Pro-inflammatory gene expression in solid glioblastoma microenvironment and in hypoxic stem cells from human glioblastoma

Background: Adaptation to hypoxia and consequent pro-inflammatory gene expression of prostate and breast carcinomas have been implicated in the progression toward cancer malignant phenotype. Only partial data are available for the human tumor glioblastoma multiforme (GBM). The aim of our study was to analyze the hypoxic and pro-inflammatory microenvironment in GBMs and to demonstrate that in a stem/progenitor cell line derived from human glioblastoma (GBM-SCs), hypoxia activates a coordinated inflammatory response, evidencing an invasive and migratory phenotype. Methods: From each of 10 human solid glioblastomas, clinically and histopathologically characterized, we obtained three surgical samples taken from the center and the periphery of the tumor, and from adjacent host normal tissue. Molecular and morphological analyses were carried out using quantitative real-time PCR and western blot (WB). GBM stem and differentiated cells were incubated under hypoxic conditions and analyzed for pro-inflammatory gene expression and for invasive/migratory behavior. Results: A panel of selected representative pro-inflammatory genes (RAGE and P2X7R, COX2, NOS2 and, PTX3) were analyzed, comparing tumor, peritumor and host normal tissues. Tumors containing leukocyte infiltrates (as assessed using CD45 immunohistochemistry) were excluded. Selected genes were overexpressed in the central regions of the tumors (i.e. in the more hypoxic areas), less expressed in peripheral regions, and poorly expressed or absent in adjacent normal host tissues. Western blot analysis confirmed that the corresponding pro-inflammatory proteins were also differently expressed. Hypoxic stem cell lines showed a clear time-dependent activation of the entire panel of pro-inflammatory genes as compared to differentiated tumor cells. Biological assays showed that invasive and migratory behavior was strengthened by hypoxia only in GBM stem cells. Conclusions: In human solid glioblastoma we have observed a coordinated overexpression of a panel of pro-inflammatory genes as compared to host normal tissue. We have also evidenced a similar pattern of overexpressed genes in GBM-SCs after hypoxic treatment, showing also a gain of invasive and migratory function that was lost when these stem cells differentiated. We suggest that, as has been previously described for prostatic and mammary carcinoma, in human glioblastoma acquisition of a proinflammatory phenotype may be relevant for malignant progression.