Advertisement
Translational Cancer Mechanisms and Therapy

TMPRSS2-ERG Controls Luminal Epithelial Lineage and Antiandrogen Sensitivity in PTEN and TP53-Mutated Prostate Cancer

Alexandra M. Blee, Yundong He, Yinhui Yang, Zhenqing Ye, Yuqian Yan, Yunqian Pan, Tao Ma, Joseph Dugdale, Emily Kuehn, Manish Kohli, Rafael Jimenez, Yu Chen, Wanhai Xu, Liguo Wang and Haojie Huang
DOI: 10.1158/1078-0432.CCR-18-0653 Published September 2018

Abstract

Purpose: Deletions or mutations in PTEN and TP53 tumor suppressor genes have been linked to lineage plasticity in therapy-resistant prostate cancer. Fusion-driven overexpression of the oncogenic transcription factor ERG is observed in approximately 50% of all prostate cancers, many of which also harbor PTEN and TP53 alterations. However, the role of ERG in lineage plasticity of PTEN/TP53–altered tumors is unclear. Understanding the collective effect of multiple mutations within one tumor is essential to combat plasticity-driven therapy resistance.

Experimental Design: We generated a Pten-negative/Trp53-mutated/ERG-overexpressing mouse model of prostate cancer and integrated RNA-sequencing with ERG chromatin immunoprecipitation-sequencing (ChIP-seq) to identify pathways regulated by ERG in the context of Pten/Trp53 alteration. We investigated ERG-dependent sensitivity to the antiandrogen enzalutamide and cyclin-dependent kinase 4 and 6 (CDK4/6) inhibitor palbociclib in human prostate cancer cell lines, xenografts, and allografted mouse tumors. Trends were evaluated in TCGA, SU2C, and Beltran 2016 published patient cohorts and a human tissue microarray.

Results: Transgenic ERG expression in mice blocked Pten/Trp53 alteration–induced decrease of AR expression and downstream luminal epithelial genes. ERG directly suppressed expression of cell cycle–related genes, which induced RB hypophosphorylation and repressed E2F1-mediated expression of mesenchymal lineage regulators, thereby restricting adenocarcinoma plasticity and maintaining antiandrogen sensitivity. In ERG-negative tumors, CDK4/6 inhibition delayed tumor growth.

Conclusions: Our studies identify a previously undefined function of ERG to restrict lineage plasticity and maintain antiandrogen sensitivity in PTEN/TP53–altered prostate cancer. Our findings suggest ERG fusion as a biomarker to guide treatment of PTEN/TP53-altered, RB1-intact prostate cancer. Clin Cancer Res; 24(18); 4551–65. ©2018 AACR.

Translational Relevance

Prostate cancer resistance to androgen deprivation and AR-targeted therapies remains a pressing clinical obstacle, partly explained by lineage plasticity and transition to AR-independent tumor types in response to these therapies. A comprehensive understanding of genetic prostate tumor subtypes and the unique response of each mutational subtype to AR-targeted therapies is necessary to develop new, subtype-specific therapeutic strategies that overcome therapy-induced lineage plasticity. Our results demonstrate that E-twenty-six transformation specific (ETS)-related gene (ERG) prevents PTEN- and tumor protein 53 (TP53)-negative tumor cell lineage plasticity and antiandrogen resistance by blocking E2F1-mediated expression of lineage switch genes. These findings also reveal the efficacy of targeting retinoblastoma (RB)/E2F1 activity with palbociclib in ERG-negative, PTEN/TP53-altered tumors. This study redefines the role of ERG in a specific tumor subtype and may guide evaluation of the status of concomitant ERG fusion, PTEN/TP53 alteration, and RB1 when selecting therapeutic strategies.

Introduction

Castration-resistant prostate cancers respond to current antiandrogen therapies with variable levels of success (1), in part, due to extensive genetic heterogeneity (2–4). While mechanisms of androgen receptor (AR) pathway restoration and compensation are well documented, adenocarcinoma cell lineage plasticity and reprogramming to AR independence represents an additional resistance mechanism (5). Interestingly, the incidence of AR-independent tumor progression after castration and antiandrogen treatment has increased since the advent of enzalutamide and abiraterone use in the clinic, highlighting that prostate cancer lineage plasticity is an increasingly important barrier to overcome (6). Recent studies have identified a few key molecular events involved in AR-independent tumor progression, such as RB1/PTEN/TP53 loss, MYCN/AURKA amplification, and altered epigenetic regulators including EZH2 (7). However, the molecular basis underlying prostate cancer lineage plasticity and antiandrogen resistance remains poorly understood due to extensive patient tumor heterogeneity and model limitations.

PTEN loss frequently overlaps with TP53 mutation or loss in drug-resistant, morphologically distinct, reprogrammed tumors (8–11). A significant proportion of both primary and castration-resistant tumors with PTEN/TP53 alteration also have AR-dependent, TMPRSS2 fusion–driven overexpression of the ETS family transcription factor ERG (2–4). ERG alone has been shown to repress a neural gene expression signature (12) as well as partially rescue the AR pathway under PTEN loss conditions (13), but the mechanistic role of ERG in the clinically relevant context of both PTEN/TP53 alteration remains uncharacterized.

To address these gaps in the field, we generated a mouse model of prostate cancer that encompasses Pten deletion, Trp53 mutation, and ERG overexpression. Notably, we revealed a novel function of ERG to repress expression of a subset of cell cycle–related genes and block RB hyperphosphorylation in Pten/Trp53-altered, Rb1-intact tumors. As a result, ERG-positive, Pten/Trp53-altered tumors had minimal expression of E2F1 downstream targets involved in a mesenchymal cell lineage switch. We extended these findings to both preclinical xenograft and allograft models of tumor progression and demonstrated that ERG overexpression maintained AR positivity and sensitivity to enzalutamide. In stark contrast, ERG-negative, Pten/Trp53–altered tumors were resistant to enzalutamide treatment and instead developed a reliance on the RB/E2F1 pathway, which was effectively targeted with a CDK4/6 inhibitor, palbociclib. This study emphasizes the importance of evaluating the individual genetic profile of tumors when designing therapeutic strategies, with particular emphasis on ERG fusion, RB1, and PTEN/TP53 status.

Materials and Methods

Cell lines, cell culture, and drug treatment

LNCaP, HEK293T, VCaP, and PC-3 cells were obtained from the ATCC. C4-2 cells were purchased from Uro Corporation. LNCaP-RF cells were described previously (14). HEK293T cells were maintained in DMEM supplemented with 10% FBS. VCaP cells were maintained in DMEM supplemented with 13% FBS. C4-2, LNCaP, LNCaP-RF, and PC-3 cells were maintained in RPMI1640 medium supplemented with 10% FBS. All cell lines were authenticated (karyotyping, mutations in p53 and ERG fusions, and AR, PTEN, p53, and ERG protein expression) and used within 6 months of thawing. No mycoplasma contamination was detected in these cell lines by testing with the Lookout Mycoplasma PCR Detection Kit (Sigma-Aldrich). Charcoal-stripped serum (CSS) was purchased from Thermo Fisher Scientific-Gibco (#12676029). Enzalutamide was kindly provided by Medivation. LNCaP-RF cells were treated with 10 μmol/L of enzalutamide for 72 hours unless otherwise noted. Palbociclib (PD-0332991) was obtained from ApexBio. LNCaP-RF cells were treated with 1 μmol/L of palbociclib for 72 hours unless otherwise noted. For combination treatment, LNCaP-RF cells were treated with 10 μmol/L enzalutamide and 1 μmol/L palbociclib for 72 hours.

Cell transfection and lentivirus transduction

For lentiviral shRNA or stable plasmid expression, HEK293T cells were transiently transfected with pTsin-HA-ERG FL, pTsin-HA-ERG-T1-E4, pTsin-EV, pLKO-shNT, pLKO-shRB, pLKO-shERG, pLKO-shPTEN, or pLKO-shE2F1 as indicated using Lipofectamine 2000 (Thermo Fisher Scientific) following manufacturer's instructions. Virus-containing supernatant was collected 48 hours posttransfection and indicated cells were infected with virus-containing supernatant and 8 μg/mL polybrene. Selection was performed with 1.5 μg/mL puromycin. Sequences of gene-specific shRNAs are listed in Supplementary Table S1. Two shRNAs per gene were tested.

Coimmunoprecipitation and Western blotting

Coimmunoprecipitation and subsequent Western blotting was performed as described previously (15). The following antibodies were used: anti-ERG (ab92513, Abcam; CM421C, Biocare Medical), anti-PTEN (CST9559L, Cell Signaling Technology), anti-p53 (sc126, Santa Cruz Biotechnology), anti-AR (sc816, Santa Cruz Biotechnology), anti-NKX3.1 (NB100-1828, Novus Biologicals), anti-RB (554136, BD Biosciences), anti-pRB S795 (CST9301S, Cell Signaling Technology), anti-SKP2 (32-3300, Life Technologies), anti-CCND1 (sc718, Santa Cruz Biotechnology), anti-CDK1 (sc54, Santa Cruz Biotechnology), anti-TWIST (sc6269, Santa Cruz Biotechnology), anti-CDH1 (610181, BD Biosciences), anti-VIM (sc73258, Santa Cruz Biotechnology), anti-ERK2 (sc1647, Santa Cruz Biotechnology), anti-CDK2 (sc6248, Santa Cruz Biotechnology), anti-E2F1 (sc193, Santa Cruz Biotechnology), anti-pAKT S473 (CST4060L, Cell Signaling Technology), and anti-AKT (CST9272, Cell Signaling Technology).

qRT-PCR

qRT-PCR was performed as described previously (15). All quantifications were normalized to the level of endogenous GAPDH gene. Primers used are listed in Supplementary Table S2.

Chromatin immunoprecipitation and qPCR

Chromatin immunoprecipitation (ChIP) was performed as described previously (16). DNA was pulled down with indicated primary antibodies (anti-ERG, ab92513; anti-E2F1, sc193) or nonspecific IgG. Primers to amplify DNA by real-time qPCR are listed in Supplementary Table S3.

Cell proliferation assays

LNCaP-RF, VCaP, or PC-3 cells were seeded in 96-well plates (∼3,000 cells/wells) and treated as indicated. Cells were fixed at indicated timepoints (day 0–5) and cell growth was measured using sulfohodamine B (SRB) assay (n = 5) as described previously (17).

Hematoxylin and eosin staining

Four micron–thick sections were cut from formalin-fixed paraffin-embedded (FFPE) tumor samples from indicated samples. Xylene washes were used to deparaffinize the tissue, followed by graded ethanol washes to rehydrate tissue. Tissue was stained with hematoxylin, washed, and counterstained with 1% eosin. Stained tissue was dehydrated with graded ethanol washes and a final xylene wash before mounting and sealing with coverslips.

IHC and immunofluorescent cytochemistry

Four micron–thick sections were cut from FFPE tumor samples from indicated mice, xenografts, or human tissue microarrays. Tissue was deparaffinized with xylene and rehydrated through graded ethanol washes. Antigen retrieval and immunostaining was performed as described previously (18, 19). Antibodies for IHC and IFC include: anti-AR (ab108341), anti-ERG (ab92513), anti-CD31 (ab28364), anti-CKAE1/3 (ab27988), anti-RB pS795 (ab85607), anti-Ki67 (ab15580), anti-pAKT S473 (CST4060L), anti-CK8/18 (ab531826), anti-CK5 (ab52635), anti-Vimentin (CST5741S). Ki67 and pRB S795 staining of mouse and xenograft tissues was quantified by counting the number of positive cells out of 100 cells in five random fields of view at 400× per mouse. Staining intensity and percentage for ERG and AR staining of human tissue microarrays were graded using a set of criteria. Intensity was graded 0–3: 0 no staining, 1 low staining, 2 medium staining, and 3 strong staining. A staining index score for each tissue biopsy was obtained by multiplying the staining intensity and percentage values, and used for Pearson product–moment correlation analysis.

Gene set enrichment analysis

Gene set enrichment analysis (GSEA) was performed with a preranked list of the target genes identified by integrated analysis of RNA-seq and ChIP-seq data against curated datasets including HALLMARK_E2F_TARGETS, HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION, CHARAFE_BREAST_CANCER_ LUMINAL_VS_MESENCHYMAL_DOWN, CHARAFE_BREAST_ CANCER_LUMINAL_VS_MESENCHYMAL_UP from the Broad Institute (20).

Samples from patients with prostate cancer

The advanced prostate cancer dataset was generated from patients undergoing standard-of-care clinical biopsies at Mayo Clinic (Rochester, MN). A tissue microarray was constructed from the FFPE samples of metastatic prostate cancer, identified after a search of pathologic and clinical databases of archival tissues. The Mayo Clinic institutional review board approved the experimental protocols for retrieving pathology blocks/slides and for accessing electronic medical records. The human tissue microarray contained 157 cores (16 0.6 mm and 141 1.0 mm cores) resulting from 53 samples (20 bone metastases and 33 nonbone metastases) from 51 patients. Cores in which greater than 50% of the tissue was lost during IHC were excluded from analysis.

Meta-analysis of publicly available datasets

ERG fusion and genetic alterations of PTEN and TP53 for TCGA (n = 333) and SU2C (n = 150) cohorts were downloaded from cBioPortal (http://www.cbioportal.org/; refs. 21, 22). ERG fusion status for Beltran cohort (metastatic tumor specimens = 114) was downloaded from Supplementary Table S5 of the original study (23). Only ERG fusions with RNA-seq or NanoString evidence were included into the analysis. ORs were calculated in cBioPortal where OR > 1 indicates cooccurrence and OR < 1 indicates mutual exclusivity, followed by two-tailed Fisher exact tests to determine significance of the cooccurrence or mutual exclusivity, as described previously (24).

RNA-seq and data analysis

Total RNA was isolated from mouse prostates by homogenization of frozen tissue and purified using the RNeasy Plus Mini Kit (Qiagen). Two-hundred micrograms of high-quality total RNA was used to generate the RNA sequencing library. cDNA synthesis, end-repair, A-base addition, and ligation of the Illumina indexed adapters were performed according to the TruSeq RNA Sample Prep Kit v2 (Illumina). The concentration and size distribution of the completed libraries was determined using an Agilent Bioanalyzer DNA 1000 chip and Qubit fluorometry (Invitrogen). Paired-end libraries were sequenced on an Illumina HiSeq 4000 following Illumina's standard protocol using the Illumina cBot and HiSeq 3000/4000 PE Cluster Kit. Samples were sequenced in biological triplicates and each sample yielded 60–90 million paired-end reads (2 × 50 nucleotide read length). Base calling was performed using Illumina's RTA software (version 2.5.2). Paired-end RNA-seq reads were aligned to the mouse reference genome (GRCm38/mm10) using RNA-seq spliced read mapper Tophat2 (v2.0.6; ref. 25). Pre- and postalignment quality controls, gene-level raw read count, and normalized read count (i.e., FPKM) were performed using RSeQC package (v2.3.6) with NCBI mouse RefSeq gene model (26). Differential gene expression analyses were conducted using edgeR (version 3.6.8) and the built-in “TMM” (trimmed mean of M values) normalization method were used (27). Differentially expressed genes were determined on the basis of the false discovery rate (FDR) threshold 0.01.

ChIP-seq data analysis

ERG, H3K4me1, and H3K4me3 ChIP-seq data in mouse prostate tissue was downloaded from NCBI Gene Expression Omnibus (GEO) with accession number GSE47119 (13). To be compatible with our RNA-seq analysis results, raw reads were realigned to the mouse reference genome (GRCm38/mm10) using bowtie2 (version 2.2.9; ref. 28). MACS2 (version 2.0.10) was used to identify peaks with input samples used as background and a P value cutoff 1E−5 (macs2 callpeak –bdg –SPMR -f BAM; ref. 29). ChIP-seq tag intensity tracks (bedGraph files) were generated by MACS2, and then were converted into bigWig files using UCSC “wigToBigWig” tool. The association of peaks to target genes was performed by Genomic Regions Enrichment of Annotations Tool (GREAT; ref. 30). ERG ChIP-seq data in VCaP cells (GSE14092; ref. 31), E2F1 ChIP-seq data in PC-3 cells (GSE77448; ref. 32), and H3K4me3 ChIP-seq data in LNCaP cells (GSE43791; ref. 33) were downloaded from GEO. ChIP-seq analysis procedure was the same as described above after mapping reads to the human reference genome (GRCh37/hg19).

Generation of Pten/Trp53/ERG-altered mouse model and genotyping

All animal studies were approved by the Mayo Clinic Institutional Animal Care and Use Committee (IACUC). All mice were housed in standard conditions with a 12-hour light/12-hour dark cycle and access to food and water ad libitum. The indicated groups of mice were generated by crossing the following mice: Probasin (Pb)-driven Cre4 recombinase transgenic mice, acquired from the National Cancer Institute (NCI) Mouse Repository and originally generated in the laboratory of Dr. Pradip Roy-Burnam at University of Southern California (Los Angeles, CA; ref. 34); transgenic ERG mice purchased from the Jackson Laboratory (010929), originally generated in the laboratory of Dr. Valeri Vasioukhin at Fred Hutchinson Cancer Research Center (Seattle, WA; ref. 35); Pten loxp/loxp conditional mice, acquired from Jackson Laboratory (004597) and originally generated in the laboratory of Dr. Hong Wu at University of California (Los Angeles, CA; ref. 36); Trp53 loxp/loxp conditional mice, acquired from the NCI Mouse Repository and originally generated in the laboratory of Dr. Tyler Jacks at Massachusetts Institute of Technology (Cambridge, MA; ref. 37); and Trp53 loxp-STOP-loxp-R172H conditional mice, acquired from the NCI Mouse Repository and originally generated in the laboratory of Dr. Tyler Jacks at Massachusetts Institute of Technology (Cambridge, MA; ref. 37). PCR genotyping primers are listed in Supplementary Table S4.

Generation and treatment of prostate cancer cell line xenografts and mouse-derived allografts

All animal studies were approved by the Mayo Clinic IACUC. All mice were housed in standard conditions with a 12-hour light/12-hour dark cycle and access to food and water ad libitum. NOD-SCID IL2 receptor γ-null (NSG) mice were generated in house and at 6 weeks of age, were randomly divided into different experimental treatment groups as indicated (six mice per group). For prostate cancer cell line xenografts, 5 × 106 LNCaP-RF cells per injection were suspended in 0.1 mL of 50% PBS and 50% Corning Matrigel Matrix and implanted by subcutaneous injection into the left flank of each NSG mouse (one implantation per mouse) using a 16 gauge needle. LNCaP-RF cells were tested and ensured to be mycoplasma-free prior to injection using the Lookout Mycoplasma PCR Detection Kit purchased from Sigma-Aldrich, and were stably expressing either pTsin-EV or pTsin-HA-ERG-T1-E4. For mouse-derived allografts, three ARlow/KRTlow DMT prostate tumors and three ARhigh/KRThigh TMT prostate tumors were homogenized and 200 μL of tissue per NSG mouse was implanted by subcutaneous injection into the left flank of each NSG mouse (one implantation per mouse) using a 16 gauge needle. Once the implanted cells grew to reach a size of 100 mm3 measured externally with calipers (approximately 4–5 weeks posttransplantation), drug treatment began. Mice were treated with vehicle (100 μL sodium lactate), enzalutamide (30 mg/kg/day), palbociclib (100 mg/kg/day), or combination by oral gavage, once daily five days per week for three weeks. Mouse weight and tumor size was measured every three days by measuring tumor length (L) and width (W) using a caliper, and tumor volume (TV) was calculated with the following formula: TV = (L × W2)/2. Posttreatment, xenografted tissue was harvested and collected for subsequent study.

Data availability

The datasets generated and/or analyzed during the current study are available in the following repositories. The Cancer Genome Atlas (TCGA) and Stand Up To Cancer (SU2C) datasets analyzed in Fig. 1 and Supplementary Fig. S1 were accessed from cBioPortal (http://www.cbioportal.org/; refs. 21, 22). The Beltran cohort analyzed in Fig. 5 was downloaded from Supplementary Table S5 of the original study (23). The ERG, H3K4me1, and H3K4me3 ChIP-seq datasets analyzed in Fig. 3 were accessed from the NCBI Gene Expression Omnibus (GEO) with accession number GSE47119 (13), the ERG and H3K4me3 ChIP-seq datasets analyzed in Fig. 4 were accessed from the NCBI GEO with the accession numbers GSE14092 (31) and GSE43791 (33), and the E2F1 and H3K4me3 ChIP-seq datasets analyzed in Supplementary Fig. S5 were accessed from the NCBI GEO with the accession numbers GSE77448 (32) and GSE43791 (https://www.ncbi.nlm.nih.gov/geo/; ref. 33). The HALLMARK_E2F, HALLMARK_EPITHELIAL_TO_MESENCHYMAL, and CHARAFE_BREAST_CANCER datasets for GSEA in Fig. 3 were accessed from the Broad Institute (http://software.broadinstitute.org/gsea/index.jsp; ref. 20). The RNA-seq data generated from mouse prostate tissues in Fig. 3 is accessible from the NCBI GEO with the accession number GSE103871.

Figure 1.
Figure 1.

ERG tempers PTEN/TP53 alteration–induced loss of ARhigh luminal epithelial cells. A, Oncoprint image with percentage of ERG, PTEN, TP53, and AR genetic alterations in 333 primary prostate cancer patient samples (top, TCGA cohort; ref. 3) and 150 advanced mCRPC patient samples (bottom, SU2C cohort; ref. 4). B, Contingency tables used by Fisher exact test (two-tailed) to examine association between ERG fusion and PTEN/TP53 alterations in primary TCGA (left) and mCRPC SU2C (right) cohorts. C, Histologic characterization of mouse prostate tissue from 16–20 weeks of age. Wild-type n = 8, Pb-ERG n = 9, DMT n = 10, TMT n = 12. Top, hematoxylin and eosin (H&E) staining. Subsequent rows, IHC for AR, ERG, CD31, Pan-KRT, KRT8/18, KRT5, and Vimentin. CD31 as an endothelial cell marker to distinguish between endogenous endothelial versus transgenic ERG. Asterisk in Vimentin IHC tissue indicates a stromal compartment that is distinct from the Vimentinlow adenocarcinoma.

Statistical analysis

All data are shown as mean values ± SE for experiments performed with at least three replicates. Differences between two groups were analyzed using paired Student t tests unless otherwise noted. P values < 0.05 were considered significant.

Results

Generation and characterization of a clinically relevant Pten/Trp53/ERG triple-mutant mouse model

By mining the whole-exome sequencing data from TCGA patients with primary prostate cancer (PRPC; N = 333; ref. 3), we revealed significant cooccurrence (P = 1.11 × 10−6, OR = 3.01; 95% CI = 1.89–4.84) of PTEN/TP53 deletions or mutations with ERG gene fusions, one of the most frequent genetic alterations in prostate cancer (ref. 38; Fig. 1A and B). In contrast, while a similar trend was observed in the SU2C metastatic castration-resistant prostate cancer (mCRPC) patients (N = 150; ref. 4), the correlation (P = 0.043, OR = 2.04; 95% confidence interval = 0.98–4.33) was much weaker than that in TCGA PRPC patients (Fig. 1A and B). Given that AR is more commonly expressed in PRPC compared with mCRPC, especially neuroendocrine CRPC (NEPC; refs. 23, 39), these data suggest that ERG fusions are prone to cooperate with PTEN and TP53 gene alterations in the pathogenesis of AR-positive prostate cancer. It is important to note that in the mCRPC SU2C cohort, only 3.6% of samples displayed neuroendocrine (ARlow/KRTlow) features (4), which may partly explain the apparent under-representation of AR loss samples in the SU2C dataset (Fig. 1A). To genetically test this hypothesis in vivo, we generated four cohorts of mice recapitulating the genetic alterations most frequently occurring in human prostate cancers (such as R175H in TP53; refs. 3, 4, 40; Supplementary Fig. S1): (i) “wild-type” (Cre-negative littermates); (ii) ERG transgenic alone, with Met33 N-terminally truncated ERG driven by the AR-dependent probasin (Pb) promoter (hereafter termed Pb-ERG); (iii) prostate-specific Pten deletion and Trp53 deletion and mutation (Ptenpc−/−;Trp53pcR172H/−, hereafter termed double mutant or DMT) where Trp53 R172H is the mouse equivalent to human TP53 R175H; and (iv) Ptenpc−/−;Trp53pcR172H/−;Pb-ERG (hereafter termed triple mutant or TMT; Supplementary Fig. S2A). We generated these four groups of mice by using Pb-driven Cre recombinase (Pb-Cre4; ref. 34), Pb-ERG (35), Ptenloxp/loxp (36), and Trp53loxp-stop-loxp-R172H/loxp (37) as breeders. For comparison, we also generated prostate-specific Pten deletion (Ptenpc−/−), Pten and Trp53 double deletion (Ptenpc−/−;Trp53pc−/−) (hereafter termed double knockout or DKO) mice, as well as prostate-specific Pten and Trp53 double deletion plus ERG transgenic (Ptenpc−/−;Trp53pc−/−;Pb-ERG) mice (Supplementary Fig. S2A–S2C). Pten/Trp53 DKO mice have been shown to develop plastic, dedifferentiated tumors (8–10, 41) and served as controls for comparison purposes with the Trp53-mutant lines, which represent an unstudied portion of patients with PTEN deletion/TP53 mutation.

At the age of 8–10 weeks, 100% of both Ptenpc−/−;Trp53pc−/− (DKO) and Ptenpc−/−;Trp53pcR172H/− (DMT) mice developed well-differentiated adenocarcinomas with high expression of AR proteins (ARhigh; Supplementary Fig. S2B). In contrast, AR expression was dramatically reduced (ARlow) in prostate tumors in approximately 90% of DKO and DMT mice at the age of 16–20 weeks (Fig. 1C; Supplementary Fig. S2B). Consistent with the reduced AR expression, the level of pan-keratin (pan-KRT) in tumors, used as an indicator of epithelial cells as opposed to mesenchymal cells, was also markedly reduced (KRTlow) in both DKO and DMT mice at the age of 16–20 weeks compared with mice 8–10 weeks younger (Supplementary Fig. S2B). In addition, DMT tumor cells from mice at the age of 16–20 weeks were also negative for both luminal epithelial cell marker KRT8/18 and basal epithelial cell marker KRT5, but positive for vimentin (VIM), a mesenchymal cell marker (Fig. 1C). These results suggest that tumors in DKO and DMT mice at the age of 16–20 weeks transitioned to minimal luminal epithelial phenotypes and were less differentiated compared with tumors in mice at younger ages, as indicated by the comparatively weak but detectable pan-keratin and AR levels. These data provide support to the previous observation that loss of Pten and Trp53 induces lineage plasticity in mouse prostate cancer (8–10, 41).

In striking contrast, at the same age (16–20 weeks), approximately 50% of Ptenpc−/−;Trp53pcR172H/−;Pb-ERG (TMT) mice developed well-differentiated ARhigh/KRThigh adenocarcinomas while the other 50% developed ARlow/KRTlow tumors reminiscent of those in DMT mice (Fig. 1C). Notably, ARlow/KRTlow tumors in TMT mice lacked transgenic ERG expression in the majority of tumor cells, but as expected, endogenous ERG was highly expressed in CD31-positive endothelial cells of blood vessels (Fig. 1C). Importantly, lack of epithelial expression of transgenic ERG correlated with decreased expression of AR proteins in these Pten/Trp53–altered tumors (Fig. 1C; Supplementary Fig. S2C). This observation is supported by the previous report that ERG knockdown decreases the AR-positive luminal cell population in TMPRSS2-ERG–expressing VCaP prostate cancer cells (12). Together, these findings reveal that PTEN/TP53 alteration induces loss of the ARhigh luminal epithelial cell lineage in prostate cancer and this phenomenon is disrupted in the presence of ERG expression.

Compared with Pten wild-type prostate tissues (“wild-type” and Pb-ERG genotypes), Pten-null PIN lesions in Ptenpc−/− mice or tumors in DMT and TMT mice had increased, but comparable levels of phosphorylated AKT (pAKT S473; Fig. 2A and B; Supplementary Fig. S3A), reinforcing the concept that PTEN loss is a key driver of initial tumorigenesis in these models (42–44). Intriguingly, plasma membrane expression of pAKT S473 was detected in the luminal epithelial cells of ARhigh/KRThigh tumors in TMT mice, whereas no typical plasma membrane staining of pAKT S473 was detected in ARlow/KRTlow tumor cells in DMT and TMT mice (Fig. 2A; Supplementary Fig. S3A and S3B), a phenomenon reminiscent of prostate-specific Pten/Rb1 double KO tumors (8). At present, the exact cause-and-effect of altered cellular localization of phosphorylated AKT remains to be elucidated. Previous study has shown that in the presence of PTEN loss, ERG partially rescues AR function (13). However, further analyses showed that ARhigh/KRThigh tumors in TMT mice had lower levels of phosphorylated RB (pRB S795) and lower expression of a subset of cell cycle–promoting genes compared with ARlow/KRTlow tumors in DMT and TMT mice (Fig. 2C and D, see Fig. 3 below). Furthermore, expression of cell lineage regulators commonly associated with epithelial-to-mesenchymal transition (EMT) and neuroendocrine cell lineage was also much lower in ARhigh/KRThigh TMT tumors than that in ARlow/KRTlow tumors (Fig. 1C; see Fig. 3 below). These data argue that ERG-induced preservation of the late-stage ARhigh/KRThigh phenotype was not solely mediated by restored AR activity in the PTEN loss context, but may require additional drivers.

Figure 2.
Figure 2.

ERG prevents Pten/Trp53 alteration–induced proliferation and loss of membrane-localized phosphorylated AKT in mouse prostate tumors. A, IHC for pAKT S473 in mouse prostate tissues from 16–20 weeks of age. Wild-type n = 8, Pb-ERG n = 9, DMT n = 10, TMT n = 12, Ptenpc−/− n = 8. B, Protein levels of pAKT S473, total AKT, and AR in mouse prostate tissues at 16–20 weeks of age. Both blots for each protein of interest were exposed and developed on the same piece of film. ERK2 as a loading control. Band intensity was quantified and normalized to ERK2 for each lane. Asterisk = outlier samples with significantly low levels of total protein. C, IHC for pRB S795 in mouse prostate tissues from 16–20 weeks of age as described in A. D, Quantification of pRB S795 staining as shown in C. E, IHC for Ki67 in mouse prostate tissues from 16–20 weeks of age as described in A. F, Quantification of Ki67 IHC as shown in E.

Figure 3.
Figure 3.

ERG expression downregulates key cell-cycle–driving genes and maintains both AR pathway and epithelial gene expression in mouse prostate tumors. A, Venn diagram indicating overlap between up- or downregulated genes in DMT ARlow/KRTlow (n = 2) versus TMT ARhigh/KRThigh (n = 3) tumors and ERG target genes identified by ChIP-sequencing (13). Fisher exact test (assuming human genome code for 27,000 genes, estimated from RefSeq) for ERG ChIP-seq versus downregulated genes: P < 0.001; for ERG ChIP-seq versus upregulated genes: P < 9.088e−23. B, Gene Ontology analysis of 972 ERG target genes downregulated in TMT ARhigh/KRThigh tumors, ranked by P value. C, GSEA for all 2,595 up- or downregulated ERG target genes (mouse gene names converted to human homologs using NCBI Homology Map). D, Heatmap showing differentially expressed genes between two DMT ARlow/KRTlow and three TMT ARhigh/KRThigh tumors, highlighting a subset of genes involved in cell cycle, AR pathway, and epithelial-to-mesenchymal transition (EMT). E, RNA-seq and ChIP-seq track views from UCSC Genome Browser for two ERG predicted target genes (Ccnd1, Cdk1) with ERG-binding peaks and two predicted passenger genes (Sox11, Twist1) without ERG-binding peaks. Peaks underlined with black bars and boxed with a dashed red line indicate significant ERG ChIP-seq peaks with P = 1e−5 or lower as determined by MACS. ERG ChIP-seq input tracks shown as a control for true ERG peaks. H3K4me3 peaks shown to indicate gene promoter regions. H3K4me1 peaks shown to indicate gene enhancer regions. F, qRT-PCR of n = 2 DMT ARlow/KRTlow and n = 3 TMT ARhigh/KRThigh tumors for Ccnd1, Cdk1, Twist1, and Sox11. Relative to Gapdh.

Loss of RB function has been implicated in development of plastic, antiandrogen-resistant prostate tumors in Pten and Trp53-deficient mice (8, 10). In agreement with functional loss of RB as reflected by increased RB phosphorylation (Fig. 2C and D), cell proliferation as indicated by Ki67 staining was much higher in ARlow/KRTlow dedifferentiated tumors in both DMT and TMT mice compared with that in ARhigh/KRThigh tumors in TMT mice (Fig. 2E and F). It is worth noting that proliferation in ARhigh/KRThigh tumors in TMT mice was still much higher than that in malignant and nonmalignant prostate tissues in Pten KO alone and ERG transgenic alone mice, respectively (Fig. 2E and F), reinforcing the concept that ERG is an oncogenic protein that promotes prostate tumorigenesis by cooperating with other lesions. Nevertheless, these data suggest that ERG may regulate the cell cycle and subsequently, RB activity.

ERG downregulates a subset of cell cycle–promoting genes in Pten/Trp53–altered mouse prostate tumors

To define the molecular mechanisms by which ERG modulates prostate cancer cell lineage plasticity, we performed RNA sequencing (RNA-seq) analysis in three ARlow/KRTlow tumors from DMT mice and three ARhigh/KRThigh tumors in TMT mice. We selected DMT ARlow/KRTlow tissues rather than TMT ARlow/KRTlow tissues for this analysis to ensure no possible contamination from any tumor cells that may have low levels of ERG expression. Although we did not observe strongly positive ERG-expressing tumor cells by IHC in the TMT ARlow/KRTlow tissues (Fig. 1C), the presence of the Pb-ERG transgene in these mice would not allow us to eliminate that possibility. RNA-seq data for one DMT tumor was excluded from further analysis due to its poor correlation with the other two biological replicates (Supplementary Fig. S4A). Differential gene expression analyses revealed 1,281 and 1,598 genes that were significantly down- and upregulated by ERG, respectively, in ARhigh/KRThigh TMT prostate tumors in comparison with ARlow/KRTlow DMT tumors (Fig. 3A; Supplementary Fig. S4B). After integrating RNA-seq data with ERG ChIP-coupled sequencing (ChIP-seq) data obtained from prostate tumors in Rosa26 TMPRSS2-ERG mice (13), we found 76% (972 of 1,281) of ERG-downregulated genes and 82% (1,314 out of 1,598) of ERG-upregulated genes contained ERG ChIP-seq peaks in their promoter and/or enhancer regions (Fig. 3A), suggesting that they are putative ERG target genes.

Gene ontology (GO) analysis of the 972 ERG-downregulated genes demonstrated a significant enrichment of genes that regulate the cell cycle (Fig. 3B), in agreement with our finding that ARhigh/KRThigh tumors in TMT mice display decreased RB phosphorylation and cell proliferation in comparison with ARlow/KRTlow tumors in DMT and TMT mice (Fig. 2C–F). GSEA revealed that ERG-downregulated genes significantly overlap with hallmark E2F targets and EMT genes (Fig. 3C). Additional comparison demonstrated that ERG-downregulated genes also correlated with luminal epithelial-to-mesenchymal changes in other cancer types. For examples, these genes were also significantly overlapped with genes downregulated in the luminal breast cancer cell type as compared with the mesenchymal-like breast cancer cell type, while ERG-upregulated genes were significantly overlapped with genes upregulated in the luminal breast cancer cell type (ref. 45; Fig. 3C). Further analysis of RNA-seq profiles between ARlow/KRTlow tumors in DMT mice and ARhigh/KRThigh tumors in TMT mice revealed that ERG expression resulted in drastic upregulation of AR pathway genes (e.g., Ar and Nkx3.1) and luminal epithelial lineage genes (e.g., Cdh1 and Krt8), and robust downregulation of cell-cycle genes (e.g., Ccnd1 and Cdk1) and nonluminal epithelial (mesenchymal and neuroendocrine) lineage-regulatory genes (ref. 11; e.g., Twist1 and Sox11; Fig. 3D).

ERG-downregulated genes are exemplified in Fig. 3E and were further confirmed by reverse transcription–coupled quantitative PCR (qRT-PCR; Fig. 3F). qRT-PCR analysis of key cell-cycle and EMT genes and Western blot analysis of AR proteins further confirmed that ARlow/KRTlow tumors in DMT and TMT mice shared similar molecular traits (Fig. 2B; Supplementary Fig. S4C). It is important to note that although TMT ARlow/KRTlow tissues were not analyzed by RNA-seq, the trends in gene expression observed in DMT ARlow/KRTlow tissues were seemingly conserved in the TMT ARlow/KRTlow tissues (Fig. 2B; Supplementary Fig. S4C). ERG ChIP-seq data clearly showed ERG-binding peaks in the promoter region of cell-cycle genes such as Ccnd1, and Cdk1, but not cell lineage–regulatory genes such as Twist1 and Sox11 (Fig. 3E), suggesting that cell-cycle–related genes are likely direct targets of ERG while Twist1 and Sox11 are not. These data suggest that in the context of Pten/Trp53 alteration, ERG transcriptionally downregulates a subset of key cell-cycle–promoting genes and maintains AR signaling.

VCaP cells harbor intact PTEN, one allele loss of TP53 and a gain-of-function mutation R248W, which is a hotspot mutation in CRPC (40) (Supplementary Fig. S1). ERG is also overexpressed in this cell line due to TMPRSS2-ERG fusion (TMPRSS2 exon 1 fused with ERG exon 4 or termed T1-E4 ERG). Importantly, knockdown of ERG in the presence of PTEN depletion increased CCND1, CDK1, and SKP2 protein levels in VCaP cells (Supplementary Fig. S5A). Thus, these data provide futher support to the hypothesis and the validation of ERG regulation of a few representative gene targets defined by the integrated RNA-seq and ChIP-seq analyses. Our data that ERG bound to the promoter of Ccnd1 and Cdk1 genes and repressed their expression suggests ERG is a potent upstream regulator of RB hypophosphorylation and activation. This notion is further supported by a recent report that in spite of the androgen-stimulating effect of RB hyperphosphorylation in TMPRSS2-ERG–negative LNCaP cells, RB remains hypophosphorylated in TMPRSS2-ERG–positive VCaP cells even after androgen stimulation (46).

Human prostate cancer cell lines recapitulate ERG-mediated repression of the cell cycle through the RB pathway

To delineate the relationship between ERG expression and PTEN/TP53 alteration in tumor cell proliferation, cellular identity, and antiandrogen resistance, we surveyed human prostate cancer cell lines including VCaP, C4-2, LNCaP, LNCaP-RF, and PC-3 (Fig. 4A). Among the cell lines surveyed, VCaP cells had the highest level of AR protein, hypophosphorylated RB, minimal expression of cell-cycle–promoting proteins CCND1, CDK1, and SKP2, and low levels of mesenchymal-related proteins TWIST1 and VIM (Fig. 4A), similar to ARhigh/KRThigh tumors in TMT mice (Fig. 1C). Consistent with ARlow/KRTlow DMT or DKO tumors, PTEN- and ERG-negative CRPC cell lines PC-3 and LNCaP-RF, which lack or express very low levels of functional p53, respectively, displayed little to no expression of AR, but increased hyperphosphorylated RB and augmented expression of cell-cycle–driven proteins and mesenchymal-specific proteins (Fig. 4A). Overexpression of full-length or fusion (T1-E4) ERG in LNCaP-RF and PC-3 cells partially reversed these trends in a dose-dependent manner (Fig. 4B; Supplementary Fig. S5B) and decreased cell proliferation (Fig. 4C), as detected in ERG-positive, Pten/Trp53-mutated mouse prostate tumors (Fig. 2E and F). Conversely, concomitant knockdown of endogenous TMPRSS2-ERG and PTEN in TP53-mutated VCaP cells, mimicking the situation in ARlow/KRTlow tumors in DMT or TMT mice, resulted in increased expression of cell-cycle–related proteins, hyperphosphorylation of RB, upregulation of nonepithelial cell markers TWIST1 and VIM, and decreased expression of AR and the epithelial cell markers CDH1 and NKX3.1 (Supplementary Fig. S5A and S5C).

Figure 4.
Figure 4.

ERG binds to the promoter and regulates expression of cell-cycle genes in human PTEN/TP53–altered prostate cancer cells. A, Western blot analysis of expression of key AR pathway, cell-cycle, and EMT-related proteins in five prostate cancer cell lines. B, Western blot analysis of expression of key AR pathway, cell-cycle, and EMT-related proteins in LNCaP-RF and PC-3 cell lines after lentiviral-mediated expression of full-length (ERG-FL) or ERG (T1-E4). C, Cell proliferation as measured by SRB assay for LNCaP-RF and PC-3 cell lines with lentiviral-mediated expression of ERG-FL or ERG (T1-E4). D, Top, coimmunoprecipitation of endogenous ERG and RB in VCaP cells. ERG and AR coimmunoprecipitation shown as positive control. Bottom, coimmunoprecipitation of ERG and RB in PC-3 cells stably expressing ERG (T1-E4). E, Western blot analysis of expression of key AR pathway and EMT-related genes in LNCaP-RF and PC-3 cell lines after lentiviral-mediated expression of ERG (T1-E4) with or without RB1 knockdown. F, RT-qPCR of CCND1, CDK1, SKP2, FOXM1, TWIST1, TWIST2, SOX11, and TGFB2 in PC-3 cells with or without lentiviral-mediated expression of ERG (T1-E4). Relative to GAPDH. G, ERG ChIP-seq tracks in VCaP cells (GSE14092; ref. 31) and H3K4me3 (histone mark of promoters) ChIP-seq tracks in LNCaP cells (GSE43791) (33) from UCSC genome browser for CCND1, CDK1, SKP2, FOXM1, TWIST1, TWIST2, SOX11, and TGFB2. ERG ChIP-seq input tracks shown as a control for true ERG peaks. Peaks underlined with black bars and boxed with a dashed red line indicate significant ERG ChIP-seq peaks with P = 1e−5 or lower as determined by MACS. Asterisk indicates ERG peak that could not be validated by ChIP-qPCR. H, ERG ChIP-qPCR of CCND1, CDK1, SKP2, FOXM1, TWIST1, TWIST2, SOX11, and TGFB2 in PC-3 cells with lentiviral-mediated expression of ERG (T1-E4).

Previous study has indicated that hypophosphorylated RB can be recruited by AR to repress cell-cycle genes (46). Coimmunoprecipitation assay in VCaP cells demonstrated that similar to previous study (31), ERG interacted with AR (Fig. 4D). However, no interaction was detected between ERG and RB, and similar results were obtained in PC-3 cells stably expressing T1-E4 ERG (Fig. 4D), excluding the possibility that ERG may recruit RB to repress cell-cycle genes in a manner similar to AR (46). However, because key cell-cycle regulators such as CCND1 and CDK1 were identified as transcriptionally repressed target genes of ERG, it is possible that ERG causes a reduction in RB phosphorylation and cell-cycle progression by directly downregulating cell-cycle genes. This hypothesis is consistent with decreased expression of a subset of cell-cycle genes in ARhigh/KRThigh adenocarcinomas in TMT mice (Fig. 3) and in human LNCaP-RF and PC-3 cells stably expressing ERG (Fig. 4B). Moreover, ERG-mediated upregulation of AR and downregulation of EMT genes were reversed by depletion of RB in ERG-expressing LNCaP-RF cells (Fig. 4E). Similar effects were observed in PC-3 cells (Fig. 4E). These results along with the ERG ChIP-seq data (Fig. 3) suggest that ERG functions as an upstream activator of RB by specifically binding to the promoter and repressing expression of a subset of cell-cycle–driving genes.

E2F1 activates expression of EMT-promoting factors in ERG-negative, PTEN/TP53–altered tumor cells

It has been reported recently that E2F1 promotes prostate cancer cell metastasis and enhanced mesenchymal-like phenotypes (increased migration and invasion) by binding to the promoter and upregulating expression of the RHAMM gene (HMMR; ref. 47). By analyzing E2F1 ChIP-seq data obtained in PC-3 cells (32), we found robust binding of E2F1 proteins in the loci of ERG-suppressed mesenchymal lineage-driving genes including SNAI1, TGFB2, TWIST1, TWIST2, and HMMR (Supplementary Fig. S5D). This observation was further confirmed by ChIP-qPCR (Supplementary Fig. S5E). Most importantly, the effect of ERG and PTEN double knockdown on expression of EMT-promoting genes and epithelial and mesenchymal cell markers in VCaP cells was abrogated by concomitant knockdown of E2F1 by two independent shRNAs (Supplementary Fig. S5C). These data suggest that E2F1 mediates repression of downstream mesenchymal lineage genes in the context of ERG+/PTEN−/p53null/mutant cells. In further support of the transcriptome results in DMT and TMT mouse tumors, ectopic expression of T1-E4 ERG in PC-3 cells reduced expression of CCND1, CDK1, TWIST1, and SOX11, as well as other key cell cycle, EMT, and neuroendocrine-related genes (Fig. 4F; Supplementary Fig. S6A–S6C). ERG ChIP-seq in VCaP cells and ChIP-qPCR in PC-3 cells stably expressing T1-E4 ERG confirmed cell-cycle genes such as CCND1 and CDK1 as direct ERG targets in human prostate cancer cells, but E2F1 ChIP-seq and ChIP-qPCR in PC-3 cells demonstrated TWIST1 and other cell lineage–regulatory factors as downstream gene targets of E2F1 (Fig. 4G and H; Supplementary Fig. S5D and S5E; Supplementary Fig. S6D and S6E). It should be noted that there was an observed ERG ChIP-seq peak at the TGFB2 locus, although this binding could not be validated by ChIP-qPCR (Fig. 4G and H). Thus, our data cannot completely rule out the possibility that ERG may also potentially regulate expression of this locus. Together, these data suggest that ERG directly binds to and regulates expression of a subset of cell-cycle genes in human PTEN/TP53–mutated prostate cancer cells, which in turn leads to RB hypophosphorylation and inhibition of E2F1-mediated transcription of mesenchymal-promoting genes.

ERG and AR expression is positively associated in human prostate tumors

In agreement with our observations in mouse models and human prostate cancer cell lines, genome analysis of CRPC adenocarcinomas (CRPC-Ad) and neuroendocrine tumors (CRPC-NE; ref. 23) revealed a significant association of ERG fusion gene expression with CRPC-Ad (ARhigh), but not CRPC-NE tumors (ARlow; P = 0.0485; OR = 0.14; 95% CI = 0.003–1.06; Fig. 5A). Because CRPC-NE tumors have low or absent AR expression and TMPRSS2-ERG gene fusion expression is driven by AR, our analyses were only focused on those samples with expression of ERG gene fusion as demonstrated by RNA-seq or NanoString data (23, 39). In addition, we performed IHC analysis on a human tissue microarray with 157 cores constructed from 51 patients with metastatic CRPC undergoing standard-of-care clinical biopsies at Mayo Clinic (Rochester, MN). This analysis confirmed a strong association between AR and ERG expression (P = 2.98e−7, correlation = 0.41; 95% CI = 0.263–0.536; Fig. 5B and C; Supplementary Table S5).

Figure 5.
Figure 5.

ERG expression correlates with AR expression in human patient datasets. A, Fisher exact test to determine association between ERG fusion and CRPC-adenocarcinoma (CRPC-Ad) or CRPC-neuroendocrine (CRPC-NE) tumor subtypes from the Beltran cohort (23). B, Representative IHC images for AR and ERG from human tissue microarray of metastatic CRPC obtained from Mayo Clinic. C, Pearson product–moment correlation between AR and ERG IHC staining in clinical biopsies of metastatic CRPC in B.

ERG expression in PTEN/TP53 tumors regulates prostate tumor response to antiandrogen and anti-RB/E2F1 pathway drugs

The above findings prompted us to hypothesize that ERG+/ARhigh/KRThigh (TMT) adenocarcinoma cells would respond to enzalutamide treatment, but ERG−/ARlow/KRTlow (DMT) tumor cells would not. Instead, ERG−/ARlow/KRTlow (DMT) tumor cells may rely heavily on RB hyperphosphorylation to maintain cell proliferation, dedifferentiation, and antiandrogen-resistant phenotypes, and therefore this type of tumor may be highly responsive to RB-targeted therapy such as CDK4/6 inhibitors. Palbociclib (PD-0332991) is a CDK4/6 inhibitor that has been shown to be effective in preclinical models of prostate and other cancer types, and was recently approved by the FDA for treatment of breast cancer (46–50). Treatment of control LNCaP-RF cells (ERG−/ARlow) with enzalutamide alone had no overt effect on expression of cell-cycle genes, RB phosphorylation, and cell proliferation (Fig. 6A and B; Supplementary Fig. S7A and S7B), confirming the antiandrogen-resistant nature of ERG−/ARlow cells. However, palbociclib treatment, either alone or in combination with enzalutamide, significantly decreased expression of cell-cycle genes and inhibited proliferation in ERG−/ARlow cells (Fig. 6A and B; Supplementary Fig. S7A and S7B). It is interesting to note that combination of enzalutamide and palbociclib significantly inhibited proliferation of the ERG−/ARlow LNCaP-RF cells compared with palbociclib treatment alone (Fig. 6B), suggesting that palbociclib treatment may resensitize these cells to enzalutamide treatment. As expected, LNCaP-RF cells stably expressing ERG (ERG+/ARhigh) responded favorably to enzalutamide alone, and such effect was not enhanced in combination with palbociclib (Fig. 6A and B; Supplementary Fig. S7A and S7B). In further support of the finding that RB1 knockdown in ERG-positive cells abrogates the subsequent downregulation of cell lineage genes (Fig. 4E), RB1 knockdown in LNCaP-RF-ERG T1-E4 cells also abolished enzalutamide sensitivity (Supplementary Fig. S7C and S7D). Sensitivity to enzalutamide in ERG+/ARhigh cells (LNCaP-RF-ERG T1-E4 and VCaP), but not ERG−/ARlow (LNCaP-RF-EV) cells, was abrogated by androgen deprivation of culture media (Supplementary Fig. S7E and S7F), confirming AR pathway dependence in ERG-positive cells.

Figure 6.
Figure 6.

Differential responses of ERG-positive and ERG-negative human xenograft and mouse allograft tumors with PTEN/TP53 alterations to enzalutamide and palbociclib. A, Western blot analysis of expression of key AR pathway, cell-cycle, and EMT-related proteins in LNCaP-RF cells with or without lentiviral-mediated ERG (T1-E4) expression after treatment with vehicle, enzalutamide (ENZ, 10 μmol/L), palbociclib (PD, 1 μmol/L), or combination (ENZ + PD). B, Cell proliferation as measured by SRB assay for LNCaP-RF cells with or without lentiviral-mediated ERG (T1-E4) expression after treatment with vehicle, ENZ (10 μmol/L), PD (1 μmol/L), or combination. C, LNCaP-RF xenograft tumor volume with or without lentiviral-mediated ERG (T1-E4) expression during 3-week treatment with vehicle, ENZ (30 mg/kg/day), PD (100 mg/kg/day), or combination. Six xenografts (n = 6) per cell line, per drug treatment. D, ERG−/ARlow/KRTlow DMT and ERG+/ARhigh/KRThigh TMT allograft tumor volume during 3 weeks of treatment with vehicle, ENZ (30 mg/kg/day), PD (100 mg/kg/day), or combination. Five allografts (n = 5) per genotype, per drug treatment. E, Characterization of allograft tumors from (D) after 3 weeks of treatment. Top, H&E. Subsequent rows, IHC for ERG, AR, pRB S795, and Ki67. F, A hypothetical model. In prostate cancer cells without the TMPRSS2-ERG fusion, PTEN deletion/mutation and TP53 deletion/mutation favor cell-cycle gene expression, CDK activation, and RB inhibition (hyperphosphorylation), which in turn lead to E2F1 activation and luminal-epithelial-to-mesenchymal cell identity transition, antiandrogen resistance, and increased CDK4/6 inhibitor sensitivity. In contrast, in prostate cancer cells harboring the TMPRSS2-ERG fusion, overexpression of ERG results in decreased expression of a subset of cell-cycle–promoting genes and RB activation (hypophosphorylation), thereby leading to E2F1 inhibition and maintenance of luminal epithelial cell identity, increased antiandrogen sensitivity, but CDK4/6 inhibitor resistance.

We further examined the responsiveness of ERG-positive prostate cancer to antiandrogen therapy using in vivo models. Similar to the findings in vitro, ERG−/ARlow LNCaP-RF xenograft tumors were resistant to enzalutamide treatment (Fig. 6C; Supplementary Fig. S8A–S8C). In contrast, treatment of these tumors with palbociclib significantly decreased tumor volume, Ki67 staining, and RB phosphorylation (Fig. 6C; Supplementary Fig. S8A–S8C). Similar to the LNCaP-RF cell line study, combination of enzalutamide and palbociclib significantly decreased ERG−/ARlow LNCaP-RF xenograft tumor volume compared with palbociclib treatment alone (Fig. 6C; Supplementary Fig. S8A), which further highlights the potential efficacy of combination treatment in these tumors. In ERG+/ARhigh LNCaP-RF xenograft tumors, palbociclib treatment alone exerted little to no effect but both enzalutamide treatment alone and in combination with palbociclib significantly reduced the tumor volume, Ki67 staining, and expression of pRB S795 (Fig. 6C; Supplementary Fig. S8A–S8C).

We attempted to perform similar studies using DMT and TMT spontaneous tumor models. However, we found it was quite challeging to crossbreed five different alleles together to simultaneously generate large cohorts of DMT and TMT mice at the same ages. Because of this technical difficulty, we performed similar drug treatment studies using allografts derived from ERG−/ARlow/KRTlow DMT and ERG+/ARhigh/KRThigh TMT tumors and confirmed the findings from LNCaP-RF xenografts (Fig. 6D and E; Supplementary Fig. S8D and S8E). Collectively, these data highlight that PTEN/TP53–altered tumors with hyperphosphorylated RB are resistant to enzalutamide, but are sensitive to CDK4/6 inhibition alone or in combination with enzalutamide. In contrast, ERG expression maintains antiandrogen sensitivity in tumors even with PTEN/TP53 alteration and this effect is related to ERG-induced inhibition of cell-cycle gene expression and restored AR signaling.

Discussion

The findings in this study emphasize that the unique combination of genetic mutations present within a single prostate tumor can greatly affect response to androgen- and AR-targeted therapies. In particular, our study of the novel Pten/Trp53/ERG triple-mutant mouse model of prostate cancer recapitulates a trio of genetic events that cooccur in a significant subtype of prostate tumors. Previous studies demonstrated that loss of Pten and Trp53 induces lineage plasticity in mouse prostate cancer, where prostate-specific Pten and Trp53 double KO mice develop prostate adenocarcinoma at a young age and further evolve into ARlow/KRTlow tumors (8–10, 41). These data and ours support the hypothesis that Pten/Trp53-altered tumors may transition from an ARhigh/KRThigh adenocarcinoma to an altered ARlow/KRTlow state (Fig. 6F, left).

Further analysis of the novel Pten/Trp53/ERG model revealed that ERG binds to chromatin loci of a subset of cell-cycle–driving genes and decreases their expression in Pten/Trp53–altered mouse prostate tumors, thereby preventing loss of RB activity and E2F1-mediated cellular reprogramming (Fig. 6F, right). Studies in human prostate cancer cell lines also supported these findings. Most importantly, similar results were obtained through analysis of patient datasets and clinical samples. Although previous studies have suggested a potential role for ERG in repressing neuroendocrine differentiation and partially rescuing AR function (12, 13), this study represents the first to demonstrate ERG-mediated protection of the epithelial adenocarcinoma cell lineage in a clinically relevant mouse model with Pten/Trp53 mutations (Fig. 6F).

We further demonstrated in Pten/Trp53–mutated mouse prostate cancer and xenograft models that while ERG-positive tumors are sensitive to antiandrogen treatment, ERG-negative tumors have no overt response to antiandrogens and instead respond well to the CDK4/6 inhibitor palbociclib. These findings were recapitulated in human cell lines. Together, these data reveal a previously undefined role of ERG in maintaining neoplastic epithelial cell identity and antiandrogen sensitivity in PTEN/TP53–mutated prostate cancer and highlight that different therapeutic strategies are needed for PTEN/TP53–altered tumors with or without ERG (Fig. 6F).

Despite a previous finding that ERG overexpression alone is sufficient for focal prostatic intraepithelial neoplasia (PIN) formation in mice with 129/Sv background (35), we did not observe any PIN lesions in Pb-ERG mice during the course of our previous study (51) and this report (Fig. 1C) perhaps due to the different genetic (C57BL6/129) background of mice we used. ERG expression levels in Pb-ERG mouse prostate tissue were comparable with that in the TMPRSS2-ERG fusion-positive human VCaP cell line (see Supplementary Fig. S2D). Nevertheless, our findings are consistent with other reports that ERG alone is not sufficient to promote prostate tumorigenesis in mice within the studied time frame (51–53). The Pb-ERG transgenic mouse model also provides the unique ability to study AR-dependent transgenic expression of ERG, which mimics AR-driven TMPRSS2-ERG expression in human prostate tumors (35). For these reasons, this model system is particularly relevant and analogous to human prostate cancer. However, it is important to note that the reduced AR expression observed in approximately 50% of the tumors in TMT mice could contribute to absence of AR-dependent, Pb-promoter-driven expression of ERG. The exact underlying molecular mechanism warrants further investigation, and future studies will explore the exact cause-and-effect of reduced AR expression and absence of ERG expression in the ARlow/KRTlow subset of TMT mice. Additional studies with a more robust knock-in model of ERG (13) would be particularly useful to better characterize this mechanism, although slightly less physiologically relevant.

These findings in prostate cancer also raise the larger question of whether the mechanism defined in the current study might be applicable to other RB alteration–related cancer types such as lung cancer (genomic loss of RB1 promotes the transition from adenocarcinoma to small-cell lung cancer; ref. 54) and breast cancer (functional loss of RB due to HER2 amplification leads to formation of nonluminal breast cancer). Nevertheless, our findings support the evaluation of ERG fusion as a viable biomarker to guide antiandrogen and RB pathway–targeted therapies for PTEN/TP53–mutated, RB1-intact prostate cancer. Studies such as these will be essential to combat lineage plasticity-mediated therapy resistance in prostate cancer as well as other cancers.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Authors' Contributions

Conception and design: W. Xu, H. Huang

Development of methodology: A.M. Blee, Y. He, Y. Chen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.M. Blee, Y. He, Y. Yang, J. Dugdale, M. Kohli, R. Jimenez

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.M. Blee, Y. He, Z. Ye, Y. Yan, T. Ma, Y. Chen, L. Wang

Writing, review, and/or revision of the manuscript: A.M. Blee, Y. He, M. Kohli, R. Jimenez, Y. Chen, L. Wang, H. Huang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.M. Blee, Y. Pan, E. Kuehn

Study supervision: A.M. Blee, W. Xu, H. Huang

Acknowledgments

This work was supported in part by grants from NIH (CA134514, CA130908, CA203849, and CA193239; to H. Huang) and DOD (W81XWH-14-1-0486; to H. Huang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Footnotes

  • Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

  • Received February 26, 2018.
  • Revision received May 4, 2018.
  • Accepted May 23, 2018.
  • Published first May 29, 2018.

References

  1. 1.↵
    1. Watson PA,
    2. Arora VK,
    3. Sawyers CL
    . Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer. Nat Rev Cancer 2015;15:701–11.
    OpenUrlCrossRefPubMed
  2. 2.↵
    1. Barbieri CE,
    2. Baca SC,
    3. Lawrence MS,
    4. Demichelis F,
    5. Blattner M,
    6. Theurillat J-P,
    7. et al.
    Exome sequencing identifies recurrent SPOP, FOXA1 and MED12 mutations in prostate cancer. Nat Genet 2012;44:685–9.
    OpenUrlCrossRefPubMed
  3. 3.↵
    Cancer Genome Atlas Research Network. The molecular taxonomy of primary prostate cancer. Cell 2015;163:1011–25.
    OpenUrlCrossRefPubMed
  4. 4.↵
    1. Robinson D,
    2. Van Allen EM,
    3. Wu YM,
    4. Schultz N,
    5. Lonigro RJ,
    6. Mosquera JM,
    7. et al.
    Integrative clinical genomics of advanced prostate cancer. Cell 2015;161:1215–28.
    OpenUrlCrossRefPubMed
  5. 5.↵
    1. Rickman DS,
    2. Beltran H,
    3. Demichelis F,
    4. Rubin MA
    . Biology and evolution of poorly differentiated neuroendocrine tumors. Nat Med 2017;23:1–10.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Bluemn EG,
    2. Coleman IM,
    3. Lucas JM,
    4. Coleman RT,
    5. Hernandez-Lopez S,
    6. Tharakan R,
    7. et al.
    Androgen receptor pathway-independent prostate cancer is sustained through FGF signaling. Cancer Cell 2017;32:474–89e6.
    OpenUrl
  7. 7.↵
    1. Akamatsu S,
    2. Inoue T,
    3. Ogawa O,
    4. Gleave ME
    . Clinical and molecular features of treatment-related neuroendocrine prostate cancer. Int J Urol 2018;25:345–51.
    OpenUrl
  8. 8.↵
    1. Ku SY,
    2. Rosario S,
    3. Wang Y,
    4. Mu P,
    5. Seshadri M,
    6. Goodrich ZW,
    7. et al.
    Rb1 and Trp53 cooperate to suppress prostate cancer lineage plasticity, metastasis, and antiandrogen resistance. Science 2017;355:78–83.
    OpenUrlAbstract/FREE Full Text
  9. 9.↵
    1. Martin P,
    2. Liu YN,
    3. Pierce R,
    4. Abou-Kheir W,
    5. Casey O,
    6. Seng V,
    7. et al.
    Prostate epithelial Pten/TP53 loss leads to transformation of multipotential progenitors and epithelial to mesenchymal transition. Am J Pathol 2011;179:422–35.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Mu P,
    2. Zhang Z,
    3. Benelli M,
    4. Karthaus WR,
    5. Hoover E,
    6. Chen CC,
    7. et al.
    SOX2 promotes lineage plasticity and antiandrogen resistance in TP53- and RB1-deficient prostate cancer. Science 2017;355:84–8.
    OpenUrlAbstract/FREE Full Text
  11. 11.↵
    1. Zou M,
    2. Toivanen R,
    3. Mitrofanova A,
    4. Floc'h N,
    5. Hayati S,
    6. Sun Y,
    7. et al.
    Transdifferentiation as a mechanism of treatment resistance in a mouse model of castration-resistant prostate cancer. Cancer Discov 2017;7:736–49.
    OpenUrlAbstract/FREE Full Text
  12. 12.↵
    1. Mounir Z,
    2. Lin F,
    3. Lin VG,
    4. Korn JM,
    5. Yu Y,
    6. Valdez R,
    7. et al.
    TMPRSS2:ERG blocks neuroendocrine and luminal cell differentiation to maintain prostate cancer proliferation. Oncogene 2015;34:3815–25.
    OpenUrl
  13. 13.↵
    1. Chen Y,
    2. Chi P,
    3. Rockowitz S,
    4. Iaquinta PJ,
    5. Shamu T,
    6. Shukla S,
    7. et al.
    ETS factors reprogram the androgen receptor cistrome and prime prostate tumorigenesis in response to PTEN loss. Nat Med 2013;19:1023–9.
    OpenUrlCrossRefPubMed
  14. 14.↵
    1. Zegarra-Moro OL,
    2. Schmidt LJ,
    3. Huang H,
    4. Tindall DJ
    . Disruption of androgen receptor function inhibits proliferation of androgen-refractory prostate cancer cells. Cancer Res 2002;62:1008–13.
    OpenUrlAbstract/FREE Full Text
  15. 15.↵
    1. Blee AM,
    2. Liu S,
    3. Wang L,
    4. Huang H
    . BET bromodomain-mediated interaction between ERG and BRD4 promotes prostate cancer cell invasion. Oncotarget 2016;7:38319–32.
    OpenUrl
  16. 16.↵
    1. Boyer LA,
    2. Lee TI,
    3. Cole MF,
    4. Johnstone SE,
    5. Levine SS,
    6. Zucker JP,
    7. et al.
    Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 2005;122:947–56.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. He Y,
    2. Peng S,
    3. Wang J,
    4. Chen H,
    5. Cong X,
    6. Chen A,
    7. et al.
    Ailanthone targets p23 to overcome MDV3100 resistance in castration-resistant prostate cancer. Nat Commun 2016;7:13122.
    OpenUrl
  18. 18.↵
    1. Zhang H,
    2. Pan Y,
    3. Zheng L,
    4. Choe C,
    5. Lindgren B,
    6. Jensen ED,
    7. et al.
    FOXO1 inhibits Runx2 transcriptional activity and prostate cancer cell migration and invasion. Cancer Res 2011;71:3257–67.
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    1. Tammachote R,
    2. Hommerding CJ,
    3. Sinders RM,
    4. Miller CA,
    5. Czarnecki PG,
    6. Leightner AC,
    7. et al.
    Ciliary and centrosomal defects associated with mutation and depletion of the Meckel syndrome genes MKS1 and MKS3. Hum Mol Genet 2009;18:3311–23.
    OpenUrlCrossRefPubMed
  20. 20.↵
    1. Subramanian A,
    2. Tamayo P,
    3. Mootha VK,
    4. Mukherjee S,
    5. Ebert BL,
    6. Gillette MA,
    7. et al.
    Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005;102:15545–50.
    OpenUrlAbstract/FREE Full Text
  21. 21.↵
    1. Cerami E,
    2. Gao J,
    3. Dogrusoz U,
    4. Gross BE,
    5. Sumer SO,
    6. Aksoy BA,
    7. et al.
    The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov 2012;2:401–4.
    OpenUrlAbstract/FREE Full Text
  22. 22.↵
    1. Gao J,
    2. Aksoy BA,
    3. Dogrusoz U,
    4. Dresdner G,
    5. Gross B,
    6. Sumer SO,
    7. et al.
    Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 2013;6:pl1.
    OpenUrlAbstract/FREE Full Text
  23. 23.↵
    1. Beltran H,
    2. Prandi D,
    3. Mosquera JM,
    4. Benelli M,
    5. Puca L,
    6. Cyrta J,
    7. et al.
    Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer. Nat Med 2016;22:298–305.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Le X,
    2. Antony R,
    3. Razavi P,
    4. Treacy DJ,
    5. Luo F,
    6. Ghandi M,
    7. et al.
    Systematic functional characterization of resistance to PI3K inhibition in breast cancer. Cancer Discov 2016;6:1134–47.
    OpenUrlAbstract/FREE Full Text
  25. 25.↵
    1. Kim D,
    2. Pertea G,
    3. Trapnell C,
    4. Pimentel H,
    5. Kelley R,
    6. Salzberg SL
    . TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol 2013;14:R36.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Wang L,
    2. Wang S,
    3. Li W
    . RSeQC: quality control of RNA-seq experiments. Bioinformatics 2012;28:2184–5.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Robinson MD,
    2. McCarthy DJ,
    3. Smyth GK
    . edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 2010;26:139–40.
    OpenUrlCrossRefPubMed
  28. 28.↵
    1. Langmead B,
    2. Salzberg SL
    . Fast gapped-read alignment with Bowtie 2. Nat Methods 2012;9:357–9.
    OpenUrlCrossRefPubMed
  29. 29.↵
    1. Wang L,
    2. Chen J,
    3. Wang C,
    4. Uuskula-Reimand L,
    5. Chen K,
    6. Medina-Rivera A,
    7. et al.
    MACE: model based analysis of ChIP-exo. Nucleic Acids Res 2014;42:e156.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. McLean CY,
    2. Bristor D,
    3. Hiller M,
    4. Clarke SL,
    5. Schaar BT,
    6. Lowe CB,
    7. et al.
    GREAT improves functional interpretation of cis-regulatory regions. Nat Biotechnol 2010;28:495–501.
    OpenUrlCrossRefPubMed
  31. 31.↵
    1. Yu J,
    2. Yu J,
    3. Mani R-S,
    4. Cao Q,
    5. Brenner CJ,
    6. Cao X,
    7. et al.
    An integrated network of androgen receptor, polycomb, and TMPRSS2-ERG gene fusions in prostate cancer progression. Cancer Cell 2010;17:443–54.
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. He Y,
    2. Zhao Y,
    3. Wang L,
    4. Bohrer LR,
    5. Pan Y,
    6. Wang L,
    7. et al.
    LSD1 promotes S-phase entry and tumorigenesis via chromatin co-occupation with E2F1 and selective H3K9 demethylation. Oncogene 2018;37:534–43.
    OpenUrl
  33. 33.↵
    1. Chen Z,
    2. Lan X,
    3. Thomas-Ahner JM,
    4. Wu D,
    5. Liu X,
    6. Ye Z,
    7. et al.
    Agonist and antagonist switch DNA motifs recognized by human androgen receptor in prostate cancer. EMBO J 2015;34:502–16.
    OpenUrlAbstract/FREE Full Text
  34. 34.↵
    1. Wu X,
    2. Wu J,
    3. Huang J,
    4. Powell WC,
    5. Zhang J,
    6. Matusik RJ,
    7. et al.
    Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation. Mech Dev 2001;101:61–9.
    OpenUrlCrossRefPubMed
  35. 35.↵
    1. Klezovitch O,
    2. Risk M,
    3. Coleman I,
    4. Lucas JM,
    5. Null M,
    6. True LD,
    7. et al.
    A causal role for ERG in neoplastic transformation of prostate epithelium. Proc Natl Acad Sci U S A 2008;105:2105–10.
    OpenUrlAbstract/FREE Full Text
  36. 36.↵
    1. Wang S,
    2. Gao J,
    3. Lei Q,
    4. Rozengurt N,
    5. Pritchard C,
    6. Jiao J,
    7. et al.
    Prostate-specific deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer Cell 2003;4:209–21.
    OpenUrlCrossRefPubMed
  37. 37.↵
    1. Olive KP,
    2. Tuveson DA,
    3. Ruhe ZC,
    4. Yin B,
    5. Willis NA,
    6. Bronson RT,
    7. et al.
    Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome. Cell 2004;119:847–60.
    OpenUrlCrossRefPubMed
  38. 38.↵
    1. Tomlins SA,
    2. Rhodes DR,
    3. Perner S,
    4. Dhanasekaran SM,
    5. Mehra R,
    6. Sun XW,
    7. et al.
    Recurrent fusion of TMPRSS2 and ETS transcription factor genes in prostate cancer. Science 2005;310:644–8.
    OpenUrlAbstract/FREE Full Text
  39. 39.↵
    1. Kumar A,
    2. Coleman I,
    3. Morrissey C,
    4. Zhang X,
    5. True LD,
    6. Gulati R,
    7. et al.
    Substantial interindividual and limited intraindividual genomic diversity among tumors from men with metastatic prostate cancer. Nat Med 2016;22:369–78.
    OpenUrlCrossRef
  40. 40.↵
    1. Shi XB,
    2. Nesslinger NJ,
    3. Deitch AD,
    4. Gumerlock PH,
    5. deVere White RW
    . Complex functions of mutant p53 alleles from human prostate cancer. Prostate 2002;51:59–72.
    OpenUrlCrossRefPubMed
  41. 41.↵
    1. Chen Z,
    2. Trotman LC,
    3. Shaffer D,
    4. Lin HK,
    5. Dotan ZA,
    6. Niki M,
    7. et al.
    Crucial role of p53-dependent cellular senescence in suppression of Pten-deficient tumorigenesis. Nature 2005;436:725–30.
    OpenUrlCrossRefPubMed
  42. 42.↵
    1. Di Cristofano A,
    2. Pesce B,
    3. Cordon-Cardo C,
    4. Pandolfi PP
    . Pten is essential for embryonic development and tumour suppression. Nat Genet 1998;19:348–55.
    OpenUrlCrossRefPubMed
  43. 43.↵
    1. Podsypanina K,
    2. Ellenson LH,
    3. Nemes A,
    4. Gu J,
    5. Tamura M,
    6. Yamada KM,
    7. et al.
    Mutation of Pten/Mmac1 in mice causes neoplasia in multiple organ systems. Proc Natl Acad Sci U S A 1999;96:1563–8.
    OpenUrlAbstract/FREE Full Text
  44. 44.↵
    1. Suzuki A,
    2. de la Pompa JL,
    3. Stambolic V,
    4. Elia AJ,
    5. Sasaki T,
    6. del Barco Barrantes I,
    7. et al.
    High cancer susceptibility and embryonic lethality associated with mutation of the PTEN tumor suppressor gene in mice. Curr Biol 1998;8:1169–78.
    OpenUrlCrossRefPubMed
  45. 45.↵
    1. Charafe-Jauffret E,
    2. Ginestier C,
    3. Monville F,
    4. Finetti P,
    5. Adelaide J,
    6. Cervera N,
    7. et al.
    Gene expression profiling of breast cell lines identifies potential new basal markers. Oncogene 2006;25:2273–84.
    OpenUrlCrossRefPubMed
  46. 46.↵
    1. Gao S,
    2. Gao Y,
    3. He HH,
    4. Han D,
    5. Han W,
    6. Avery A,
    7. et al.
    Androgen receptor tumor suppressor function is mediated by recruitment of retinoblastoma protein. Cell Rep 2016;17:966–76.
    OpenUrl
  47. 47.↵
    1. Thangavel C,
    2. Boopathi E,
    3. Liu Y,
    4. Haber A,
    5. Ertel A,
    6. Bhardwaj A,
    7. et al.
    RB loss promotes prostate cancer metastasis. Cancer Res 2017;77:982–95.
    OpenUrlAbstract/FREE Full Text
  48. 48.↵
    1. Bollard J,
    2. Miguela V,
    3. Ruiz de Galarreta M,
    4. Venkatesh A,
    5. Bian CB,
    6. Roberto MP,
    7. et al.
    Palbociclib (PD-0332991), a selective CDK4/6 inhibitor, restricts tumour growth in preclinical models of hepatocellular carcinoma. Gut 2017;66:1286–96.
    OpenUrlAbstract/FREE Full Text
  49. 49.↵
    1. Comstock CE,
    2. Augello MA,
    3. Goodwin JF,
    4. de Leeuw R,
    5. Schiewer MJ,
    6. Ostrander WF Jr.,
    7. et al.
    Targeting cell cycle and hormone receptor pathways in cancer. Oncogene 2013;32:5481–91.
    OpenUrlCrossRefPubMed
  50. 50.↵
    1. Finn RS,
    2. Martin M,
    3. Rugo HS,
    4. Jones S,
    5. Im SA,
    6. Gelmon K,
    7. et al.
    Palbociclib and letrozole in advanced breast cancer. N Engl J Med 2016;375:1925–36.
    OpenUrlCrossRef
  51. 51.↵
    1. Yang Y,
    2. Blee AM,
    3. Wang D,
    4. An J,
    5. Pan Y,
    6. Yan Y,
    7. et al.
    Loss of FOXO1 cooperates with TMPRSS2-ERG overexpression to promote prostate tumorigenesis and cell invasion. Cancer Res 2017;77:6524–37.
    OpenUrlAbstract/FREE Full Text
  52. 52.↵
    1. Carver BS,
    2. Tran J,
    3. Gopalan A,
    4. Chen Z,
    5. Shaikh S,
    6. Carracedo A,
    7. et al.
    Aberrant ERG expression cooperates with loss of PTEN to promote cancer progression in the prostate. Nat Genet 2009;41:619–24.
    OpenUrlCrossRefPubMed
  53. 53.↵
    1. King JC,
    2. Xu J,
    3. Wongvipat J,
    4. Hieronymus H,
    5. Carver BS,
    6. Leung DH,
    7. et al.
    Cooperativity of TMPRSS2-ERG with PI3-kinase pathway activation in prostate oncogenesis. Nat Genet 2009;41:524–6.
    OpenUrlCrossRefPubMed
  54. 54.↵
    1. Niederst MJ,
    2. Sequist LV,
    3. Poirier JT,
    4. Mermel CH,
    5. Lockerman EL,
    6. Garcia AR,
    7. et al.
    RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer. Nat Commun 2015;6:6377.
    OpenUrlCrossRefPubMed
View Abstract
Back to top
View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
TMPRSS2-ERG Controls Luminal Epithelial Lineage and Antiandrogen Sensitivity in PTEN and TP53-Mutated Prostate Cancer
Alexandra M. Blee, Yundong He, Yinhui Yang, Zhenqing Ye, Yuqian Yan, Yunqian Pan, Tao Ma, Joseph Dugdale, Emily Kuehn, Manish Kohli, Rafael Jimenez, Yu Chen, Wanhai Xu, Liguo Wang and Haojie Huang
Clin Cancer Res September 15 2018 (24) (18) 4551-4565; DOI: 10.1158/1078-0432.CCR-18-0653
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement