NAi-induced AR silencing in these cells produced only a limited or a non statistically significant effect. In contrast, Zegarra-Moro et al. showed that the microinjection of AR antibodies almost completely arrested C4-2 cells’ proliferation, whereas Liao et al. reported that AR silencing induced their massive apoptosis. In our hands, AR silencing produced a time-dependent inhibition of cells’ proliferation in C4-2 as in 22RV1 cells, but Gynostemma Extract web triggered apoptosis in C4-2 cells only, at least during the time-course of this experiment. Using a lentiviral AR-shRNA expressing vector, Chen et al stably silenced AR in LAPC4 and in castration-resistant LNCaP cells, which then failed to 7938165 develop tumors once grafted into SCID mice. However, as the authors did not report if the cells’ proliferation was affected or not before the grafting, the in vivo data are difficult to interpret. From all the above studies, it was therefore not possible to foretell whether AR was a key target in established resistant tumors. We used to address this question one ADCaP and two unrelated CRCaP models and checked during at least 5 days that each tumor was exponentially growing before randomization for siRNA treatment. AR-siRNA induced in all three tumor models a strong inhibitory effect. The tumors that were smaller at the start of the treatment responded better than the larger ones. It is well established that AR regulates the VEGF production in the normal 
prostate as in androgen-dependent tumors. The VEGF level is very tightly controlled, at the transcriptional and post transcriptional levels, and AR is likely involved in several of these mechanisms. Tumors that became resistant to hormonal manipulations are very angiogenic and express high VEGF levels. Importantly, we demonstrate here for the first time that, in vitro as in vivo, AR still controls the VEGF production in CRCaP. AR silencing in vivo reduces VEGF synthesis, triggering a reduction in the blood vessels density, although other modulators of angiogenesis regulated by androgens could also participate in this process. Therefore, part of the antitumoral effects produced by AR-siRNA likely result from the inhibition of angiogenesis in the treated tumors. Our results demonstrate that androgen dependent, but also and more importantly, exponentially growing and vascularized castration-resistant prostate tumors, are still dependent on the expression of a functional AR for their growth and angiogenesis. Importantly, siRNA-directed silencing of AR allows inhibiting the synthesis of overexpressed wild type receptors, as well as mutated 26023119 or stabilized isoforms that can be encountered in advanced prostate carcinomas. The strong antitumoral effects triggered by siRNA-directed silencing of AR, which results from a combination of antiproliferative and antiangiogenic effects, opens exciting therapeutic perspectives for treating advanced prostate tumors that became resistant to hormonal manipulations, for which treatments available today are only palliative. agonist, or with the ARantagonist bicalutamide or with vehicle. Cell proliferation and viability were quantified either by counting cells excluding the trypan blue or using the MTT assay. Caspases activities were measured using a fluorimetric quantification assay. siRNA Synthetic siRNA were purchased from Dharmacon. Sequences are indicated in table S1. Animals, siRNA injection and tumorigenicity assays Studies involving animals, including housing and care, method of euthana