ETS transcription factors control transcription of EZH2 and epigenetic silencing of the tumor suppressor gene Nkx3.1 in prostate cancer

Abstract

Background: ETS transcription factors regulate important signaling pathways involved in cell differentiation and development in many tissues and have emerged as important players in prostate cancer. However, the biological impact of ETS factors in prostate tumorigenesis is still debated.

Methodology/principal findings: We performed an analysis of the ETS gene family using microarray data and real-time PCR in normal and tumor tissues along with functional studies in normal and cancer cell lines to understand the impact in prostate tumorigenesis and identify key targets of these transcription factors. We found frequent dysregulation of ETS genes with oncogenic (i.e., ERG and ESE1) and tumor suppressor (i.e., ESE3) properties in prostate tumors compared to normal prostate. Tumor subgroups (i.e., ERG(high), ESE1(high), ESE3(low) and NoETS tumors) were identified on the basis of their ETS expression status and showed distinct transcriptional and biological features. ERG(high) and ESE3(low) tumors had the most robust gene signatures with both distinct and overlapping features. Integrating genomic data with functional studies in multiple cell lines, we demonstrated that ERG and ESE3 controlled in opposite direction transcription of the Polycomb Group protein EZH2, a key gene in development, differentiation, stem cell biology and tumorigenesis. We further demonstrated that the prostate-specific tumor suppressor gene Nkx3.1 was controlled by ERG and ESE3 both directly and through induction of EZH2.

Conclusions/significance: These findings provide new insights into the role of the ETS transcriptional network in prostate tumorigenesis and uncover previously unrecognized links between aberrant expression of ETS factors, deregulation of epigenetic effectors and silencing of tumor suppressor genes. The link between aberrant ETS activity and epigenetic gene silencing may be relevant for the clinical management of prostate cancer and design of new therapeutic strategies.

Figure 1. ETS gene signatures in prostate cancer.

A. Expression of ERG, ESE1 and ESE3 determined by qRT-PCR in normal prostate and prostate tumors. Tumors are grouped according to the predominantly expressed ETS factor. B. Principal component analysis. Dots represent individual samples with their location determined by the principal components of the transcriptome. C. Number of differentially expressed genes with Q≤0.1 in each ETS subgroup. D-E. Venn diagrams showing shared and distinct differentially expressed genes among the indicated tumor subgroups.

Figure 2. Functional analysis of the transcriptional programs of prostate tumor subgroups.

A. Number of unique, similar and common features among the differentially expressed genes compared to normal prostate in ERG^high, ESE3^low, ESE1^high and NoETS tumors according to Metacore. B. Commonly affected GeneGo Pathway Maps in the tumor subgroups. C. Differentially affected GeneGo Pathway Maps in ERG^high, ESE3^low, ESE1^high and NoETS tumors.

Figure 3. ERG induces EZH2 expression.

A. EZH2 expression in tissue samples. Microarray data are presented as log2 ratios compared to the reference. B. EZH2 expression in NoETS and ERG^high tumors in the Wallace et al. microarray dataset. C. EZH2 expression in LNCaP and 22Rv1 cells transiently transfected with empty (ev) or ERG expression (Erg) vector determined by RT-PCR (left) and Western blot (right). D. EZH2 level in control and stable ERG-expressing 22Rv1 and LNCaP cells determined by RT-PCR (upper) and Western blot (bottom). E. EZH2 level in VCaP cells transiently transfected with control and ERG-specific (siERG) siRNA determined by RT-PCR (left) and Western blot (right). F. ChIP in the indicated cell lines with ERG antibody and qPCR with primer sets encompassing the EBS in the EZH2 promoter. Positive (MMP3 promoter) and negative (ETS2) controls are shown in Fig. S6A-B and 7A, respectively. G Tissue specimens of ERG^high and NoETS tumors were subjected to ChIP with anti-ERG antibody and analyzed by qPCR. ETS2 was used as negative control (Fig. S7B). *, P<0.01; **, P<0.0001.

Figure 4. ERG represses Nkx3.1 expression.

A. Nkx3.1 expression in normal and tumor tissue samples. B. Nkx3.1 level in LNCaP cells transiently transfected with empty (ev) or ERG expression (Erg) vector determined by Western blot (upper panel), Nkx3.1 level in VCaP cells transiently transfected with siERG and control siRNA determined by RT-PCR and Western blot (middle panel), Nkx3.1 level in control and stable ERG-expressing 22Rv1 and LNCaP cells analyzed by RT-PCR and Western blot (bottom panel). C. Binding of ERG to the Nkx3.1 promoter determined by ChIP and qPCR in indicated cell lines. Negative controls are shown in Fig. S7A. D. Tissue specimens of ERG^high and NoETS tumor were subjected to ChIP with anti-ERG antibody and analyzed by qPCR. Negative controls are shown in Fig. S7B. E. VCaP, parental (control) and ERG expressing (ERG 18) LNCaP cells transfected with EZH2 specific (siEZH2) and control siRNA and analyzed by RT-PCR. F. ChIP with an antibody for methylated H3K27 and qPCR with primer sets encompassing the EBS (upper panel) and an androgen responsive enhancer (ARE) (bottom panel) in the Nkx3.1 gene. ETS2 was used as negative control (Fig. S8A). G. Nkx3.1 promoter activity in LNCaP cells transfected with human Nkx3.1 promoter reporter along with the indicated expression vectors. Luciferase reporter activity was measured after 24 h. H. Nkx3.1 protein level in LNCaP cells transiently transfected with empty (−) or either ERG (pERG) or EZH2 (pEZH2) expression vector determined by Western blot. I. Nkx3.1 promoter methylation was assessed by bisulfite-treated DNA sequencing. Empty and filled circles represent unmethylated and methylated CpG sites, respectively. *, p<0.01; **, p<0.001.

Figure 5. ESE-3 regulates EZH2 and Nkx3.1 expression.

A. EZH2 level in normal prostate, ESE3^low and NoETS tumors. B. EZH2 level in NoETS and ESE3^low tumors in Wallace et al. microarray dataset. C. EZH2 level in control (sh-) and stable ESE3 knock-down (sh 4, 6, 7) LNCaP clones determined by RT-PCR. D. EZH2 level in control (sh-) and ESE3 knock-down (sh17, 28) LHS clones determined by RT-PCR. E. Nkx3.1 level in normal prostate, ESE3^low and NoETS tumors. F. Nkx3.1 level in control (sh-) and ESE3-knock-down LNCaP and LHS cells determined by RT-PCR and Western blot (right bottom). G. Nkx3.1 promoter activity in PC3 cells transfected with human NKx3.1 promoter reporter along with the indicated expression vectors. Luciferase reporter activity was measured after 24 h. *p<0.01; **, p<0.001.

Figure 6. Silencing of Nkx3.1 is mediated by EZH2 in ESE3 knock-down cells.

A. RNA was extracted from control (sh-) and ESE3 knock-down (sh7) LNCaP cells 48 h post-transfection with siEZH2 and control siRNA and analyzed by RT-PCR. B. H3K27 methylation was assessed in sh- and sh7 LNCaP cells by ChIP and qPCR with primers encompassing the EBS and ARE in the Nkx3.1 gene. ETS2 was used as negative control (Fig. S8B) C. ESE3 binding to the EZH2 and Nkx3.1 promoter was assessed by ChIP and qPCR in sh- and sh7 LNCaP cells. ETS2 was used as negative control (Fig. S9A). D. ESE3 binding to the EZH2 and Nkx3.1 promoter in control and ERG-expressing (ERG-18) LNCaP cells assessed by ChIP and qPCR. ETS2 was used as negative control (Fig. S9B). *, p<0.01.

Figure 7. Effects of deregulated expression of ERG and ESE3 in prostate cancer cells.

A. ERG over-expression and ESE3 knock-down (B) enhance cell migration. Cells were grown until confluence and starved for 24 h when a scratch was performed on the monolayer. Pictures were taken at 0 and 72 h. Representative photographs of triplicate experiments are shown. C. Control and ERG over-expressing LNCaP cell clones were plated in polyhema coated 96-well plates and cell viability was measured using a colorimetric assay at the indicated time points. D. Control and ESE3 knock-down LNCaP cells were plated in polyhema and assayed as described above. E. VCaP cells were transfected with siRNAs against ERG or EZH2 and plated in polyhema after 24 h. Cell viability was measured as described above. F. Parental and ERG-expressing (ERG-18) LNCaP cells were transfected with siRNAs against EZH2 and assayed as indicated above. G. Parental and ERG-expressing (ERG-18) LNCaP cells were transfected with full length Nkx3.1 expression vector and assayed as indicated above. H. Parental (sh-) and ESE3 knock-down (sh-7) LNCaP cells were transfected with full length Nkx3.1 expression vector and assayed as indicated above. Data are presented as mean ± SD of triplicate experiments. *, p<0.01; **, p<0.001.

Figure 8. Model of the reciprocal regulation of EZH2 and Nkx3.1 by competing ETS factors.

ESE3 controls expression of EZH2 and Nkx3.1 in normal prostate epithelial cells. ERG over-expression or loss of ESE3 leads to abnormal expression of these key genes and promotes cell transformation.