Gene therapy is founded on the concept that tissue-specific promoters can express heterologous genes for molecular imaging or therapeutic applications. The engineering of cell-specific enhancers to improve potency and the development of two-step transcriptional activation (TSTA) approaches have significantly improved the efficacy of transgene expression. Here we combine these technologies to create a robust, titratable, androgen-responsive system targeted to prostate cancer cells. Our "chimeric" TSTA system uses a duplicated variant of the prostate-specific antigen (PSA) gene enhancer to express GAL4 derivatives fused to one, two, or four VP16 activation domains. We targeted the resulting activators to cells with reporter templates bearing one, two, or five GAL4 binding sites upstream of firefly luciferase. We monitored activity via firefly luciferase assays in transfected cell extracts and in live nude mice using a cooled charge-coupled device (CCD) imaging system. In this system, we found that firefly luciferase expression in prostate cancer cells can be varied over an 800-fold range. We also found that a single plasmid bearing the optimized enhancer, GAL4-VP16 derivative, and reporter expressed firefly luciferase at 20-fold higher levels than the cytomegalovirus enhancer. We discuss the implications of this strategy and its application to molecular imaging and therapy.

The current understanding of the response of androgen receptor to pharmacologic inhibitors in prostate cancer is derived primarily from serum prostate-specific antigen (PSA) levels. In this study, we test whether a novel androgen receptor -specific molecular imaging system is able to detect the action of the antiandrogen flutamide on androgen receptor function in xenograft models of prostate cancer. Adenoviruses bearing an optical imaging cassette containing an androgen receptor -responsive two-step transcriptional amplification system were injected into androgen-dependent and hormone-refractory tumors of animals undergoing systemic time-controlled release of the antiandrogen flutamide. Imaging of tumors with a cooled charge-coupled device camera revealed that the response of AdTSTA to flutamide is more sensitive and robust than serum PSA measurements. Flutamide inhibits the androgen signaling pathway in androgen-dependent but not refractory tumors. Analysis of androgen receptor and RNA polymerase II binding to the endogenous PSA gene by chromatin immunoprecipitation revealed that flutamide treatment and androgen withdrawal have different molecular mechanisms. The application of imaging technology to study animal models of cancer provides mechanistic insight into antiandrogen targeting of androgen receptor during disease progression. [Mol Cancer Ther 2005;4(11):1662-9]

Mitogen-activated protein kinases (MAPK) play important roles in malignancy. The ability to detect and quantitate MAPKs in live animal models of cancer will facilitate an understanding of disease progression. We have developed a gene expression-based imaging system that detects and quantifies MAPK activity in prostate cancer tumors implanted into severe combined immunodeficient mice. The imaging technology uses a modified version of two-step transcriptional amplification (TSTA). The tissue specificity of gene expression is imparted by an enhanced version of the prostate-specific antigen regulatory region that expresses GAL4-ELK1. GAL4-ELK1 confers MAPK specificity by activating a firefly luciferase (FLuc) reporter gene when the Ets-like transcription factor (ELK) 1 activation domain is phosphorylated by MAPK. FLuc activity in live animals was detected using the Xenogen In vivo Imaging System. We validated the TSTA-ELK1 system by analyzing its response to epidermal growth factor treatment in transfected tissue culture cells and in adenovirus (AdTSTA-ELK1)-injected prostate cancer xenograft tumors. We measured MAPK activity in two well-characterized xenograft models, CWR22 and LAPC9. Although no significant differences in MAPK levels were detected between androgendependent and androgen-independent xenografts, the CWR22 models display significantly higher levels of AdTSTA-ELK1 activity versus LAPC9. Western blots of tumor extracts showed that the elevated imaging signal in CWR22 xenografts correlated with elevated levels of phosphorylated extracellular signal-regulated kinase 1/2 but not p38 or c-Jun NH 2 -terminal kinase. We conclude that a gene expression-based optical imaging system can accurately detect and quantify MAPK activity in live animals. (Cancer Res 2006; 66(22): 10778-85)