Mesenchymal stem cells (MSCs) possess self-renewal and multipotential differentiation abilities, and they are thought to be one of the most reliable stem cell sources for a variety of cell therapies. Recently, cell therapy using MSCs has been studied as a novel therapeutic approach for cancers that show refractory progress and poor prognosis. MSCs from different tissues have different properties. However, the effect of different MSC properties on their application in anticancer therapies has not been thoroughly investigated. In this study, to characterize the anticancer therapeutic application of MSCs from different sources, we established two different kinds of human MSCs: umbilical cord blood-derived MSCs (UCB-MSCs) and adipose-tissue-derived MSCs (AT-MSCs). We used these MSCs in a coculture assay with primary glioblastoma multiforme (GBM) cells to analyze how MSCs from different sources can inhibit GBM growth. We found that UCB-MSCs inhibited GBM growth and caused apoptosis, but AT-MSCs promoted GBM growth. Terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick-end labeling assay clearly demonstrated that UCB-MSCs promoted apoptosis of GBM via tumor necrosis factor-related apoptosis-inducing ligand (TRAIL). TRAIL was expressed more highly by UCB-MSCs than by AT-MSCs. Higher mRNA expression levels of angiogenic factors (vascular endothelial growth factor, angiopoietin 1, platelet-derived growth factor, and insulin-like growth factor) and stromal-derived factor-1 (SDF-1/CXCL12) were observed in AT-MSCs, and highly vascularized tumors were developed when AT-MSCs and GBM were cotransplanted. Importantly, CXCL12 inhibited TRAIL activation of the apoptotic pathway in GBM, suggesting that AT-MSCs may support GBM development in vivo by at least two distinct mechanisms-promoting angiogenesis and inhibiting apoptosis. The opposite effects of AT-MSCs and UCB-MSCs on GBM clearly demonstrate that differences must be considered when choosing a stem cell source for safety in clinical application.