Suppression of androgen receptor-mediated gene expression by a sequence-specific DNA-binding polyamide

Abstract

Androgen receptor (AR) is essential for the growth and progression of prostate cancer in both hormone-sensitive and hormone-refractory disease. A DNA-binding polyamide that targets the consensus androgen response element binds the prostate-specific antigen (PSA) promoter androgen response element, inhibits androgen-induced expression of PSA and several other AR-regulated genes in cultured prostate cancer cells, and reduces AR occupancy at the PSA promoter and enhancer. Down-regulation of PSA by this polyamide was comparable to that produced by the synthetic antiandrogen bicalutamide (Casodex) at the same concentration. Genome-wide expression analysis reveals that a similar number of transcripts are affected by treatment with the polyamide and with bicalutamide. Direct inhibition of the AR-DNA interface by sequence-specific DNA binding small molecules could offer an alternative approach to antagonizing AR activity.

Fig. 1.

Targeting the ARE with DNA-binding polyamides. (A) Model of the AR transcription complex. (B) Consensus ARE. (C) Structures and ball-and-stick models of polyamides 1, designed to bind the consensus ARE, and 2, a 2-bp mismatch. Imidazole and pyrrole monomer units are represented by filled and open circles, respectively. The isophthalic acid tail moiety is represented by an hexagon.

Fig. 2.

Binding of 1 and 2 to the ARE in the PSA promoter. (A) Illustration of pAR and partial sequence of the PSA promoter. (B) Quantitative DNase I footprint titration experiments for polyamides 1 and 2 on the 5′-end-labeled PCR product of plasmid pAR-PSA: lane 1, intact DNA; lane 2, A reaction; lane 3, G reaction; lane 4, DNase I standard; lanes 5–15, 1 pM, 3 pM, 10 pM, 30 pM, 100 pM, 300 pM, 1 nM, 3 nM, 10 nM, 30 nM, 100 nM polyamide, respectively. (C) Isotherm for 1 binding to the ARE half-site 5′-AGAACA-3′. Polyamide 1 has a K_a = 8.3 ± 1.7 × 10⁹ for this site. Polyamide 2 shows no measurable binding in the footprinted region. (D) EMSA of DHT-stimulated LNCaP cell nuclear extract (NE) binding to a 31-bp oligonucleotide duplex containing the PSA promoter ARE in the presence of 1 and 2.

Fig. 3.

Inhibition of DHT-induced PSA and FKBP5 expression by 1 and 2. (A) Induction of PSA mRNA in the presence of 1 and 2 and bicalutamide, B, measured by quantitative real-time PCR. 1 and bicalutamide inhibit expression of PSA in a dose-dependent manner up to ≈70% at 10 μM. 2 has a more modest effect. (B) Secreted PSA protein measured by ELISA. (C) Chromatin immunoprecipitation assays with anti-AR or mock antibody treatment expressed as fold-enrichment (specific/mock) of DNA sequences at the PSA promoter and enhancer. AR occupancy at the PSA promoter and enhancer is decreased in the presence of 1 (10 μM) but not 2. (D) Induction of FKBP5 mRNA in the presence of 1 and 2 and bicalutamide, B. (E) Chromatin immunoprecipitation assays with anti-AR at the FKBP5 fifth intron enhancer. Polyamide concentrations are 10 μM.

Fig. 4.

Global effects on transcripts interrogated by using Affymetrix high-density Human Genome U133 Plus 2.0 arrays. (A) Divisive clustering of all measured transcripts under the four specified conditions: no treatment control, bicalutamide (B, 10 μM), 1 (10 μM), and 2 (10 μM). Clustering was based on an error-weighted Pearson correlation of intensity ratios for each treatment as compared with DHT-induced controls. (B) Ven diagrams representing transcripts down- and up-regulated (|fold change| ≥ 2.0, P ≤ 0.01) by bicalutamide and 1. Numbers inside the intersections represent transcripts affected by both treatments. Of the 122 transcripts down-regulated by both bicalutamide and 1, 117 are also observed to be induced by DHT at the same thresholds. (C) Agglomerative clustering of expression changes of the 199 transcripts induced or repressed 4-fold (P ≤ 0.01) or more by 1 nM DHT under the designated treatment conditions. Of the DHT-induced set, 70 were inhibited by polyamide 1, 20 were inhibited by 2, and 186 were inhibited by bicalutamide (|fold change| ≥ 2.0, P ≤ 0.01). Clustering parameters were the same as in A. Treatments reported are an error-weighted average from three experiments, except the noninduced control, which was an average from two experiments.