TAS 3681, an androgen receptor antagonist, prevents drug resistance driven by aberrant androgen receptor signaling in prostate cancer

VCaP (RRID:CVCL_2235), COS‐7 (RRID:CVCL_0224), and HCC1806 (RRID:CVCL_1258) cells were purchased from the American Type Culture Collection (Rockville, MD, USA); HEK293 (RRID:CVCL_0045) cells were purchased from Health Science Research Resources Bank (Osaka, Japan); MIAPaCa‐2 (RRID:CVCL_0428), SK‐OV‐3 (RRID:CVCL_0532), MCF‐7 (RRID:CVCL_0031), and T‐47D (RRID:CVCL_0553) cells were purchased from Dainippon Sumitomo Pharma Co., Ltd. (Osaka, Japan). 22Rv1 (RRID:CVCL_1045) cells were obtained from European Collection of Cell Cultures (Salisbury, UK). LNCaP (RRID:CVCL_0395) cells were obtained from the Institute of Microbial Chemistry and the Microbial Chemistry Research Foundation (Tokyo, Japan). Enzalutamide‐resistant cells, SAS MDV No. 3‐14, were established as previously reported [8]; G418‐selected U‐2 OS tomato‐AR cells stably expressing red fluorescent protein (tdTomato) fused AR, and hygromycin B‐selected LNCaP cells stably expressing F877L mutant AR protein were established in‐house. All human cell lines were reauthenticated by means of short tandem repeat‐based DNA profiling and all experiments were performed with mycoplasma‐free cells. All cell lines were authenticated by STR profiling, performed at Biologica, Co. (Nagoya, Japan) after the study was completed. VCaP and COS‐7 cells were cultured in RPMI1640 containing 10% FBS. HEK293 were cultured in MEM containing 10% FBS. MCF‐7 cells were cultured in MEM containing 10% FBS, 0.1 mm NEAA, and 1 mm Na‐pyruvate. T‐47D cells were cultured in RPMI1640 containing 10% FBS and 0.2 I.U. insulin. MIAPaCa‐2 were cultured in DMEM including 10% cFBS and 2.5% horse serum. HCC1806 cells were cultured in RPMI1640, including 10 mmN‐2‐hydroxyethylpiperazine‐n′‐(2‐ethanesulfonic acid) (HEPES), 1 mm sodium pyruvate, and 10% FBS; SK‐OV‐3 cells were cultured in McCoy's 5A, including 15% FBS. The 22Rv1 cells were cultured in phenol red‐free RPMI 1640 (RPMI‐PR‐) containing 10% FBS. LNCaP cells and LNCaP cells stably expressing F877L mutant AR were cultured in RPMI 1640 medium containing 5% FBS. U‐2 OS tomato‐AR cells were cultured in McCoy's 5A medium containing 10% FBS. SAS MDV No. 3‐14 cells were maintained in RPMI‐PR‐ containing 5% FBS and 10 μm enzalutamide.

The vector plasmid expressing androgen receptor (AR; wild‐type, L702H, V716M, W742C, W742L, H875Q, H875Y, F877L, T878A, D891Y, Q903H, H875Y/T878A, F877L/T878A, T878A/S889G, and T878A/D891H) were obtained from Genewiz, Inc. (Beijing, China). The pGL4.36 [luc2P/MMTV/Hygro] and pGL4.75 [hRluc/CMV] vector plasmids were obtained from Promega Corporation (Madison, WI, USA). The pGLPE plasmid vector was generated by cloning the prostate‐specific antigen (PSA) promoter and enhancer fragments into the pGL3 plasmid vector [9]. TAS3681, 2‐chloro‐4‐[4‐[[5‐[2‐hydroxypropan‐2‐yl]pyridin‐2‐yl]amino]‐5,8‐dihydropyrido[3,4‐d]pyrimidin‐7(6H)‐yl]benzonitrile, and enzalutamide were synthesized by Taiho Pharmaceutical Co., Ltd. (Tsukuba, Japan) and obtained from AstaTech, Inc. (Bristol, PA, USA). Bicalutamide was obtained from Fidia Farmaceutici S.p.A. (Abano Terme, Italy). Apalutamide was obtained from Haoyuan Chemexpress Co., Ltd. (Shanghai, China). Darolutamide was obtained from MedChemExpress Co., Ltd. (Monmouth Junction, NJ, USA), TOK‐001 was obtained from Sundia MediTech Co., Ltd. (Shanghai, China). Dihydrotestosterone (DHT), cycloheximide (CHX), actinomycin D (Act D), and 17‐AAG were obtained from Sigma‐Aldrich (Tokyo, Japan). Camptothecin and paclitaxel were obtained from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). Fetal bovine serum (FBS) and cFBS were obtained from Gibco (Grand Island, NY, USA) or HyClone (Logan, UT, USA); [³H]methyltrienolone was obtained from Perkin‐Elmer, Inc (Shelton, CT, USA). CellTiter‐Glo®, Bright‐Glo®, and Dual‐Glo® reagents were obtained from Promega KK (Tokyo, Japan). WST‐8 and 4′, 6‐diamidino‐2‐phenylindole (DAPI) solution were purchased from Dojindo Laboratories (Tokyo, Japan).

Androgen receptor ligand‐binding studies were conducted in vitro with whole‐cell lysate, including wild‐type AR or T878A mutant AR [10], which were prepared according to previously reported methods [11, 12] and [³H]‐methyltrienolone. Cell lysates and [³H]‐methyltrienolone were combined with an incubation buffer (25 mm HEPES, pH 7.4, 10% glycerol, 1 mm EDTA, and 10 mm sodium molybdate) and incubated for 20 h at 4 °C. The bound radio‐ligand was then separated from the free radio‐ligand by vacuum filtration through 0.3% polyethyleneimine‐pretreated GF/B filter mats (Perkin‐Elmer, Inc.). The filter mats were washed four times with washing buffer (25 mm HEPES, pH 7.4, and 1 mm EDTA). The radioactivity was measured using a liquid scintillation counter. Non‐specific binding was estimated in the presence of 10 μm testosterone.

Luciferase reporter assays were performed as described previously [9]. For the luciferase reporter assay of COS‐7 cells, the PSA‐promoter firefly luciferase plasmid, the Renilla luciferase reporter plasmid, and the wild‐type AR‐expressing plasmid were transfected using FuGENE® HD Transfection Reagent (Promega). For the luciferase reporter assay of LNCaP cells, the AR‐responsive firefly luciferase alone was transfected using the Amaxa™ Cell Line Nucleofector™ Kit (Lonza Ltd., Basel, Switzerland), as previously reported [9]. For the luciferase reporter assay of VCaP cells, the wild‐type AR‐expressing plasmid and AR‐responsive firefly luciferase reporter plasmid were transfected using Lipofectamine® 2000 or Lipofectamine® 3000 (Life Technologies Corporation, Waltham, MA, USA). For the luciferase reporter assay of HEK293 cells, AR‐responsive firefly luciferase reporter plasmid and the wild‐type AR or mutant AR‐expressing plasmid were transfected using FuGENE® HD Transfection Reagent. Luciferase activity was measured using Dual‐Glo® reagent or Bright‐Glo® reagent, following the manufacturer's instructions (Promega). AR transcriptional activity was measured in mammalian cells expressing AR‐dependent reporter constructs. DHT was used for the AR antagonist assays.

Proliferation assays were performed as described previously [9, 10]. For the proliferation assay of SAS MDV No. 3‐14 cells, the suspended cells in RPMI‐PR‐ containing 10% cFBS were seeded in 96‐well plates (3 × 10³ cells/well) and the plate incubated overnight in a humidified incubator with 5% CO₂ at 37 °C for attachment. The cells were incubated with 3, 5, 7, and 10 μm of TAS3681 and enzalutamide for 3, 6, or 9 days in the absence of DHT. The culture medium was changed every 3 days. Cellular proliferation was assessed 3, 6, and 9 days after treatment by incubation with CellTiter‐Glo® reagent. Luminescence was measured using Multilabel Counter (Wallac Arvo‐HTS; Perkin‐Elmer). For proliferation assay of LNCaP and VCaP cells, each cell line was suspended in RPMI‐PR‐ containing 5% or 10% cFBS, respectively, and separately plated in 96‐well plates (3 × 10³ cells/well for LNCaP, 1 × 10⁴ cells/well for VCaP) and incubated overnight in a humidified incubator with 5% CO₂ at 37 °C for attachment. They were treated with 0.003, 0.01, 0.03, 0.1, 0.3, 1, 3, and 10 μm of TAS3681, enzalutamide, bicalutamide, or apalutamide for 3 days in the presence of 0.1 μm DHT. A colorimetric growth assay using the WST‐8 solution was performed according to the manufacturer's instructions. For the proliferation assay of AR‐negative cells (MIAPaCa‐2, HCC1806, SK‐OV‐3 and DU145), each cell line was suspended in culture medium and seeded at a density of 3 × 10³ cells/well in 96‐well plates and incubated overnight in a humidified incubator with 5% CO₂ at 37 °C for attachment. Cells were incubated with 0.003, 0.01, 0.03, 0.1, 0.3, 1, 3, 10 μm of TAS3681, or 0.1 μm of paclitaxel for 3 days. A crystal violet assay was performed as previously described [13].

Cell proteins were extracted using ice‐cold M‐PER Mammalian Protein Extraction Reagent (Thermo Fisher Scientific Inc., Waltham, MA, USA). Nucleic and cytoplasmic proteins were extracted according to the instructions of the corresponding extraction kit (Promega). Tumor proteins were extracted using ice‐cold T‐PER Mammalian Protein Extraction Reagent containing a protease inhibitor cocktail (Sigma‐Aldrich, Milan, Italy). Western blotting (WB) was performed according to the previously reported standard method to analyze protein expression [8]. First, equal amount of protein was loaded onto the gel. After running the gel, transferring the protein to a membrane (Bio‐Rad Laboratories, Inc. (Hercules, CA, USA) or Life Technologies Corporation), and incubating it with antibodies, the protein image was acquired using a chemiluminescence detection system (ImageQuant LAS‐3000 or ImageQuant LAS‐4010; GE Healthcare, Chalfont, UK or Amersham™ Imager 600 QC; Cytiva, Chalfont, UK). The primary antibodies used were as follows: human AR (clone D6F11; Cell Signaling Technology, Danvers, MA, USA or clone N‐20; Santa Cruz Biotechnology, Dallas, TX, USA), AR‐V7 (clone RM7; RevMAb Biosciences (Burlingame, CA, USA) and clone E3O8L; Cell Signaling Technology), estrogen receptor α (clone D8H8; Cell Signaling Technology), glucocorticoid receptor (clone D8H2; Cell Signaling Technology), progesterone receptor A/B (clone D8Q2J; Cell Signaling Technology), human glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH; Catalog #2275‐PC‐100; Trevigen, Gaithersburg, MD, USA), HDAC1 (Catalog #2062), β‐tubulin (Catalog #2126), and cleaved‐poly (ADP‐ribose) polymerase (cPARP; Catalog #5625; Cell Signaling Technology).

RNA was extracted using an RNeasy® mini kit (QIAGEN, Maryland, MD, USA) according to the manufacturer's instructions. Two and a half micrograms of total RNA were used for cDNA synthesis using the SuperScript VILO cDNA Synthesis Kit (Life Technologies Corporation). To detect indicated genes, each cDNA sample was amplified using TaqMan® Fast Universal PCR Master Mix (Life Technologies Corporation) using an ABI7900HT Real‐Time qPCR system (Thermo Fisher Scientific Inc.). The gene‐specific primers for cDNA and SYBR Green dye detection or TaqMan MGB or VIC probe were used (Table S1). Relative mRNA levels were calculated using the comparative threshold cycle method (2–ΔΔCt), where the mRNA level of each gene of interest was normalized to that of the endogenous housekeeping gene GAPDH.

U‐2 OS tomato‐AR cells, suspended in RPMI‐PR‐ including 5% cFBS, were seeded on chamber slides (Thermo Fisher Scientific Inc.) or a 96‐well plate (1 × 10³ cells/well). The next day, the cells were treated with 10 μm of TAS3681, enzalutamide, and bicalutamide in the absence of DHT or were treated with 0.003, 0.01, 0.03, 0.1, 0.3, and 1 μm of TAS3681, enzalutamide, and bicalutamide in the presence of DHT for 2 h at 37 °C in a CO₂ incubator. After fixation with a 4% PFA phosphate buffer solution, the cells were washed with PBS, and nuclear staining was performed using ProLong® Gold antifade mountant with DAPI. For imaging, coverslips were mounted onto cells, and a nuclear image of an AR was obtained using a confocal microscope (Olympus Corporation, Tokyo, Japan). For quantitative analysis, four images were acquired of each well, and measurement of AR fluorescence intensity in the nuclear region (N) and in the cytoplasmic region (C) was performed with the IN Cell Analyzer1000 (GE Healthcare) and the IN Cell Developer Toolbox software (GE Healthcare). The N/C Ratio Mean was calculated from data on the ratio between the nuclear intensity and cytoplasmic fluorescence intensity. The N/C Ratio Mean for each well was determined from the four images of the well.

Chromatin immunoprecipitation (ChIP) assay was performed as previously described [14]. VCaP cells grown in androgen‐depleted media with 10% cFBS were treated with 3 nm DHT or 10 μm enzalutamide and TAS3681 combined with 3 nm DHT for 4 h. The cells were cross‐linked for 10 min and processed for ChIP using an AR antibody (N‐20; Santa Cruz Biotechnology). Real‐time PCR quantification of immunoprecipitated kallikrein‐related peptidase 3 (KLK3) and transmembrane protease, serine 2 (TMPRSS2) enhancer is performed using SYBR Green dye detection and is shown as (% input). Primers (Sigma‐Aldrich Co. LLC, Irvine, CA, USA) used for qPCR of kallikrein‐related peptidase 3 (KLK3) [15] and transmembrane protease, serine 2 (TMPRSS2) [16] enhancers are shown in Table S1.

CHX and Act D chase assays were performed as previously reported [17]. For the CHX chase assay, LNCaP and SAS MDV No. 3‐14 cells were cultured in RPMI‐PR‐ containing 5% cFBS. On the next day, LNCaP and SAS MDV No. 3‐14 cells were treated with 5 or 10 μm of TAS3681 or 1 μm of 17‐AAG for 0, 4, 8, or 24 h in the presence of 10 μg·mL⁻¹ of CHX. Expression of AR protein was evaluated using WB. For the Act D chase assay, LNCaP and SAS MDV No. 3‐14 cells were cultured in RPMI‐PR‐ containing 5% cFBS. On the next day, LNCaP cells were treated with 5 or 10 μm of TAS3681 or 30 μg·mL⁻¹ of CHX for 0, 4, 8, or 24 h in the presence of 5 μg·mL⁻¹ Act D. Expression of AR protein and mRNA was evaluated using WB and gene expression analysis, respectively.

Five‐week‐old male CB‐17/Icr‐scid/scidJcl mice (SCID mice) were purchased from CLEA Japan, Inc. (Tokyo, Japan). The mice were castrated at 6 weeks of age and subcutaneously injected with SAS MDV No. 3‐14 cells [8]. To evaluate the antitumor activity, animals bearing SAS MDV No. 3‐14 cell xenografts were randomized on day 1 according to tumor volume, once the mean tumor volume had reached ~ 130–250 mm³, and administered with 7.5, 15, and 22.5 mg·kg⁻¹ body weight of TAS3681 or vehicle control (0.5% hydroxypropyl methylcellulose; HPMC) twice daily by oral gavage for 14 days. The tumor volume and the body weight of mice were measured twice a week. The tumor volume (mm³) was defined as (A × B²)/2, where A and B were the longest and the shortest tumor diameter, respectively. Relative tumor volume (RTV) on day 15 was calculated as the ratio of tumor volume (TV) on day 15 to that on day 0. On days 1 and 15, blood was collected from the facial vein, and serum was separated. Serum prostate‐specific antigen (PSA) concentrations were assessed using a PSA ELISA kit (Catalog #DKK300; R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions, and the data were used to calculate the PSA ratio. The PSA ratio was calculated by dividing the PSA concentration of each sample on day 15 by that on day 1.

The mice were kept in the specific pathogen free animal facility with 12/12 h light/dark cycle. The animal facility is regularly tested for standard pathogens. Mice were fed no more than five per cage with free access to sterile water and food. Animal studies were reviewed and approved by the Institutional Animal Care and Use Committees at Taiho Pharmaceutical Co., Ltd. (Approval No. AE‐2016‐101) and carried out according to their guidelines for animal experiments.

Results are reported as mean ± standard deviation (SD) for in vitro experiments or mean ± standard error of the mean (SEM) for in vivo experiments. Statistical analyses and determination of the half‐maximal inhibitory concentration (IC₅₀) values were performed using sas version 9.2 and exsus version 8.0.0 (CAC Exicare Corporation, Tokyo, Japan). Statistical significance was assessed using the Student's t‐test or Wilcoxon test for two groups and the Dunnett's test, nonparametric Dunnett's test, or Williams' test for more than three groups. In all analyses, statistical significance was set at P < 0.05.

Through modification of N‐arylpiperazine‐1‐carboxamide compounds [18], we discovered TAS3681 as a novel nonsteroidal compound in a screening campaign using AR transactivation assay, DHT‐dependent cell proliferation assay, and another cellular assay designed to detect compounds with the capacity to downregulate AR expression in LNCaP cells, and subsequent medicinal chemistry optimization (Fig. 1A). In the present study, a competitive AR‐binding assay was performed to investigate the binding affinity of TAS3681 to wild‐type AR and AR T878A mutants. The inhibition constant (K_i) values of TAS3681 against wild‐type AR and AR T878A mutant were 7.39 and 23.8 nm, respectively (Table S2). These values were comparable to those of enzalutamide (7.11 nm for wild‐type AR and 22.6 nm for AR T878A mutant) and apalutamide (6.01 nm for wild‐type AR and 115.0 nm for AR T878A mutant) that were tested under the same conditions (Table S2). AR transactivation assays were performed using an AR‐responsive reporter along with a wild‐type AR in COS‐7 cells and VCaP cells that express wild‐type AR [19]. TAS3681 was a potent antagonist for wild‐type AR with IC₅₀ values of 52.7 and 60.9 nm, respectively (Fig. 1B, Fig. S1 and Table S3). These values were comparable to those of enzalutamide (59.7 and 117.7 nm). Bicalutamide showed weaker antagonist activity (IC₅₀ values of 429 and 662.5 nm for wild‐type AR) and could not completely block the transcriptional activity of wild‐type AR even at 1000 nm because of its partial agonist profile [20] (Fig. 1B, Fig. S1 and Table S3). Moreover, AR transactivation assays were performed using an AR‐responsive reporter along with LNCaP that express T878A mutant [21]. TAS3681 also shown potent antagonist activity (IC₅₀ value of 10.1 nm) similar to enzalutamide (IC₅₀ value of 12.5 nm) (Fig. 1B and Table S3). To determine whether the anti‐AR transcriptional activity of TAS3681 is accompanied by reduced cell proliferation, the number of viable VCaP and LNCaP cells was quantified after incubation with TAS3681. TAS3681 suppressed the DHT‐induced proliferation of VCaP cells in a dose‐dependent manner, with an IC₅₀ of 170 nm, which was comparable to that of enzalutamide (IC₅₀ of 180 nm) (Fig. 1C and Table S4). TAS3681, as well as enzalutamide and bicalutamide, also dose‐dependently suppressed LNCaP cell proliferation, with an IC₅₀ of 18 nm. Enzalutamide and bicalutamide had IC₅₀ values of 55 and 340 nm, respectively (Fig. 1C and Table S4). Consistent with the results of the antagonist assay, bicalutamide showed weaker inhibition of cell proliferation in VCaP and LNCaP, and the IC₅₀ values for bicalutamide in each cell line were approximately 3‐ and 19‐fold higher, respectively, than those for TAS3681. Therefore, TAS3681 inhibited DHT‐induced cell proliferation more potently than bicalutamide, and with activity similar to enzalutamide.

No significant effect of TAS3681 was observed on the viability of AR‐negative cells, such as MIAPaCa‐2 (human pancreatic carcinoma), HCC1806 (human breast cancer), SK‐OV‐3 (human ovarian cancer), and DU145 (human prostate cancer) indicating that the antiproliferative effect of TAS3681 in AR‐positive PCa cells is mediated through AR antagonism (Figsand). S2 S3

We evaluated the effect of TAS3681 on AR protein levels using WB with an AR antibody‐N‐terminal. LNCaP cells were treated with TAS3681 for 1 day; TAS3681, but not enzalutamide, reduced AR expression in a dose‐dependent manner (Fig. 1D). Similar results were observed for VCaP cells (Fig. S4). In addition, TAS3681 downregulated AR protein, but had little effect on the protein levels of other nuclear receptors, such as estrogen receptor, glucocorticoid receptor, and progesterone receptor (Fig. S5). Overall, the findings suggest that TAS3681 has the same potential as enzalutamide, a representative second‐generation compound, as an AR pure antagonist, but also has unique AR reduction activity.

Androgen receptor movement from the cytoplasm to the nucleus upon androgen binding is an essential step in AR‐mediated gene transcription. Second‐generation small‐molecule ARSIs, including enzalutamide, block the nucleo‐cytoplasmic translocation of AR and DNA binding without any AR agonist activity. As shown in Fig. 2A, AR was predominantly localized in cytoplasm in the absence of androgen, and DHT addition increased the nuclear/cytoplasmic (N/C) ratio of fluorescent intensity (N/C ratio: approximately 2.2), indicating the translocation of AR from the cytoplasm to the nucleus (Fig. 2B).

TAS3681 potently and dose‐dependently blocked AR nuclear translocation in U‐2 OS tomato‐AR cells stimulated with DHT (Fig. 2B). Enzalutamide also showed dose‐dependent inhibition of AR nuclear translocation (Fig. 2B). Conversely, bicalutamide showed little inhibition of AR nuclear translocation in those cells (Fig. 2A,B). The IC₅₀ of TAS3681 for blocking DHT‐induced AR nuclear translocation was estimated to be 63.7 nm, whereas that of enzalutamide was 103 nm (Table S5). Bicalutamide did not suppress AR nuclear translocation (Table S5). We also examined the effect of antiandrogens on AR subcellular localization in the absence of androgen. Nuclear translocation of AR was not observed after treating cells with TAS3681 or enzalutamide (Fig. S6A). The N/C ratios of cells treated with 10 μm TAS3681 or enzalutamide were calculated to be 1.360 and 1.325, respectively, which were similar to those of cells treated with DMSO indicating that TAS3681 and enzalutamide do not induce AR nuclear translocation and have a pure antagonist profile (Fig. S6B). In contrast, nuclear translocation of AR was observed after the treatment of cells with 10 μm bicalutamide, as previously reported [20]. The N/C ratio of cells treated with bicalutamide was 1.858, indicating that bicalutamide induces AR nuclear translocation and has partial agonist activity (Fig. S4B).

Next, we examined the effect of TAS3681 on AR binding to DNA in PCa cells. ChIP‐qPCR assays revealed that AR was efficiently recruited to the enhancer regions of KLK3 and TMPRSS2 in VCaP cells in the presence of DHT. Treatment with TAS3681 or enzalutamide impaired the AR binding to these regions (Fig. 2C). This finding suggests that TAS3681 blocks the nucleo‐cytoplasmic translocation of AR and DNA binding without any AR agonist activity.

Mutations in the AR are known to promote resistance to antiandrogen therapies, including enzalutamide [22, 23]. AR mutations are more common in patients progressing to these therapies than in responders [24]. The effects of antiandrogens on mutant ARs such as T878A, AR W742C, and AR F877L were studied using cell‐based transient transactivation assays with an androgen‐responsive luciferase reporter gene construct. In the absence of DHT, agonist mode transactivation assays showed that enzalutamide acted as an agonist for the AR F877L mutant, whereas bicalutamide and hydroxyflutamide were agonistic for mutants AR W742C and AR T878A, respectively (Fig. 2D). However, TAS3681 alone did not show agonistic profiles for all tested mutant ARs in the absence of DHT (Fig. 2D). We obtained similar results by transactivation assays in the presence of DHT, and only TAS3681 functioned as a pure antagonist for all tested mutant ARs (Fig. S7). Moreover, we expanded the list of functionally characterized AR‐LBD mutants with an additional 10 variants (mutated ARs L702H, V716M, W742L, H875Q, H875Y, D891Y, Q903H, H875Y/T878A, T878A/S889G, and T878A/D891H) [25, 26]. In addition to the three AR mutations tested earlier, we evaluated the response of these 10 AR mutants to increasing concentrations of DHT to determine the optimal concentration for each AR mutant, following a process similar to that for other mutants. The effects of antiandrogens (TAS3681, enzalutamide, apalutamide, and darolutamide) on the transactivation of 13 AR mutants were studied, and IC₅₀ values were calculated (Fig. S8 and Table S6). Enzalutamide and apalutamide showed agonistic effects on the mutated ARs F877L/T878A, H875Y/T878A, and F877L (Fig. S8 and Table S6). While darolutamide showed an agonistic effect on the mutated ARs V716M and H875Y, TAS3681 did not show an agonistic effect on any additional mutant ARs in this study (Fig. S8 and Table S6). Next, we investigated the effectiveness of TAS3681 in inhibiting the proliferation of LNCaP cells stably expressing F877L AR (F877L AR cells). TAS3681 treatment in the presence of DHT potently and dose‐dependently suppressed the proliferation of both the parent and F877L AR cells (Fig. 2E). Enzalutamide and apalutamide suppressed the proliferation of parent cells but failed to suppress the proliferation of F877L AR cells at any concentration, as previously reported (Fig. 2E) [7]. Taken together, these findings suggest that TAS3681 functions as an AR pure antagonist against wild‐type and clinically relevant LBD mutations.

Since TAS3681 showed AR reduction activity in a previous cell‐based study, we examined whether TAS3681 also reduces the expression of both AR‐FL and AR‐V7 in SAS MDV No. 3‐14 cells [8]. As shown in Fig. 3A, Figs S9 and S10, TAS3681 effectively and dose‐dependently reduced not only AR‐FL but also AR‐V7 in SAS MDV No. 3‐14 cells. TAS3681 also reduced the protein levels of both ARs in 22Rv1 cells, another AR‐V7 positive enzalutamide‐resistant cell line (Fig. S11) [27]. TAS3681 effectively and dose‐dependently suppressed the proliferation of SAS MDV No. 3‐14 cells expressing AR and AR‐V7 (Fig. 3B). In contrast, enzalutamide did not suppress the growth of SAS MDV No. 3‐14 cells at any of the concentrations tested (Fig. 3B).

Next, to examine whether TAS3681 affected the transcriptional activity of AR‐V7, real‐time qPCR analysis was conducted to quantify mRNA levels of AR‐V7‐dependent endogenous genes, including CDC20, CDK1, CCNA2, and UBE2C, in SAS MDV No. 3‐14 cells [28]. DHT did not affect the mRNA levels of the genes in SAS MDV No. 3‐14 cells (Fig. 3C and Fig. S8). TAS3681 considerably decreased the expression of AR‐V7 target genes in the presence and absence of DHT, whereas enzalutamide did not (Fig. 3C and Fig. S12). Next, to further explore the subcellular locations of AR‐FL and AR‐Vs, cytoplasmic and nuclear protein fractions of SAS MDV No. 3‐14 cells were analyzed. Under androgen‐depleted conditions, AR‐FL was detected in both the cytoplasm and nucleus, whereas AR‐Vs were localized predominantly in the nucleus, as previously reported [29]. TAS3681 reduced the expression of both AR‐Vs and AR‐FL within their respective localizations (Fig. 3D). In contrast, enzalutamide did not affect the expression of AR‐Vs in the nucleus and that of AR‐FL in the cytoplasm and nucleus (Fig. 3D).

The downregulation of AR‐V7 by antisense oligonucleotides is accompanied by suppression of androgen‐independent cell growth and apoptosis induction in PCa cells [30]. We found through WB analysis that cPARP, an apoptosis‐related protein, was expressed in SAS MDV No. 3‐14 cells treated with TAS3681 (Fig. 3E). In contrast, cPARP was not expressed after enzalutamide treatment (Fig. 3E). Altogether, TAS3681 suppressed the proliferation and induced apoptosis accompanied by downregulation of AR‐FL and AR‐V7 in enzalutamide‐resistant SAS MDV No. 3‐14 cells.

Androgen receptor overexpression and amplification have been reported as mechanisms of enzalutamide resistance [24, 26]. Next, we examined the effect of AR overexpression on the AR inhibitory activity of TAS3681 and enzalutamide. VCaP cells transfected with the wild‐type AR showed higher AR protein levels, and as a result, AR transcriptional activation and proliferation were more efficiently enhanced by DHT in AR‐overexpressing cells than in parental cells (Fig. 4A–C). Enzalutamide completely blocked DHT‐induced AR transcriptional activation and proliferation in parental cells but did not completely suppress them in AR‐overexpressing cells (Fig. 4A–C). In contrast, TAS3681 effectively blocked DHT‐induced AR transcriptional activation and proliferation in both parental and AR‐overexpressing cells. These observations of TAS3681 were consistent with the attenuation of AR levels in AR‐overexpressing cells treated with TAS3681 (Fig. 4A–C). The addition of AR‐lowering effects to AR antagonists might be an effective strategy to address the issues of AR amplification and overexpression.

Next, we investigated whether the mechanism of AR downregulation by TAS3681 caused an increase in the rate of AR protein degradation or a decrease in AR protein synthesis. A CHX chase assay was used to analyze AR protein stability. Treatment of LNCaP cells with TAS3681 in the presence of CHX did not increase the rate of AR degradation as well as β‐actin degradation compared with the vehicle control, in contrast to HSP90 inhibitor 17‐AAG (Fig. 5A and Fig. S13). Next, we examined AR mRNA and protein levels in the presence of Act D, an inhibitor of mRNA synthesis. AR mRNA levels were evaluated by RT‐qPCR after treatment with TAS3681 in the presence of Act D. Treatment with TAS3681 did not significantly reduce AR mRNA level compared to vehicle treatment; therefore, TAS3681 did not affect AR mRNA stability (Fig. 5B). The AR protein level was evaluated using WB analysis after the treatment of TAS3681 in the presence of Act D. The results showed that the AR protein level reduced over time after TAS3681 treatment (Fig. 5C lower left panel and Fig. S14). GAPDH protein expression did not decrease over time after TAS3681 treatment (Fig. 5C lower right panel and Fig. S14). Conversely, a non‐selective protein synthesis inhibitor, CHX, promoted the reduction of both AR and GAPDH levels in the presence of Act D (Fig. 5C upper panel and Fig. S14). TAS3681 did not show significant effect on AR mRNA levels in LNCaP cells compared to the vehicle (Fig. S15). Similarly, we investigated the mechanism of AR‐Vs and AR‐V7 downregulation by TAS3681 using SAS MDV No. 3‐14 cells. Treatment with TAS3681 did not prompt degradation of AR‐Vs protein in the presence of CHX. In the presence of Act D, AR‐V7 mRNA stability was not affected by TAS3681, while AR‐Vs protein level was reduced by TAS3681 compared to vehicle treatment (Fig. S16A–C). TAS3681 did not show significant effect on AR‐V7 mRNA levels in SAS MDV No. 3‐14 cells compared to the vehicle (Fig. S17). Considering the present findings, it can be deduced that TAS3681 reduces the rates of AR and AR‐Vs protein synthesis.

To assess the in vivo efficacy of TAS3681 in an enzalutamide‐resistant CRPC mouse model, castrated male SCID mice with subcutaneously injected SAS MDV No. 3‐14 cells were treated orally with TAS3681 twice daily for 14 days [8]. The half‐life of TAS3681 in mice was not optimal (3.63 h); therefore, TAS3681 was orally dosed using bids. The ratios of the mean RTV in each TAS3681‐treated group (7.5, 15, or 22.5 mg·kg⁻¹·day⁻¹) to those in the vehicle control group (T/C ratio, %) were 64%, 41%, and 24%, respectively. At all doses, the difference in the mean RTV between the vehicle group and each TAS3681 group was statistically significant (P < 0.025) (Fig. 6A and Fig. S18). Tumor regression was observed in one of ten mice treated with 15 mg·kg⁻¹·day⁻¹ of TAS3681, and three of ten mice treated with 22.5 mg·kg⁻¹·day⁻¹ of TAS3681 (Fig. 6B). Administration of TAS3681 caused a dose‐dependent decrease in serum PSA levels in castrated SCID mice bearing SAS MDV No. 3‐14 tumors. At 15 and 22.5 mg·kg⁻¹·day⁻¹ of TAS3681, the difference in mean fold PSA change from baseline after 14 days of treatment between the vehicle group and both TAS3681 groups was statistically significant (P < 0.025) (Fig. 6C and Fig. S19). In the present study, the expression of AR‐FL and AR‐Vs proteins in harvested tumors (on day 15) was analyzed by WB. As shown in Fig. 6D and Fig. S20, the expression of AR‐FL and AR‐Vs in xenograft tumors was dose‐dependently downregulated by TAS3681. Altogether, these results indicate that TAS3681 exerts significant antitumor effects against enzalutamide‐resistant PCa xenografts, accompanied by a reduction of AR‐FL and AR‐Vs levels in tumors.

All relevant data are within the paper and itsfiles. Supporting Information