Comparison of different methods of CAIX quantification in relation to hypoxia in three human head and neck tumor lines

Purpose: In head and neck cancer, it has been shown that hypoxic tumors respond poorly to therapy. Methods to identify hypoxic tumors are, therefore, of importance to select patients for oxygenation modifying or other intensified treatments. The aim of this study was to compare tumor cell hypoxia assessed by the hypoxic cell marker pimonidazole (PIMO) with expression of the endogenous hypoxia-related marker carbonic anhydrase IX (CAIX) in three human head and neck tumor lines. Material and methods: Forty-five tumors of three human head and neck tumor lines, SCCNij3, SCCNij59 and MEC82, xenografted in athymic mice, were used. CAIX was quantified by biodistribution (% injected dose/g tumor) after injecting 3-5 mu l In-111-labeled G250 mouse antibody 3 days prior to euthanizing. In a tissue section from the same tumor, fractions of tumor area positive for PIMO, CAIX and Hoechst 33342 (perfusion marker) were assessed after immunohistochemical staining, using a digital image analysis system. Results: SCCNij3 and MEC82 were relatively hypoxic tumor lines with fractions of tumor area positive for pimonidazole of 0.16 and 0.15, respectively. SCCNij59 was a better-oxygenated tumor line with a PIMO-fraction of 0.03. The three tumor lines showed different levels and patterns of CAIX immunohistochemical staining, but only in MEC82 there was a good correlation between PIMO-fraction and CAIX-fraction (r(2) =0.92, P<0.0001). Correlations between In-111-G250 uptake and CAIX-fraction or PIMO-fraction within tumor lines were weak or absent. Conclusions: Assessment of CAIX expression depends largely on the techniques and tumor lines used. Furthermore, the immunohistochemical staining pattern of CAIX relative to PIMO differs between human tumor lines of similar anatomical origin. Therefore, the use of CAIX as endogenous marker of tumor hypoxia remains questionable. (C) 2005 Elsevier Ireland Ltd.