Colonizing the heart from the epicardial side

The clinical use of stem cells, such as bone marrow-derived and, more recently, resident cardiac stem cells, offers great promise for treatment of myocardial infarction and heart failure. The epicardium-derived cells have also attracted attention for their angiogenic paracrine actions and ability to differentiate into cardiomyocytes and vascular cells when activated during cardiac injury. In a recent study, Chong and colleagues have described a distinct population of epicardium-derived mesenchymal stem cells that reside in a perivascular niche of the heart and have a broad multilineage potential. Exploring the therapeutic capacity of these cells will be an exciting future endeavor.

Congestive heart failure, the leading cause of morbidity and mortality worldwide, is characterized by the irreversible loss of cardiac cells and function. Implantation of exogenous stem cells or activation of endogenous progenitors holds great potential for the therapeutic re genera tion of lost cardiomyocytes and supporting vascular cells [1]. With the progress in the fi eld, it is becoming evident that profound understanding of cardiac developmental biology will be necessary for the discovery of cell types and conditions that can lead to successful cardiac repair.
Recently, epicardium of the heart has received a lot of attention as a potential source of cardiovascular progenitors [2][3][4]. Traditionally, embryonic epicardium has been known for its paracrine roles in myocardial growth and diff erentiation and its contribution to cells of cardiac connective tissue and coronary vasculature [5]. Recent lineage tracing studies using epicardial markers Wt1 and Tbx18 have suggested that murine embryonic epicardium-derived cells (EPDCs) can also diff erentiate into cardiomyocytes [2,3]. Unlike in the embryo, the epicardium in the healthy adult heart remains quiescent; however, upon myocardial infarction, it gives rise to new EPDCs that diff erentiate into mesenchymal cells expressing fi broblast and smooth muscle markers. Although EPDCs do not adopt cardiomyocyte or endothelial cell fate, they still secrete paracrine factors that strongly stimulate angiogenesis in subepicardial layers [6]. Interestingly, if the murine heart is preconditioned with thymosin β4 (Tβ4) prior to infarction, the adult EPDCs can migrate into myocardium and give rise to new cardiomyocytes [4]. Unfortunately, in a clinically relevant scenario of Tβ4 post-conditioning, the EPDCs remain in the thickened epicardium and do not attain cardiogenic fate [7].
In line with these exciting developments, Chong and colleagues [8] have recently reported a new type of mesen chymal stem cell of epicardial origin found in embryonic and adult mouse heart that, similar to mesenchymal stem cells from bone marrow (BM), can form clonal colonies and undergo a long-term expansion for approximately 40 passages. In the adult heart, these cardiac colony-forming-unit fi broblasts (cCFU-Fs) reside in a perivascular adventitial niche and can be isolated from the non-cardiomyocyte fraction of the heart based on the expression of the surface markers platelet-derived growth factor receptor-α and Sca-1. Th e cCFU-Fs lack expression of cKit, CD31, Flk1, CD45, Nkx2-5, NG2, and the pluripotency genes Rex1, Sox2, Klf4 and Nanog. Under optimized conditions, they diff er entiate in vitro and in vivo into diff erent mesodermal lineages (for example, bone, cartilage, fat, smooth muscle, endo thelium, cardiomyocytes), and when co-trans planted with embryonic stem cells, also attain non-mesodermal fates (for example, liver, neuronal, lung epithelium). Under the 5-aza-2'deoxycytidine treatment or in non-contact cocultures with neonatal rat cardio myocytes, 1 to 2% of the cCFU-Fs attain cardiac diff eren tiation markers, but despite some sarcomeric organiza tion, do not spontaneously contract, presumably due to paracrine inhibition by surrounding non-cardiomyocytes [8].
In the elegant Cre-recombinase lineage tracing studies, Chong and colleagues further show that cCFU-Fs originate from mesoderm (expressing Mesp1) but not from neural crest (expressing Wnt1) or cardiomyocytes (expressing Nkx2.5 and Myl2). Importantly, the use of the epicardial-specifi c Cre lineage tags Wt1 and Gata5 showed that cCFU-Fs derive from proepicardium and

Abstract
The clinical use of stem cells, such as bone marrowderived and, more recently, resident cardiac stem cells, off ers great promise for treatment of myocardial infarction and heart failure. The epicardium-derived cells have also attracted attention for their angiogenic paracrine actions and ability to diff erentiate into cardiomyocytes and vascular cells when activated during cardiac injury. In a recent study, Chong and colleagues have described a distinct population of epicardium-derived mesenchymal stem cells that reside in a perivascular niche of the heart and have a broad multilineage potential. Exploring the therapeutic capacity of these cells will be an exciting future endeavor. epicardium, and may be similar or hierarchically related to previously identifi ed EPDCs. In addition, the cCFU-Fs appear to be distinct from previously reported cKIT + progenitors found in epicardium as these cells also expressed Nkx2-5 [9]. Th ey also appear mostly diff erent from pericytes based on no expression of NG2 or other pericyte markers.
In a previous study [10], the cCFU-Fs were found to be transcriptionally similar, but not identical, to BM CFU-Fs, suggesting their similar early origins, but potentially diff erent tissue-specifi c functions. To investigate the potential contribution of BM cells to the cCFU-F com partment, Chong and colleagues transplanted green fl uorescent protein (GFP) + BM cells into lethally irradiated mice and examined their contribution to cCFU-F colonies after myocardial infarction and 2 to 8 months of aging. Th ey found that both cCFU-Fs and BM cells were irradiation-sensitive and while GFP + BM cells were able to reestablish the niche for BM CFU-Fs, they failed under all conditions (including mobilization by granulocyte colony stimulat ing factor) to contribute to cCFU-Fs. Th e majority of GFP + cells that entered the heart postmyocardial infarc tion were cKit + /CD45 + but none formed cCFU-Fs, suggesting a distinct phenotype of cCFU-Fs from those of BM-derived cKit + cells. Finally, none of the Wt1-, Gata5-, or Mesp1-tagged cells participated in the forma tion of BM cCFU-Fs, further suggesting distinct post-gastrulation origins of BM cells and cCFU-Fs.
Th e cutting-edge studies by Chong and colleagues open multiple avenues for future investigations to answer several important questions: are the cCFU-Fs solely derived from the epicardium, or does the endocardium contribute a fraction of these cells? How diff erent are these cells from other resident stem cells identifi ed in the adult heart [11]? How does aging aff ect frequency and potency of the cCFU-Fs? Do the endogenous cCFU-Fs diff erentiate into functional cardiomyocytes in vivo? If separated from the non-cardiomyocyte cCFU-F derivatives (perhaps via an α-MHC driven tag), will the in vitro diff erentiated cardiomyocytes become electromechanically active? Could the robust proliferative capacity of cCFU-Fs make them a good candidate for in vitro or in situ direct reprogramming to the cardiomyocyte fate using transcription factors [12] or microRNAs? Would priming these cells with Tβ4 enhance their cardiogenic potential? And most importantly, to what extent are these fi ndings from mice applicable to a human setting? Answer ing these questions may bring us a few steps closer to making cardiac regenerative therapies a reality.

Competing interests
The author declares that they have no competing interests.