Influence of hypoxia and irradiation on osteopontin expression in head and neck cancer and glioblastoma cell lines

Background Tumor hypoxia is a known risk factor for reduced response to radiotherapy. The evaluation of noninvasive methods for the detection of hypoxia is therefore of interest. Osteopontin (OPN) has been discussed as an endogenous hypoxia biomarker. It is overexpressed in many cancers and is involved in tumor progression and metastasis. Methods To examine the influence of hypoxia and irradiation on osteopontin expression we used different cell lines (head and neck cancer (Cal27 and FaDu) and glioblastoma multiforme (U251 and U87)). Cells were treated with hypoxia for 24 h and were then irradiated with doses of 2 and 8 Gy. Osteopontin expression was analyzed on mRNA level by quantitative real-time RT-PCR (qPCR) and on protein level by western blot. Cell culture supernatants were evaluated for secreted OPN by ELISA. Results Hypoxia caused an increase in osteopontin protein expression in all cell lines. In Cal27 a corresponding increase in OPN mRNA expression was observed. In contrast the other cell lines showed a reduced mRNA expression under hypoxic conditions. After irradiation OPN mRNA expression raised slightly in FaDu and U87 cells while it was reduced in U251 and stable in Cal27 cells under normoxia. The combined treatment (hypoxia and irradiation) led to a slight increase of OPN mRNA after 2 Gy in U251 (24 h) and in U87 (24 and 48 h) cell lines falling back to base line after 8 Gy. This effect was not seen in Cal27 or in FaDu cells. Secreted OPN was detected only in the two glioblastoma cell lines with reduced protein levels under hypoxic conditions. Again the combined treatment resulted in a minor increase in OPN secretion 48 hours after irradiation with 8 Gy. Conclusion Osteopontin expression is strongly modulated by hypoxia and only to a minor extent by irradiation. Intracellular OPN homeostasis seems to vary considerably between cell lines. This may explain the partly conflicting results concerning response prediction and prognosis in the clinical setting