Not only do hearts recover poorly after injury, the prevalence of and morbidity from heart damage is high. Not surprisingly, cardiac regeneration is seen as among the most important applications of stem cell research. But although cell replacement therapy, or the successful engraftment of stem cells, works for bone marrow transplantation, this kind of cell therapy will be much more difficult in solid organs, at least according to results, presented at the June 2007 meeting of the ISSCR. Indeed, not only does heart tissue fail to promote integration of transplanted cardiomyocytes, it may even provide a hostile environment.

Infarcted hearts form scar tissue that can't contribute to contraction, thus impeding the ability of the heart to pump. Many researchers are differentiating embryonic or tissue-specific stem cells into cardiomyocytes and injecting them into the infarct, hoping the cells can form grafts with host tissue that contribute to pumping activity. Christine Mummery of University Medical Center Utrecht in the Netherlands showed how difficult it is to coax engrafted cells to integrate functionally into heart.

Mummery made a couple of key observations after labeling cardiomyocytes expressing green-fluorescent protein (GFP) into infarcted hearts of mice. She identified two reasons why apparently encouraging results may not be so.

First, many cells in the heart that give a bright green signal are not expressing GFP. Because a significant background of green cells exists in the host animal, the full emission spectrum of a sample should be presented when GFP is used to identify injected cardiomyocytes. Otherwise, engraftment potential will be overestimated. Previous studies using GFP to label engrafted cardiomyocytes should be re-examined.

Second, Mummery probed cell-cell connections in treated hearts and saw little evidence that host and engrafted cells were communicating. Grafts were surrounded by large quantities of extracellular matrix, but few desmosomes, which form connections between the plasma membranes of the host and grafted cells.

To see whether heart function improved even with disconnected cells, Mummery and colleagues examined whether ES-derived cardiomyocytes ameliorated acute myocardial infarction in an immune-compromised mouse. Four weeks after injection, the ejection fraction had increased significantly. However, this improvement decreased over time and disappeared after 12 weeks. Moreover, surprisingly, even non-cardiomyocyte cells injected as controls had a beneficial effect.

Gordon Keller of the McEwen Centre for Regenerative Medicine in Toronto noted that the heart seems to be a hostile environment for new cells. If engraftment is to be effective, he believes, survival factors must be identified.

Dawn Delo, from Anthony Atala's laboratory at Wake Forest University, showed a way to isolate cardiomyocytes from mouse amniotic fluid stem cells, isolated from amniocentesis1. The researchers achieved superior results in culture by growing cells in a bioreactor that provided cyclic strain resembling the contractile environment of the heart. As a result of the enriched environment, the cardiomyocytes showed better cellular organization and contractility.

When these cells were injected into mice with hypertrophic hearts, the cells spread throughout the myocardium. Delo also noticed that, as in Mummery's study, undifferentiated cells provide a functional benefit. For 30 days ejection fraction was maintained. However the fraction decreased after 60 days, probably because the grafted cells died.

The common conclusion of these labs is that cardiomyocytes aren't prone to long-term engraftment and that transient improvements occur even if no functional connections are formed with the host tissue. In clinical studies, such transient benefits could obscure meaningful benefits. In short, encouraging early results mean little for robust, long-term improvements.

The good news is that real, beating cardiomyocytes can be grown from undifferentiated stem cells, and large quantities of these cells with distinctly human characteristics can be obtained from human embryonic stem cells, as Mummery's talk proved. These differentiated cells will permit screens for drugs that bolster the numbers of cardiomyocytes produced and that help cardiomyocytes engraft and survive. Thus, although this progress may not signal the arrival of effective therapies, it may mark the true beginning of their development.