The EDCs described here share several characteristics of cardiac explant derived progenitors described by other investigators
. In agreement with previous works, EDCs are a distinct population of small, round highly refractile cells which appeared 1–3 weeks after cardiac explants were cultured despite some differences in culture conditions. The appearance of EDCs was also preceded by the formation of a layer of fibroblast-like smooth muscle actin positive cells. Transmission EM analyses revealed the presence of interstitial cells within the explant with ultrastructural attributes very similar to the EDCs, suggesting that these cells track through the interstitium towards surface layer, and then emerge from the explant. Preliminary expression studies detected GATA-4 mRNA (a cardiomyogenic transcription factor) and vimentin and sarcomeric actinin protein, consistent with the cardiomyogenic activity reported for the other explant-derived cells
Given these promising attributes, it was rather disappointing that analyses with the MLC2v-Cre/ZEG reporter system failed to support the presence of cardiomyogenic activity in cultured EDCs (as evidenced by the absence of cre-mediated recombination of the reporter locus). To eliminate confounding effects of autofluorescence, DAB immunostaining was performed with adequate negative controls. There was no background staining where the primary antibody was omitted. Backgound endogenous peroxidase activity was also excluded by peroxidase pre-treatment. The presence of GFP as well as vimentin and sarcomeric actinin immune reactivity, appeared to result from phagocytotic activity. This view was supported by ultrastructural analyses which demonstrated the presence of mitochondria, cardiac myocyte sarcomeric structures, and anti-GFP immunogold staining within the endocytic vesicles of the EDCs. Although the EDCs appeared to exhibit phagocytic activity, they did not express leukocyte or macrophage markers. It is also possible that the observed GFP fluorescence resulted from nanotube formation between EDCs and cardiomyocytes, as was recently reported for cardiomyocyte/endothelial progenitor cell co-cultures
, although it is unlikely that molecules transported in this manner would localize to endocytic vesicles.
The notion that EDCs are macrophage-like cells is also supported by the observation that explants generated from perfused hearts gave rise to fibroblast-like out-growths but not to EDCs. Thus, the EDCs are more likely to arise from circulating cells which migrated into the heart, rather than from a static component of the cardiac interstitium. A recent study has suggested a similar origin for c-kit+ cardiac progenitors
. This idea was further supported by another study, suggesting that hematopoietic progenitor cells trafficked through blood and other peripheral tissues as part of immunosurveillance
. The main difference between EDCs & cells described in the latter study
is that EDCs did not express c-kit, Sca-1 nor CD45. Although it is possible that the loss of EDCs from perfused heart explants could be due to the wash-out of a requisite growth factor, this is unlikely given the high serum content employed in the culture media.
The absence of electrically-evoked calcium transients in EDCs following transplantation into infarcted hearts provided additional evidence for the lack of cardiomyogenic potential. Since the cell culture techniques used to generate the EDCs was similar to those employed in earlier studies, which obtained positive results
, it is unlikely that in vitro
culturing per se
resulted in a loss of cardiomyogenic potential. It is however possible that culturing in suspension (i.e., generation of so-called cardiospheres) is required to unmask cardiomyogenic potential. Indeed, cardiomyogenesis proceeds much more efficiently when ES cells are differentiated in suspension culture rather than in monolayer culture. Regardless of such theoretical activities, the inability of EDCs to form functional myocardium following transplantation into injured hearts strongly argues that their normal activity in vivo
does not entail direct cardiomyogenic activity.
The initial observation of GATA-4 transcripts in EDCs from non-transgenic mice and GFP fluorescence in EDCs from MLC2v-Cre/ZEG reporter mice underscores potential difficulties encountered when examining cardiomyogenic differentiation in vitro
. In the case of GATA-4 expression, it is noteworthy that previous analyses of mesenchymal derived cells showed partial activation of “myocyte specific” transcription in the absence of cardiomyogenic activity
. Thus, transcriptional induction of a limited number of target genes is insufficient evidence of lineage induction. Although phagocytosis offers a ready explanation for the presence of GFP epifluorescence in the absence of reporter gene recombination in EDCs isolated from MLC2v-Cre/ZEG reporter mice, the absence of beta-galactosidase activity is problematic. This phenomenon of failed beta-galactosidase activity in the ZEG mouse lines has been previously described: beta-galactosidase activity was absent in the liver and lungs of ZEG lines, but robust GFP fluorescence was observed following cre-mediated recombination
. The author of that study suggested that low or absent expression of β-galactosidase may be an occasional phenomenon related to the poor expression of prokaryotic genes in eukaryotes. A similar mechanism could be at play in our model.
In conclusion, the present study demonstrated that, under our conditions, EDCs are not cardiac progenitors and raises caution about using cells from this source in human clinical trials. This study has also shown that the use of a limited number of parameters may not be sufficient to correctly identify the source and fate of cardiomyogenic progenitor cells.