Glucocorticoid receptor confers resistance to antiandrogens by bypassing androgen receptor blockade

Abstract

The treatment of advanced prostate cancer has been transformed by novel antiandrogen therapies such as enzalutamide. Here, we identify induction of glucocorticoid receptor (GR) expression as a common feature of drug-resistant tumors in a credentialed preclinical model, a finding also confirmed in patient samples. GR substituted for the androgen receptor (AR) to activate a similar but distinguishable set of target genes and was necessary for maintenance of the resistant phenotype. The GR agonist dexamethasone was sufficient to confer enzalutamide resistance, whereas a GR antagonist restored sensitivity. Acute AR inhibition resulted in GR upregulation in a subset of prostate cancer cells due to relief of AR-mediated feedback repression of GR expression. These findings establish a mechanism of escape from AR blockade through expansion of cells primed to drive AR target genes via an alternative nuclear receptor upon drug exposure.

Figure 1. GR mRNA and protein is expressed in resistant tissues

A. Most differentially expressed genes in a pilot cohort of LnCaP/AR xenograft tumors with acquired resistance to ARN-509 (n=6) or RD162 (n=9) compared to control (n=3) determined by microarray (Affymetrix Ex1.0). Mice with resistant tissues were continued on drug treatment through time of harvest. In vitro androgen-induced or -repressed genes are annotated (See also Supplementary Table 1B). B. Mean tumor volumes +/− s.e.m of LnCaP/AR xenografts in validation cohort. Days tumors were harvested are annotated on x-axis (long hash mark). C. RT-qPCR analysis of GR and AR mRNA expression in a validation cohort of LnCaP/AR xenograft tumors from mice treated with vehicle (control, n=10), 4 days of anti-androgen (n=8), or with acquired resistance to 10 mg/kg enzalutamide (n=8) or 10 mg/kg ARN-509 (n=8). See also Supplementary Table 1B. D. Western blot analysis of GR and AR protein expression in a subset of tissues also analyzed in B. Control (n=6), 4 day (n=5), Resistant (n=13). Resistant samples were loaded for protein analysis from highest to lowest GR levels based on corresponding mRNA analysis (See also Supplementary Table 1C.) E. Intracellular GR flow cytometric analysis of LnCaP/AR, CS1, and LREX’, cells passaged in vitro, under standard passage conditions (see methods). See also Figure S1.

Figure 2. GR is necessary for resistance in the LREX’ xenograft model

A. Mean tumor volume +/− s.e.m. of LREX’ (n=20) or LnCaP/AR (n=14) cells in castrate mice treated with 10 mg/kg enzalutamide B. Mean tumor volumes +/− s.e.m. of CS1 in castrate mice treated with vehicle (n=10) or 10 mg/kg ARN-509 (n=10). C. GR IHC of enzalutamide (10 mg/kg)-treated LREX’ tumors and vehicle-treated LnCaP/AR xenograft tissues. Blue arrow = endothelial/stromal cells, Black arrow = epithelial cell. D. Mean tumor volumes +/− s.e.m of LREX’ xenografts in 10 mg/kg enzalutamide-treated castrate mice after infection with a non-targeting (n=14) or GR-targeting (n=12) hairpin. Comparison is by Mann-Whitney test. E. Tumor growth curve of CS1 in castrate mice after infection with the non-targeting (n=20) or GR-targeting (n=20) hairpin. F. Western blot analysis of GR expression in LREX’ cells prior to implantation and of available tissues from D at day 49. See also Figure S1.

Figure 3. GR induction in disseminated tumor cells is associated with poor clinical response to enzalutamide and persistence of PSA

A. Schematic of sample acquisition timeline and response groups. B. Number of good or poor responders who achieved PSA decline greater than 50%. C. Examples of GR IHC images from matched samples at baseline and 8 weeks. D. Percent GR positive epithelial cells in all tissue available at 0 and 8 weeks or E. matched samples obtained from the same patient at 0 and 8 weeks +/− s.e.m. Comparisons are by Mann-Whitney test. See also Figure S2.

Figure 4. Variable expression of AR target genes in LREX’, in vivo, and after glucocorticoid treatment, in vitro

A. Normalized expression array signal (Illumina HT-12) of a suite of 74 AR target genes in control (n=10), 4 day (n=8), and LREX’ (n=8, right) xenograft tumors. Genes are ranked by degree of restoration of expression in resistant tissue ((Res-4 day) / (Control-4 day)). All resistant tissues were continued on anti-androgen treatment through time of harvest. B. Fractional restoration values of each of the 74 AR targets in LREX’ xenografts (n=8) or resistant tissues from the validation cohort (n=12, see also Figure S3). C. GR mRNA in resistant tissues used in B. D. Expression of AR target genes in the LREX’ cell line in steroid depleted media after 8 hours of treatment with the indicated agonists, in vitro. Enzalutamide = 10 micromolar, V = Vehicle. +/− s.e.m. See also Figures S3, S4.

Figure 5. Comparative AR and GR transcriptome and cistrome analysis in LREX’

A. Venn diagram of AR and GR signature gene lists. AR or GR signatures were defined as all genes showing >1.6 (or <−1.6) fold change (FDR <. 05) after 8 hours of addition of DHT (1nM) or Dex (100nM) to charcoal stripped media, respectively. B. Heat map depiction of expression changes of AR signature genes (left) or GR signature genes (right) associated with the indicated treatment. Enzalutamide = 10 micromolar. C. Expression of AR- or GR-induced signature genes (as defined in A.) were compared in DHT (1nM) or Dex (100nM) treated samples. GR signature genes that also had higher expression in Dex samples (>1.1 fold, FDR <.05) were designated as GR-selective (n=67) and AR signature genes that showed higher expression in DHT samples (>1.1 fold, FDR <.05) were designated as AR-selective (n=39). D. Expression of AR- and GR-selective genes in LREX’ and control tumors in vivo compared by GSEA. E. AR cistrome defined by AR ChIP-seq after DHT (1nM) treatment of LREX’ in vitro in charcoal stripped media. Percent of AR defined peaks that overlap with GR peaks found by GR ChIP-seq after Dex (100nm) treatment of LREX’ in vitro are shown in pie graph. Top binding motifs in AR-unique and AR/GR overlap peaks are indicated below. F. GR cistrome defined by GR ChIP-seq after Dex treatment of LREX’ in vitro in charcoal stripped media. Percent of GR peaks that overlap with AR peaks found by AR ChIP-seq after DHT (1 nM) treatment of LREX’ in vitro are shown in pie graph. Top binding motifs in GR-unique and AR/GR overlap peaks are indicated below. See also Figure S5.

Figure 6. GR activity is sufficient to confer enzalutamide resistance in VCaP

FOR ALL PANELS: VCaP cells do not tolerate charcoal stripped media and were cultured in standard culture conditions (fetal bovine serum with endogenous hormones). Enz=10 micromolar, Dex = 100 nM, CMP 15 = 1 micromolar. A. Western blot analysis of prostate cancer cell lines. B, C and D. Cell viability assessed by CellTiter-Glo (Promega) assay and normalized to day 1 value after indicated treatments +/− s.e.m. E. Confirmation of GR knock-down by western blot after infection with GR targeting shRNA. F. Apoptosis as assessed by cPARP western blot after 3 days of indicated treatment. G. A suite of AR targets relevant to VCaP was defined (see methods) and normalized expression of each gene after 24 hours of indicated drug treatments is depicted by heat map and ranked by degree of induction with Dex. H. Expression of the top two genes from B. (KLK2 and FKBP5) after 24 hours of indicated treatments +/− s.e.m. See also Figure S6.

Figure 7. Resistant cells are primed for GR induction upon AR inhibition

A. GR mRNA in LREX’ xenografts. Tumors were injected into castrated mice and immediately treated with 10 mg/kg enzalutamide (n=20) for 7 weeks . Half of the mice were then continued on 10 mg/kg enzalutamide (n=10) or discontinued for 8 days (n=10). B. LREX’ are maintained in vitro in the presence of enzalutamide 1 micromolar. GR mRNA was assessed in LREX’ cell line after passage for indicate number of days in standard fetal bovine serum containing media without enzalutamide. C. GR mRNA in LREX’ cultured in charcoal stripped media for 48 hours and then treated for 8 hours with vehicle or DHT with or without 10 micromolar enzalutamide. D. AR ChIP-qPCR with LREX’ cultured in charcoal stripped media and then treated for 1 hour with DHT (1nM) or Dex (100nM) at an intronic enhancer site +/− s.d. E. Intracellular GR flow cytometric analysis of indicated cells at indicated times points. AUC = area under curve. Enzalutamide = 1 micromolar F. Plotted median fluorescence (minus background) values from E and Figure S7C. For both LREX plots, R² values for non-linear regression analysis is >.98. G. Model of GR induction in resistant tissues. See also Figure S7.