The vertebrate heart is composed of a plethora of different cell types. Around 70% of these cells reside in the cardiac interstitium (CI), the extracellular space between cardiomyocytes (Bergmann et al., 2015; Pinto et al., 2016). All these cells are jointly referred to as cardiac interstitial cells (CICs) and are known to play an essential role in myocardial embryonic development, and adult homeostasis (Krenning et al., 2010; Pérez-Pomares and de La Pompa, 2011; Pogontke et al., 2019), and in the adaptive responses of the heart to pathological conditions (Takeda et al., 2011; Ruiz-Villalba et al., 2015).

The CIC population is highly heterogeneous. It comprises multiple cell types such as endothelial cells (ECs), vascular smooth muscle cells (vSMCs), cardiac fibroblasts (CFs), pericytes, and circulating cells among others (Ruiz-Villalba et al., 2015; Pinto et al., 2016; Sampaio-Pinto et al., 2020). These different cellular pools are not, however, equally represented. Indeed, ECs, which approximately represent 58% of all CICs, are very abundant. This finding is in accordance with extensive myocardial vascularization by coronary vessels, including arteries, veins, and a massive capillary bed (Tomanek et al., 2010). Furthermore, the coronary endothelium has a key scaffolding role during embryonic interstitium formation (Pogontke et al., 2019) since coronary EC interaction with vSMCs, pericytes, and CFs is necessary for the building of periendothelial coronary domains. Moreover, due to the cellular complexity, a niche role for the periendothelial coronary milieu has been proposed (Fioret et al., 2014; Pogontke et al., 2019). Unlike EC, other cell types are far less frequent in the CI. For example, the small numbers of blood-borne cells have been described to permanently or transiently reside in the adult cardiac interstitium in the absence of evident pathological stimuli (Epelman et al., 2014; Molawi et al., 2014). The functions played by these cells in the healthy heart are poorly known, and data on the specific diversity and the spatiotemporal distribution of these cells are scarce (Pinto et al., 2016; Farbehi et al., 2019).

Other sparse cells holding an intrinsic multipotent differentiation potential were reported to be present in the cardiac interstitium. These cells, frequently dubbed cardiac stem cells (CSCs), were thought to be the origin of de novo cardiomyocyte differentiation in the adult heart. A bone marrow origin for these cells was discarded by some authors based on the absence of CD45 expression (Limana et al., 2005; Bearzi et al., 2007) but not by other researchers (Zhou et al., 2010). CSCs were originally identified based on their characteristic expression of molecules like Bmi1, Abcg2, Isl-1, Sca1, or c-Kit (Martin-Puig et al., 2012; Anversa et al., 2013), all of which are known to be expressed by other organ-resident, well-characterized stem or progenitor cells (Bradfute et al., 2005; Thoren et al., 2008). Of these, c-Kit- and Sca1-expressing CSCs have been the most widely studied. Both molecules are associated with stem/progenitor cell properties and are often considered markers for these cell types. The membrane-bound stem cell factor (SCF) receptor c-Kit (a.k.a. CD117) is expressed by hematopoietic cells but also by other cell types, many of which do not display stem/progenitor cell properties (Lennartsson and Rönnstrand, 2012; Liang et al., 2013). Similarly, Sca1 (Stem Cell Antigen 1) is expressed in different cell populations, including blood and cancer cells, and only a fraction of Sca1-expressing cells has been unambiguously shown to represent a stem/progenitor cell fraction (Holmes and Stanford, 2007).

As indicated, since 2003, cardiac c-kit^POS cells have been thought to be a self-renewing, clonogenic, and multipotent population of CSC. These cells were described to differentiate into a minimum of three different cardiogenic cell lineages (myocytes, smooth muscle cells, and endothelial cells) in vitro (Beltrami et al., 2003) and, when grown in non-adherent in vitro assays, c-Kit^POS CSCs were shown to form cardiospheres (Vicinanza et al., 2017). Most importantly, these cells were characterized, isolated, expanded, and used in cell-based experimental therapies to treat the diseased heart (Gude and Sussman, 2018; Keith and Bolli, 2015). Intensive recent research, however, has revealed that the myocardiogenic potential of these c-kit^POS CSCs is negligible (Van Berlo et al., 2014; Sultana et al., 2015), so their actual multipotency has been refuted. As for cardiac Sca1^POS cells, different studies have shown that they robustly differentiate into endothelial cells in vitro after treatment with VEGF or PDGF BB (CD31^POS, vWF^POS, caveolin^POS), and form endothelial tubules when cultured in 3D matrices (e.g., Matrigel) (Takamiya et al., 2011; Wang et al., 2006). More recently, these cells have also been associated with cardiac endothelial cells and fibroblasts in vivo using a transgenic model for lineage tracing (Tang et al., 2018).

Regardless of their multipotent potential, the existence of cardiac cells expressing molecular markers classically associated with stem cells such as c-Kit and Sca1 has not been disputed. However, their dynamics along with embryonic development and adulthood, specific origin, nature, and function remain obscure. In this work, we aimed to progress in the characterization of these elusive cells by the evaluation of a possible epicardial origin for, at least, part of them, also considering the bone marrow/circulation as a plausible source for cardiac c-Kit^POS and Sca1^POS cells. This working hypothesis is supported by the well-known contribution of the embryonic epicardium to the adult heart resident CICs pool (Ruiz-Villalba et al., 2015; Volz et al., 2015; Cano et al., 2016). In order to tackle this objective, permanent genetic tracing of epicardial cell derivatives has been carried out taking advantage of strong expression of the Wilms’ tumor suppressor gene (Wt1) in embryonic epicardial cells and by carefully analyzing tissues using both immunohistochemistry and FACS.

Cardiac interstitial cells (CICs) are a dynamic and heterogeneous cell population. This poorly studied group of cells has been historically overshadowed by the size and functional relevance of cardiomyocytes. CICs, however, largely exceed cardiac muscle cells in number and their enormous importance in the diseased heart has recently attracted much attention. Indeed, CICs contain progenitors and quiescent forms of cardiac fibroblasts (CFs), which are central to the reparative fibrotic response that characterizes the pathophysiology of multiple cardiac diseases, from ischemic, dilated, and arrhythmogenic cardiomyopathies to the aging and diabetic heart. Increasing our understanding of this unique group of cells will open new avenues for the early diagnosis of and effective therapy for these ailments.

The ontogenetic origin of CICs has been shown to be very diverse, but two of the main sources for these cells have been identified, namely, the epicardial-derived mesenchymal cell progeny often referred to as epicardial-derived cells or EPDCs and the blood cell lineages (for a comprehensive review see Pogontke et al., 2019). In both cases, the cellular colonization of the interstitial space takes place progressively during the perinatal stages (Epelman et al., 2014; Moore-Morris et al., 2015; Ruiz-Villalba et al., 2015). The main aim of this study was to study two specific CICs fractions (c-Kit^POS and Sca1^POS) in relation to epicardial-derived and hematopoietic/blood cell lineages. Since the beginning of this century, both c-Kit^POS and Sca1^POS have been considered to represent multipotent resident cardiac stem cell (CSC) populations (Beltrami et al., 2003; Vicinanza et al., 2017). This concept was first intensely debated and then recently refuted, so that the existence of this CSCs as they were originally described is no longer accepted (Van Berlo et al., 2014; Sultana et al., 2015). The existence of CICs expressing c-Kit and Sca1 remains, however, an undisputed fact, but their real nature is far from having been deciphered.

Our FACS analysis first revealed that very few c-Kit^POS cells are related to cardiac Wt1-lineage-derived cells, which can be genetically tagged by their original embryonic expression of the Wilms’ tumor suppressor gene 1 (Wt1). This analysis was specifically carried out in the CD31^NEG/CD45^NEG CICs fraction to exclude endothelial and blood-borne cells from the screening. We nonetheless identified a small proportion of c-Kit^POS/YFP^POS CICs whose identity is not still clear. The presence of c-Kit^POS, EPDCs homologous mesothelial-derived Cajal cells in the intestinal tract (Carmona et al., 2013) strongly suggested that cardiac c-Kit^POS/YFP^POS cells are not an artifact of the genetic cell tracing technology but a real biological entity. In any case, immunohistochemistry unambiguously shows that c-Kit^POS/YFP^POS is normally located in the subepicardial space or in contact with the wall of coronary blood vessels, two domains massively contributed by epicardial cells (Cano et al., 2016). The sharp postnatal decrease in c-Kit^POS/YFP^POS cell numbers within the CD31^NEG/CD45^NEG CICs fraction observed by FACS may reflect a progressive dilution of a poorly proliferating cell subset with respect to other expanding cardiac cell types and also reflects the difficulties of finding significant numbers of these cells in adult cardiac tissues using immunohistochemical approaches. Alternatively, it could be argued that the decrease in the number of c-Kit^POS/YFP^POS cells is due to the loss of this cell surface marker during cell differentiation/maturation. More research is needed to clarify this point.

As the majority of cardiac c-Kit^POS cells are also YFP^NEG in Wt1Cre-YFP mice, we then decided to evaluate a possible origin for these cells from hematopoietic/blood lineages. In order to do so, we searched for the specific pan-leukocyte marker CD45 in CICs. The results from this inspection show that CD45^POS/c-Kit^POS cells are abundant among CICs (up to 65% at E15.5). As expected, virtually none of these cells was YFP^POS, excluding a possible relation between blood-borne cells and Wt1 cell lineages. Of note, nearly all c-Kit^POS CICs co-expressed CD34, an immature endothelium and stem cell marker (Carmona et al., 2020). Furthermore, c-Kit^POS cells expressing the endothelial and hematopoietic cell marker CD31^POS were also identified in our immunohistochemical analysis. Other authors have already identified c-Kit^POS/CD45^NEG CICs as endothelial progenitors in the adult (Sandstedt et al., 2010). However, as described above, we found a relevant contribution of c-Kit^POS cardiac cells in the CD45^POS CIC fraction, suggesting that these cells are therefore more likely to be blood-borne cells than endothelial cells. In accordance with this conclusion, CD45^NEG/c-Kit^POS/YFP^NEG CICs were found to encompass cells expressing markers for macrophages (F4/80), lymphoid (CXCR3), and myeloid (CD11b) cell types, a result that endorses an extracardiac origin for these CIC. Finally, a marked decrease in the number of CD45^POS/c-Kit^POS CICs was observed between embryonic and postnatal stages, a finding that can be interpreted either as a cell dilution similar to that previously described for c-Kit^POS/YFP^POS cells or as an actual decrease of blood circulating cell recruitment to the adult heart.

The second part of our study focused on the study of Sca1^POS CICs. FACS characterization of Wt1Cre-YFP CICs clearly indicates that the majority of CD31^NEG/CD45^NEG/Sca1^POS CICs were YFP^POS at P7, a finding that relates these cells with the Wt1-derived cardiac mesenchymal lineage. The progressive, sustained increase of Sca1^POS CICs from developmental to postnatal stages (rising from 0% to 2% to almost 70%) is evidence of the fast expansion of these cells after birth. Further FACS characterization indicates that an important number of these cells are CD31^POS endothelial ones (Tang et al., 2018, 2020). Indeed, we show that 30% of CD31^POS ventricular endothelial cells also expressed Sca1 at E18.5, reaching 80% in the adult, indicating that the contribution of Sca1^POS/CD31^POS cells to the cardiac interstitium is highly dynamic over time. Relevant to this discussion, 90% of YFP^POS/CD31^POS cells and 60% of YFP^POS/CD31^NEG/mEFSK4^POS CFs expressed Sca1, suggesting that the expression of this molecule is related to the epicardial cell lineages. This finding could be relevant due to the importance of epicardial cells in the context of cardiac repair (Ruiz-Villalba et al., 2015; Chen et al., 2018). However, since Sca1^POS/YFP^POS and Sca1^POS/YFP^NEG cells included similar percentages of mesenchymal stem-like cells, such as CD73, CD90, and CD105, more research is needed to determine potential differences between both cell subpopulations.

We have shown that postnatal CD31^NEG/CD45^NEG/Sca1^POS/YFP^POS CICs are likely to arise through the proliferation of prenatal cell populations, although the possibility of this increase results from the de novo expression of Wt1 in the Sca1^POS population cannot be directly discarded. To ensure this was not the case, we used an inducible epicardial driver line (Wt1Cre^ERT2) to activate the promoter in the postnatal stages. At P7, we did not identify YFP^POS CICs, confirming that CD31^NEG/CD45^NEG/Sca1^POS/YFP^POS CICs derive from embryonic founder cells.

Our study provided novel evidence on the relationship existing between cardiac interstitial c-Kit^POS and Sca1^POS cell populations and the Wt1 cell lineage, also contributing relevant information on the spatiotemporal distribution of these cells in the embryonic and postnatal heart. Since the majority of cardiac Wt1 lineage cells (E11-E17) are known to derive from the embryonic epicardium (Wessels et al., 2012; Cano et al., 2016), we generally assumed that YFP cells are epicardial derivatives. In accordance with this view, recent publications based on single-cell RNA-Seq revealed the existence of an epicardial transcriptomic signature related to Wt1 lineage-derived cells in the adult heart (Quijada et al., 2020; Hesse et al., 2021). For this reason, we considered that adult YFP cardiac cells are mostly derived from the embryonic epicardium. Notwithstanding, this conclusion does not rule out the possibility of Wt1 lineage cells deriving from other alternative cellular sources (King-Underwood et al., 2005).

As it can be inferred from the previous discussion, Sca1^POS/YFP^POS CICs are very likely to be epicardial derivatives. The enrichment of Sca1 in a major part of adult putative epicardial-derived cells suggested that this molecule may be involved in the regulation of epicardial biology. The colonization of the cardiac interstitium by these cells starts during embryonic development and is highly dynamic. In this respect, since the turnover rate for CICs is unlikely to be steady throughout adult life, the specific cellular dynamics and composition of the cardiac interstitium have been proposed to be a sensor for the health status of the heart (Pogontke et al., 2019).

In summary, increasing our knowledge of the origin, diversity, and functions of CICs will be instrumental to the development of a diagnostic and prognostic test for the evaluation of cardiac homeostasis and the cardiac interstitium´s response to pathologic stimuli. However, further, systematic research on CICs biology is needed to bring about significant progress in our understanding of interstitial responses to pathologic conditions.