Different Effects of Human Umbilical Cord Mesenchymal Stem Cells on Glioblastoma Stem Cells by Direct Cell Interaction or Via Released Soluble Factors

Glioblastoma (GBM), the most common primary brain tumor in adults, is an aggressive, fast-growing and highly vascularized tumor, characterized by extensive invasiveness and local recurrence. In GBM and other malignancies, cancer stem cells (CSCs) are believed to drive invasive tumor growth and recurrence, being responsible for radio- and chemo-therapy resistance. Mesenchymal stem cells (MSCs) are multipotent progenitors that exhibit tropism for tumor microenvironment, mediated by cytokines, chemokines and growth factors. Initial studies proposed that MSCs might exert inhibitory effects on tumor development, although to date contrasting evidence have been provided. Different studies supported either MSC anti-tumor activity or their support to tumor growth. Here, we examined the effects of umbilical cord (UC)-MSCs on in vitro GBM-derived CSCs growth, by direct cell-to-cell interaction or indirect modulation, via the release of soluble factors. We demonstrate that UC-MSCs and CSCs exhibit reciprocal tropism when co-cultured as 3D spheroids and their direct cell interaction reduces the proliferation of both cell types. Contrasting effects were obtained by UC-MSC released factors: CSCs cultured in the presence of conditioned medium collected from UC-MSCs promoted CSC proliferation through transient ERK1/2 and Akt phosphorylation/activation. Analysis of cytokine profile released by UC-MSCs in the conditioned medium revealed a strong production of molecules involved in inflammation, angiogenesis, cell migration and proliferation such as IL-8, GRO, ENA-78 and IL-6. Since CXCR2, a common receptor for several of these ligands, is expressed in GBM CSCs, we evaluated its involvement in CSC proliferation induced by UC-MSC-conditioned media. Using the CXCR2 antagonist SB225002 we observed a partial but statistically significant inhibition of CSC proliferation and migration induced by the UC-MSC-released cytokines. Conversely, CXCR2 blockade did not reduce the reciprocal tropism of CSCs and UC-MSCs grown as spheroids. In conclusion, we show that direct (cell-to-cell contact) or indirect (via the release of soluble factors) interactions between GBM CSCs and UC-MSCs in co-culture produce divergent effects on cell growth, invasion and migration, with the former mainly causing an inhibitory response and the latter a stimulatory one, involving a paracrine activation of CXCR2.