Janni Mirosevich

Src family kinases (SFK) are currently being investigated as targets for treatment strategies in various cancers. The novel SFK/Abl inhibitor, dasatinib (BMS-354825), is a promising therapeutic agent with oral bioavailability. Dasatinib has been shown to inhibit growth of Bcr-Abl-dependent chronic myeloid leukemia xenografts in nude mice. Dasatinib also has been shown to have activity against cultured human prostate and breast cancer cells. However, the molecular mechanism by which dasatinib acts on epithelial tumor cells remains unknown. In this study, we show that dasatinib blocks the kinase activities of the SFKs, Lyn, and Src, in human prostate cancer cells at low nanomolar concentrations. Moreover, focal adhesion kinase and Crk-associated substrate (p130(CAS)) signaling downstream of SFKs are also inhibited at similar concentrations of dasatinib. Consistent with inhibition of these signaling pathways, dasatinib suppresses cell adhesion, migration, and invasion of prostate cancer cells at low nanomolar concentrations. Therefore, dasatinib has potential as a therapeutic agent for metastatic prostate cancers harboring activated SFK and focal adhesion kinase signaling.

Androgens and mesenchymal factors are essential extracellular signals for the development as well as the functional activity of the prostate epithelium. Little is known of the intraepithelial determinants that are involved in prostatic differentiation. Here we found that hepatocyte nuclear factor-3 alpha (HNF-3 alpha), an endoderm developmental factor, is essential for androgen receptor (AR)-mediated prostatic gene activation. Two HNF-3 cis-regulatory elements were identified in the rat probasin (PB) gene promoter, each immediately adjacent to an androgen response element. Remarkably, similar organization of HNF-3 and AR binding sites was observed in the prostate-specific antigen (PSA) gene core enhancer, suggesting a common functional mechanism. Mutations that disrupt these HNF-3 motifs significantly abolished the maximal androgen induction of PB and PSA activities. Overexpressing a mutant HNF-3 alpha deleted in the C-terminal region inhibited the androgen-induced promoter activity in LNCaP cells where endogenous HNF-3 alpha is expressed. Chromatin immunoprecipitation revealed in vivo that the occupancy of HNF-3 alpha on PSA enhancer can occur in an androgen-depleted condition, and before the recruitment of ligand-bound AR. A physical interaction of HNF-3 alpha and AR was detected through immunoprecipitation and confirmed by glutathione-S-transferase pull-down. This interaction is directly mediated through the DNA-binding domain/hinge region of AR and the forkhead domain of HNF-3 alpha. In addition, strong HNF-3 alpha expression, but not HNF-3 beta or HNF-3 gamma, is detected in both human and mouse prostatic epithelial cells where markers (PSA and PB) of differentiation are expressed. Taken together, these data support a model in which regulatory cues from the cell lineage and the extracellular environment coordinately establish the prostatic differentiated response.

We have previously shown that a forkhead transcription factor Foxa1 interacts with androgen signaling and controls prostate differentiated response. Here, we show the mouse Foxa1 expression marks the entire embryonic urogenital sinus epithelium (UGE), contrasting with Shh and Foxa2, which are restricted to the basally located cells during prostate budding. The Foxa1-deficient mouse prostate shows a severely altered ductal pattern that resembles primitive epithelial cords surrounded by thick stromal layers. Characterization of these mutant cells indicates a population of basal-like cells similar to those found in the embryonic UGE, whereas no differentiated or mature luminal epithelial cells are found in Foxa1-deficient epithelium. These phenotypic changes are accompanied with molecular aberrations, including focal epithelial activation of Shh and elevated Foxa2 and Notch1 in the null epithelium. Perturbed epithelial-stromal interactions induced by Foxa1-deficient epithelium is evident, as demonstrated by the expansion of surrounding smooth muscle and elevated levels of stromal factors (Bmp4, Fgf7, Fgf10 and Gli). The prostatic homeobox protein Nkx3.1, a known proliferation inhibitor, was downregulated in Foxa1-deficient epithelial cells, while several prostate-specific androgen-regulated markers, including a novel Foxa1 target, are absent in the null prostate. These data indicate that Foxa1 plays a pivotal role in controlling prostate morphogenesis and cell differentiation.

The molecular mechanism(s) for prostate cancer progression to androgen independence are poorly understood. We have recently shown that Foxa1 and Foxa2 proteins are differentially expressed in epithelial cells during murine prostate development, growth, and adult function. Currently, the role of Foxa proteins in prostate cancer development and progression is unknown. Foxa protein expression was investigated in the LPB-Tag LADY mouse prostate cancer models, in human prostate cancer specimens, and various prostate cancer cell lines using Western blot and immunostaining analysis. In vitro transient transfection, studies were performed to investigate Foxa/prostate-specific gene regulation. Foxa1 was strongly expressed in areas of prostatic intraepithelial neoplasia (PIN) in both the androgen dependent 12T-7f and in the metastatic, androgen independent 12T-10 LADY models. Prominent Foxa1 and Foxa2 expression was observed in 12T-10 invasive undifferentiated neuroendocrine carcinomas, in the hormone independent and metastasizing 12T-10 derived, NE-10 allograft tumors, and in all metastatic lesions isolated from 12T-10 mice. Foxa1 protein expression was always observed in human prostate carcinomas, regardless of Gleason grade score, while Foxa2 was only detected in neuroendocrine small cell carcinomas and in some high Gleason score adenocarcinomas. Foxa proteins were also differentially expressed in three prostate cancer cell lines. Importantly, in vitro functional assays demonstrated that Foxa2 could activate androgen-dependent prostate-specific genes in an androgen receptor and ligand-independent manner. These results suggest that Foxa proteins are important in prostate carcinogenesis. In particular, Foxa2 may be involved in progression of prostate cancer to androgen independence. As such, Foxa proteins may represent novel targets for therapeutic intervention.