Measurement of hypoxia using invasive oxygen-sensitive electrode, pimonidazole binding and 18F-FDG uptake in anaemic or erythropoietin-treated mice bearing human glioma xenografts

Abstract

Relationship between haemoglobin levels and tumour oxygenation has been already reported. The purpose of this work was to compare in human malignant glioma-bearing mice the sensitivity of two well established techniques of tumour hypoxia assessment, especially their ability to detect expected weak variations of tumour oxygenation status associated to haemoglobin level modifications. The relationship between tumour hypoxia and glucose metabolism was also investigated. Experiments were performed on a human malignant glioma (GBM Nan1) xenografted into nude mice. Twenty-four hours after tumour implantation, animals were randomized into three groups: 'Anaemia' for mice subjected to repeated blood samplings, 'Control', and 'rHuEPO' for mice receiving recombinant human erythropoietin. Once the tumours reached a volume of 300+/-100 mm(3), tumour hypoxia was assessed both using the pO(2)-Histograph, Eppendorftrade mark and the pimonidazole binding assay. Glucose metabolism was evaluated by (18)F-FDG autoradiography and compared with the pimonidazole binding distribution pattern. Repeated blood samplings significantly reduced mean haemoglobin levels (10.9+/-2.0 g/dl), inducing chronic anaemia in mice, while daily administration of rHuEPO led to increase of haemoglobin levels (15.8+/-2.0 g/dl). Oxygenation status evaluated by a microelectrode was worsened in anaemic mice (mean pO(2) in tumour = 6.9+/-0.8 mmHg) and improved in rHuEPO-treated animals (mean pO(2)in tumour = 11.4+/-1.2 mmHg). No correlation was observed between the oxygen-sensitive probe and pimonidazole labelling results: both techniques give different but complementary information about tumour hypoxia. Areas of high pimonidazole binding and areas of high (18)F-FDG uptake superimposed well. Present results confirm that modification of haemoglobin levels leads to alteration of tumour oxygenation status. These variations were detectable using the oxygen-sensitive electrode but not the pimonidazole binding assay. The strong correlation between pimonidazole labelling and (18)F-FDG uptake suggests a positive relationship between hypoxia and increased glucose metabolism in this tumour model.