Acquired resistance to the second-generation androgen receptor antagonist enzalutamide in castration-resistant prostate cancer

Abstract

Enzalutamide (MDV3100) is a second generation Androgen Receptor (AR) antagonist with proven efficacy in the treatment of castration resistant prostate cancer (CRPC). The majority of treated patients, however, develop resistance and disease progression and there is a critical need to identify novel targetable pathways mediating resistance. The purpose of this study was to develop and extensively characterize a series of enzalutamide-resistant prostate cancer cell lines. Four genetically distinct AR-positive and AR-pathway dependent prostate cancer cell lines (CWR-R1, LAPC-4, LNCaP, VCaP) were made resistant to enzalutamide by long-term culture (> 6 months) in enzalutamide. Extensive characterization of these lines documented divergent in vitro growth characteristics and AR pathway modulation. Enzalutamide-resistant LNCaP and CWR-R1 cells, but not LAPC-4 and VCAP cells, demonstrated increased castration-resistant and metastatic growth in vivo. Global gene expression analyses between short-term enzalutamide treated vs. enzalutamide-resistant cells identified both AR pathway and non-AR pathway associated changes that were restored upon acquisition of enzalutamide resistance. Further analyses revealed very few common gene expression changes between the four resistant cell lines. Thus, while AR-mediated pathways contribute in part to enzalutamide resistance, an unbiased approach across several cell lines demonstrates a greater contribution toward resistance via pleiotropic, non-AR mediated mechanisms.

Keywords: androgen receptor; castration-resistance; enzalutamide; prostate cancer.

Figure 1. Morphology, growth rates and sensitivity to docetaxel of enzalutamide resistant and parental cell lines

(A) Phase contrast and Fluorescent images taken at 60x of Parental and respective enzalutamide resistant (Enz^R) cell lines. Rhodamine conjugated Phalloidin (Red) which stains the F-actin cytoskeleton [50] is used to outline cellular morphology, and DAPI (blue) is a DNA stain used to outline the nuclei [51]. Cross signifies statistical significance to both parental and short-term treated cells, *indicated p-value < 0.05. (B) Relative growth rates of Parental, Enz^R (grown in 10 μM enzalutamide) and cells acutely treated (7 days) with 10 μM enzalutamide measured by MTT. (C) Relative growth rates of Parental and Enz^R (grown in 10 μM enzalutamide) cells grown in increasing concentrations (0.5, 1, 5 and 10 nM) of the microtubule stabilization chemotherapeutic agent, Docetaxel.

Figure 2. Enzalutamide resistant LNCaP and CWR-R1 cells display increased castration resistant tumor take, and increased metastatic colonization to bone

(A) Kaplan-Meier survival curves illustrating time to tumor take (measured as > 100 mm³) in a castrated male nude mice. Parental and Enz^Rcells were injected subcutaneously on the flanks of castrated nude mice: LNCaP Parental (N = 20) vs LNCaP Enz^R (N = 20) (Left Panel; p < 0.0001); CWR-R1 Parental (N = 10) vs. CWR-R1 Enz^R (N = 10) (Right Panel; p = 0.0041). (B) Kaplan-Meier survival curves illustrating overall survival of mice post-intracardiac injection to the veterinarian approved endpoint. LNCaP parental (N = 8) vs. LNCaP Enz^R (N = 13) cells (Left Panel; p = 0.0076); CWR-R1 parental (N = 12) vs. CWR-R1 Enz^R (N = 9) cells (Right Panel; p = 0.0015). (C) Representative images of in vivo bioluminescent imaging of the metastatic colonization of the LNCaP Parental and Enz^R intracardiac injected mice (Left Panel, Dorsal side shown), and CWR-R1 Parental and Enz^R intracardiac injected mice (Right Panel, Ventral side shown). (D) Representative H & E images (20× magnification) of histological sections of metastases obtained from CWR-R1 Enz^R injected mice. Metastases were confirmed by a pathologist in the Bone (Top image, Tibia), Brain (Middle image) and Adrenal glands (Bottom image) (additional images in Supplementary Figures S2 and S3). Scale bar represents 100 μm.

Figure 3. The status of the androgen receptor (AR) in parental and respective enzlutamide-resistant (Enz^R) cell lines

(A) Analysis of AR mRNA expression from Parental, Enz^R (grown in 10 μM enzalutamide) and cells acutely treated (48 hours) with 10 μM enzalutamide to identify splice variants and truncations. One set of primers spans the junction between exons 1 and 2 to identify total AR expression, another is designed to identify most full length AR transcripts, and finally primers were designed to AR-V7 within cryptic exon (CE) 3b, to gain a broader sense of total, full length and splice variant expression. Expression was normalized to GAPDH using the ΔΔCT method. (B) Western blot analysis to determine AR protein levels. Whole cell lysates from Parental, Enz^R (grown in 10 μM enzalutamide) and cells acutely treated (48 hours) with 10 μM enzalutamide. We investigated using three different AR antibodies: polyclonal targeting the N-terminus (N-20), monocolonal targeting the N-terminus (XP) and polyclonal targeting the C-terminus (C-19), and AR splice variant V7 expression (precision antibody). β-Actin was used as a loading control. Red Bars demarcate 100kDA molecular weight (AR N-20, XP, and C-19); Blue bars demarcate 75 kDA molecular weight (AR-V7). (C) Nuclear localization of AR protein in Enz^R cells. Western blot analysis of AR (N-20) from Parental, Enz^R(grown in 10 μM enzalutamide) and cells acutely treated (48 hours) with 10 μM enzalutamide in cells that were biochemically fractionated to isolate Nuclear and Cytosolic compartments, as well as whole cell lysates. Lamin A/C served as a nuclear control, and GAPDH served as a cytosolic marker. Each lane represents 100,000 cells worth of lysate.

Figure 4. The expression of the AR target genes PSA and TMPRSS2 and the status of AR signaling in the Enz^R cell lines

(A) RNA analyses of canonical AR-target genes, Prostate Specific Antigen (PSA), or KLK3 [35], and TMPRSS2 (Transmembrane protease serine 2) [36] using custom designed primers and cDNA from Parental, Enz^R (grown in 10 μM enzalutamide) and cells acutely treated (48 hours) with 10 μM enzalutamide. (B) PSA protein expression of conditioned media from Parental, Enz^R (grown in 10 μM enzalutamide) and cells acutely treated (48 hours) with 10 μM enzalutamide grown for 48 hours.

Figure 5. Analysis of adaptive altered AR-Regulated gene expression in Enz^R cells

(A) Schematic illustrating how AR- regulated restored genes were identified. Gene expression analysis to identify AR-pathway promoted and repressed gene expression differences between Parental and cells acutely treated (48 hours) with 10 μM enzalutamide, and which genes were then “restored” in the Enz^R cells (grown in 10 μM enzalutamide), or altered back to the expression levels normally found in the Parental lines (Dark Red and Blue Lines). (B) Overall numbers of genes promoted or repressed by enzalutamide treatment (light blue, and light red, respectively), then restored to the levels normally found in the respective parental lines (dark blue or dark red, respectively). (C) Venn Diagrams illustrating the overlap analysis of the “restored” gene subsets. Only one gene, TMEFF2 (Transmembrane protein with EGF-like and two follistatin-like domains-2), is a common gene restored across all cell lines. (D) Ingenuity Pathway Analysis^® (IPA) performed on the “restored” gene subsets that are shared by at least two of the four lines prioritizes both the AR and β-Catenin (CTNNB1) pathways (both AR and CTNNB1 genes not altered by mRNA but imputed by IPA). Proteins outlined in bold red represent genes highly expressed, while genes shown in bold green have lower expression.

Figure 6. Analysis of adaptive altered non-AR-associated gene expression in Enz^R cells

(A) Schematic illustrating how novel and non-AR-pathway associated genes that promote enzalutamide resistance were identified. Non-AR-pathway associated genes altered in Enz^R cells were genes that did not change significantly with short term AR antagonism (48 hrs), but were significantly changed upon acquisition of resistance. The green line represents these novel genes that are increased during enzalutamide resistance (Activated), and the gray line represents those genes that are decreased during enzalutamide resistance (Repressed). (B) Overall numbers of genes Non-AR-Associated genes identified in each cell line (Center Number). Subsets of the numbers in green represent the novel genes that are upregulated during enzalutamide resistance (Activated), and the numbers in gray represent those genes that are downregulated during enzalutamide resistance (Repressed). (C) Venn Diagrams illustrating the overlap analysis of the novel non-AR gene subsets in each of the four cell lines. Only one gene is commonly altered restored across all four cancer cell lines: MT2A (Metallothionein 2A). (D) Ingenuity Pathway Analysis^® (IPA) of the non-AR associated resistance genes. IPA was performed on both the genes that shared by at least two of the four lines prioritizes both microRNA 16 (miR-16) and Oncostatin M (OSM) pathways. Proteins outlined in red represent genes highly expressed, while genes shown in green have lower expression. Blue arrows represent a positive regulation, which red lines represent a negative regulation, dashed lines represent an indirect interaction.