Background: Focal application of MSCs of which differentiation has been observed in vivo at extremely low rates could create fixed tissue heterogeneity that may predispose the heart to ventricular arrhythmia. We hypothesized that the modification of MSCs to cardiogenic cell prior to delivery could help prevent post-infarct arrhythmia by acquiring tissue homogeneity. Methods: We screened 41 characterized inhibitors of six major protein kinase subfamilies to alter the orchestration of multiple signaling involved in differentiation of MSCs. To differentiate MSCs into cardiogenic cells, MSCs were treated with protein kinase C activator, phorbol myristate acetate. Immunocytochemical characterization of MSCs was performed prior to the delivery. MSCs (n=19) and differentiated MSCs (dMSCs, n=12) were respectively administrated at border zone in rats with MI. To assess cardiac function and electrophysiological change including conduction velocity (CV), we performed in vivo LV catheterization and optical mapping using di4-ANEPSS in Langendorff pefused hearts. Results: Immunocytochemical staining of dMSCs confirmed differentiation of MSCs into cardiogenic cells by increased fluorescence of cardiac-specific markers such as cTnT, MLC, and MHC. At 7-10 days after injury and administration, pressure-volume loop analyses showed more improved LV contractile function in dMSCs group than MSCs group. Ex vivo pacing protocol induced VT or VF in 13% of dMSC and 44.4 % of MSC group (p<0.05). CV maps revealed marked restoration of CV in the dMSCs-engrafted lesion with recovery of local CV, in contrast to partial restoration of CV in the MSC-engrafted lesion (0.84 짹 0.15 m/s vs. 0.43 짹 0.1 m/s, p<0.05). The direction of the CV vector was homogeneous in the dMSCs-engrafted lesion, whereas it was heterogeneous in the MSC-engrafted lesion. The rise time of action potential phase 0 was shorter in the dMSC than in the MSCs engrafted lesion (2.9짹0.1 ms vs. 3.3짹0.2 ms, p<0.05), suggesting elimination of passive electrotonic current flow by acquiring cardiogenic properties. Incidence of sudden deaths was much lower in dMSCs group than MSCs group(8.3% in CPM group vs. 31.6% in MSC group, p<0.05). Histologic analysis revealed expression of connexin43 between dMSCs and surrounded cardiomyocytes in the infarcted region, few inflammatory cells, and low fibrotic markers in myocardium implanted with MSCs. Conclusions: We found that engraftment of dMSCs at border zone enhance conduction and suppress VT. It suggests that the application of dMSCs in infarcted myocardium can be more optimal therapy for preventing ventricular arrhythmia by reducing tissue inhomogeneity.