Background: Myoblasts and stem cells implanted into myocardium can differentiate into myocytes and may functionally improve impaired ventricles. For implanted cells to actually contribute to the synchronous contractions of the heart, however, they must be anatomically integrated with the existing native myocardial fibers. Methods: Isogenic Lewis rats were used as donors and recipients to simulate clinical autotransplantation. Either skeletal myoblasts or marrow stromal stem cells were isolated from donors, culture expanded, and labeled with 4′,6-diamidino-2-phenylindole (4′,6-diamidino-2-phenylindole). Labeled cells were then injected into the myocardium of recipients. At intervals specimens were obtained, sectioned, and stained with hematoxylin and eosin and immunohistochemically against connexin-43 to demonstrate gap junctions (intercalated discs). Results: At 1 week the labeled cells were still undifferentiated, but early expression of connexin-43 could be detected at contact points between the implanted cells and the native myocytes. By 4 to 6 weeks, labeled, fully differentiated myocytes could be seen to interconnect among themselves and with native cardiomyocytes by means of intercalated discs. In sections parallel to the myofibers, full integration of new labeled myocytes with the native myofiber cells could be observed. Furthermore, the labeled myofibers were in parallel with the native, unlabeled fibers. We postulate that such supracellular structural integration was enhanced by fiber stretching during cardiac contractions, sending signals for cellular reorientation and incorporation, in which the cytoskeletal system may play an important role. Conclusion: We conclude that implanted precursor cells can be integrated into native myocardial structure so as to contribute to myocardial function. Direct cell-to-cell contact seems to be an important signaling mechanism, which has implications for cellular implantation strategies.J Thorac Cardiovasc Surg 2002;124:584-90