Fibroblast Growth Factor Receptor-1 (FGFR1) is commonly overexpressed in advanced prostate cancer (PCa). To investigate causality, we utilized an inducible FGFR1 (iFGFR1) prostate mouse model. Activation of iFGFR1 with chemical inducers of dimerization (CID) led to highly synchronous, step-wise progression to adenocarcinoma that is linked to an epithelial-to-mesenchymal transition (EMT). iFGFR1 inactivation by CID withdrawal led to full reversion of prostatic intraepithelial neoplasia, whereas PCa lesions became iFGFR1-independent. Gene expression profiling at distinct stages of tumor progression revealed an increase in EMT-associated Sox9 and changes in the Wnt signaling pathway, including Fzd4, which was validated in human PCa. The iFGFR1 model clearly implicates FGFR1 in PCa progression and demonstrates how CID-inducible models can help evaluate candidate molecules in tumor progression and maintenance.

The development of safe vectors for gene therapy requires fail-safe mechanisms to terminate therapy or remove genetically altered cells. The ideal ''suicide switch'' would be nonimmunogenic and nontoxic when uninduced and able to trigger cell death independent of tissue type or cell cycle stage. By using chemically induced dimerization, we have developed powerful death switches based on the cysteine proteases, caspase-1 ICE (interleukin-1␤ converting enzyme) and caspase-3 YAMA. In both cases, aggregation of the target protein is achieved by a nontoxic lipid-permeable dimeric FK506 analog that binds to the attached FK506-binding proteins, FKBPs. We find that intracellular cross-linking of caspase-1 or caspase-3 is sufficient to trigger rapid apoptosis in a Bcl-x L -independent manner, suggesting that these conditional proapoptotic molecules can bypass intracellular checkpoint genes, such as Bcl-x L , that limit apoptosis. Because these chimeric molecules are derived from autologous proteins, they should be nonimmunogenic and thus ideal for long-lived gene therapy vectors. These properties should also make chemically induced apoptosis useful for developmental studies, for treating hyperproliferative disorders, and for developing animal models to a wide variety of diseases.

A number of "suicide genes" have been developed as safety switches for gene therapy vectors or as potential inducible cytotoxic agents for hyperproliferative disorders, such as cancer or restenosis. However, most of these approaches have relied on foreign proteins, such as HSV thymidine kinase, that primarily target rapidly dividing cells. In contrast, novel artificial death switches based on chemical inducers of dimerization (CIDs) and endogenous proapoptotic molecules function efficiently in both dividing and nondividing cells. In this approach, lipid-permeable, nontoxic CIDs are used to conditionally cross-link target proteins that are fused to CID-binding domains (CBDs), thus activating signaling cascades leading to apoptosis. In previous reports, CID-regulated Fas and caspases 1, 3, 8, and 9 were described. Since the maximum efficacy of these artificial death switches requires low basal and high specific activity, we have optimized these death switches for three parameters: (1) extent of oligomerization, (2) spacing between CBDs and target proteins, and (3) intracellular localization. We describe improved conditional Fas and caspase 1, 3, 8, and 9 alleles that function at subnanomolar levels of the CID AP1903 to trigger apoptosis. Further, we demonstrate for the first time that oligomerization of the death effector domain of the Fas-associated protein, FADD, is sufficient to trigger apoptosis, suggesting that the primary function of FADD, like that of Apaf-1, is oligomerization of associated caspases. Finally, we demonstrate that nuclear-targeted caspases 1, 3, and 8 can trigger apoptosis efficiently, implying that the cleavage of nuclear targets is sufficient for apoptosis.

To develop an inducible and progressive model of mammary gland tumorigenesis, transgenic mice were generated with a mouse mammary tumor virus–long terminal repeat–driven, conditional, fibroblast growth factor (FGF)–independent FGF receptor (FGFR)1 (iFGFR1) that can be induced to dimerize with the drug AP20187. Treatment of transgenic mice with AP20187 resulted in iFGFR1 tyrosine phosphorylation, increased proliferation, activation of mitogen-activated protein kinase and Akt, and lateral budding. Lateral buds appeared as early as 3 d after AP20187 treatment and initially consisted of bilayered epithelial cells and displayed apical and basolateral polarity appeared after 13 d of AP20187 treatment. Invasive lesions characterized by multicell-layered lateral buds, decreased myoepithelium, increased vascular branching, and loss of cell polarity were observed after 2–4 wk of treatment. These data indicate that acute iFGFR1 signaling results in increased lateral budding of the mammary ductal epithelium, and that sustained activation induces alveolar hyperplasia and invasive lesions.

By an unknown mechanism, β-thymosins are extracellular modulators of angiogenesis, inflammation, wound healing, and development. We were interested in identifying β-thymosin interactors and determining their importance in β-thymosins signaling in human vein endothelial cells (HUVECs). We performed pulldown experiments with biotinylated thymosin β-4 (Tβ4) in comparison to neutravidin beads alone and used mass spectrometric analysis to identify differentially interacting proteins. By this method, we identified F1-F0 ATP synthase, a known target of antiangiogenic angiostatin. By surface plasmon resonance, we determined for Tβ4 binding to the β subunit of ATP synthase a K(D) of 12 nM. Blocking antibodies and antagonists (oligomycin, IC(50) ∼1.8 μM; piceatannol, IC(50) ∼1.05 μM; and angiostatin, IC(50) ∼2.9 μg/ml) of ATP synthase inhibited the Tβ4-induced increase in cell surface ATP levels, as measured by luciferase assay, and the Tβ4-induced increase in HUVEC migration, as measured by transwell migration assay. Silencing of the ATP-responsive purinergic receptor P2X4 with siRNA also blocked Tβ4-induced HUVEC migration in a transwell assay. Furthermore, in silico we identified common amphiphilic α-helical structural similarities between β-thymosins and the inhibitory factor 1 (IF1), an inhibitor of ATP synthase hydrolysis. In summary, we have identified an extracellular signaling pathway where Tβ4 increases cell surface ATP levels via ATP synthase and have shown further that ATP-responsive P2X4 receptor is required for Tβ4-induced HUVEC migration.