DESCRIPTION (provided by applicant): Bladder cancer kills 15,000 annually and because of few therapeutic advances there is a need for innovation. This project has discovered suppressors of bladder tumor growth and metastasis and shown that low expression of AGL, an enzyme involved in glycogenolysis, and RhoGDI2, an inhibitor of multiple GTPases, is associated with cancer recurrence and death. The Goal of this continuation project is to understand how reduced AGL levels result in aggressive bladder cancer and to translate this knowledge into novel therapies. We found bladder cancer cells with low AGL express elevated levels of Hyaluronic Acid (HA) synthase 2 (HAS2), and its extracellular product HA, and have activated RhoC and Rac1, known drivers of tumor progression. We demonstrated RhoC/Rac1 activity is suppressed by RhoGDI2, linking the actions of the two metastasis suppressors. Cells with low AGL also exhibit increased glycolysis, glucose import and autophagy. When either glucose import or autophagy are blocked, low AGL tumor cells show greater reduction of in vitro growth than cells with higher AGL levels, exposing a vulnerability of these aggressive cells. These data support an innovative Hypothesis: Tumor cells with low AGL generate growth and metastatic signals via production of HA, activation of its receptors CD44 and RHAMM, and downstream effectors such as RhoC and Rac1. For these signals to be actualized, sufficient energy needs to be provided via enhanced glucose import and autophagy. Three Specific Aims will test this hypothesis. 1) Evaluate impact of HA signaling on aggressiveness of bladder cancers with low AGL. We will determine if HA drives tumor progression via its receptors CD44 and RHAMM and effectors RhoC and Rac1 that are inhibited by RhoGDI2. Experiments will use RNAi and receptor inhibition in human xenografts. 2) Determine role of the AGL signaling network in bladder cancer development and progression. We will evaluate the independent predictive value of tumor suppressor genes and components of the AGL signaling network in a large panel of annotated human tumors, to define a predictive multigene biomarker signature and possible new therapeutic targets. We will use full and conditional (bladder urothelium) AGL knockout mice in a chemical carcinogenesis model that generates spontaneous visceral metastases to test whether lack of AGL makes urothelium more susceptible to carcinogenesis and resulting tumors more aggressive. 3) Target molecular vulnerabilities of bladder cancers with low AGL expression. Here we test whether blocking glucose import and/or autophagy reduces in vivo bladder tumor growth. We will use shRNA and pharmacologic agents already in preclinical/clinical testing. We will screen shRNA libraries to metabolic and autophagy genes in cells with low AGL to discover novel genes essential to the aggressive tumor phenotype, with the expectation that some of these will become therapeutic targets. We will evaluate...