Androgen Receptor, Although Not a Specific Marker For, Is a Novel Target to Suppress Glioma Stem Cells as a Therapeutic Strategy for Glioblastoma

Abstract

Targeting androgen receptor (AR) has been shown to be promising in treating glioblastoma (GBM) in cell culture and flank implant models but the mechanisms remain unclear. AR antagonists including enzalutamide are available for treating prostate cancer patients in clinic and can pass the blood-brain barrier, thus are potentially good candidates for GBM treatment but have not been tested in GBM orthotopically. Our current studies confirmed that in patients, a majority of GBM tumors overexpress AR in both genders. Enzalutamide inhibited the proliferation of GBM cells both in vitro and in vivo. Although confocal microscopy demonstrated that AR is expressed but not specifically in glioma cancer stem cells (CSCs) (CD133+), enzalutamide treatment significantly decreased CSC population in cultured monolayer cells and spheroids, suppressed tumor sphere-forming capacity of GBM cells, and downregulated CSC gene expression at mRNA and protein levels in a dose- and time-dependent manner. We have, for the first time, demonstrated that enzalutamide treatment decreased the density of CSCs in vivo and improved survival in an orthotopic GBM mouse model. We conclude that AR antagonists potently target glioma CSCs in addition to suppressing the overall proliferation of GBM cells as a mechanism supporting their repurposing for clinical applications treating GBM.

Keywords: AR antagonist; androgen receptor; cancer stem cells; enzalutamide; glioblastoma.

Figure 1

Histological studies of androgen receptor (AR) expression in human brain/GBM tissue. (A–C) H&E staining of a GBM slide from a male patient (40×, 100× and 400× magnification, respectively). (D–F) Positive AR expression (brown) in a GBM slide from the same patient as A (40×, 100× and 400× magnification, respectively). (G–I) H&E staining of a GBM slide from a female patient (40×, 100× and 400× magnification, respectively). (J–L) AR staining in a GBM slide from the same patient as G (40×, 100× and 400× magnification, respectively). Enriched AR positive cells at peri-vascular area are best shown in K (arrow). Majority of AR staining (brown) is in nuclei in a subset of the cells (arrow in L). (M) H&E staining of a GBM specimen showing the endothelial proliferation of the vessels (arrow). (P) AR staining of the GBM specimen from the same patient as M. (N) H&E staining of a normal human brain autopsy specimen. (Q) Negative AR staining of the normal brain autopsy specimen from the same patient as N, at 400× magnification. (O) H&E staining and (R) scattered weakly positive nuclear staining of AR in a temporal lobectomy surgical specimen from a patient with epilepsy, at 400× magnification.

Figure 2

AR antagonists inhibit proliferation of human and mouse GBM cell lines in vitro. (A–D) GBM cell lines were treated with indicated concentrations of enzalutamide for 2 days before cell titer blue assays were performed. (A) U87MG cell line (human) (IC50 = 41 μM). (B) U138MG cell line (human) (IC50 = 45 μM); (C) Ln229 cell line (human) (IC50 = 41 μM). (D) MGPP3 cell line (mouse) (IC50 = 56 μM); (E–H) GBM cell lines were treated with indicated concentrations of bicalutamide for 2 days before cell titer blue assays. (E) U87MG cell line (IC50 = 121 μM). (F) U138MG cell line (IC50 = 122 μM). (G) Ln229 cell line (IC50 = 72 μM). (H) MGPP3 cell line (IC50 = 42 μM). (I) Western blotting assays demonstrate that AR and c-Myc protein levels decrease in the U87MG, MGPP3 and LnCap cells (a prostate cancer cell line for positive control) after the treatment of enzalutamide (40 µM) time dependently. 8, 24, 48 h: cells were treated with the drug for 8, 24 or 48 h. All experiments were performed with three independent replicates.

Figure 3

Confocal immunofluorescence studies demonstrate that enzalutamide decreases AR (red) and c-Myc (green) in GBM cells cultured in vitro (U87MG and MGPP3). (A) U87MG cells were treated with DMSO (negative control) or various concentrations of enzalutamide (20 µM or 40 µM). (B) MGPP3 cells were treated with DMSO or various concentrations of enzalutamide (20 µM or 40 µM). (C) LnCap (prostate cancer cell line) served as a positive control. (D) Quantification of the immunofluorescence signals of AR and c-Myc in U87MG and MGPP3 cell lines after enzalutamide treatment (20 µM or 40 µM) for 24 h. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 4

AR antagonists decrease the cancer stem cell population in U87MG GBM cells in vitro. U87MG spheroids were cultured in low serum medium (0.5%) for 2 days and then treated with DMSO (negative control), enzalutamide (120 µM), or bicalutamide (180 µM) for another 4 days with diameters of spheroids measured daily. (A) Representative figures of the U87MG spheroids before and after the treatment of enzalutamide (120 µM) or bicalutamide (60 µM). (B) Measured diameters of the spheroids normalized to before the treatment during drug treatment with increasing concentrations of enzalutamide (40, 120, and 240 µM) or bicalutamide (60, 180, and 360 µM). (C) Representative flow cytometry gating strategy with dissociated cells from U87MG spheroids. Upper left: no drug treatment and no CD133 antibody staining; upper right: DMSO only (no drug treatment control); lower left: enzalutamide (120 µM); lower right: bicalutamide (180 µM). (D) CSC populations (CD133+) were compared in dissociated cells from spheroids with or without treatment of AR antagonists using flow cytometry as shown in (C) Eight replicates per group were used for flow cytometry with mean values presented in the histogram. (E) Western blotting assays showed decreased expression levels of CSC markers, Nanog and Oct4, after enzalutamide treatment for 24 and 72 h. (F) Quantification of Nanog and Oct4 protein levels before and after the treatment of enzalutamide for 24 and 72 h. All experiments were performed with three independent replicates. (G) Extreme Limiting Dilution Assays (ELDA) in vitro to estimate the frequency of tumor sphere-forming stem-like cells in U87MG and MGPP3 cell lines. Plot of the log (e) fraction of wells without tumor spheres as a function of plated U87MG/MGPP3 cell number/well. The more vertical the line, the higher the percentage of tumor sphere-forming cells or CSCs. (H) Extreme Limiting Dilution Assays (ELDA) in vivo to estimate the frequency of CSCs. Plot of the log (e) fraction of mice without tumor growth as a function of implanted MGPP3 cells. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 5

AR positively correlates with CSC marker genes at mRNA expression level. (A) Gene Set Enrichment Analyses (GSEA) against the KEGG pathway database showing the top 15 pathways most significantly regulated in the U87MG cells with differentially expressed genes (DEGs) after the treatment of enzalutamide for 48 h. (B) The positive correlations of mRNA expressions between CSC marker genes and AR from in vivo RNA-seq results in GBM patients from TCGA database. GAPDH, a housekeeping gene, showed no correlation with AR (R = -0.03). (C) mRNA expression levels of various CSC marker genes in U87MG cells cultured in vitro before and after the treatment of enzalutamide (80 µM) from RNA-seq results from our laboratory. CD133 is an alias name for Prom1. (D) Quantitative RT-PCR results of the representative CSC genes after the treatment of enzalutamide. (E) Heatmap of the genes specific for cancer stem cell, TGF-β signaling pathway, cell cycle, and cell proliferation responding to enzalutamide treatment in U87MG cells at different time points. The scale bar at the bottom represents the normalized read counts of the genes from RNA-seq results. *p < 0.05.

Figure 6

Enzalutamide suppresses CSC marker gene expression, GBM tumor growth in vivo and prolongs overall survival in treated mice with MGPP3 cells implanted in brain. (A) Fold changes of the bioluminescent signals of the tumors in brain after the treatment of vehicle only or enzalutamide (20 mg/kg, three times per week, IP). (B) Representative IVIS images of the tumor growth in mouse brain after vehicle only or enzalutamide treatment. Wk 0, 3 and 5: IVIS imaging taken prior to (week 0), 3 and 5 weeks after initiating of drug injection. Mann–Whitney U tests were performed in both (A, B) to compare between groups. (C) Weight changes of the mice after the treatment. (D) Overall survival was significantly improved in mice treated with enzalutamide comparing with vehicle only. (E) Representative images of the mouse brain tissue after IHC staining for CD133 (100×), c-Myc (200×), and AR (200×). Corner image (left upper corner of the left panel): high magnification image showing tumor cells with positive AR staining with nuclear localization without enzalutamide treatment. Corner image (left upper corner of the right panel): high magnification image showing tumor cells with positive AR staining but more cytosol distribution after enzalutamide treatment. (F) Quantifications of positive cells in the brain tumors from IHC staining after vehicle only or enzalutamide treatment. Two-tailed student t tests were performed for statistical comparisons. *p < 0.05; **p < 0.01.

Figure 7

Immunofluorescence confocal microscopy on mouse GBM FFPE specimen (5 µm thickness) showing the subcellular localizations of AR (green), CD133 (pink) and DAPI-stained nuclei (blue). (A) Low magnification images (10×) and (B) high magnification images (63×) which were scanned from areas in the dash line boxes in (A) demonstrated co-expression of CD133 in AR-expressing (AR+) cells with high staining intensity (arrowheads). AR expression can be seen in CD133 negative cells but with low AR staining intensity (thin arrows). (C) High magnification images (63×) demonstrated co-expression of Nanog in AR+ cells with high AR staining intensity (arrowheads). Cells with negative Nanog staining could be AR+ but showed weaker AR staining intensities than Nanog+ cells (thin arrows). (D) High magnification images (63×) with slides stained with secondary antibodies only as negative controls. (E) Schematic figure showing the proposed mechanisms of AR antagonists inhibiting both cancer stem cells maintenance and cancer proliferation from more differentiated cells. CSC, cancer stem cell; PDCC, partially differentiated cancer cell; DCC, differentiated cancer cell.