November 2009 | Vol 6 | N.º 11 | CNIC-14 [PDF (1.008 KB)]

The epicardium: development, differentiation and its role during heart regeneration

Juan Manuel González-Rosa and Nadia Mercader *
Fundación Centro Nacional de Investigaciones Cardiovasculares (CNIC), Madrid, Spain

Correspondence: N. Mercader, Department of Cardiovascular Developmental Biology, C/ Melchor Fernandez Almagro 3, E-28029 Madrid, Spain, Thelephone: +34-914531287
Email nmercader@cnic.es

Competing interests: The authors declare no competing interests.

Acknowledgment
We thank Simon Bartlett for editing assistance. N.M. is supported by RYC-2006-001694 and BFU2008-00212/BMC from the Ministerio de Educación y Ciencia of Spain. J.M.G.R. was supported by a Ramón Areces postgraduate fellowship and is currently supported by FPU fellowship AP2008-00546 from the Ministerio de Educación y Ciencia, Spain.

Abstract
In the past few decades, the mortality rate after myocardial infarction has been significantly reduced through the use of thrombolytic therapy, primary percutaneous coronary intervention and bypass surgery to restore coronary circulation. However, reconstitution of the damaged coronary vessels is usually not achieved, leading to myocardial damage. Lost myocardium is replaced with scar tissue instead of healthy cells, impairing cardiac function. The human heart has classically been considered a postmitotic organ, incapable of self-repair. However, new evidence demonstrates that, to a limited extent, cardiomyocytes are renewed during adulthood and after myocardial damage^1-3 . The mechanisms controlling cardiomyocyte proliferation and coronary vessel regeneration and the development of therapeutic methodologies to enhance these processes are the focus of current research.The coronary vasculature is partly derived from the epicardium, an epithelial layer enclosing the myocardium. Investigation of the epicardial structure has become a higher priority in the past decade, because it was shown to be a source of progenitor cells and to promote cardiomyocyte proliferation during heart development and regeneration^4-6 . This review summarises the most recent advances in epicardial development and its role during heart regeneration.

Origin of the epicardium
Coelomic mesothelial cells lining the pericardial cavity form an aggregate at the base of the inflow tract, which is named the proepicardium (for review see ref. ^4-6) (Fig. 1A).

In chicks, the proepicardium has a characteristic cauliflower-like structure, while in other species, such as dogfish and teleost fish, it is less morphologically distinct. Most of the epicardium covering the heart is derived from the proepicardium; however a small portion enveloping the distal outflow tract originates from the cephalic pericardium⁷. Recent lineage tracing experiments have shown that the proepicardium is derived from a common cardiac precursor pool that generates most of the cell types that form the heart. The early cardiac marker gene encoding the transcription factor GATA4 is expressed in the proepicardium and its loss of function disrupts proepicardium formation⁸. Early cardiac precursor cells are also defined by the expression of the transcription factor genes Nkx2.5 and Isl1. Nkx2.5 and Isl1 are not expressed in proepicardial (PE) cells. However, PE cellsare descendents of Nkx2.5⁺ Isl1⁺ cells^9-10. The proepicardium does not form correctly in Nkx2.5 null mice; in contrast, Isl1 is not necessary for proepicardial development¹⁰.

Some of the earliest genes expressed in the proepicardium are those encoding Retinaldehyde-dehydrogenase 2 (Raldh2),¹¹,¹² the T box transcription factor 18 (Tbx18),¹³,¹⁴ Wilm’s tumour protein (Wt1)¹⁵,¹⁶ and the extracellular matrix protein podoplanin^17,18 (Fig.1A’ and data not shown). None of these genes is exclusively expressed in the proepicardium and epicardium. Raldh2 is expressed in posterior heart precursors at early stages of heart tube formation and is later expressed in the epicardium, the pericardium and the endothelial cells of the atrioventricular cushions. Loss of Raldh2 function, which is involved in retinoic acid (RA) synthesis, leads to severe heart malformations. Epicardium-specific ablation of Raldh2 has not been reported. However, inactivation of the RA receptor, RXRa, using a GATA5-enhancer–driven Cre construct, which is predominantly active in the epicardium, disrupts epicardial layer formation¹⁹. Tbx18 is expressed in the coelomic wall of the pericardium and the myocardium of the sinus horns and is required for the development of the sinus horns but not for epicardium development¹³, ¹⁴, ²⁰. Sites of Podoplanin expression include the coelomic wall of the pericardium, the cardiac conduction system and sinus venosus18, and Podoplanin null mutants display defective epicardium development as well as alterations in the sinus venosus and sinoatrial myocardium¹⁷. Similar to Tbx18 and Podoplanin, Wt1 expression is not confined to PE cells and the epicardium, but can also be detected in the mesothelial cells lining the pericardial cavity¹⁵,¹⁶. Wt1 mutants also display a defective formation of the epicardium and its derivatives²¹,²².

The formation of the proepicardium at the venous pole of the heart seems to be controlled by opposing actions of Fibroblast Growth Factor 2 (FGF2) and Bone Morphogenetic Protein 2 (BMP2)²³. In the chick, FGF2 is expressed in PE cells, whereas BMP2 is expressed in myocardial cells at the heart inflow tract. Inhibition of FGF2 signalling blocks PE formation. Instead, precursor cells become incorporated into the inflow tract (Fig. 1A’’).

PE cells proliferate and adhere to the myocardial surface, forming an enveloping epithelial layer. Two mechanisms for PE cell translocation to the myocardium have been proposed (Fig. 1B). In avian animal models, attachment of the proepicardium to the myocardial surface at the level of the atrioventricular boundary allows the transfer of PE cells to the myocardial surface²⁴-²⁶. Increased extracellular matrix accumulation at the contact site favours adhesion; and treatment with heparinase impedes the attachment of the proepicardium to the ventricular surface²⁵. In dogfish and mouse, PE cell aggregates adhere progressively to the myocardium²⁷-²⁸. This phenomenon has been proposed to be driven by the heartbeat: the beating heart is suggested to promote alternating attachment and detachment between proepicardium and myocardium, leading to the formation of PE cell aggregates that remain adhered to the myocardial wall or temporarily free-float in the pericardial cavity until they reattach²⁹. The possibility that both mechanisms could coexist in some species cannot be excluded. Furthermore, alternative mechanisms might be involved, since in the axolotl mutant lethal, which lacks a heartbeat, PE cells still attach to the myocardium although their spreading seems to be impaired³⁰. Little is known no the molecular basis of PE cell migration and adhesion to the myocardium. The vascular cell adhesion molecule, VCAM-1, and its receptor, a4b1 integrin, mediate adhesion of a4b1-integrin-positive epicardial cells to VCAM-1 expressing cells on the myocardial surface³¹,³². The T-box transcription factor, Tbx5, has been suggested to control PE cell migration, since Tbx5 gain or loss of function impairs PE cell migration and survival³³.

An extracellular-matrix rich space, the subepicardial space, is established between the epicardium and myocardium. Some epicardial cells undergo an epithelial-mesenchymal transition (EMT) and delaminate from the epicardium to invade this space (Fig. 1C). Some of these epicardial-derived cells (EPDCs) remain as mesenchymal cells or form blood islands; others migrate further into the myocardium and endocardial cushions, where they differentiate into a subset of cell types described below^4-6. Epicardial cells express the ATP-dependent SWI/SNF chromatin-remodelling component, BAF180, and analysis of BAF180 null mutant mice identified a regulatory role for this gene in epicardial cell EMT³⁴. Proper establishment of gap junctions is essential for correct EMT and epicardial cell migration, which was revealed by gain and loss of function analysis of Connexin43 mutants³⁵. EMT and EPDC migration are mainly controlled by paracrine signals from the myocardium, which are generated through interaction of the transcription factors GATA4 and FOG-2, among others³⁶, ³⁷. Several studies have shown that Transforming Growth Factor b (TGFb), via its receptor, Alk5, promotes epicardial cell EMT³⁸, ⁴⁰. FGF Receptor 1 (FGFR1) signalling is important for myocardial invasion by EPDCs, since FGFR1 knockdown in PE cells results in a reduced number of coronary vessels in the myocardium and their accumulation in the subepicardial space⁴¹. A similar phenotype was observed in mice deficient for the actin-binding peptide Thymosin b4 (Tb4)⁴², which is normally expressed in the myocardium and promotes EPDC migration.

Epicardial derived cell differentiation
Much knowledge about the formation of the epicardium and its derivatives comes from studies in the chick. Analysis of quail-chick chimeras and retroviral labelling experiments suggest a primitive epicardium origin for the mesenchymal cells of the subepicardial layer, the endothelium and smooth muscle cells (SMCs) of the coronary vasculature, the perivascular and intermyocardial fibroblasts and some cells in the atrioventricular valves⁴, ⁶ (Fig. 1E).

More recent studies in mouse using the Cre-lox technique examined the fate of cells expressing the epicardial marker genes Wt1 and Tbx189,⁴³ (Fig. 1E’). These studies showed that Wt1/Tbx18 descendants contribute to the coronary vessel SMCs and the fibroblasts of the myocardial interstitium. Unlike reports in the chick, these studies found that in mice only a small population of coronary vascular endothelial cells is derived from the epicardium. Interestingly, a Tbx18-negative population expressing the endothelial precursor cell marker Flk1 was found in the proepicardium, suggesting that different mesothelial-derived cell types coexist in this structure⁴³. Alternatively, the Flk1-positive cell population might be derived from a different source. In chicks, the proepicardium has been reported to contain a mixture of hepatic endothelial precursors and mesothelial cells⁴⁴. The authors suggest that the hepatic derived endothelial cell precursors give rise to the endothelial cells of the coronary vasculature, whereas the mesothelial component of the proepicardium contains the SMC progenitors. Contrasting with the observations in chicks, the mouse epicardium is suggested to make a considerable contribution to the myocardium⁹,⁴³. However, this suggestion is questionable, since the supposed epicardium-derived cardiomyocytes might instead be a Tbx18-positive population residing within the myocardium⁴⁵.

EPDC differentiation is controlled by epicardial and myocardial signals. A wave of epicardial Sonic hedgehog (Shh) expression precedes the formation of the coronary vasculature (for review see ref. ⁴⁶). Epicardial Shh activates the production of proangiogenic factors, such as Vascular Endothelial Growth Factor A (VEGFA), VEFGB and Angiopoietin 2 in the myocardium and VEGFC in the perivascular mesenchyme. Platelet Derived Growth Factor b (PDGFb) is involved in smooth muscle cell differentiation and PDGFb null mice have defective coronary arteriogenesis, with reduced numbers of pericytes and SMCs surrounding endothelial capillaries⁴⁷,⁴⁸. Correct angiogenesis of the coronary vasculature also involves Wnt signalling, since epicardial-specific suppression of b-catenin impairs myocardial invasion by EPDCs and their differentiation into SMCs⁴⁹.

The epicardium as a signalling source
Experiments in the chick, where the proepicardium was removed or PE-cell adhesion to the myocardium was prevented, demonstrated that correct development of the myocardium requires the epicardium. In the absence of a healthy epicardial layer, the myocardium develops as a thin compact layer, and the development of the atrioventricular (AV) valves and intraventricular septum is impaired⁵⁰-⁵² (Fig. 1D). Defective epicardium also leads to a bulging of the ventricular myocardium, suggesting that the epicardium controls the passive mechanical properties of the myocardial wall⁵². The effects of epicardium ablation in chicks are similar to the phenotype of Wt1 null mutant mice, which have a thin myocardial wall, impaired AV valve development and incomplete intraventricular septation²¹-²².

There is evidence that epicardium-derived secreted molecules signal to the underlying myocardium. FGF9, 16 and 20 are expressed in the epicardium and myocardium-specific inactivation of FGF receptors 1 and 2 leads to defects in cardiomyocyte development⁵³. Epicardial-derived RA triggers a signalling cascade that upregulates myocardial FGF2, thereby leading to cardiomyocyte proliferation19. EPDCs are also a source of paracrine signals for cardiomyocytes. Intermyocardial fibroblasts, which are at least partly epicardially derived, promote cardiomyocyte proliferation during development⁵⁴.

Role of the epicardium during heart regeneration
The role of the epicardium during myocardial regeneration has attracted growing interest within the scientific community in recent years. In a zebrafish heart regeneration model, resection of the ventricular apex leads to complete regrowth of the amputated region, including the coronary vasculature, myocardium and endocardial tissues⁵⁵. Up to 20% of the heart can be regenerated in this species. The epicardium has been suggested to play a role during heart regeneration in zebrafish⁵⁶. Amputation of heart tissue induces a strong expression of epicardial marker genes and vascular endothelial precursor cells accumulate at the amputation site. To date, it is unclear whether the epicardium contributes in a paracrine manner, stimulating myocardial cell proliferation, or as a cell source, providing EPDCs that populate the regrowing myocardial wall. Studies of the epicardium in the zebrafish suggest that EPDCs are incorporated into the heart not only during regeneration, but also during continuous growth of the adult heart, since DiI-labelled epicardial cells are incorporated into the compact layer of the myocardium⁵⁷. The supportive action of the epicardium during regeneration is not exclusive to teleosts; it has also been reported in mammals. Cells positive for the stem cell markers c-kit and CD34 have been detected in the subepicardial space of foetal and adult human hearts⁵⁸. In culture, these cells can acquire myocardial marker gene expression and an endothelial phenotype (Fig. 1E’’). In a mouse myocardial infarction model, c-kit-positive and CD34-positive cells in the subepicardial space proliferate and differentiate into myocardial, endothelial and smooth muscle cell lineages59; further, human epicardial cells also can differentiate in vitro into smooth muscle cells⁵⁹. Moreover, myocardial infarction in mice reactivates Wt1 expression in the epicardium and coronary vessels⁶⁰; suggesting that tissue damage in adult animals reactivates the developmental gene regulatory network. In a recent study, transplantation of human EPDCs into a mouse heart after coronary artery occlusion resulted in improved cardiac function, demonstrating that there is a genuine therapeutic potential for EPDCs to restore heart function after myocardial infarction⁶¹. However, while hEPDCs were clearly incorporated into the heart, it is uncertain as to which exact cell type they gave rise to.

The potential for ectopic factors to promote EPDC migration, or proliferation, has also been investigated. Administration of Tb4 enhances epicardial marker gene expression and slightly improves capillary formation at the infarction border zone in adult mice following coronary artery ligation⁶². The effect of FGF and VEGF administration on angiogenesis after myocardial ischemia-reperfusion injury are also being tested (reviewed in ref. ⁶³). In addition, activation of Hedgehog signalling in mice or Shh gene therapy promotes neovascularisation and protects against ischemic injury in the adult heart⁶⁴, ⁶⁵.

Future perspectives
Given that regeneration mechanisms in a mature organ generally recapitulate the developmental processes of its formation, understanding the developmental mechanisms that underlie the formation of the epicardial layer, trigger EMT by EPDCs and drive their differentiation into diverse cell types will be crucial for the development of strategies to manipulate these events in adults. Functional genetic studies and lineage tracing experiments in animals have been hampered by the lack of a suitable epicardial-specific enhancer. To date, most studies have relied on reporter lines that are not exclusive to the epicardium, thus making interpretation of the results challenging. Identification of a more suitable enhancer will therefore assist greatly. In light of the differences between chick and mouse EPDC lineages, it will be interesting to analyse the contribution of EPDCs to the growing heart in other species, such as teleost fish. This will also clarify whether the extraordinary regenerative capacity of the teleost heart is reflected in the differences in the composition, or responsiveness to injury, of their epicardium or EPDCs. In suitable models, such as zebrafish, in vivo imaging of the processes through which PE cells attach and spread over myocardial tissue will provide novel insights and help to confirm in vitro findings. Small molecule screens to find new factors involved in epicardial development might lead to the discovery of pharmacological compounds with therapeutic applications for the ischemic heart.

It is important to continue efforts to define ways to promote neovascularisation by mobilising resident cells or by administering cells or factors ectopically. However, off-target effects of administered factors need to be carefully examined to exclude deleterious consequences. It will also be important to promote the replacement of damaged tissue by newly formed myocardium that is functionally coupled to the rest of the myocardium. In the context of the immense efforts being made to identify resident stem cells in the adult heart⁶⁶, it seems appropriate to consider the epicardium as another potential source of pluripotent cells.