IκBα mediates prostate cancer cell death induced by combinatorial targeting of the androgen receptor

Abstract

Background: Combining different clinical agents to target multiple pathways in prostate cancer cells, including androgen receptor (AR) signaling, is potentially an effective strategy to improve outcomes for men with metastatic disease. We have previously demonstrated that sub-effective concentrations of an AR antagonist, bicalutamide, and the histone deacetylase inhibitor, vorinostat, act synergistically when combined to cause death of AR-dependent prostate cancer cells.

Methods: In this study, expression profiling of human prostate cancer cells treated with bicalutamide or vorinostat, alone or in combination, was employed to determine the molecular mechanisms underlying this synergistic action. Cell viability assays and quantitative real time PCR were used to validate identified candidate genes.

Results: A substantial proportion of the genes modulated by the combination of bicalutamide and vorinostat were androgen regulated. Independent pathway analysis identified further pathways and genes, most notably NFKBIA (encoding IκBα, an inhibitor of NF-κB and p53 signaling), as targets of this combinatorial treatment. Depletion of IκBα by siRNA knockdown enhanced apoptosis of prostate cancer cells, while ectopic overexpression of IκBα markedly suppressed cell death induced by the combination of bicalutamide and vorinostat.

Conclusion: These findings implicate IκBα as a key mediator of the apoptotic action of this combinatorial AR targeting strategy and a promising new therapeutic target for prostate cancer.

Fig. 1

Characterization of cell death caused by the combination of bicalutamide and vorinostat. LNCaP cells (2 × 10⁴ cells per well in 24-well plates) were cultured in triplicate wells with either vehicle control [VEH] bicalutamide [BIC] (2.5 μM) or vorinostat [VOR] (1 μM), individually and in combination [BIC + VOR], in RPMI 1640 supplemented with 10 % FBS. VCaP cells (5 × 10⁴ cells per well in 24-well plates) were cultured in triplicate wells with bicalutamide (1.25 μM) or vorinostat (0.5 μM), individually and in combination, in DMEM supplemented with sodium pyruvate, non-essential amino acids and 10 % FBS. a LNCaP cells were counted at 2, 4 and 6 days of culture with bicalutamide and vorinostat, while VCaPs were counted at 4, 6 and 8 days of culture. b LNCaP cells were cultured with treatment one (1.) for 24 h, which was removed and replaced with treatment two (2.) for 48 h. (C) LNCaP cells were pre-treated for 1 h with cycloheximide (10 μM), and then cultured for 3 d with indicated treatments (HIGH BIC = 50 μM, HIGH VOR = 7.5 μM). d LNCaP cells were cultured with indicated treatments, and at 1, 2, 4, 6, 8, 16, or 24 h treatment medium was removed and replaced with normal culture medium not containing either agent until a total of 72 h of culture. At each end point, cells were counted using a haemocytometer, and assessed for viability using trypan blue dye exclusion. Cell death is expressed as a percentage of total cell number. Values indicated are the mean of triplicate wells ± SEM, and are representative of three independent experiments. * = p < 0.05 using one-way ANOVA with Bonferroni post-hoc test

Fig. 2

Summary of genes expression changes after treatment with bicalutamide, vorinostat, and the combination. LNCaP cells were treated with vehicle control, bicalutamide, vorinostat and bicalutamide and vorinostat in combination, and RNA was extracted at 6 h of treatment. Microarray analysis was performed on Affymetrix Human GeneChip 1.0 ST Arrays, with six biological replicates per sample. a Venn diagram of genes with mRNA levels significantly changed by vorinostat (light grey), bicalutamide (dark grey) or the doses of vorinostat and bicalutamide in combination (white), when compared with vehicle control. Circles represent both significantly up-regulated and significantly down-regulated genes over vehicle control, and sizes are proportional to the number of genes. b Numbers of genes significantly up-regulated and significantly down-regulated by each treatment when compared with vehicle control

Fig. 3

Changes in gene expression induced by bicalutamide and vorinostat in prostate cancer cells are inversely related to androgen-induced gene profiles. a Venn diagram of the overlap between genes significantly changed by the combination of bicalutamide and vorinostat compared with vehicle control (white) and genes significantly changed by treatment with 100 nM DHT for 16 h from the Wang [33] dataset (grey). Circles represent both up-regulated and down-regulated genes and are proportional to the number of genes. b Genes regulated by treatment with the combination of bicalutamide and vorinostat (compared with vehicle control) are negatively correlated with genes induced by DHT in prostate cancer cells, as assessed by GSEA. Sextuplicate sets of expression profiles were compared between LNCaP cells treated with vehicle control, vorinostat, bicalutamide or the combination of vorinostat and bicalutamide. Probe sets in the data were collapsed to gene level, assigned a score based on a signal-to-noise ratio algorithm and rank-ordered by this score. DHT-induced genes in the ordered data set are shown as black lines (middle), and the running enrichment score is plotted (bottom). The change in expression of each gene in response to the combination treatment is shown as a heatmap (top). c Flowchart of the refinement process involved in determining genes involved in the cell death caused by the combination. The refined list included the 216 genes that were significantly altered (p < 0.05; ANOVA with FDR adjustment) by the combination when compared to both of the individual agents. d Venn diagram of genes with mRNA levels significantly changed by the combination of vorinostat and bicalutamide [BIC + VOR vs individual agents] when compared with vehicle control, bicalutamide and vorinostat individually (white) and genes regulated by 100 nM DHT from the Wang [33] dataset (grey). Circles represent both up-regulated and down-regulated genes and are proportional to the number of genes

Fig. 4

Quantitative analysis of androgen regulated genes altered by combinatorial AR targeting. a Quantitative real-time PCR analysis of an independent RNA sample set, generated by treatment of LNCaP cells with vehicle control [VEH], 1 μM vorinostat [VOR], 2.5 μM bicalutamide [BIC] or the combination of 1 μM vorinostat and 2.5 μM bicalutamide [BIC + VOR] in triplicate for 6 h. The expression of KLK2, KLK3, NKX3-1, IGF1R, NFKBIA and C1orf116 was normalised to GUSB and HPRT1. Fold changes are expressed relative to vehicle control. Values indicated are the mean of technical and biological replicates ± SEM, and are representative of three independent experiments. * = p < 0.05 using one-way ANOVA with Bonferroni post-hoc test, compared with vehicle control. # = p < 0.05 using t-test compared with vorinostat, and ^ = p < 0.05 using t-test compared with bicalutamide. b Western blot analysis of lysates from LNCaP cells cultured with either vehicle control [VEH], 1 μM vorinostat [VOR], 2.5 μM bicalutamide [BIC] or the combination of 1 μM vorinostat and 2.5 μM bicalutamide [BIC+VOR] for 12, 48, or 96 h. Steady state levels of PSA (KLK3) are shown, and hsp90 is shown as a loading control

Fig. 5

The effect of combinatorial AR targeting on NFKBIA levels and p53 signaling. a The expression of NFKBIA over a timecourse was normalized to GUSB and HPRT1. Relative expression is the fold change expressed relative to vehicle control. Values indicated are the mean of technical and biological replicates ± SEM, and are representative of three independent experiments. * = p < 0.05 using one-way ANOVA with Bonferroni post-hoc test, compared with bicalutamide or vorinostat individually. b Western blot analysis of lysates from LNCaP cells cultured with either vehicle control [VEH], 1 μM vorinostat [VOR], 2.5 μM bicalutamide [BIC] or the combination of 1 μM vorinostat and 2.5 μM bicalutamide [BIC+VOR] for 12, 48, or 96 h. Steady state levels of IκBα are shown, and tubulin is shown as a loading control. c mRNA expression levels of TP53INP1 and CDKN1A are shown as for NFKBIA. * = p < 0.05 using one-way ANOVA with Bonferroni post-hoc test

Fig. 6

NFKBIA is a critical mediator of death induced by combinatorial AR targeting. LNCaP cells (1 × 10⁵ cells per well in 12-well plates) or VCaP cells (2 × 10⁵ cells per well in 12-well plates) were transfected in triplicate wells with 10 nM of either non-specific, scrambled siRNA [N.S.] or specific siRNAs [siRNA #1 and siRNA #4] targeting NFKBIA. LNCaP cells were transfected in triplicate wells with 1.5 μg of plasmid expressing either GFP or NFKBIA+GFP, for 4 h. The transfection medium was then overlaid with vehicle control [VEH], or the combination [BIC + VOR] to give a final concentration of 2.5 μM bicalutamide and 1 μM vorinostat in either PRF-RPMI 1640 supplemented with 10 % FBS and 1 % l-glutamine (LNCaP) or DMEM supplemented with non-essential amino acids, 1 % l-glutamine and 10 % FBS (VCaP). a Western blot analysis of lysates from LNCaP and VCaP cells transfected with either non-specific [N.S.] or specific [#1 and #4] NFKBIA siRNA. Steady state levels of IκBα at day three (LNCaP) or day six (VCaP) of treatment are shown, with hsp90 as a loading control. b Cells were counted at 3 days (LNCaP) or 6 days (VCaP) of treatment, and assessed for viability using trypan blue dye exclusion. Cell death is expressed as a percentage of total cell number. Values indicated are the mean of triplicate wells ± SEM, and are representative of three independent experiments. c Western blot analysis of lysates from LNCaP cells transfected with plasmids expressing GFP only [GFP] or NFKBIA and GFP [NFKBIA], and then treated with vehicle control [VC] or combination [BIC+VOR]. Steady state levels of IκBα at day three of treatment are shown, with hsp90 as a loading control. d Cells were counted at 3 days of treatment, and assessed for viability using trypan blue dye exclusion. Cell death is expressed as a percentage of total cell number. Values indicated are the mean of triplicate wells ± SEM, and are representative of three independent experiments. * = p < 0.05 using t-test