Identification of modulated genes by three classes of chemopreventive agents at preneoplastic stages in a p53-null mouse mammary tumor model

Abstract

Genetically engineered mouse cancer models are among the most useful tools for testing the in vivo effectiveness of the various chemopreventive approaches. The p53-null mouse model of mammary carcinogenesis was previously characterized by us at the cellular, molecular, and pathologic levels. In a companion article, Medina et al. analyzed the efficacy of bexarotene, gefitinib, and celecoxib as chemopreventive agents in the same model. Here we report the global gene expression effects on mammary epithelium of such compounds, analyzing the data in light of their effectiveness as chemopreventive agents. SAGE was used to profile the transcriptome of p53-null mammary epithelium obtained from mice treated with each compound versus controls. This information was also compared with SAGE data from p53-null mouse mammary tumors. Gene expression changes induced by the chemopreventive treatments revealed a common core of 87 affected genes across treatments (P < 0.05). The effective compounds, bexarotene and gefitinib, may exert their chemopreventive activity, at least in part, by affecting a set of 34 genes related to specific cellular pathways. The gene expression signature revealed various genes previously described to be associated with breast cancer, such as the activator protein-1 complex member Fos-like antigen 2 (Fosl2), early growth response 1 (Egr1), gelsolin (Gsn), and tumor protein translationally controlled 1 (Tpt1), among others. The concerted modulation of many of these transcripts before malignant transformation seems to be conducive to predominantly decrease cell proliferation. This study has revealed candidate key pathways that can be experimentally tested in the same model system and may constitute novel targets for future translational research.

FIGURE 1

Deregulated transcripts by treatment with chemoprentive agents in the p53-null mammary model. (A) Scatter-plot representation of differentially expressed genes between bexarotene, gefitinib and celecoxib treatments and control SAGE libraries (p< 0.05). (B) Gene ontology (GO) classification of differentially expressed transcripts as a result of each chemopreventive agent treatment. Relative representation of the deregulated transcripts with specific GO term annotations related to Biological processes or molecular function.

FIGURE 2

Co-occurring differentially expressed genes among bexarotene, gefitinib and celecoxib treatments in p53-null `normal' mammary epithelium. Eighty-seven genes were identified modulated by more than one treatment. (A) Heat map of the 87 deregulated transcripts. Color scale at the bottom depicts the approximate fold change in expression for each transcript and library relative to control mammary gland. Negative fold change (e.g.: transcripts with decreased expression in bexarotene treatment) is represented in green, and positive fold change (e.g.: transcripts with over-expression in bexarotene treatment) is represented in red. Aquamarine lines on left: co-ocurring transcripts modulated both by bexarotene and gefitinib treatments. (B) Venn diagram showing the overlap between transcripts modulated by bexarotene, gefitinib and celecoxib treatments. Statistical analysis showed a significant number of overlapping genes between treatments (p< 0.001). Hatched area with blue lines: number of genes commonly modulated by both bexarotene and gefitinib treatments. (C) Gene ontology classification of the 87 transcripts deregulated by the chemopreventive treatments.

FIGURE 3

Transcripts identified as deregulated in p53-null mammary tumors that were observed to be modulated in the opposite direction as the result of treatment with chemopreventive agents in normal mammary epithelium (i.e. up in tumors, down in the treated epithelium or viceversa). (A) Scatter-plot representation of differentially expressed genes between p53-null `normal' epithelium and p53-null tumors SAGE libraries (p< 0.05). (B) Heat maps of the transcripts modulated in the opposite direction in tumors vs treated normal epithelium: p53-null tumors (Black cluster) and bexarotene treated normal p53 null mice epithelium (aquamarine cluster), gefitinib treated (fuchsia cluster) and celecoxib treated (orange cluster). Color scale at the bottom depicts the approximate fold change in expression for each transcript and library relative to control mammary gland. Negative fold change is represented in green, and positive fold change is represented in red.