Abstract 703: Verification of a novel approach to identify disease mechanisms in breast cancer patient subsets and application of this approach to indicating and characterizing tamoxifen resistance

Abstract

Abstract Breast cancer is a heterogeneous disease that can be divided into distinct clinical and molecular subtypes. Key biomarkers such as the expression of HER2 or ER serve as a basis for subdividing the disease into distinct clinical entities and to distinguish patients who will most likely respond to specific targeted therapies, such as Herceptin and tamoxifen, respectively. Further understanding the molecular mechanistic differences among and within the breast cancer subtypes can facilitate the identification of treatment regimens that can be even more tailored to an individual's disease. If such a strategy is implemented, efficacy rates for cancer treatments can be improved, and more importantly, patients will be spared from any collateral effects of those treatments that will not combat their specific disease. We have taken a systems biology approach to segregate breast cancer patients into subgroups by semi-quantitatively evaluating the literature-derived molecular gene expression footprints of over 2100 mechanisms in our Knowledgebase. This approach effectively enables the stratification of breast cancer patients by their individual levels of signaling amplitude for these mechanisms. Using publicly available breast cancer data sets from GEO, we validated the stratification capability of these inferred mechanism activities against immunohistochemical staining for ER and HER2, and sequencing-based identification of TP53 mutation. In addition, applying this semi-quantitative evaluation of mechanism molecular footprints based on gene expression data from ER+ breast cancer patients before tamoxifen treatment (GSE17705), we developed a tamoxifen-sensitivity classifier a priori of knowledge of clinical outcome. This classifier enabled prediction of 16 of the 19 patients from an independent test set (GSE6532) who had gone on to develop distant metastases. Beyond identifying specific breast cancer sub-groups, which has been well established by clustering sets of genes, the systems biology-based approach presented here enables us to identify the specific mechanisms defining the separation of patient groups, and offers a means of either selecting the right patient population for a drug candidate or indicating additional/alternative disease pathways that could be targeted for clinical efficacy. This strategy is universal and can be applied to a variety of other cancers and diseases. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 703. doi:1538-7445.AM2012-703