We report that healthy mesenchymal stem cells are capable of rescuing post-ischemic cardiomyoblasts from cell death through a mechanism not yet implicated in the effects of stem cells after ischemic conditions. Thus, the beneficial effect of stem cell grafting may be based not only on improved neovascularisation and replacement of lost cells but on rescuing the damaged cells of the host as well.
The most likely explanation of the beneficial effects of MSC co-culture is that these cells improve the chances of the damaged H9c2 cells to restore their function and prevent later cell death. Ethidium homodimer has been shown to stain not only dead but damaged cells as well . Thus, although most H9c2 cells were stained by ethidium homodimer after OGD (Figure 1B), many cells were probably only reversibly damaged. An alternative explanation of our results could be an increased replication of surviving H9c2 cells. However, the nearly ten-fold difference in the number of viable H9c2 cells between our experimental groups 24 hr after OGD (Figure 3C) and the normal doubling time of these cells make this possibility unlikely to explain the difference.
The used in vitro ischemia model demonstrates that the beneficial effects of MSC co-culture seem to be dependent on direct cell-to-cell connections and intercellular nanotubes.
Nanotube formation has already been shown to occur among endothelial progenitor cells, cardiomyocytes [20, 24], immune cells and other lineages [25, 26]. The characterization of nanotubes revealed that these filaments contain actin and in some cases, microsomes or mitochondria . We found that this phenomenon occurs frequently between cardiomyoblasts and mesenchymal stem cells. Stem cells failed to rescue post-ischemic cardiomyoblasts when intercellular connections were blocked by a physical barrier. These observations indicate that intercellular connections work toward the survival of cells both during and after ischemia, however, the underlying mechanisms may be slightly different.
One plausible mechanism for the rescuing effect is that transplanted cells improve regeneration through secreting paracrine factors [14, 28–31]. However, results from our experiments with the plate insert show that paracrine factors secreted by the cells are probably too low in our system to have any beneficial effect on these severely damaged cells. This in vitro experimental setup allows the investigation of cell-to-cell contacts, however, it cannot rule out that paracrine effects play a significant role in a more physiological in vivo setting. The time frame of the experimental protocol is also important. In our experiments we added the cells at an early time point and terminated the experiment before significant differentiation can occur. During a later time point the effect of paracrine factors is probably much more important especially in the differentiation process as shown by several other authors [6, 11].
Cell fusion is another phenomenon which is frequently observed in co-culture studies and in some cases, in in vivo experiments as well [15, 32, 33]. Several studies have shown that cell fusion can result in transdifferentiation, thus offering an alternative mechanism by which grafted cells improve the infarcted myocardium. Using videomicroscopy we also found several double labeled, double nuclei cells indicative of cell fusion. However, cell fusion showed high variations among different culture and detection techniques, and therefore extensive cell fusion as an in vitro artefact cannot be ruled out [14, 34, 35]. During our investigations we only observed a few unquestionable cell fusions which cannot account for the rescue of the high number of damaged cardiomyoblasts .
We also found double labeled cells without double nuclei in the co-culture of cardiomyoblasts and stem cells after 24 hours. The double labeling of these cells may be the result of direct cell-to cell connections. During these periods of connection, cells are able to exchange membrane parts and Vybrant dye molecules can drift from one cell to another. Movement of dye molecules from one cell to another through gap junction connections is precluded because the lipophilic Vybrant dyes are high molecular weight stains and cannot permeate through gap junctions . Driesen et al.  showed that low molecular weight tracers such as calcein-AM get from one cell to another through gap junctions, and high molecular weight tracers by partial cell fusion, thus the conclusion may be drawn that dye transfers after 24 hours in our experiment are most probably the results of direct cell-to-cell connections. Still, gap junctions may create an opportunity for grafted cells to interact with the host tissues [37, 38]. In the present experimental model most of cell-to-cell interactions were short-lived tubular connections, which formed a constantly changing web between the two investigated cell types.
Our experimental model was devised to investigate acute effects with high temporal and spatial resolution, therefore ruling out differentiation, which occurs over time. Moreover, an in vitro transplantation model in a cell culture system cannot mimic the 3-dimensional tissue where cell-to-cell connections are different. These circumstances obviously limit the conclusions drawn from our results. On the other hand, this experimental setting was necessary and favorable to investigate short-term cellular interactions.