Androgen deprivation causes truncation of the C‐terminal region of androgen receptor in human prostate cancer LNCaP cells

Abstract

The androgen receptor (AR) acts as a ligand‐dependent transcription factor, whereas mutant AR lacking the C‐terminal ligand‐binding domain functions in a ligand‐independent manner. In the present study we report that the C‐terminal truncated AR, which we named AR‐NH1 (the N‐terminal fragment of AR cleaved in the neighborhood of helix 1 of the ligand‐binding domain), is produced in LNCaP prostatic carcinoma cells. The AR‐NH1 of ~90 kDa was observed in an androgen‐independent LNCaP subline and was further accumulated by the proteasome inhibitor MG132. MG132 treatment caused the accumulation of AR‐NH1 even in parent LNCaP cells. AR‐NH1 was produced in the absence of ligand or in the presence of the AR antagonist bicalutamide, whereas AR agonists suppressed its production. AR‐NH1 was detected with different AR antibodies recognizing amino acid residues 1–20 and 300–316 and was also generated from exogenous AR. Both siRNA‐mediated AR knockdown and treatment with a serine protease inhibitor (4‐(2‐aminoethyl)‐benzenesulfonyl fluoride) reduced AR‐NH1 levels. According to the predicted cleavage site (between amino acid residues 660–685) and its nuclear localization, it is assumed that AR‐NH1 functions as a constitutively active transcription factor. These data suggest that AR‐NH1 is produced under hormone therapy and contributes to the development of castration‐resistant prostate cancer due to its ligand‐independent transcriptional activity. (Cancer Sci 2012; 103: 1022–1027)

Prostate cancer is the most frequently diagnosed carcinoma and the second leading cause of cancer death among men in North America and Europe.1 Male hormone signaling acting through the androgen receptor (AR) plays a dominant role in prostate cancer development.2 Therefore, surgical or chemical castration (also called androgen‐deprivation therapy or hormone therapy) is used for its treatment. However, subsequent to these medical castrations, the majority of prostate cancers eventually relapse.3 Such castration‐resistant prostate cancers (CRPCs) have a poor prognosis and remain incurable. Understanding the mechanisms underlying the progression to CRPC is essential to evaluate therapeutic targets for development of future therapies.4

AR is a member of the nuclear receptor superfamily of ligand‐activated transcription factors.2 AR localizes in the cytoplasm with chaperone proteins such as HSP90. After binding to testosterone or its active metabolite, 5α‐dihydrotestosterone (DHT), AR translocates from the cytoplasm to the nucleus where it controls the transcription of specific genes by recruiting the coactivators on target gene promoters. Prostate‐specific antigen is the most characterized AR target gene and is used as a marker for clinical diagnosis of prostate cancer. AR is composed of (in order from the N‐terminus to the C‐terminus) an N‐terminal domain, a DNA‐binding domain, a hinge region, and a ligand‐binding domain (LBD).5 The N‐terminal domain possesses ligand‐independent activation function (AF)‐1 and the LBD possesses ligand‐dependent AF‐2.6, 7 AF‐1 is considered to be more potent than AF‐2 in the transcriptional activation of AR, because AR lacking the C‐terminal LBD (the constitutively active form) shows strong transcription due to its AF‐1 activity alone.8

The development of CRPC still depends on AR function.3, 9, 10 The expression level of prostate‐specific antigen, which is decreased by hormonal therapy, is restored in CRPC. Thus, it is essential to identify the molecular switch that triggers the transcription of AR even in the absence of ligand in order to understand ligand‐independent but AR‐dependent CRPC development. Recently, an abnormal accumulation of AR lacking the C‐terminal LBD was found in CRPC patients.11 The appearance of constitutively active forms of AR (lacking the LBD) can account for the AR‐dependent but hormone‐independent growth of CRPC because they have ligand‐independent activity. Although CRPC develops during hormone therapy, the relationships between the production of ARs lacking the LBD and hormone therapy are not clear. In the present study, we report that proteolysis‐mediated truncation of the C‐terminus of AR was observed in the absence of ligand and in the presence of antagonist, that is, under conditions similar to hormone therapy. Our results illustrate the transition to androgen‐independent growth of CRPC after hormonal therapy.

Materials and Methods

Cell culture

The AR‐positive and AR‐negative human prostatic cancer cell lines, LNCaP and PC‐3, respectively, were cultured in RPMI‐1640 medium supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin and 100 μg/mL streptomycin as described previously.12 Dextran‐coated charcoal‐treated FBS (5%) and phenol red‐free RPMI‐1640 medium was used when steroid‐free medium was prepared for experiments using parent LNCaP or PC‐3 cells. Cells were maintained at 37°C in a 5% CO₂/95% air atmosphere at 100% humidity unless otherwise indicated. LNCaP and PC‐3 cells were obtained from the Cell Resource Center for Biomedical Research, Institute of Development, Aging and Cancer, Tohoku University (Miyagi, Japan).

Isolation of androgen‐independent LNCaP (AI‐LNCaP) cell line

AI‐LNCaP cells were used for the in vitro model of CRPC. We established AI‐LNCaP cells that grow under androgen‐independent conditions using the method previously described.13 Briefly, LNCaP cells were incubated with 1 μM bicalutamide in the steroid‐free medium containing 10% dextran‐coated charcoal‐treated FBS for 3 months, and formed colonies were isolated. The AI‐LNCaP cells were cultured in the steroid‐free medium. The AR nucleotide sequence in AI‐LNCaP cells was identical to that in parent LNCaP cells (data not shown).

Plasmids

To generate N‐terminal 3xFlag‐tagged human wild‐type AR, AR cDNA was subcloned from pcDNA3.1‐AR12 into p3xFlag‐Myc‐CMV vectors (Sigma, St. Louis, MO, USA). For expression of the C‐terminal truncated form of AR, PCR products encoding the AR (1–660) and AR (1–685) were inserted into p3xFlag‐CMV vectors. Amino acid numbers of AR in the present study are based on GenBank accession No. AAA51729. Luciferase reporter vector, pGL4‐ARE2‐TATA‐Luc, was constructed by inserting the two tandem repeats of an androgen‐response element of rat C3 (1) gene together with an adenovirus E1b TATA sequence into pGL4.14 vector.12

siRNA

Control siRNA was obtained from Sigma (SIC001). Double‐strand siRNA for AR was chemically synthesized. The sense sequence of the AR siRNA was as follows: 5′‐GACUCAGCUGCCCCAUCCA(dTdT)‐3′.14 LNCaP cells that had been cultured in steroid‐free medium were transiently transfected with 10 nM of control or AR siRNA for 24 h using Opti‐MEM and Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA). Medium was exchanged to fresh steroid‐free medium and the cells were further incubated for an additional 24 h in the presence or absence of 10 μM MG132.

Western blot analysis

LNCaP and PC‐3 cells that had been cultured in steroid‐free medium were transiently transfected with AR expression vector with HilyMax reagent (Dojindo, Kumamoto, Japan) for 24 h, followed by medium exchange. After incubation with MG132 or PS‐341, inhibitors of 26S proteasome, together with ligand, the cells were lysed in 20 mM HEPES‐NaOH, pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.5% Nonidet P‐40, 10 mM sodium pyrophosphate, 2 mM EDTA, 1 mM 4‐(2‐aminoetyl)benzenesulfonyl fluoride (AEBSF), 10 μg/mL leupeptin, and 1 μg/mL aprotinin. The sonicated lysate was centrifuged at 20 000 g for 10 min and the supernatant was subjected to SDS‐PAGE, followed by Western blot analysis with anti‐AR (N20 and AR441; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or anti‐Flag (M2; Sigma) antibodies. After immunoreaction to HRP‐conjugated secondary antibodies (Bio‐Rad, Hercules, CA, USA), the immunoreactive bands were developed with Immobilon Western Chemiluminescent HRP Substrate (Millipore, Bedford, MA, USA) or Super Signal West Femto Chemiluminescent Substrate (Pierce, Rockford, IL, USA) and were detected with an LAS4000 imager (Fujifilm, Tokyo, Japan).

Subcellular fractionation

PC‐3 cells were transiently transfected with pcDNA3.1‐AR12 or pcDNA3.1‐AR(L26A/F27A)15 using HilyMax reagent for 8 h. Medium was exchanged to fresh medium and cells were incubated for an additional 20 h. Subcellular fractionation was carried out as described previously.15 Briefly, cells were suspended in sucrose buffer (10 mM HEPES‐NaOH, pH 7.5, 250 mM sucrose, 1 mM EDTA, 1 mM DTT, 1 mM AEBSF, 10 μg/mL leupeptin, and 1 μg/mL aprotinin) and homogenized. The homogenate was subjected to differential centrifugation as described previously,16 and nuclear and post‐nuclear fractions were prepared. Each fraction (10 μg) was analyzed by Western blotting with anti‐AR (N20) antibody.

Reporter assay

PC‐3 cells that had been cultured in steroid‐free medium were transiently transfected with 0.01 μg pGL4.74[hRluc/TK], 0.1 μg pGL4‐ARE2‐TATA‐Luc, and 0.1 μg AR expression vectors (p3xFlag‐AR, p3xFlag‐AR[1–660], p3xFlag‐AR[1–685]) for 24 h using HilyMax reagent. Medium was exchanged to fresh medium and cells were incubated in the presence or absence of 10 nM DHT for an additional 24 h. Luciferase reporter activities were determined using a dual‐luciferase reporter assay system (Promega, Madison, WI, USA) as described previously.12 Data are expressed as relative light units.

Results

Production of a ~90 kDa protein in AI‐LNCaP cells and its degradation by the proteasome system

We newly established AI‐LNCaP cells for an in vitro model of CRPC and used them to analyze AR by Western blotting with N20 antibody (epitope: amino acids 1–20). An immunoreactive protein of ~90 kDa (designated p90 and indicated by an arrow in the figures) was observed only in AI‐LNCaP cells, even when the bands were developed for a short time. In contrast, p90 was not detected in parent LNCaP cells even when the bands were developed for a long time (Fig. 1A). Treatment of AI‐LNCaP cells with MG132 to prevent their degradation by the ubiquitin–proteasome system17 markedly increased p90 in AI‐LNCaP cells (Fig. 1B). MG132 treatment caused the accumulation of p90 even in parent LNCaP cells (Fig. 1C). p90 also accumulated in the presence of PS341 (another proteasome inhibitor also known as bortezomib). These data suggest that p90 production is associated with the acquisition of androgen independence and is degraded by the ubiquitin‐proteasome system.

Figure 1.

Expression of a ~90 kDa protein and its accumulation by a proteasome inhibitor. (A) Parent LNCaP prostate cancer cells and androgen‐independent (AI)‐LNCaP cells were incubated in the absence of ligand. Cell lysates from parent LNCaP cells and AI‐LNCaP were prepared and subjected to SDS‐PAGE, followed by Western blotting with anti‐androgen receptor (AR) (N20) antibody. The bands were developed for a short (upper panel) or a long (lower panel) time. (B) AI‐LNCaP cells were incubated in the presence or absence of MG132 for 24 h. The cell lysates were analyzed by Western blotting with anti‐AR (N20) antibody. (C) Parent LNCaP cells that had been cultured in steroid‐free medium for 60 h were further incubated with 10 μM MG132 or PS341 for an additional 24 h. The cell lysates were analyzed by Western blotting with anti‐AR (N20) antibody. The graph is representative of three independent experiments. Authentic AR and ~90 kDa protein are indicated by the arrowheads and arrows, respectively.

Accumulation of p90 in ligand‐regulated manners

Accumulation of p90 by MG132 was suppressed by testosterone (100 nM) and DHT (10 nM) (Fig. 2A) but little affected by the androgen antagonist bicalutamide (10 μM) in parent LNCaP cells (Fig. 2B). In addition, the decrease in the production of p90 by DHT was reversed by bicalutamide. Similar results were obtained with AI‐LNCaP cells (Fig. 2C). Agonists (e.g., testosterone and DHT) reduced p90 levels, whereas bicalutamide had little effect on p90 levels. Moreover, bicalutamide reversed agonist‐reduced p90 levels. These results show that p90 production is regulated by AR ligands and suggest that p90 is selectively produced under conditions similar to hormone therapy.

Figure 2.

Effects of ligands on the production of ~90 kDa protein. (A) Parent LNCaP prostate cancer cells that had been grown in steroid‐free medium for 60 h were treated with 100 nM testosterone or 10 nM 5α‐dihydrotestosterone (DHT) in the presence or absence of 10 μM MG132 for 24 h. (B) Parent LNCaP cells that had been grown in steroid‐free medium for 60 h were incubated with androgen antagonist bicalutamide (10 μM) with or without 10 nM DHT in the presence or absence of 10 μM MG132 for 24 h. (C) AI‐LNCaP cells were incubated in the presence of ligand for 15 h. Cell lysates were subjected to SDS‐PAGE, followed by Western blotting with anti‐androgen receptor antibody. The graph is representative of three independent experiments. Authentic AR and ~90 kDa protein are indicated by arrowheads and arrows, respectively.

Identification of p90 as the N‐terminal part of AR

We investigated whether p90 was the N‐terminal region of AR. The siRNA‐mediated knockdown of AR not only decreased the authentic AR but also reduced the MG132‐dependent accumulation of p90 in parent LNCaP cells (Fig. 3A). p90 was recognized not only with N20 antibody but also with AR441, another anti‐AR antibody (epitope: amino acid residues 300–316) (Fig. 3B). When N‐terminal Flag‐tagged AR was exogenously expressed in LNCaP cells, Flag‐tagged p90 was also detected with anti‐Flag antibody (Fig. 3C). Detection of Flag‐tagged p90 could not have been the result of alternative splicing of AR because exogenous AR cDNA does not have any introns. MG132‐dependent accumulation of p90 was also observed in AR‐negative PC‐3 prostate cancer cells when exogenous AR was expressed (Fig. 3D). Overexpression of mutant AR(L26A/F27A) substituting for Ala at Leu26 and Phe27 within a target of the ubiquitin–proteasome system18 (e.g., ²³FQNLF²⁷ motif) caused an accumulation of L26A/F27A‐mutated p90 even in the absence of MG132. Subsequently, PC‐3 cells overexpressing AR(L26A/L27A) that had been incubated in the absence of MG132 were fractionated. Full‐length AR(L26A/F27A) was distributed both in the nuclear and post‐nuclear fractions, whereas the L26A/F27A mutated p90 was mostly localized in the nuclear fraction (Fig. 3E). Together, the antibody staining results indicate that p90 is the N‐terminal part of AR and locates in the nucleus. Moreover, p90 is expected to be produced by the processing of authentic AR because it was generated from exogenous AR. A weak band of p90 was also observed even in the absence of MG132 when exogenous Flag‐tagged AR was expressed in the cells, indicating that p90 was naturally produced as a part of the life cycle of AR protein.

Figure 3.

Identification of the ~90 kDa proteins as N‐terminal fragment of androgen receptor (AR). (A) Parent LNCaP prostate cancer cells seeded on 35‐mm dishes with steroid‐free medium were transiently transfected with 10 nM each of control siRNA or AR siRNA for 24 h. Medium was then exchanged to fresh medium and cells were further incubated in the presence of 10 μM MG132 for an additional 24 h. (B) Parent LNCaP cells were incubated with 10 μM MG132 in the absence of ligand for 24 h. (C) Parent LNCaP cells were transiently transfected with p3xFlag‐Mock or p3xFlag‐AR vector for 4 h. After incubation for 20 h, cells were grown for an additional 24 h in the presence of 10 μM MG132. (D,E) PC‐3 cells were overexpressed with wild‐type AR or AR (L26A/L27A) and incubated in the presence or absence of 10 μM MG132 for 24 h. Whole cell lysates (D) or fractionated extracts (E) were prepared. Prepared cell lysates were subjected to SDS‐PAGE, followed by Western blot analysis with anti‐AR or anti‐Flag antibodies. Each image is representative of three independent experiments. Authentic AR and ~90 kDa AR are indicated by arrowheads and arrows, respectively.

Suppression of the production of p90 by a serine protease inhibitor

A serine protease inhibitor, AEBSF, suppressed the MG132‐induced accumulation of p90 in parent LNCaP cells (Fig. 4A). In addition, AEBSF selectively decreased p90 levels in AI‐LNCaP cells (Fig. 4B). These data suggest that p90 is proteolytically processed from authentic AR by a serine protease.

Figure 4.

Effect of serine protease inhibitor on the production of the ~90 kDa androgen receptor (AR). (A) Parent LNCaP prostate cancer cells that had been incubated in steroid‐free medium for 60 h were cultured in the presence of 10 μM MG132 and 500 μM 4‐(2‐aminoethyl)benzenesulfonyl fluoride (AEBSF) for an additional 24 h. (B) Androgen‐independent (AI)‐LNCaP cells were incubated in 500 μM AEBSF for the indicated time in the absence of MG132. Cell lysates were prepared and subjected to SDS‐PAGE, followed by Western blotting with anti‐AR antibody. The graph is representative of three independent experiments. Authentic AR and ~90 kDa AR are indicated by arrowheads and arrows, respectively.

Prediction of the cleavage site and function of p90

The molecular mass of p90 was found to be between the molecular masses of C‐terminal truncated mutants AR (1–660) and AR (1–685), indicating that the cleavage site is located between residues 660 and 685 (Fig. 5A). Because helix 1 of the AR LBD is located in amino acid residues 670–677, the cleavage site is within or close to helix 1. Accordingly, p90 was renamed AR‐NH1 (the Ṉ‐terminal fragment of AR cleaved in the neighborhood of ẖelix 1 of the LBD).

Figure 5.

Prediction of the cleavage site and transcriptional activity of ~90 kDa androgen receptor (AR). (A) Parent LNCaP prostate cancer cells that had been cultured in steroid‐free medium were transiently transfected with Flag‐tagged AR expression vectors for 24 h. Cells were incubated in the presence or absence of 10 μM MG132 for an additional 24 h. Cell lysates were prepared and subjected to SDS‐PAGE, followed by Western blotting with anti‐Flag antibody. Authentic AR and ~90 kDa AR are indicated by the arrowhead and arrow, respectively. (B) PC‐3 cells that had been incubated in steroid‐free medium were transiently transfected by Flag‐tagged AR expression vectors together with pGL4‐ARE2‐TATA‐Luc and pGL4.74[hRluc/TK] for 24 h. Medium was exchanged to fresh medium and cells were incubated in the presence or absence of 10 nM 5α‐dihydrotestosterone (DHT) for an additional 24 h. Cells were lysed and luciferase activities were determined. Values are indicated as the mean ± SD. Each graph is representative of three independent experiments.

To evaluate the possibility that AR‐NH1 functions as a ligand‐independent transcription factor, we addressed the transcriptional activities of AR (1–660) and AR (1–685) because AR‐NH1 is composed of amino acid residues 1 to 660–685. The PC‐3 cells that had been transfected with AR expression vector and luciferase‐reporter vector were incubated in the presence or absence of 10 nM DHT, and their luciferase activities were determined. As might be expected, the transcriptional activity of the wild‐type was stimulated by DHT, whereas both AR (1–660) and AR (1–685) exerted their transcriptional activities even in the absence of DHT (Fig. 5B). Taken together, these results strongly suggest that AR‐NH1 functions as a ligand‐independent transcription factor.

Discussion

CRPC develops during hormone therapy which is one of the most effective treatments for the hormone‐sensitive early stage. Restoration of AR target genes are observed in CRPC, revealing the dominant role of AR in the recurrence.3, 9, 10 The characteristics of C‐terminal truncated AR, which acts as a ligand‐independent transcription factor,19 well fit the model for the ligand‐independent but AR‐dependent growth pattern of CRPC. Notably, AR protein lacking the C‐terminal LBD is observed in CRPC patients.11 Alternative splicing that produces AR lacking the LBD accounts for the generation of the C‐terminal truncated AR.20 In fact, an AR variant lacking exons 5–7 was found in a human prostate cancer.22 As summarized in Figure 6, the present results indicate that C‐terminal truncated AR, AR‐NH1, was expressed in an AI‐LNCaP in vitro model of CRPC and that AR‐NH1 was proteolytically produced in the absence of ligand and in the presence of antagonist, that is, under conditions similar to hormone therapy. The ligand‐dependent AR proteolysis model accommodates the production of AR lacking the LBD in hormone therapy.

Figure 6.

Proposed mechanisms underlying the development of castration‐resistant prostate cancer (CRPC). Prostate cancer initially develops in an androgen‐dependent manner. After hormone therapy, antagonist‐bound or unliganded androgen receptor (AR) is cleaved between amino acid residues 660 and 685, and this cleavage is inhibited by a serine protease inhibitor. The N‐terminal fragment after cleavage, AR‐NH1, is normally degraded by the ubiquitin–proteasome system but is accumulated in the nucleus under some biological conditions. Considering the cleavage site and its nuclear localization, AR‐NH1 is predicted to act as a ligand‐independent transcription factor, which causes ligand‐independent but AR‐dependent development of CRPC. ARE, androgen‐response element.

AR‐NH1 was naturally produced as a part of the life cycle of authentic AR and was accumulated under certain conditions. AR‐NH1 protein was detected in AR‐overexpressing parent LNCaP cells even in the absence of MG132. Therefore, augmentation of AR mRNA, which is commonly found in CRPC patients,23, 24 will lead to an excess accumulation of AR‐NH1. On the other hand, AR‐NH1 was accumulated by proteasome inhibitor and by mutation in the ²³FQNLF²⁷ site, which is a target of the ubiquitin–proteasome system,18 indicating that AR‐NH1 is degraded by the ubiquitin–proteasome system. Suppression of proteasome activity can also cause an excess accumulation of AR‐NH1. Although proteasome activity is required for cancer cell survival,25 proteasome activity is decreased in lung cancer stem‐like cells.21 Therefore, AR‐NH1 may be preferentially produced in cancer stem cells.

AR‐NH1 was proteolytically processed from authentic AR. Some AR mutants lacking the C‐terminal LBD were generated by alternative splicing.20, 22 In contrast, the production of AR‐NH1 was observed when AR cDNA was introduced into the LNCaP cells. This result indicates that AR‐NH1 was not the product of alternative splicing. Moreover, the production of AR‐NH1 was inhibited by the serine protease inhibitor AEBSF. Our data suggest that serine protease inhibitors are candidates for the treatment of CRPC.

The production of the AR‐NH1 was observed in the absence of ligand and in the presence of antagonists whereas agonists suppressed its production, indicating that an agonist‐induced conformational change inhibited the generation of AR‐NH1. This notion is supported by the fact that AR forms different structures depending on the binding to agonists or antagonists.26 The cleavage site that produces AR‐NH1 was predicted to be in the amino acid region 660–685 overlapping with helix 1 of the LBD. Helix 1 of the thyroid receptor or estrogen receptor‐α largely alters their conformation by binding to their ligands,5, 27 and the LBD of the nuclear receptors is more highly conserved in overall structure than the primary sequence.28 Therefore, our data suggest that a ligand‐related conformational change also occurs in the region surrounding helix 1 of AR and is involved in the C‐terminal truncation of AR.

Proteasome inhibitors are considered to be candidate cancer drugs.25 For example, PS341 has been approved for treating multiple myeloma.29 The use of a proteasome inhibitor might produce AR‐NH1 in the prostate. However, proteasome activity is required for the transactivation of AR.30 Because the proteasome inhibitor decreased not only the ligand‐dependent transcriptional activity of AR but also the ligand‐independent activity of AR (1–660) (data not shown), it probably suppresses the function of AR‐NH1.

Like AR‐NH1, the LBD‐lacking ARs produced by alternative splicing are also considered to be involved in the development of CRPC.22 Due to the lack of the LBD, the transcriptional activities of these mutant ARs could not be downregulated by antagonists. In contrast, resveratrol (a polyphenol contained in grape) or EPI‐001 (harmless metabolite of bisphenol A) can suppress the transcriptional activity of LBD‐lacking ARs through mechanisms other than an antagonist.31, 32, 33 Therefore, non‐antagonistic anti‐androgens such as resveratrol or EPI‐001 are notable for the prevention or treatment of CRPC.

In the present study, we found AR‐NH1 proteins was expressed in an established androgen‐independent CRPC model of LNCaP cells even in the absence of proteasome inhibitor. Our results suggest that AR‐NH1 was selectively produced under conditions similar to hormone therapy and acts as a ligand‐independent transcription factor, which contributes to the development of CRPC.

Disclosure Statement

The authors have no conflict of interest.

Abbreviations

AEBSF

4‐(2‐aminoethyl)benzenesulfonyl fluoride

AF

activation function

AI‐LNCaP

androgen‐independent lymph node carcinoma of the prostate cells

AR

androgen receptor

ARE

androgen‐response element

CRPC

castration‐resistant prostate cancer

DHT

5α‐dihydrotestosterone

FBS

fetal bovine serum

LBD

ligand‐binding domain