Abstract P2-10-43: The combination of immunohistochemistry for predicting TP53 mutation is useful prognostic marker in breast cancer.

Abstract

Abstract The comprehensive expression analysis classified subtypes of breast cancer which is clinically applied for immunohistochemistry (IHC) with ER, HER2 and Ki67. On the other hand, recent whole genome sequencing reveals a certain quantity of TP53 mutation in breast cancer, especially with ER negative tumor. Therefore, it is important for risk evaluation that TP53 mutation is applied for IHC. However, p53 IHC is not well correlated with TP53 mutation because of false negative (deletion mutant) and false positive (stabilization of p53). This study is initiated to improve accuracy to combine the IHC of p53 and p53-regulated protein for predicting TP53 mutation. First, we explored p53-dependent down regulated proteins to rescue the false negative of p53 IHC cases. Using the microarray analysis and DNA sequencing of TP53 gene in 37 breast cancer patients, we extracted 97genes which were overexpressed in TP53 mutation breast cancer cases compared with TP53 wild cases. Furthermore, using the microarray analysis by two tet-inducible p53 cell lines, 597 genes were extracted for p53-dependent repressed genes. There were 29 common genes between both of them. In 29genes, we picked up three genes, BIRC5, BUB1 and PLK1 which were confirmed with p53-dependent repression by RT-PCR or luciferase assay. Next, the correlation of TP53 status and IHC of these genes products and p53 were evaluated. In 37 breast cancer patients, TP53 mutations were detected in 19 cases (51%). There were 11 missense mutations, 5 protein-truncating mutations and two splicing-site mutations. p53 IHC positive (>10%) were 12 cases (32.4%) which included only 9 TP53 mutations. Five protein-truncating TP53 mutations were all negative staining of p53. Among BIRC5, BUB1 and PLK1 proteins, positive staining of PLK1 is the most correlated with TP53 mutations, sensitivity 0.78, specificity 0.84 and positive predictive value 0.78 (p < 0.0001). Third, we tested well known p53-tageted gene product p21 IHC whether it is used for decreasing false positive of p53 IHC. Although strong staining of p21 is only 3 cases (8.1%), all of three were wild type TP53. Therefore IHC of p53 + and p21 – (to decrease false positive of p53) or p53- and PLK1+ (to save false negative of p53) were defined as IHC-TP53 mutation+. The IHC-TP53 mutation + were strongly correlated with TP53 mutations, sensitivity 0.89, specificity 0.95 and positive predictive value 0.94 (p < 0.0001). Furthermore, p53 IHC, TP53 mutation and IHC-TP53 mutation were evaluated as prognostic marker. In only 37 breast cancers, TP53 mutation and IHC-TP53 mutation+ were statistically significant for poor prognosis (log rank test, p = 0.005, p = 0.023, respectively) whereas p53 IHC was not. We next carried out IHC of other 157 breast cancers to validate IHC-TP53 mutation as prognostic marker. In this cases, p53 IHC had also statically significant for prognosis (p = 0.009), however IHC-TP53 mutation were more strong relation to prognosis (p = 0.0006). The combination of p53, p21 and PLK1 IHC are well predicted TP53 mutations and is a useful biological, prognostic marker such as ER or HER2. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-10-43.