Given the loss of func tional PgR in MCF7 miR155 cells compared to vector

In this model, a clonal tumor cell subpopulation displays non CSC characteristics, including poor anchorage independent growth in vitro and limited tumorigenic or metastatic potential in vivo, whereas a second subpopu lation is enriched in CSCs INNO-406 ic50 as inferred from vigorous anchorage independent growth and tumor formation and lung colonization potential in vivo. We had previ ously found that the non CSC subpopulation enhances the ability of the CSC subpopulation for metastatic dis semination, and produced preliminary evidence that this activity involved, at least in part, diffusible factors secreted by S cells. In the present study, we have confirmed and extended our original observations, including the in duction of epithelial mesenchymal transition in the CSC enriched epithelial PC 3M subpopulation upon culture with conditioned medium from the non CSC PC 3S sub population, and the dependence of this effect on specific signaling pathways.<br><br> We have applied a comparative proteomics approach in an effort to identify proteins differentially secreted by S vs. M cells as candidates to explain the observed paracrine effects. Immunodepletion, spe cific transcript knockdown and complementation experi ments have led us to conclude that, of the secreted proteins with the strongest differential secretion LBH589 between S and M cells, the matricellular protein SPARC, abun dantly secreted by S cells, explains most of the pro invasive effects of S conditioned medium on M cells.<br><br> Knock down of SPARC abrogated not only the pro invasive activity of the conditioned medium from the non CSC S subpopulation on オーダー LY2109761 the CSC enriched M cells, but also the overall pro invasive and pro metastatic activ ities observed upon co culture in vitro or co inoculation in vivo. Of note, conditioned medium from non CSC S cells required specific signaling pathways for the enhance ment of the invasive behavior of M cells, including PI3K, MAPK, NF κB and non receptor tyrosine kinase path ways, which are also involved in cellular responses elicited by SPARC. The observed induction by SPARC in M cells of the MAPK and PI3K AKT signaling pathways fur ther supports the role played by these pathways in our model. This underlines the relevance of paracrine interac tions between tumor cell subpopulations displaying non CSC and CSC properties to modulate the phenotypic out comes of the tumor.<br><br> Other laboratories performing un biased secretome analyses have shown that SPARC also mediates cooperation between tumor cell subpopulations with different invasive potentials in ovarian and bladder cancer models. Although our knockdown and complementation experi ments strongly support that SPARC is indeed the key fac tor mediating invasive and metastatic cooperation in our interacting CSC vs. non CSC neoplastic subpopulation model, other molecules produced by neoplastic or non neoplastic cells, such as PAI 1, could also participate in this process. In this regard, we observed that knockdown of SPARC in S cells was accompanied with a decrease in PAI 1 expression. It is known that the expression of PAI 1 is modulated by changes in the expression levels of SPARC. Our subsequent immunodepletion experi ments suggest that PAI 1 may not play a major role in the enhanced invasiveness of M cells by S CM.