LAUSR

Introduction Glioblastoma (GB) are known to be highly hypoxic and both hypoxia inducible factors, HIF-1 and HIF-2, have been implicated in their growth and resistance to treatments. Recently, it was proposed that a severe chronic hypoxia would enhance HIF-2 expression at the expense of HIF-1 through HAF (hypoxia-associated factor). HAF is expressed in many cancers including high and low grade glioma. However, its implication in GB growth and its treatments is poorly documented (Koh et al., 2011). In this context, we conducted in vivo and in vitro studies to identify whether HAF expressed by glioma cells modulates tumour growth and efficacy of GB conventional treatments, i.e. chemotherapy (temozolomide, TMZ) and radiotherapy (X-rays). Material and methods Stable inhibition of HAF expression was established in human glioblastoma cells by RNA interference (U251shHAF). Orthotopic GB models were developed in mice (8/group) for U251shHAF and U251Sc cells (scrambled-shRNA infected cells) as control. Tumour development was assessed with 7T MRI (T2w sequence). At the end of the experiments, an immunohistology study was performed to characterise the vascularisation (PECAM), glial (GFAP) and inflammatory (CD68) reactions. In vitro, the radio- and chemosensitivity of U251shHAF were studied by clonogenic assay and cell cycle analysis following X-rays irradiation (X-RAD 225Cx) or TMZ exposition. Annexin-V binding and propidium iodure uptake followed by flow cytometry was used to quantify apoptotic and necrotic cells. Results and discussions The stable inhibition of HAF expression in U251 cells leads to around 70% of its extinction in either normoxia or hypoxia (1% O2). Accordingly, the expression of VEGFA and CAIX, both known as HIF-1 and HIF-2 dependent genes, was decreased in U251shHAF cultured in hypoxia (1% O2) compared to U251Sc cells. Loss of function of HAF leads to a significant growth delay of U251shHAF tumours of 3 weeks compared to U251Sc tumours, although both tumours display similar vascularisation, glial and inflammation reactions. In other hand, HAF silencing in glioma cells does not modified their sensitivity to X-rays or TMZ as suggested by the similar results obtained for both U251shHAF and U251Sc cells, through clonogenic assay, cell cycle and apoptosis analyses. Conclusion Our results suggest that HAF might be of poor prognosis for GB since its inhibition in glioma cells reduces tumour growth without alleviating glioma cell chemo- and radioresistance.