Inhibition of androgen receptor by decoy molecules delays progression to castration-recurrent prostate cancer

Abstract

Androgen receptor (AR) is a member of the steroid receptor family and a therapeutic target for all stages of prostate cancer. AR is activated by ligand binding within its C-terminus ligand-binding domain (LBD). Here we show that overexpression of the AR NTD to generate decoy molecules inhibited both the growth and progression of prostate cancer in castrated hosts. Specifically, it was shown that lentivirus delivery of decoys delayed hormonal progression in castrated hosts as indicated by increased doubling time of tumor volume, prolonged time to achieve pre-castrate levels of serum prostate-specific antigen (PSA) and PSA nadir. These clinical parameters are indicative of delayed hormonal progression and improved therapeutic response and prognosis. Decoys reduced the expression of androgen-regulated genes that correlated with reduced in situ interaction of the AR with androgen response elements. Decoys did not reduce levels of AR protein or prevent nuclear localization of the AR. Nor did decoys interact directly with the AR. Thus decoys did not inhibit AR transactivation by a dominant negative mechanism. This work provides evidence that the AR NTD plays an important role in the hormonal progression of prostate cancer and supports the development of AR antagonists that target the AR NTD.

Fig 1. Lentivirus delivery of decoy AR_1-558 delays the time to castration-recurrence.

A, PSA nadir or percent drop in serum PSA levels in response to castration of mice bearing LNCaP xenografts and treated with lentivirus for mock (vehicle control), decoy AR_1-558 (ARN), GFP, or GFP-AR_1-558 (GFP-ARN). B, The time to reach pre-castration levels of serum PSA was doubled in animals injected with decoys. C, The time for the tumor volume to double was increased by decoys. Tumors were inoculated 5 days before castration and subsequently injected every 5 days until the duration of the experiment. Student t-test: *, p<0.05; **, p<0.01.

Fig 2. Levels of expression of decoys and endogenous AR in vivo.

A, Western blot analysis for GFP with a representative animal showing extremely high expression of GFP delivered by lentivirus to the xenograft, yet non-detectable levels of expression in the spleen, liver, lung, heart, and kidney of the same animal. Similar levels of protein (40μg) from whole cell lysates. The membrane was stripped and re-probed for β-actin as a loading control. B, AR, AR_1-558 (ARN), GFP, and GFP-AR_1-558 (GFP-ARN) protein levels in harvested xenografts. A non-specific diffuse band migrates slightly slower than AR_1-558 and was most apparent in the mock-treated lysates.

Fig 3. Decoy AR_1-558 does not prevent nuclear localization of the AR.

A, Xenografts were harvested at the duration of the experiments and sections were stained for AR NTD (441) or the LBD (C19). B, Fluorescent microscopy of GFP-AR in LNCaP cells stably expressing vector (left) or decoy (right) and treated with R1881 (10nM) for 30 minutes.

Fig 4. Decoys block the expression of androgen-regulated genes.

Real-time qPCR was performed using total RNA isolated from: A, LNCaP cells stably transfected with vector (Vec) or decoys (ARN) and treated for 24 hours with 10nM R1881; or B, xenografts injected with mock, AR_1-558 (ARN), GFP, and GFP-AR_1-558 (GFP-ARN). Transcript levels of HMGCR, KLK2, KLK3/PSA, MAF, RHOU, and SAT, were normalized to levels of GAPDH. The ratio of each transcript to GAPDH is plotted as fold-change. The bars represent the mean ± SD (n = 3). Student t-test: * p<0.05; **: p<0.01.

Fig 5. Decoys block AR-ARE interaction.

ChIP was performed in non-transfected and stably transfected (vector or decoy) cells in response to androgen. Parental LNCaP cells (untransfected) or stables expressing vector or decoy were treated ± androgens (R1881, 10 nM) for 3 hours and used in ChIP analyses with rabbit IgG (no antibody negative control) and anti-AR (mAb; C19 antibody to AR LBD). Eluted DNA fragments were purified and used for qPCR with primers designed to amplify the PSA ARE. Bars show the percentage input as the mean ± SD (n = 3). A representative result from repeated experiments is shown.

Fig 6. Decoy molecules do not interact with the AR.

A, LNCaP cells stably expressing vector (V) or decoy AR_1-558 (D) were incubated in serum (FBS) or with R1881 (1nM) for 3 h followed by immunoprecipation of the AR using an antibody to the LBD (Santa Cruz C19). Whole cell lysates (lanes 1 and 2), supernatant (lanes 3 and 4), wash (lanes 5 and 6), and immunoprecipitated complex-IP elution (lanes 7 and 8) were analyzed by Western blot using antibody to the AR NTD (Santa Cruz 441) to detect both AR and the decoy AR_1-558. B, Decoy AR_1-558 blocked ligand-independent activation of the AR by forskolin while AR_1-233 and AR_392-558 did not. LNCaP cells were transiently transfected with PSA(-630/+12)-luciferase reporter and expression vectors for His-tag, His-AR_1-558, His-AR_1-233, and His-AR_392-558 and treated with forskolin (50μM) for 48 h under serum-free conditions. The percent induction of PSA-luciferase activities relative to values achieved with expression of His-tag is shown. Bars represent the mean ± SE of three separate experiments. C, Western blot analysis using an antibody to His-tag with whole cell lysates from LNCaP cells transiently transfected with expression vectors for His-tag, His-AR_1-558, His-AR_1-233, and His-AR_392-558.

Fig 7. Effects of decoy on androgen-repressed genes.

Real-time qPCR was performed using total RNA isolated from LNCaP cells stably transfected with vector or decoy (ARN) and treated for 24 hours with 10nM R1881. Transcript levels of CAM2KN1, MMP16, SESN2, SLC44A1, ST7, TMEM144, UGT2B15, and UGT2B17 were normalized to levels of GAPDH. The ratio of each transcript to GAPDH is plotted as percent activity relative to vector control. The bars represent the mean ± SD (n = 3).