Expression and prognostic significance of tissue hypoxia markers in head and neck squamous cell carcinomas (HNSCC)

Abstract

5559 Background: Previous investigators have reported that tumor hypoxia is related to prognosis in HNSCC. We and others have recently identified several hypoxia-induced proteins using gene array and proteomic analyses in a HNSCC cell line. Here, we investigate the in-vivo expression of these proteins on a HNSCC tissue array and their relationship to pO2 and prognosis. Methods: We constructed a tumor tissue array from HNSCC patients who had pretreatment pO2 measurements using the Eppendorf microelectrode. We performed immunohistochemical staining of the tissue arrays for known hypoxia-upregulated proteins (CA IX, cortactin, XBP-1, EphrinA-1, HIG-2, DHFR, Galectin-1 and IKK-β). Mann-Whitney tests and logistic regression method were used to determine the relationship between the expression of these markers and tumor pO2. Cox proportional hazard model and log rank tests were used to determine the relationship between the markers and prognosis. Results: 101 patients formed the cohort of this study, of which 95 had evaluable pO2. Markers that correlated with median tumor pO2 were XPB-1 (p = 0.04), Galectin-1 (p = 0.05) and IKK-β (p = 0.04). EphrinA-1 was of borderline significance (p = 0.07). CA IX did not correlate with tumor pO2. On logistic regression analysis, the only factor that had borderline correlation with tumor pO2 was ephrinA-1 (p = 0.06). CA-IX and EphrinA-1 were prognostic factors for freedom from relapse (FFR), cause-specific survival (CSS) and overall survival (OS); galectin-1 for FFR and CSS; and IKK-β and DHFR for FFR. However, none of these factors maintained prognostic significance on multivariate analysis when adjusted for tumor site, stage and treatment. Conclusions: Although we found that the expression of several proteins correlated with tumor hypoxia and treatment outcomes, we did not find that the expression of a single protein predicted for all of these parameters. More accurate prediction of clinical outcomes will likely be achieved by evaluating the expression of a panel of markers rather than individual makers on tissue arrays. No significant financial relationships to disclose.