**4. Discussion**

AR-V7 is a major splice variant expressed in human prostate cancer that is associated with the development and progression of CRPC [39,40]. AR-V7 potentially contributes to the resistance to enzalutamide and abiraterone in CRPC [11,41,42]. The involvement of AR-V7 in taxane resistance is not well understood. The LNCaP95 cell line was derived from the LNCaP cell line and it has acquired resistance to androgen depletion conditions. LNCaP95 cells express full-length AR and AR-V7, but the level of AR-V567es is negligible [43]. Proliferation of LNCaP95 cells is driven by AR-V7, despite the endogenous expression of functional FL-AR [39,44]. These cells are resistant to enzalutamide [45]. To elucidate whether AR-V7 plays a role in the acquired resistance to taxanes, the LNCaP95-DR cell line was developed and used as a model for CRPC.

Resistance to taxanes can be associated with the overexpression of P-gp [16,17], which has been confirmed in CRPC patients [18]. Levels of P-gp expression are higher in docetaxel-resistant TaxR and DU145-DTXR cells as compared to that in docetaxel-sensitive, parental C4-2B and DU145 cells, respectively [46,47]. Similarly, P-gp is overexpressed in docetaxel-resistant DU145R and CWR22rv1R cells derived from the parental DU145 and CWR22rv1 cells, respectively [48]. Consistent with these reports, we showed that P-gp was overexpressed in LNCaP95-DR cells, and that tariquidar treatment restored sensitivity to docetaxel. Therefore, overexpression of P-gp played a major role in the acquired resistance of LNCaP95 to docetaxel. Tariquidar is an anthranilic acid-derived third-generation P-gp inhibitor. Its efficacy has been evaluated in several clinical trials on different types of cancer including lung cancer, but there are no reports on its use for prostate cancer. Our study was the first report using tariquidar for prostate cancer cell lines.

Cabazitaxel is a next-generation semisynthetic taxane chemotherapeutic agen<sup>t</sup> that is effective in patients with docetaxel-resistant CRPC [49]. In the TROPIC clinical trial, cabazitaxel significantly improved the overall survival in CRPC patients during or after docetaxel treatment, but the survival benefit was limited to 2.4 months [20]. In our in vitro study, LNCaP95-DR cells were resistant to high doses of docetaxel with an IC50 of >400 nM; however, these cells maintained some sensitivity to cabazitaxel with an IC50 of approximately 70 nM (Figure 1C,D). The data was consistent with that observed clinically with the resistance to docetaxel. Importantly, tariquidar restored sensitivity to docetaxel, as well as to cabazitaxel, thereby indicating that the cross-resistance between docetaxel and cabazitaxel was mediated by the overexpression of P-gp.

Expression of AR-V7 is clinically important and has been proposed for the assessment of which patients should receive inhibitors of the androgen receptor or taxanes [24]. In this study, we showed that the expression of AR-V7-regulated genes was increased in LNCaP95-DR, but that knockdown of AR-V7 did not restore sensitivity to docetaxel and cabazitaxel. These data support the idea that AR-V7 was not involved in taxane resistance. Contrary to these data, Tadani-Mulero M et al. reported that the expression of AR-V7 resulted in taxane resistance in a mouse model of CRPC due to the absence of the AR hinge region, which appears to be critical for microtubule binding [25]. This report compared FL-AR with AR-V567es-expressing LuCaP86.2 tumor xenografts and FL-AR with AR-V7-expressing LuCaP23.1 tumor xenografts. That report concluded that AR-V7, but not AR-V567es, was important for resistance to docetaxel which was not supported by the work presented here in the LNCaP95-DR cells. Consistent with data presented in this study, are those obtained from clinical studies that have shown that the detection of AR-V7 in circulating tumor cells from men with metastatic CRPC was not associated with primary resistance to taxane chemotherapy [23,24,50].

Some recent reports sugges<sup>t</sup> that taxanes can inhibit AR signaling in prostate cancer cells [51–53]. However, in our study, neither docetaxel or cabazitaxel decreased the expression of genes regulated by FL-AR nor were there any effects on the transcriptional activity of the AR. Zhu ML and Darshan MS examined the effect of taxanes on the androgen/AR axis using very high concentrations of paclitaxel (1 μM and 100 nM, respectively), which were over the clinically effective range of paclitaxel [54,55]. In the present study, we tested docetaxel and cabazitaxel at the concentrations close to their IC50 (5 nM and 10 nM, respectively), which was clinically feasible [56,57]. Our data suggested that taxane chemotherapy did not affect the androgen/AR axis in LNCaP95-C and LNCaP95-DR when used at the clinically feasible concentration.

AR-NTD is essential for the transcriptional activities of both FL-AR and AR-Vs. Therefore, AR-NTD-targeting therapy has benefits over the drugs targeting the AR-LBD. EPI-002 targets the NTD of the AR and can block the signaling induced by the FL-AR and AR-Vs [58]. In this study, we showed that the inhibitory effect by EPI-002 on the proliferation of LNCaP95-DR cells was similar to that achieved with the parental LNCaP95 cells. LNCaP95-DR proliferation remained driven by AR-V7, suggesting that AR-NTD could be a therapeutic target for cancers such as LNCaP95-DR, with acquired resistance to taxanes and enzalutamide.
