Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 27 (6), 797-808

ERG Activates the YAP1 Transcriptional Program and Induces the Development of Age-Related Prostate Tumors

Affiliations

ERG Activates the YAP1 Transcriptional Program and Induces the Development of Age-Related Prostate Tumors

Liem T Nguyen et al. Cancer Cell.

Abstract

The significance of ERG in human prostate cancer is unclear because mouse prostate is resistant to ERG-mediated transformation. We determined that ERG activates the transcriptional program regulated by YAP1 of the Hippo signaling pathway and found that prostate-specific activation of either ERG or YAP1 in mice induces similar transcriptional changes and results in age-related prostate tumors. ERG binds to chromatin regions occupied by TEAD/YAP1 and transactivates Hippo target genes. In addition, in human luminal-type prostate cancer cells, ERG binds to the promoter of YAP1 and is necessary for YAP1 expression. These results provide direct genetic evidence of a causal role for ERG in prostate cancer and reveal a connection between ERG and the Hippo signaling pathway.

Figures

Figure 1
Figure 1. Prostate tumors in aged Tg(Pbsn-ERG)1Vv mice
(A) Western blot (WB) analysis of ERG protein expression in primary human prostate tumors, ventral prostates from 8.5 month-old wild-type (WT) and Tg(Pbsn-ERG)1Vv (Pbsn-ERG) mice and the VCaP cell line. (B) Kaplan-Meier survival curves of Tg(Pbsn-ERG)1Vv (Pbsn-ERG) and control wild-type littermate mice. n=15 for wild-type. n=24 for Pbsn-ERG. Logrank test. (C) Gross appearance of urogenital tract organs in 20 month-old Pbsn-ERG male. Arrow points to the primary prostate tumor. (D-M) Hematoxylin and Eosin (H&E, D, F), immunohistochemical (E, G-K) and immunofluorescence (L-M) staining of prostatic adenocarcinoma in 20 month-old Tg(Pbsn-ERG)1Vv (Pbsn-ERG) male. Areas outlined in D (left panel) and E (left panel) are shown at higher magnification in right panels in D and E. Staining was performed with indicated antibodies. Bar in D represents 2 mm in D left panel, 0.14 mm in D right panel, 0.2 mm in E left panel, 45 μm in E right panel, 36 μm in F-K, 12 μm in L-M. See also Figure S1, Table S1.
Figure 2
Figure 2. Large sarcomatoid carcinoma tumor in prostate gland of Tg(Pbsn-ERG)1Vv mice
(A) Histology of prostate gland from 26-month-old Tg(Pbsn-ERG)1Vv mouse. (B) Histology of prostate gland from 38-month-old wild-type mouse. (C-H) Serial sections of the tumor and the adjacent, uninvolved epithelium from Tg(Pbsn-ERG)1Vv mouse stained with hematoxylin & eosin (C), anti-E-cadherin (epithelial cell marker, D), anti-Androgen Receptor (AR, luminal cell marker, E), anti-pan-cytokeratin (epithelial cell marker, F), anti-keratin 5 (basal epithelial cell marker, G) and anti-BrdU (proliferating cell marker, H). Bar in A represents 4 mm in A, B; 80 μm in C-H.
Figure 3
Figure 3. Upregulation of Hippo pathway target gene Ctgf in prostates of Tg(Pbsn-ERG)1Vv mice
(A) RT-PCR analysis of major gene targets of the canonical Wnt, Hedgehog, TGF[.beta], Notch, and Hippo signaling pathways in prostate glands from 3- and 10-month-old Tg(Pbsn-ERG)1Vv (Pbsn-ERG) and wild-type littermate (WT) mice. (B) qRT-PCR analysis of direct Hippo pathway gene target Ctgf in prostate glands from 10 month-old Pbsn-ERG and wild-type littermate mice. Individual mice (left) and wild-type and Pbsn-ERG group (right) comparisons. Data represent mean ± standard deviation. Student's t-test. (C) Immunohistochemical analysis of CTGF expression in prostate glands from 5 month-old wild-type (left panel) and Pbsn-ERG (right panel) littermate mice. (D-E) In situ hybridization analysis of Ctgf expression in prostate glands from 5 month-old wild-type (left panels) and Pbsn-ERG (right panels) littermate mice using anti-sense (D) and control sense (E) probes. For C-E, tissue sections from two WT and two Pbsn-ERG animals were placed on the same slide and all the incubation times and treatments were identical for both genotypes. (F) Western blot analysis of total protein extracts from the prostates of 11 month-old wild-type (WT) and Pbsn-ERG mice with anti-phospho(S127)-YAP1 (P-YAP1), anti-Yap1, anti-TAZ, anti-CTGF, anti-phospho(Thr183)-MST1 (P-MST1), anti-MST1, anti-phospho(S909)-LATS1 (PLATS1), anti-LATS1, anti-ERG, and anti-β-actin antibodies. (G) Immunohistochemical staining of ventral prostates and a prostate tumor from 2.5 year-old wild-type and Pbsn-ERG mice with anti-YAP1 antibodies. (H) Confocal sections of ventral prostate from 5 month-old wild-type mouse immunostained with rabbit anti-YAP1, guinea pig anti-keratin 5, and rat anti-keratin 8. Bar in G corresponds to 40 μm in C-E’, G, and 8.5 μm in H.
Figure 4
Figure 4. ERG binds to TEAD4-interacting regions of chromatin and transactivates Hippo pathway target genes in immortalized prostate epithelial cells
(A-B) Significant overlap of gene expression changes induced by overexpression of ERG and constitutively active YAP1. Gene expression in RWPE-1 cells transduced with empty vectors (RWPE-Ctrl), ERG (RWPE-ERG) or YAP1S127A (RWPE-YAP1S127A) was analyzed by RNASeq. n=2 per each genotype. (A) Heat-map; (B) overlap between significantly (Q<0.05) upregulated genes. Two tailed Chi-square with Yates correction. (C) qRT-PCR analyses of two Hippo pathway target genes CTGF and ENC1 in RWPE-Ctrl (Ctrl), RWPE-ERG (ERG), RWPE-YAP1S127A (YAP1) and RWPE-ERG and YAP127A (ERG YAP) expressing RWPE-1 cells. Data represent mean ± standard deviation (n = 3) from one of three independent experiments. Student's t test. (D) GSEA analysis of similarities between ERG-mediated gene expression changes and YAP1S127A upregulated genes. (E-F) Brightfield images (E) and size quantitation (F) of RWPE-Ctrl (Ctrl) and RWPE-ERG (ERG) cell colonies formed after 6 days in 3D organoid culture system. The graph shows mean ±SD. n≥40. Student's t-test. Bar correspond to 100 μm. (G) Western blot (WB) analyses of RWPE-1 cells transfected with indicated siRNA oligos and analyzed with indicated antibodies. (H) Colony size quantitation of RWPE-Ctrl (Ctrl) and RWPE-ERG (ERG) cells in 3D organoid culture system transfected with indicated siRNA oligos. The graph shows mean ±SD. n≥60. Student's t-test. (I) Colony size quantitation of RWPE-Ctrl (Ctrl), RWPE-ERG (ERG), RWPE-YAP1S127A (YAP1) and RWPE-ERG + YAP127A (ERG + YAP1) cells in 3D organoid culture system. The graph shows mean ±SD. n≥44. Student's t-test. (J) UCSC genome browser views of ChIP-Seq data from RWPE-Ctrl and RWPE-ERG cells using IgG, anti-ERG and anti-TEAD4 antibodies. (K) Highly significant overlap between positions of TEAD4 and ERG peaks in RWPE-ERG cells (hypergeometric p-value=3.761525e-263 calculated by Bioconductor package “ChIPPeakAnno”). (L) Tendency toward close juxtaposition of ERG (A/C)GGAA(G/A) and TEAD4 GGAAT(G/T)(T/C) recognition motifs in ERG-TEAD4 overlapping peaks in RWPE-ERG cells. (M) qPCR analysis of ChIP experiments from RWPE-Ctrl and RWPE-ERG cells using indicated antibodies. Data represent mean ± standard deviation (n = 4) from one of three independent experiments. Student's t test. See also Figure S2, Tables S2 and S3.
Figure 5
Figure 5. ERG maintains YAP1 expression in human prostate cancer cells
(A-C) Significant concordance between ERG and YAP1/TAZ gene expression programs in VCaP cells. Gene expression determined by RNA-Seq of VCaP cells transiently transfected with non-targeting siRNA (siCtrl, n=2), siRNAs targeting ERG (siERG, n=2, using 2 independent siRNA oligos) and siRNAs targeting YAP1/TAZ (siYAP1/TAZ, n=2, using 2 independent siRNA oligo mixtures). (A) Heat-map; (B) overlap between significantly (Q<0.05) upregulated and downregulated genes; and (C) GSEA analysis of overlap between siERG-mediated gene expression changes and siYAP/TAZ downregulated genes. Two tailed Chi-square with Yates correction in B. (D) Western blot (WB) analysis of VCaP cells transfected with indicated siRNA oligos. Numbers indicate relative expression values, with value in siCtrl adjusted to 1. (E) Highly significant overlap between positions of TEAD4 and ERG peaks in VCaP cells (hypergeometric p-value=0 calculated by Bioconductor package “ChIPPeakAnno”). (F) Tendency toward close juxtaposition of ERG (A/C)GGAA(G/A) and TEAD4 GGAAT(G/T)(T/C) recognition motifs in ERG-TEAD4 overlapping peaks in VCaP cells. (G) UCSC genome browser views of ChIP-Seq data from VCaP cells using IgG, anti-ERG and anti-TEAD4 antibodies. (H) qPCR analysis of ChIP experiments from siCtrl, siERG-1 or siERG-2 oligo transfected VCaP cells using indicated antibodies. Data represent mean±standard deviation (n = 4) from one of three independent experiments. Student's t test. (I) Significant correlation between ERG expression and presence of nuclear YAP1 in primary human prostate tumors. Immunohistochemical (IHC) staining of prostate cancers (PC1 and PC2) with ERG and YAP1 antibodies. Note, YAP1 is strongly expressed in benign basal but not benign luminal epithelial cells (arrows), but is present in neoplastic luminal cells in a subset of human prostate carcinomas. Star indicates histologically benign prostatic gland. Bar corresponds to 50 μm. See also Figure S3, Tables S4 and S5.
Figure 6
Figure 6. YAP1 inhibitor Verteporfin displays prominent negative impact on the growth of pre-established ERG-positive orthotopic human prostate tumors
VCaP (ERG-positive) or PC3 (ERG-negative) human prostate cancer cells were injected into anterior lobe of prostate gland of NOD SCID mice and tumors were allowed to establish for 7 weeks without any treatment. Animals were separated into 2 random groups, which were treated with intraperitoneal injection of Verteporfin (or vehicle) every 2 days at 100mg/kg for 3 weeks. Resulting tumors were excised and analyzed. (A) VCaP gross tumor appearance. (B) Quantitation of VCaP tumor volume. Data represent mean±SD. Student's t-test. (C-D) TissueFAX images of Hematoxylin & Eosin staining of VCaP tumor sections. (E) PC3 gross tumor appearance. (F) Quantitation of PC3 tumor volume. Data represent mean±SD. Student's t-test. (G-H) TissueFAX images of Hematoxylin & Eosin staining of PC3 tumor sections. Bar in C corresponds to 2.5 mm in C and G, 50 μm in D and H. See also Figure S4.
Figure 7
Figure 7. Activation of YAP1 in prostate epithelium in vivo is sufficient for the developmen of age-related prostate tumors
(A) Model showing the development of prostate epithelium-specific YAP1-GOF mice. (B) Western blot (WB) analysis of YAP1, TAZ and [.beta]-actin expression in total proteins extracted from ventral prostates of 20 month-old control (WT) and YAP1-GOF mice. (C-I) Hematoxylin and Eosin (H&E, I) and immunohistochemical (C-H) staining of large prostatic sarcomatoid carcinoma in 14 month-old YAP1-GOF male. Staining with anti-androgen receptor (AR), anti-E-cadherin (E-cad), anti-smooth muscle actin (SMA), anti-YAP1 (YAP), anti-pancytokeratin (cytoker) and anti-Vimentin (viment) antibodies. Bar in I represents 1.4 mm in I, 0.4 mm in C-H, 40 μm in insets. (J) Prostate tumor incidence and tumor types in YAP1-GOF and control littermate mice. YAP1-GOF-triple mutant Col1a1tm1(tetO-YAP1*)Fcam/Col1a1+, Gt(ROSA)26Sortm1(rtTA,EGFP)Nagy/Gt(ROSA)26Sor+, Tg(Pbsn-cre)4Prb/0. Control - double mutant Gt(ROSA)26Sortm1(rtTA,EGFP)Nagy/Gt(ROSA)26Sor+, Tg(Pbsn-cre)4Prb/0. Numbers denote mice with indicated tumors relative to the number of mice analyzed. The prostate glands were analyzed in mice euthanized at morbidity. The Dox exposure was started at 3 months of age and varied between 12 and 16 months. Two-tailed Fisher's exact test. (K) Model showing the role of ERG and YAP1 in prostate cancer development. See also Figure S5, Tables S6-S8.

Similar articles

See all similar articles

Cited by 42 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms

Associated data

Feedback