p53 mutations and expression in ovarian cancers: correlation with overall survival

The p53 gene is altered in approximately 50% of all human malignancies. p53 overexpression, identified by immunohistochemistry, and p53 mutations, identified by single-strand conformational polymorphism (SSCP) and DNA sequencing, have been described in ovarian cancers. p53 overexpression has been correlated with poor outcome for women with ovarian cancer in some studies. With only limited data, the assumption has been made that p53 overexpression corresponds to p53 mutations. The purpose of this investigation was to assess p53 alterations in ovarian cancer to determine if p53 overexpression corresponds with mutations in the p53 gene, and to assess whether either predicts clinical outcome in ovarian carcinoma. Frozen ovarian carcinoma tumor specimens from 105 patients were analyzed by immunohistochemical staining for p53 expression. SSCP was used to screen for mutations and DNA sequencing was used to confirm the specific mutation in exons 2 to 11, encompassing the entire p53 open reading frame. Those ovarian carcinomas identified as wild-type p53 by SSCP were subjected to automated DNA sequence analysis of the entire open reading frame. Relative to DNA sequence analysis, the sensitivity of SSCP was 85% and the specificity was 98%. Immunohistochemical staining demonstrated that 72 of the 105 (69%) cases had positive immunostaining. SSCP and DNA sequencing identified and confirmed mutations in 60 of the 105 carcinomas (57%). Although there was a statistically significant association between p53 immunostaining and p53 mutations (p = 0.0002), false-negative and -positive results were identified. Tumor grade (p = 0.03), stage (p = 0.08), and overall survival (p = 0.15) were moderately associated with positive p53 immunostaining. Patients with p53 mutations and overexpression had shorter overall patient survival (p = 0.02). The findings demonstrated that, individually, p53 mutations and p53 overexpression were each related to shorter patient survival, but the strongest predictor of outcome was a combination of both mutations and overexpression. Comparisons of overall survival for women with mutations in loop 2, loop 3, and the loop-sheet-helix domains together showed a statistically significant difference in survival compared to survival of women whose ovarian cancers had other mutations (p = 0.046).