Our study provides evidence that Bmi1-derived cells contribute to de novo generation of CM after myocardial infarction, which is greater than for heart homeostasis with ageing. These findings consolidate Bmi1-CPC cells as a sound source of progenitors for cardiac repair.

The cardiac stem field was initiated by the isolation and characterization of c-KIT + CPC [9], which recently were proposed necessary and sufficient for heart repair after injury [18]. c-kit expression in vitro and in vivo has yielded disparate results, however, which probably reflects the extremely variable expression of this marker in distinct contexts and conditions [19, 20]. Three recent independent lineage-tracing studies found that c-kit + cardiac progenitor cells produce <0.008 % of the new CM in adult mouse heart, and proposed that c-kit + cells are not relevant either for homeostasis or after myocardial infarction [2123]. Our own research on Bmi1-CPC showed that these progenitors are mainly negative for c-KIT expression (0.1–1 %) [15], suggesting that the majority of de novo CM derive from Bmi1 +c-kit progeny.

Putative CPC have also been isolated based on their expression of the established hematopoietic marker SCA-1 [24], although SCA-1 appears to label a heterogeneous population with predominantly endothelial potential [25, 26]. Genetic elimination of Sca-1 affects resident CPC, which then fail to respond to pathological damage in vivo; this coincided with impaired in vitro growth and survival of these cardiac progenitor cells [27]. SCA-1 CPC contributes to CM generation in a model of pressure overload cardiac injury (transverse aortic constriction), but not after AMI [24]. Our Bmi1-CPC showed a strong response after AMI and the percentage of de novo CM generated is greater than that observed during normal cardiac homeostasis. Although Uchida et al. [24] found no major contribution by the Sca-1 + population in new CM formation after acute injury, the distinct methods and transgenic models used in these studies could explain the differences. The authors nonetheless suggested that only a small fraction of the Sca-1 + population contributes to the CM lineage [24]. Bmi1-CPC make up a fraction of the Sca-1 + population [15], and our results here suggest that Bmi1 + cells are the Sca-1 + subset involved in this new CM generation. The information gleaned from the deep-sequencing analysis of Bmi1-CPC from healthy hearts reinforces the idea that this population expresses higher levels of pluripotency markers. We consider Bmi1 to be a key transcription factor that controls stemness in the adult heart, thus defining a population of cardiac progenitors. This would be in agreement with the critical positive role of Bmi1 in fibroblast reprogramming to embryonic stem cells [28, 29] and the very recent description as a key epigenetic barrier to direct cardiac reprogramming [30].

The limited capacity of the adult mammalian heart to recover after myocardial injury is well established. A genetic fate-mapping strategy gave indirect evidence that up to 19 % of CM are replaced 3 months post-AMI, but the source of the new CM was not definitively determined [31]. Our lineage-tracing studies after cardiac infarction show that Bmi1-expressing cells contribute to significant and comparable numbers of de novo CM. This result provides direct evidence of a resident cardiac progenitor population that contributes to heart repair. RNAseq analysis of Bmi1-CPC isolated 5 days after AMI showed an increase in genes related to cell proliferation, movement, and cell cycle functions, all of which are necessary for cells to reach the infarcted area and to instigate repair. Bmi1-derived cells also showed marked upregulation of extracellular matrix proteins, growth factors, and pro- and anti-inflammatory cytokines. The data from the RNAseq study indicate that, following cardiac injury, Bmi1-CPC sense the unfavorable environment originated by the insult and respond by upregulating pathways related to key cell functions required to promote repair. Although we did not detect a defined pattern of cardiomyocyte specification markers in the Bmi1-derived cells 5 days after injury, surviving Bmi1-CPC at this early stage might need to proliferate and migrate before they can respond appropriately to injury-induced signals. Lineage tracing of Bmi1 + cells at 4 months post-AMI showed generation of 13.8 ± 5 % new YFP+ CM, which coincides with some previous reports [31] and pinpointed the Bmi1-CPC ability to contribute to myocardial repair after injury. Immunostaining of a panel of contractility-related proteins showed that these YFP+ CM were indistinguishable from YFP CM. Although we did not detect YFP+ CM within the injured area, these new CM might have an important role in supporting basal heart beat and cardiac function after AMI. In any case, the role of Bmi1 +-CPC in cardiac repair after AMI is compatible with contributions from other reported sources [24, 32]. Further characterization of the mechanisms that lead to endogenous progenitor cell activation and of the mechanisms that permit repopulation of the infarcted region could be a new opportunity for therapeutic applications.

Although the adult mammalian heart is one of the least regenerative organs in the body, different studies have described heart regeneration in lower vertebrates and neonatal mammals following apical resection [33, 34]. Lineage tracing experiments in adult zebrafish following cardiac injury suggest that the vast majority of the regenerated CM are derived from pre-existing proliferative CM rather than from a population of cardiac stem/progenitor cells [34, 35]. Additionally, the neonatal mouse heart has shown a similar ability to restore the lost myocardial tissue within the first week of postnatal life [17, 36, 37], primarily, by proliferation of pre-existing cardiomyocytes and not by putative population(s) of endogenous CPC [38, 39]. However, the ability to efficiently regenerate the heart muscle upon injury is essentially lost by postnatal day 7, coinciding with the developmental window when mouse CM mostly binucleate and withdraw from the cell cycle [40]. This compromised ability to fully repair the damaged/resected myocardium is mainly related to the lack of basal proliferation in adult CM. However, there are several reports pinpointing the relevant role of endogenous CPC leading to the reparative response in the adult mouse following cardiac injury [24, 31, 32, 41]. Although adult CPC do not fully restore a severe acutely damaged tissue, their contribution to de novo CM may contribute to maintain organ functionality. Altogether these reports, and our own data (apical resection of TM-induced day 1 Bmi1-YFP pups indicate a minor contribution of Bmi1-derived cells to de novo CM), may suggest a switch in CPC reparative role through the lifespan of the organism.

Therefore, further studies will determine and evaluate the mechanisms by which Bmi1 + cells remain less active in neonatal heart regeneration but play a more direct role in adult heart repair. Moreover, it will also be necessary to understand the hypothetic crosstalk existing between CM and Bmi1 + through the lifespan of the animal to clarify the different reparative responses that Bmi1 + exhibit following cardiac injury in the neonatal and adult heart.