USING MOUSE MODELS TO STUDY THE ROLE OF PIK3CA MUTATIONS IN MAMMARY TUMORIGENESIS

PIK3CA is the most commonly mutated oncogene in human breast cancer. Somatic missense mutations in the PIK3CA gene have been reported in 15% of human cancers and 29% of total human breast cancers, on average. In breast cancer, we have observed PIK3CA mutations in the setting of different markers of tumor development. The majority of the mutations in PIK3CA cluster in two regions, within the helical and kinase domains (HD and KD). The codons most frequently affected are E545K and H1047R ('hot-spots') in exons 9 and 20, respectively. Both mutations cause a gain of enzymatic function, upregulate the AKT pathway, and induce oncogenic transformation when expressed in primary chicken-embryo fibroblasts, mouse fibroblasts, and NIH3T3 cells. Despite these similarities, structural and biochemical studies indicate that they operate by two different and independent gain-of-function mechanisms, suggesting that these mutations represent different mechanisms of lipid kinase activation, and hence transforming activity in cancer cells. Moreover, unsupervised clustering of our human breast cancer microarray expression data indicated that tumors with HD and KD mutations were readily distinguishable. Thus, we have generated two mouse models carrying a conditionally-expressed gain-of-function mutation either in exon 9 (E545K) or in exon 20 (H1047R) of the mouse PIK3CA gene. Cre-mediated DNA excision of a 'stop' sequence has been shown to allow proper expression of each mutant allele. The advantage of our approach is that the mutation of PIK3CA is expressed via the endogenous promoter and thus mimics the expression found in human tumors. As current knowledge about the hotspot mutants' functions derive exclusively from in vitro overexpression experiments, we propose to use our mouse models to study the tumorigenic role of PIK3CA in mammary gland in vivo. It is our hypothesis that the two hotspot mutations will cause distinct forms of breast cancer, and that some cooperating genetic alterations will hasten the development of both HD and KD hotspot tumors while others will cooperate preferentially with one but not the other hotspot. Our specific aims are: (1) Evaluate genetically, through breeding programs, the malignant potential of these two hotspot mutations and their collaboration with selected signaling components such as HER2/Neu and PTEN that are known to be altered in breast cancer, and (2) comparatively analyze the mammary tumors using histopathology, immunohistochemistry, and microarray data, which may reveal signaling-specific effects leading to distinct tumor types. To achieve these goals, we will first cross both our models with strains expressing CRE specifically in mammary epithelial cells to determine if PIK3CA mutations cause spontaneous mammary carcinoma, and then through selective breeding programs with transgenic mice at our disposal, we will analyze genetically the ability of each hotspot mutation to collaborate with oncogenes and tumor suppressor genes known to be altered in human breast cancer. Overall, all mouse tumors will be monitored and analyzed using time of onset, histology, immunohistochemistry, and expression profiling. For verification of important microarray results, we will use quantitative RT-PCR and Western analyses. This study will conclusively show whether PIK3CA hotspot mutations are causally involved in the development of breast cancer, and our models will serve as valuable tools for determining the phenotypic and mechanistic consequences of each hotspot mutation either alone or in combination with other key genes involved in the biology of breast cancer. Hopefully, the information derived from the proposed research, including the comparison of mouse and human expression profiling results, will provide valuable insights into the biological consequence of PIK3CA mutations and a future platform for improving patient therapy to reduce breast cancer mortality.