The optimal cell-matrix combination for robust and sustained myocardial restoration has not been identified. The present study utilizes embryonic stem cells as the substrate of bioartificial myocardial tissue and evaluates engraftment in, and functional recovery of, the recipient heart. Collagen type I was populated with undifferentiated green fluorescent protein (GFP)-positive mouse embryonic stem cells. An intramural left ventricular pouch was fashioned after ligation of the left anterior descending artery in an athymic nude rat heterotopic heart transplant model. The bioartificial mixture (0.125 ml) was implanted in the infarcted area within the pouch. Echocardiography was performed to assess fractional shortening in: Group I, infarcted rats that received cell-matrix implants; Group II, rats given matrix implant without cells; Group III, rats given no matrix or cells; and Group IV, rats receiving transplanted hearts without ligation (n = 5/group). Hearts were stained for GFP, cardiac markers (connexin-43, alpha-sarcomeric actin), hematoxylin-eosin (H&E) and trichrome. Embryonic stem cells formed stable intramyocardial grafts that were incorporated into the surrounding area without distorting myocardial geometry, thereby preventing ventricular wall thinning (anterior wall thickness was: Group I, 1.4 +/- 0.1 mm; Group II, 1.0 +/- 0.1 mm, Group III, 0.9 +/- 0.2 mm; and Group IV, 1.3 +/- 0.2 mm). The inoculated cells expressed connexin-43 and alpha-sarcomeric actin in vivo. Fractional shortening was better in embryonic stem cell-treated animals (Group I, 21.5 +/- 3.5%; Group II, 12.4 +/- 2.8%; Group III, 8.2 +/- 2.9%; Group IV, 23.2 +/- 4.2%). Embryonic stem cells are an efficient alternative substrate for myocardial tissue engineering and can prevent myocardial wall thinning and improve contractility after implantation into injured myocardium in a 3-dimensional matrix.