Tumor hypoxia, p53, and prognosis in cervical cancers

Background: The p53 protein is involved in the regulation of initiation of apoptosis. In vitro, p53-deficient cells do not respond to hypoxia with apo ptosis as do p53-normal cells, and this may lead to a relative growth advan tage of cells without a functioning p53 under hypoxia. On the basis of this hypothesis, a selection of cells with a functionally inactive p53 may occu r in hypoxic tumors. The development of uterine cervical carcinomas is clos ely associated with infections of human papilloma viruses, which may cause a degradation of the tumor suppressor gene p53, resulting in a restriction of apoptosis. Thus, cervical cancers have often a functionally inactive p53 . The purpose of our clinical study was therefore to investigate the associ ation between p53, hypoxia, and prognosis in cervical cancers in which the oxygenation status can be determined by clinical methods. Material and Methods: Seventy patients with locally advanced squamous cell cervical cancer Stages IIB (n = 14), IIIB (n = 49), and IVA (n = 7) were in vestigated in the period from 1996 through 1999. All were treated with defi nitive radiotherapy with curative intent by a combination of external radio therapy plus high-dose-rate afterloading. Before therapy, tumor oxygenation was measured with a needle probe polarographically using the Eppendorf his tograph. Hypoxic tumors were defined as those with pO(2) measurements below 5 mm Hg (HF5). Pretreatment biopsies were taken and analyzed immunohistolo gically for p53 protein expression with the DO-7 antibody. The DNA index wa s measured by flow cytometry. The statistical data analysis was done with S PSS 9.0 for Windows. Results: The 3-year overall survival was 55% for the whole group of patient s. Clinical prognostic factors in a multivariate analysis were pretreatment hemoglobin level (3-year survival 62% for patients with a pretreatment hem oglobin greater than or equal to 11 g/dl vs. 27% for hemoglobin <11 g/dl, p = 0.006) and FIGO stage (Stage IIB: 65%; Stage IIIB: 60%; Stage IVA: 29%,p = 0.01). Sixty of the 70 tumors showed positive immunohistologic staining for p53 protein (transformed p53 = tp53), and 10/70 were negative (wild-typ e p53 = wtp53); p53 expression had no significant impact on survival (50% f or tp53 vs. 79% for wtp53,p = 0.11). FIGO stage and anemia had no impact on p53 expression. Forty-nine of 70 tumors were hypoxic (HF5+), and 21 showed no hypoxia (HF5-). Hypoxic carcinomas were more frequently positive for p5 3 as compared to nonhypoxic tumors (27% vs. 13%, p = 0.011) and showed a tr end toward a lower survival (48% vs. 70%, p = 0.07). In a further multivari ate analysis, the impact of a combination of p53 expression and hypoxia on survival was examined. After adjusting for FIGO stage and pretreatment anem ia, patients with wtp53 tumors had the best prognosis (3-year survival 79%) followed by tp53-HF5(-) patients (57%), and the most unfavorable prognosis was observed for tp53-HF5(+) patients (47%). The DNA index was higher in t p53 carcinomas compared to wtp53 tumors, 1.97 +/- 0.4 vs. 1.67 +/- 0.1, p = 0.05. The highest DNA index was found in hypoxic tumors with transformed p 53 (2.2 +/- 3.1). Conclusions: Advanced stage and pretreatment hemoglobin level are independe nt prognostic factors in cervical carcinomas. The immunohistologic detectio n of (a functionally inactive) p53 and the presence of hypoxia had no progn ostic impact, if analyzed as single parameters. However, the combination of both parameters was able to discriminate different prognostic subgroups. M oreover, hypoxic cancers were more often immunohistologically positive for tp53 protein and had a higher DNA index with the highest DNA index in tumor s with both hypoxia and tp53 protein expression. These findings in summary support the theory that the tumor's microenvironment may influence the biol ogic behavior via hypoxia. (C) 2001 Elsevier Science Inc.