer of the embryonic heart, and almost completely envelops the myocardium from embryonic day (E)11.5 in mouse. The embryonic epicardium is characteristically "active" with high proliferation, and elevated generation of mitogenic growth factors and extracellular matrix components which support cardiomyocyte propagation and maturation. As development progresses, epicardial cells undergo epithelial-mesenchymal transition (epiEMT) to provide epicardiumderived progenitors, precursors for epicardium-derived specialized cells such as cardiac fibroblasts and mural cells. Epicardial derivatives-transitory and differentiated progeny-are grouped under the term epicardiumderived cells (EPDCs). By E17.5, the "quiescing" epicardium forms a continuous layer of cells with squamous morphology, diminished proliferation, epicardial marker expression and epiEMT. Ao, aorta; LA, left atrium; LV, left ventricle; PA, pulmonary artery; RA, right atrium; RV, right ventricle; V, ventricle signals through which the epicardium communicates with CMs, endothelial/endocardial, and mesenchymal cells. R-spondin-1 and BMP4, ligands mined from this data set, were proposed to jointly stimulate proliferation of the compact myocardium. 41 It is important to note that further work will be required to experimentally validate putative cell-cell interactions, especially to delineate regionalized, protein-level crosstalk mechanisms.

| Dynamic regulation of the cardiac ECM
In addition to providing physical support for tissues, ECM molecules participate in cell-cell communication, by acting as a reservoir for ligands and essential co-receptors for signaling pathways. Thus, ECM confers the strict spatiotemporal regulation required for cardiac morphogenesis 48 and the cardiac mesenchyme is a central hub for ECM components. While epicardial contribution to ECM remodeling via its derivative mesenchymal and CFb progeny is widely accepted, the direct role of epicardial cells in ECM deposition and turnover is understudied, yet clearly important. A cell-autonomous role in formation and modification of the surrounding ECM landscape, including fibronectin fibrils, drives autocrine regulation of epicardial EMT (epiEMT) and, moreover, is postulated to signal myocardial growth and compaction, and to provide a foundation for coronary sprouting from the SV. 49 The importance of these features will be described in the next section.

| Epicardium: Native friend or foe in cardiac injury?
Prior to birth, the epicardium downregulates many of its genetic markers 18 and undergoes morphological changes, from cuboidal to squamous morphology. 50 It is generally accepted that these events, which include a steady reduction in proliferation, 51 depict the onset of epicardial quiescence ( Figure 1). Thus, with loss of markers and concealed appearance, it is almost impossible to distinguish the ventricular epicardium of the healthy adult mouse heart by histological analysis. Notably, atrioventricular sulcus and atrial epicardium were reported to sustain marker expression into adulthood, 52 the reasons for this remain unknown. Following cardiac injury, commonly Wt1 and Aldh1a2, are detected throughout the outermost layer, once again distinguishing the ventricular epicardium. 40,52,53 Albeit transient, their expression represents epicardial reactivation and increased proliferation, accompanied by subepicardial thickening that is most pronounced around the injury site ( Figure 2). The subepicardial mesenchyme is similarly transient following injury in adult mouse, as in development, and shown to originate from the reactivated epicardium. 53,54 The precise roles of the reactivated epicardium in the injured heart, and the extent to which these recapitulate developmental mechanisms, are incompletely understood. While clear differences have been identified between embryonic and post-MI responses, there is evidence that the epicardium is once again called upon for cellular contribution, paracrine signaling and ECM modulation, as discussed below.
Due to cardiac cell contribution in development, the possible recapitulation of these cell fates by the reactivated adult epicardium was largely assumed but only partially investigated. Earlier work demonstrated significant contribution to CMs, CFbs, CECs, and mural cells, suggesting that the adult epicardium preserves its cellular plasticity. 52,53 However, with revised interpretation and more cautious use of lineage tracing models, both in the neonate and adult, an epicardial origin appears unlikely, with the majority of de novo CFbs, CECs, vSMCs, and rare CMs post-MI seemingly arising from their respective preexisting resident populations. 6,8,9,11,[55][56][57] Caveats relating to the specificity of embryonic epicardial markers, as discussed previously, equally apply in the adult, 9,36,37 rendering available constitutive epicardial lineage tracing lines unreliable. Few studies demonstrate epicardium-derived cellular contribution and, without exogenous stimulation, the extent of de novo contribution is minimal. 53,54,58,59 Zhou et al 53 first utilized the inducible Wt1Cre ERT2 mouse line to trace the epicardium and its progeny post-MI. Tamoxifen was administered before MI, 53 to minimize labeling of CECs which upregulate Wt1 in response to injury. 37 However, the disadvantage of such an approach is that only a small fraction of resting epicardial cells are labeled, with the Wt1-expressing, injury-reactivated population largely unlabeled. While a proportion of labeled derivatives were found to express NG2 54 and αSMA, 52,53 ostensibly contributing pericytes and vSMCs, respectively, the subepicardial and border zone vasculature remained largely untraced. 54 Notwithstanding the inefficient labeling noted above, the scarcity of fate mapped cells implies a predominantly nonepicardial origin for neovessels of the infarcted heart.
The epicardium may be a major source of CFbs during development 60,61 but, when reactivated, seems not to be responsible for de novo CFbs, which contribute to the scar. 8,60,62 If exogenous treatments can be used to augment epicardium-derived cell differentiation, as discussed later, a greater understanding of cell fate is necessary to ensure precursors commit to beneficial cell types, rather than enhance the CFb-myofibroblast pool.
In mice, epicardial activity post-MI has been implicated in both beneficial 53  An additional role for the epicardium is emerging, in regulating crosstalk between the immune response and the injured heart, the intricacies of which will need to be better understood in order to promote beneficial, rather than detrimental, outcomes. This is borne out by a study demonstrating an important role for epicardial Hippo signaling in suppressing inflammation post-MI through enhanced Reporter line) and compromised tissue integrity. 52  Along with epicardial reactivation, injury stimulates expansion of the subepicardial mesenchyme, analogous to that which forms and functions during embryonic heart development. 18,65 Yet, its roles in cardiac repair are incompletely understood. Subepicardial tissue thickening positively correlates with cardiac function, especially when exogenously enhanced, e.g., with Tβ4 ( Figure 2). 58,59,67,68 The subepicardial space accommodates neovascularization, by vessel sprouting, 54 expansion of preexisting endothelial cells, 53 and assembly of collateral 76 and lymphatic networks. 77 In addition to revascularization, the newly synthesized matrix may temporarily stabilize the myocardial wall, akin to early reparative fibrosis 9,78 and, at least in regenerative models, provides matrix that favors CM repopulation [79][80][81] ; however, this has yet to be confirmed for the mammalian epicardium. Indeed, the de novo epicardium-derived tissue that is characteristically mesenchymal and often marks the epicardial response after cardiac injury 52,53,78 is poorly researched in adult mice. 82 Whether it is established through partial vs complete EMT or morphogenic vs fibrogenic EMT 83 87,89,105 The phenotype of these cells will ultimately be dictated by culture conditions and the extent to which they resemble human epicardium in vivo is difficult to ascertain.

| EVALUATING THE TRANSLATIONAL POTENTIAL OF THE EPICARDIUM
hPSC-derived epicardial-like cells have been powerfully applied to regenerative strategies, with their incorporation into engineered heart tissue in vitro and co-transplantation with hPSC-derived CMs into infarcted rat hearts. 105 Co-transplantation enhanced cardiac graft size and systolic function, compared with either cell type alone; however, the mechanism of the observed synergy is incompletely understood. Paracrine secretion of trophic factors from a "fetal-like" epicardium correlated with enhanced CM survival and/or proliferation, neovascularization and synthesis of a modified ECM, particularly rich in fibronectin. 105 Collectively, these properties would be expected to promote optimal repair; however, the relative contribution of each process remains to be determined. Moreover, the application of hSPC derivatives for cell therapy faces challenges, relating not only to phenotypic characterization, but to stemness, immaturity of derivatives and delivery/retention, as commonly encountered with other cell therapy candidates for cardiac regeneration. 106 Consequently, a deeper understanding of the paracrine and ECM modulatory benefits of epicardial cell therapy may allow for reproducing the effects using a cocktail of paracrine factors, to obviate the difficulties of cell transplantation.

| CONCLUSION
De novo contribution of epicardium-derived cells took central stage in the early years of research into the injury-activated adult epicardium.
However, controversies surrounding epicardial fate, both in development and injury, highlight issues with current tools and urges the field to compromise or find alternatives. Recent studies have presented strong evidence to suggest that CM and CEC contribution is unlikely and, when therapeutic strategies are applied (principally Tβ4), CM differentiation is rare and dependent upon prophylactic "priming" of the epicardium, 58,59 which may limit therapeutic application. Only with better understanding of what the epicardium can achieve might we tailor treatment and manipulate the epicardium to augment myocardial survival and neovascularization. With the future in mind, we will need to ensure that progress into epicardial-mediated regeneration is not hindered by poorly defining epicardial cells and use of outdated or unsuitable animal models. Furthermore, we should also avoid overly generalized assumptions when applying embryonic epicardial biology without establishing similarities and differences in the adult setting.
Intrinsic adult epicardial activity will consist of distinct biological processes, rather than a complete recapitulation of the embryonic program. A consensus exists in support of the reparative properties of the epicardial secretome, and future research focusing on this trait will identify new and improved targets. Moreover, recent studies have uncovered an important, but barely understood, interaction between the epicardium and the inflammatory response after injury, 40,64 which appears to strongly influence the outcome in terms of repair and functional recovery. Due to the diverse array of beneficial effects it promotes, the epicardium remains an attractive target for cardiac regeneration. However, a greater understanding is required of the endogenous repair mechanisms and of stimuli that can modulate these processes for enhanced repair. Ultimately, it may be possible to differentially control individual components of the epicardial repertoire, which may sufficiently promote repair, or may be targeted alongside therapies that promote CM proliferation and immunomodulation.