TP53 mutation status and gene expression profiles are powerful prognostic markers of breast cancer

Abstract

Introduction: Gene expression profiling of breast carcinomas has increased our understanding of the heterogeneous biology of this disease and promises to impact clinical care. The aim of this study was to evaluate the prognostic value of gene expression-based classification along with established prognostic markers and mutation status of the TP53 gene (tumour protein p53) in a group of breast cancer patients with long-term (12 to 16 years) follow-up.

Methods: The clinical and histopathological parameters of 200 breast cancer patients were studied for their effects on clinical outcome using univariate/multivariate Cox regression. The prognostic impact of mutations in the TP53 gene, identified using temporal temperature gradient gel electrophoresis and sequencing, was also evaluated. Eighty of the samples were analyzed for gene expression using 42 K cDNA microarrays and the patients were assigned to five previously defined molecular expression groups. The strength of the gene expression based classification versus standard markers was evaluated by adding this variable to the Cox regression model used to analyze all samples.

Results: Both univariate and multivariate analysis showed that TP53 mutation status, tumor size and lymph node status were the strongest predictors of breast cancer survival for the whole group of patients. Analyses of the patients with gene expression data showed that TP53 mutation status, gene expression based classification, tumor size and lymph node status were significant predictors of survival. Breast cancer cases in the 'basal-like' and 'ERBB2+' gene expression subgroups had a very high mortality the first two years, while the 'highly proliferating luminal' cases developed the disease more slowly, showing highest mortality after 5 to 8 years. The TP53 mutation status showed strong association with the 'basal-like' and 'ERBB2+' subgroups, and tumors with mutation had a characteristic gene expression pattern.

Conclusion: TP53 mutation status and gene-expression based groups are important survival markers of breast cancer, and these molecular markers may provide prognostic information that complements clinical variables. The study adds experience and knowledge to an ongoing characterization and classification of the disease.

Figure 1

Hierarchical clustering using 'intrinsic' genes. (a) Five groups were identified, namely the highly proliferating luminal (light blue), luminal A (dark blue), normal-like (green), basal-like (red) and ERBB2⁺ (magenta) groups, which were used in the survival analysis. (b) Hierarchical clustering was performed using 540 clones, representing 496 unique genes from the 'intrinsic' gene list. The individual color of the dendrogram branches reflects the correlation with centroids previously described, and tumors with low correlation (<0.2) with a specific subtype are here indicated by gray branches. Gene clusters characterizing the five groups (a) involve genes encoding, for example, (c) estrogen receptor (ER), (d) MUC1, (e) cadherin 1 (E-cadherin; CDH1), (f) FOXC1 and (g) ERBB2. Since very few genes associated with cell division and proliferation are part of the 'intrinsic' gene list, such a cluster was selected from the total list of genes (Additional file 4), clustered and organized according to (h) the 'intrinsic' dendrogram to show the difference in proliferation between the two luminal groups. (i) In the same manner, a gene cluster characteristic of the mucinous breast carcinomas was made from the total list of genes.

Figure 2

Correlation with five centroids. The expression profiles of the samples were correlated with five previously defined centroids (listed with color codes). The correlation coefficients were plotted according to the dendrogram in Figure 1 and a continuous and opposite curve-pattern of luminal A versus basal-like is evident. The subcluster of luminal samples with the second highest correlation with the luminal B centroid is named 'highly proliferating luminals' in our study.

Figure 3

Kaplan-Meier survival curves. (a) Kaplan-Meier plots of breast cancer survival for all patients. The left panel shows tumor size (T1, T2, T3+T4) and the right panel TP53 mutation status (WT, wild type; MUT, mutation). (b) Kaplan-Meier plots of breast cancer survival for patients with gene expression data; the left panel shows tumor size and the right panel TP53 mutation status. (c) Kaplan-Meier plots of breast cancer survival and recurrence-free survival according to gene expression group (LA, luminal A; NL, normal-like; ERBB2; BL, basal-like; HPL, highly proliferating luminals). The p value from the log rank test (Mantel-Cox) is shown in each panel; 'n' is the number of samples in each group. Deaths due to causes not related to breast cancer were treated as censored observations.

Figure 4

Clinical, histopathological and molecular characteristics of subgroups. Dendrogram from hierarchical clustering with distribution of clinical, histopathological and molecular markers between the five gene expression groups (highly proliferating luminals, luminal A, normal-like, basal-like, and ERBB2⁺). The color coding is as follows: red, description to the left; gray, unknown; yellow, all other. P values from cross-tabulation and Pearson X²-test are shown to the right of the panel. Relative expression of mRNA is shown for TP53 (Clone-ID: IMAGE:24415 and IMAGE:236338) and estrogen receptor (ER) (IMAGE:725321). The fraction of malignant cells in tumor tissue and histological type are shown in the lower panel: DCIS, ductal carcinoma in situ with microinvasion IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; MPC, metaplastic carcinoma; MUC, mucinous carcinoma; TLC, tubulolobular carcinoma. ILC* is the 'typical lobular' type previously described [36]. IHC, immunohistochemistry; LOH, loss of heterozygosity.

Figure 5

Gene expression pattern associated with TP53 mutations status. Hierarchical clustering of 80 samples using 80 genes selected by significance analysis of microarray analysis to be associated with TP53 mutation status. Tumor samples with TP53 mutation are labeled red and wild-type samples are labeled green (upper dendrogram). Ten genes with the highest correlation in each of the two main branches of the gene cluster (left dendrogram) are listed on the right.