The major scientific focus of this project is to determine the Ras-initiated signaling pathways and their relevant transcriptional targets that contribute to human epithelial cell transformation and metastasis. Ras is mutated in approximately one quarter of all human cancers with the highest incidence in pancreatic, lung, colon, and thyroid tumors. In addition, there is considerable experimental evidence that persistent upstream signaling in other epithelial cancers may activate Ras. Transformation functions associated with Ras effector pathways are being analyzed in experimental models of human prostate cancer. Because multiple autocrine or paracrine growth factor pathways contribute to the transformation of epithelial cells and their colonization of distant tissues during the development of metastasis, Ras signaling pathways are expected to provide broadly applicable diagnostic markers and therapeutic targets. Multiple downstream effectors mediate Ras signaling, and there is a growing appreciation that the signaling outcomes of Ras activation demonstrate species and cell context differences. We have shown that ectopic Ras activation leads to the new expression of a bone metastatic phenotype in the DU145 xenograft model of human prostate cancer, and furthermore, that bone tropism can be mediated by a specific Ras effector pathway. We have investigated the role of signaling pathways initiated by oncogenic Ras including Raf/ERK, PI3-kinase, and Ral guanine nucleotide exchange factors (RalGEFs), in stimulating tissue-specific experimental metastasis of DU145. Although the RalGEF pathway appears to be required for Ras-initiated tumorigenesis in human epithelial cells, a role in metastasis has not been addressed. Oncogenic Ras promoted metastasis of hematogenously delivered DU145 to multiple organs including bone, brain, and other soft tissues. Activation of the Raf/ERK pathway stimulated metastatic colonization of the brain, while activation of the RalGEF pathway led to bone metastases, the most common organ site for prostate cancer. The generality of RalGEF pathway activation in prostate cancer bone metastasis was confirmed in another xenograft model. In addition, shRNA-mediated knockdown of RalA in the metastatic PC3 cell line inhibited bone metastasis but did not affect subcutaneous tumor growth. Interestingly, knockdown of RalA in breast or colon cancer cells did not inhibit bone metastatic activity, suggesting some specificity for RalGEF function in the context of prostate cells. The existence of 2 model prostate cancer metastasis systems provides a strong platform to identify both general functional categories and specific gene products associated with RalGEF-mediated bone metastasis. Comparative microarray analyses have provided potential functional targets. An important question is the mechanism of RalGEF-dependent bone metastasis. Genetically manipulated loss and gain of Ral function in either the DU145 or PC3 systems suggested that Ral affects expansive growth in the bone, whereas homing and initial colonization are less affected. The role of Ras-mediated signaling with respect to osteoclast induction is being investigated. Enhanced anchorage-independent growth was a cell biological correlate of bone, but not brain, metastasis. We have taken advantage of this property to develop an in vitro DU145 cell line that is being used in a tetracycline-inducible shRNAi mediated library selection for Ral-dependent gene functions. Our data begin to identify signaling pathways relevant for organ-specific metastasis. There are very few models of metastatic prostate cancer, and our DU145 model is unique in that the initiating genetic event, i.e. Ras activation, can be manipulated. The DU145(Ras) system provides an experimental system for the identification of genes mediating metastasis to different organs. It also provides an efficient and predictable model system of tissue-specific metastases for preclinical testing of therapeutic agents. One of our goals is to assess whether Ral pathway activation occurs in clinical samples of PC bone metastases. It is not possible to directly measure Ral-GTP levels in such samples because they are available only post-mortem and generally are quite heterogeneous and include necrotic regions. Therefore, an indirect approach is to determine whether genes and their encoded products that are regulated by Ral are expressed in clinical samples of bone metastases. We have obtained a limited number of PC bone metastasis tissue sections and extracted RNA/cDNA preparations from Bob Vessella at the University of Washington. We have used two approaches, gene expression microarrays and antibody-based protein arrays, to investigate global effects of Ral pathway manipulations in PC cells. The rationale for such approaches relies on the correlation of changes in RNA and/or protein levels with phenotypic changes in the DU145 and PC3 cell lines. In order to avoid effects of clonal variation, we have produced and characterized 3 independently-derived cell lines to serve as biological replicates for the various conditions of Ral activation or inhibition assayed. In addition, 3 individual DU145(RasV12G37) bone metastasis clones were isolated, expanded, and subsequently shown to have enhanced bone metastatic activity. The microarray analyses have been most successful in identifying differential expression profiles within conditions (i.e. Ral activation or Ral depletion). For Ral activation within the DU145 system, following the application of standard statistical criteria, genes were filtered for those transcripts that are differentially expressed in DU145(RasV12G37) in addition to the 3 bone clones relative to DU145(EV). Ten genes were upregulated, and 89 genes were downregulated. These include previously described genes that have been implicated in PC bone metastasis as a result of differential expression in clinical samples and or independent experimental models. A panel of cell lines will be used to further investigate, using independent methods to assay RNA levels and encoded proteins, those candidate genes of interest identified by transcription profiling. Verified Ral-regulated genes will be further assayed in clinical samples of PC metastases. Transcription profiles have been analyzed for pathways and related gene sets using Gene Set Enrichment Analyses (GSEA) and Ingenuity Pathway Analysis software programs. No clear common pathway or related sets of differentially expressed genes were readily identified. In addition, very few consistent changes in gene expression are observed when comparing multiple replicates of PC3 populations depleted for RalA or in separate comparisons, for RalB. These data suggest that in the PC3 system, depletion of endogenous RalA or B has a modest effect upon transcription, implying that the phenotypic changes that occur following RalA depletion result from post-transcriptional regulation. Antibody protein array analyses are ongoing. For DU145(RasV12G37)B1 and consistent with its angiogenic phenotype, the concentrations of several secreted angiogenic factors (VEGFA, CXCL-5, IL-8, and uPA) are increased relative to the DU145(EV). In addition, osteoclast/osteoblast differentiating factors (BMP4, TGFβ, TNFα/β, IL-11) are increased. The generality of these observations with respect to other independently isolated clones and regulation by RalA or RalB levels will be determined. Of interest and consistent with the diminished osteoclastic response [summary truncated at 7800 characters]
