Feedback activation of NF-KB signaling leads to adaptive resistance to EZH2 inhibitors in prostate cancer cells

Abstract

Background: Prostate cancer (PCa) is the most common malignant tumor in developed countries, which has seriously threatened men's lifestyle and quality of life. The up-regulation of EZH2 is associated with advanced PCa and poor prognosis, making it a promising therapeutic target. However, the EZH2 inhibitors-based treatment is basically ineffective against PCa, which limits its clinical application.

Methods: Microarray data (GSE107779) from LNCaP cells treated with either siRNA against EZH2 or a EZH2 inhibitor EPZ6438 was analyzed by Limma R package. Western blot, real-time PCR and luciferase reporter assays were used to determine the EZH2-SOX9-TNFRSF11A axis and the activity of NF-κB signaling in PCa cells. CCK-8 assay was used to determine the viability of PCa cells following various treatments.

Results: Genetic ablation or pharmacological inhibition of EZH2 leads to feedback activation of NF-κB signaling in PCa cells. EZH2-dependent SOX9 expression regulates the activation of NF-κB signaling. TNFRSF11A, also known as receptor activator of NF-κB (RANK), is a downstream target of SOX9 in PCa cells. SOX9 recognizes two putative SOX9 response elements in the promoter region of TNFRSF11A gene to drive TNFRSF11A expression and downstream NF-κB signaling activation. Suppression of the NF-κB signaling by either TNFRSF11A silencing or BAY11-7082 treatment rendered PCa cells to EZH2 inhibitors.

Conclusion: Collectively, our finding reveals a EZH2-SOX9-TNFRSF11A axis in the regulation of activity of NF-κB signaling in PCa cells and suggests that a combination of EZH2 inhibitors and BAY11-7082 would be an effective approach for the treatment of PCa patients in the future.

Keywords: EZH2; NF-κB; Prostate cancer (PCa); SOX9; TNFRSF11A (RANK).

Fig. 1

In silico identification of EZH2-regulated genes in PCa cells. a The microarray data from LNCaP cells treated with DMSO or EPZ6438 was used to examine the differentially expressed mRNAs. 887 differentially expressed genes including 437 up-regulated and 450 down-regulated were identified by the Limma R package and selected to draw a volcano plot. b The microarray data from LNCaP cells transfected with con-siRNA or EZH2-siRNA was used to examine the differentially expressed mRNAs. 289 differentially expressed genes including 193 up-regulated and 96 down-regulated were identified by the Limma R package and selected to draw a volcano plot. c The intersection of the common up-regulated genes in LNCaP cells after EZH2 inhibition or knockdown identified 35 potential substrates of EZH2. d The KEGG pathway analysis of these potential EZH2 substrates

Fig. 2

Genetic ablation or pharmacological inhibition of EZH2 leads to feedback activation of the NF-κB signaling in PCa cells. a LNCaP cells were transfected with pGL3-NFkB-Luc and pSV40-renilla plasmids for 12 h and then treated with indicated dose of EPZ6438 for additional 24 h. The luciferase activity was then measured. **P < 0.01, ***P < 0.001. b Relative mRNA expression levels of the two NF-κB downstream target genes c-Myc and Cyclin D1 in 10 μM EPZ6438 treated LNCaP cells were determined by real-time PCR assay. ***P < 0.001. c Relative mRNA expression levels of the two NF-κB downstream inflammatory cytokines IL1B and TNFA in 10 μM EPZ6438 treated LNCaP cells were determined by real-time PCR assay. ***P < 0.001. d. LNCaP cells were transfected with indicated siRNAs for 12 h, and then co-transfected with pGL3-NFkB-Luc and pSV40-renilla plasmids for 24 h. The luciferase activity was then measured. **P < 0.01, ***P < 0.001. e LNCaP cells were transfected with indicated siRNAs for 36 h, the relative mRNA expression levels of EZH2, c-Myc and Cyclin D1 were determined by real-time PCR assay. *P < 0.05, **P < 0.01, ***P < 0.001. f LNCaP cells were transfected with indicated siRNAs for 36 h, the relative mRNA expression levels of IL1B and TNFA were determined by real-time PCR assay. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 3

Suppression of the NF-κB signaling pathway reduces the proliferation and viability of PCa cells treated with EZH2 inhibitors. a LNCaP cells were transfected with pGL3-NFkB-Luc and pSV40-renilla plasmids for 12 h and then treated with either 10 μM EPZ6438 and/or 10 μM BAY11-7082 for additional 24 h. The luciferase activity was then measured. **P < 0.01. b The cell growth curve of LNCaP cells treated with or without either 5 μM EPZ6438 and/or 5 μM BAY11-7082 for the indicated days. ***P < 0.001. c LNCaP cells were treated with or without either 10 μM EPZ6438 and/or 10 μM BAY11-7082 for 24 h. CCK-8 was then added and the absorbance at A450 was determined by a microplate reader. *** P < 0.001. d PCa cell lines PC3, 22RV1 and Du-145 were treated with or without either 10 μM GSK126/EPZ6438 and/or 10 μM BAY11-7082 for 24 h. CCK-8 was then added and the absorbance at A450 was determined by a microplate reader. ***P < 0.001. e LNCaP cells were transfected with indicated siRNAs for 24 h and then treated with or without 10 μM BAY11-7082 for additional 24 h. The luciferase activity was then measured. ***P < 0.001

Fig. 4

EZH2-dependent SOX9 expression regulates the activation of NF-κB signaling. a LNCaP cells were transfected with indicated siRNAs for 36 h, the relative mRNA expression levels of SOX9 were determined by real-time PCR assay. *P < 0.05, **P < 0.01. b LNCaP cells were treated with 10 μM EPZ6438 or GSK126 for 24 h. The relative mRNA levels of SOX9 were then determined. *P < 0.05. c LNCaP cells were transfected with indicated siRNAs for 12 h, then co-transfected with pGL3-NFkB-Luc and pSV40-renilla plasmids for 24 h. The luciferase activity was then measured. **P < 0.01. d LNCaP cells were co-transfected with indicated plasmids for 24 h. The luciferase activity was then measured. **P < 0.01. e LNCaP cells were transfected with indicated siRNAs for 24 h and then treated with or without 10 μM EPZ6438 or GSK126 for additional 24 h. The luciferase activity was then measured. **P < 0.05. f LNCaP cells were transfected with indicated siRNAs for 24 h and then treated with or without 10 μM EPZ6438 for additional 24 h. CCK-8 was then added and the absorbance at A450 was determined by a microplate reader. **P < 0.01

Fig. 5

SOX9 regulates the expression of TNFRSF11A. a The microarray data from VCAP cells transfected with con-siRNA or SOX9-siRNA was used to examine the differentially expressed mRNAs. 1118 differentially expressed genes including 409 up-regulated and 709 down-regulated were identified by the Limma R package and the top 200 differentially expressed mRNAs were selected to draw a heatmap. b The intersection of the common genes between 709 down-regulated gene caused by SOX9 downregulation and 100 SOX9 co-expression genes identified 4 putative substrates of SOX9. c The correlation of the mRNAs of SOX9 and TNFRSF11A in 492 prostate cancer tissues from the TCGA database based on the GEPIA2 website (http://gepia2.cancer-pku.cn). d LNCaP cells were transfected with indicated siRNAs for 36 h, the relative mRNA expression levels of SOX9 and TNFRSF11A were determined by real-time PCR assay. **P < 0.01, ***P < 0.001. e LNCaP cells were transfected with indicated siRNAs for 36 h, and then subjected to western blot with indicated antibodies. f LNCaP cells were transfected with vector control or Flag-SOX9 plasmids for 36 h, the relative mRNA expression levels of SOX9 and TNFRSF11A were determined by real-time PCR assay. ***P < 0.001. g LNCaP cells were transfected with vector control or Flag-SOX9 plasmids for 36 h, and then subjected to western blot with indicated antibodies. h Schematic diagram shows the putative SOX9 response elements from the human TNFRSF11A gene promoter. i The human TNFRSF11A promoter contains two SOX9 response elements. Point mutation was highlighted with black cross. SOX9 was co-transfected with indicated plasmids into LNCaP cells for 36 h. The luciferase activity was then measured. j ChIP assay shows enrichment of SOX9 at the human TNFRSF11A gene promoter in LNCaP cells. The human GAPDH promoter was served as a negative control. ***P < 0.001. The primer pairs for ChIP assay were also demonstrated in h

Fig. 6

Silencing of TNFRSF11A rendered PCa cells to EZH2 inhibitors treatment. a LNCaP cells were treated with 10 μM EPZ6438 or GSK126 for 24 h. The relative mRNA levels of TNFRSF11A were then determined. **P < 0.01, ***P < 0.001. b LNCaP cells were transfected with indicated siRNAs for 36 h, the relative mRNA expression levels of TNFRSF11A were determined. ***P < 0.001. c LNCaP cells were transfected with indicated siRNAs for 12 h, then co-transfected with pGL3-NFkB-Luc and pSV40-renilla plasmids for 12 h and treated with or with EPZ6438 for additional 24 h. The luciferase activity was then measured. ***P < 0.001. d LNCaP cells were transfected with indicated siRNAs for 12 h and treated with or with EPZ6438 for additional 24 h. CCK-8 was then added and the absorbance at A450 was determined by a microplate reader. **P < 0.01