Breast cancer metastases to bone arise as a result of intrinsic features of cancer cells making the cancer cells compatible to grow in the bone microenvironment as well as adaptive responses of bilateral cancer cell - host cell signaling, launching a vicious cycle that promotes cancer growth and lytic bone destruction. We report how comprehensive molecular profiling and high-throughput functional phenotypic screening resulted in new insights into the molecular mechanisms of breast cancer metastasis to bone, and may help to delineate druggable targets and pathways. Using MDA-MB-231 breast cancer cell line as a model, we first compared the molecular profiles of highly and weakly metastatic variants of MDA-MB-231 cells by single-gene resolution array-CGH (comparative genomic hybridization) and global gene expression profiling. Second, we carried out a high-throughput RNAi screening of the human kinome (2,500 constructs) to identify kinases whose knockdown selectively impacts the survival of the highly metastatic breast cancer cells. Data from these DNA and RNA profiling projects as well as functional RNAi screening were integrated. Genetic alterations between the highly and weakly metastatic cell line variants were relatively small but up to 3.1% of all genes were upregulated and 2.8% downregulated in the highly metastatic variant. These differentially expressed genes include some previously characterized bone metastasis genes (e.g. MMP1 and NAP1L3) as well as many novel genes whose gene ontology and pathway associations include Wnt signaling, apoptosis, cathepsin, cell cycle, and other pathways, suggesting a multitude of selection mechanisms involved in metastasis. We then compared the RNA interference profiles, i.e. the sensitivity differences of the two cell line variants to 2,000 kinase RNAi inhibitors. We identified 74 kinase siRNAs that were selectively effective against the highly metastatic cell variant. These hits include many kinases involved in cancer-relevant pathways such as cell cycle regulation, mitosis, and the Wnt signaling pathway. Our integrated profiling suggests novel mechanisms and pathways involved in breast cancer metastasis. These discovery leads from high-throughput studies will be explored further to validate the targets and compounds that could be used for novel therapeutic insights. For this purpose, we have developed an in vitro bone metastasis model, based on a co-culture of bone metastatic breast cancer cells and osteoclasts. This predictive model for in vivo mouse studies could potentially be used for exploring and intervening in the mechanisms of the ’vicious cycle’ of osteolytic bone metastasis.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]