BackgroundOsteopontin acts thru myeloid zinc finger-1 and transforming growth factor-β to drive the adoption of a cancer-associated fibroblast phenotype by local mesenchymal stem cells. Cancer-associated fibroblasts increase cancer cell stemness. MethodsMesenchymal stem cells were exposed to osteopontin or were cocultured with MB231 human breast cancer cells (high osteopontin producer) in the presence or absence of aptamer (inactivates extracellular osteopontin). Myeloid zinc finger-1 phosphorylation sites were identified, and phosphomutants of T134 (SCAN domain) and S453 (zinc finger DNA binding domain) were constructed. Transforming growth factor-β F and cancer-associated fibroblast markers (smooth muscle actin, vimentin, and tenascin-c) were measured in mesenchymal stem cells. In MB231, stemness markers Sox2, Nanog, and Oct4 were measured. ResultsMesenchymal stem cells plus osteopontin increased transforming growth factor-β and cancer-associated fibroblast markers (P < .05 vs mesenchymal stem cells alone); this was abolished by aptamer inactivation of osteopontin. In mesenchymal stem cells transfected with phosphoresistant myeloid zinc finger-1, osteopontin did not increase cancer-associated fibroblast markers or transforming growth factor-β. In contrast, phosphomimetic myeloid zinc finger-1 increased cancer-associated fibroblast markers and transforming growth factor-β (P < .05 vs mesenchymal stem cells alone). In mesenchymal stem cells plus MB231, MB231 stemness markers were increased (P < .05 vs MB231 alone). In MB231 plus mesenchymal stem cells expressing phosphoresistant myeloid zinc finger-1, MB231 stemness markers were not increased in comparison with MB231 plus mesenchymal stem cells. ConclusionMyeloid zinc finger-1 phosphorylation in mesenchymal stem cells drives the osteopontin-mediated cancer-associated fibroblast phenotype, which then increases the cancer cell stemness profile.