Improved clinical management of prostate cancer has been impeded by an inadequate understanding of molecular genetic elements governing tumor progression. Gene signatures have provided improved prognostic indicators of human prostate cancer. The TGF-β/BMP-SMAD4 signaling pathway, which induces epithelial–mesenchymal transition (EMT), is known to constrain prostate cancer progression induced by Pten deletion. Herein, cyclin D1 inactivation reduced cellular proliferation in the murine prostate in vivo and in isogenic oncogene–transformed prostate cancer cell lines. The in vivo cyclin D1–mediated molecular signature predicted poor outcome of recurrence-free survival for patients with prostate cancer (K-means HR, 3.75, P = 0.02) and demonstrated that endogenous cyclin D1 restrains TGF-β, Snail, Twist, and Goosecoid signaling. Endogenous cyclin D1 enhanced Wnt and ES cell gene expression and expanded a prostate stem cell population. In chromatin immunoprecipitation sequencing, cyclin D1 occupied genes governing stem cell expansion and induced their transcription. The coordination of EMT restraining and stem cell expanding gene expression by cyclin D1 in the prostate may contribute to its strong prognostic value for poor outcome in biochemical-free recurrence in human prostate cancer. Cancer Res; 74(2); 508–19. ©2013 AACR.

Adenocarcinoma of the prostate is the second leading cause of death in American men (1). The stratification of patient's tumors to determine outcome have included serum prostate-specific antigen levels and clinical staging with histopathological criterion, including the Gleason grade (2). Efforts to improve the predictive value of risk progression markers have led to molecular genetic marker analysis in which gene expression signatures have provided prognostic information (3, 4). Unsupervised clustering analysis identified a subset of tumors manifesting a stem cell–like signature associated with poor clinical outcome (5).

An analysis of prostate cancer genetic drivers has identified the importance of increased androgen receptor (AR) activity, loss of the PTEN tumor suppressor, gain of chromosomal fusions, such as TMRSS2-ERG, epithelial–mesenchymal transition (EMT), increased prostate cancer progenitor cells or cancer “stem cells,” and signaling modules that impact cell-cycle control and cellular survival. Androgens increase cellular proliferation of prostatic epithelial cells via the AR. Androgen-deprivation therapy is an important form of treatment for most patients with prostate cancer (6). However, many tumors regrow after 12 to 18 months. The loss of one or both copies of PTEN is commonly found in prostate cancer and TMPRSS2-ERG chromosomal fusion occurs in 50% to 60% of prostate cancer (7, 8). Molecular genetic analysis in mice has confirmed clinical observation demonstrating key genetic drivers mediating the onset and progression of prostate cancer, including the AR, Pten, c-Myc, and Ras. Analysis of the molecular genetic events constraining prostate cancer progression in Pten−/− mice identified prominent TGF-β/BMP-SMAD4 signaling and molecular analysis identified a 4-gene signature (cyclin D1, SPP1, PTEN, and SMAD4) as predictive of poor outcome (9).

The cell-cycle control protein cyclin D1, recently identified as a component of a 4-gene signature that predicted poor clinical outcome in prostate cancer (9), encodes a gene with both canonical and noncanonical functions (10). In its canonical function as the regulatory subunit of a holoenzyme, cyclin D1 leads to phosphorylation of the pRb proteins thereby inhibiting DNA synthesis and phosphorylates the NRF1 protein to inhibit mitochondrial biogenesis (11–13). The noncanonical functions of cyclin D1 include the regulation of transcription factor activity and the recruitment of transcription factors in the context of local chromatin in vivo (14). Contradictory results have been published on the effect of cyclin D1 in LNCaP cells; either inducing a proliferative or antiproliferative effect (15). Cyclin D1 overexpression in LNCaP cells enhanced S-phase entry, increased colony formation and tumor growth rate in nude mice (16), and siRNA to cyclin D1 reduced growth factor–induced cell-cycle progression (17). In contrast, transfection of an expression vector encoding cyclin D1 or a fragment of cyclin D1, encoding the previously defined “repressor domain” of cyclin D1 (18), inhibited LNCaP DNA synthesis (19).

Progression of prostate cancer includes populations of tumor-initiating cells (TIC), which have self-renewal potential, are therapy resistant, and contribute to tumor metastasis. Factors that regulate stem/progenitor function are altered in prostate cancer. The phosphoinositide 3-kinase pathway (20) and NF-κB (21) activation promote self-renewal and contribute to prostate malignancy. Wnt and Notch pathway governs the balance of progenitor self-renewal and differentiation (22) with Wnt/β-catenin promoting prostate epithelial cell hyperplasia and stem cell self-renewal (23). TIC attributes can be suppressed by EMT in prostate cancer cell lines (24). In this regard, knockdown of EMT factors in mesenchymal-like prostate cancer cells induces TIC. TGF-β/BMP-SMAD signaling is an important inducer of EMT. EMT is triggered during both embryonic development and tumor progression, by a variety of factors, including the Snail family members (Snail and Slug). Snail blocks the cell cycle and induces EMT through repression of E-cadherin transcription. Snail represses components of the cell cycle regulating G1–S transition, repressing cyclin D1 and cyclin D2, and increasing p21CIP1 (25). Several direct transcriptional repressors of E-cadherin (Snail, Slug, ZEB1, SIP1, and E47) act downstream of EMT. EMT-induced signal transduction pathways, including TGF-β, growth factors, and hypoxia, function through this EMT regulatory genetic network (26). The cell-cycle arrest that occurs during EMT creates a paradox, as tumor progression requires continued cellular growth and cyclin D1 downregulation is necessary for EMT induction in epidermoid cells.

In view of the contradictory data on the role of cyclin D1 in regulating prostate cancer cellular proliferation in tissue culture, we conducted a careful analysis of cyclin D1−/− mice in response to androgen ablation and subsequent replacement. We conducted experiments using cyclin D1 siRNA and cyclin D1 short hairpin RNA (shRNA) in isogenic murine and human prostate cancer cell lines and prostate tissue in vivo using cyclin D1−/− mice. Endogenous cyclin D1 enhanced prostate cellular proliferation in vivo and in prostate cancer cells in tissue culture. An in vivo prostate cyclin D1–mediated gene signature was defined and used to interrogate data sets of clinical outcome. The cyclin D1 signature was highly predictive of patient outcome assessed by biochemical recurrence (BCR) of prostate cancer. We show that the cyclin D1–mediated gene signature is anticorrelated within EMT signaling, whether induced by Twist, Snail, GSC, shE-cadherin, or by TGF-β. Conversely, the cyclin D1–signature, which reflects Wnt signaling, was enriched in prostatospheres and cyclin D1 shRNA reduced prostatosphere formation. These studies provide further evidence for the dynamic interaction among epithelial self-renewal and mesenchymal gene expression programs in determining TIC expansion and place cyclin D1 at the nexus between these two key processes in the prostate.

Cell culture, DNA transfection, and luciferase assays

Culture of LNCaP cells and the isogenic c-Myc and NeuT murine prostate cancer cell lines, DNA transfection, and luciferase assays were performed as previously described (27). The Nanog and SOX2 promoter luciferase reporter plasmids were described in ref. 28.

Mice, chemical reagents, and Western blotting

Experimental procedures with transgenic mice were approved by the ethics committee of Thomas Jefferson University. Cyclin D1−/− mice were in the FVB strain.

Statistical analysis

Comparisons between groups were analyzed by 2-sided t-test. A difference of P < 0.05 was considered to be statistically significant. All analyses were done with SPSS 11.5 software. Data are expressed as mean ± SEM.

RNA extraction from ventral prostate, quantitative real-time PCR

RNA samples were extracted from prostate tissues using the RecoverAll Total Nucleic Acid Isolation Kit for Formalin-fixed Paraffin Embedded (FFPE) tissues (Applied Biosystems). This was followed by RQ1 DNase I (Promega Inc.) mediated removal of contaminating DNA from RNA preparations followed by RNA clean-up using the RNAEasy Kit (Qiagen Inc.). Equal amounts of purified RNA were reverse transcribed into cDNA using the Iscript Reverse Transcriptase Kit (Bio-Rad).

Prostatosphere assays

Prostatosphere assays were conducted as previously described (29) with minor modification. LNCaP cells were cultured in Dulbecco's Modified Eagle Medium (DMEM)/F12 medium with B-27 and N2 supplement (Invitrogen), plated at 4,000 cells/mL (4 mL/plate) on 6-cm ultralow attachment plate (Corning). The prostatospheres were then cultured at 37°C, 5% CO2 for 10 days. The prostatospheres were counted under a microscope.

Endogenous cyclin D1 is required for DHT-induced prostate cellular proliferation in vivo

To examine the role of cyclin D1 in prostate cellular proliferation in vivo, a comparison was made between cyclin D1+/+ and cyclin D1−/− mice. The individual organs were weighed and the mean data demonstrated a reduction in the size of the ventral prostate and testes of the cyclin D1−/− mice (Supplementary Fig. S1A). The mean body weight was also reduced in the cyclin D1−/− mice and both the prostate and testes were unchanged when normalized for body weight (Fig. 1A). To assess further the role of cyclin D1 in dihydrotestosterone (DHT)-dependent function, animals were subjected to a castration and DHT treatment protocol (Supplementary Fig. S1B). Cyclin D1+/+ and cyclin D1−/− mice were subjected to surgical castration and after 2 weeks, treated for the subsequent 6 days with testosterone, the prostates were harvested and analyzed. The cyclin D1−/− mice showed no positivity for cyclin D1 (Fig. 1A). Immunohistochemical staining, commensurate with the increase in cell number is response to DHT, demonstrated an induction of cyclin D1 with DHT (Fig. 1A). Hematoxylin and eosin (H&E)–stained sections of age-matched cyclin D1+/+ versus cyclin D1−/− mice ventral prostates with vehicle control showed no significant difference in histology (Fig. 1B). The relative distribution of immunohistochemical staining for CK5 and CK8 markers of the basal and luminal cell type was unaltered in the cyclin D1−/− mice (data not shown). DHT treatment increased the luminal epithelial cell population; however, the response was reduced in cyclin D1−/− mice (Fig. 1B). DHT treatment induced Ki-67 staining 3-fold in the cyclin D1+/+ mice; however, there was no significant change in Ki-67 staining in DHT-treated cyclin D1−/− littermate controls (Fig. 1C and D). Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining to assess apoptosis showed an approximately 20% increase in apoptosis with DHT treatment (Fig. 1E and F). Apoptosis determined by TUNEL was increased in the cyclin D1−/− prostate compared with cyclin D1+/+ prostate. DHT reduced apoptosis in the cyclin D1−/− prostate (Fig. 1E and F). Together, these studies demonstrate an important role for cyclin D1 in prostate cellular proliferation and survival in vivo. Apoptosis may contribute to the reduced proliferation in cyclin D1−/− prostate cells.

Figure 1.

Cyclin D1 is required for DHT-induced prostate cellular proliferation in vivo. A, immunohistochemical staining for cyclin D1 in cyclin D1+/+ mice showed a robust induction of cyclin D1 by DHT in the epithelial compartment. In cyclin D1−/− mice, no cyclin D1 positive cells were observed. B, histologic evaluation of H&E-stained ventral prostates from castrated cyclin D1+/+ versus cyclin D1−/− mice treated with either vehicle or DHT. C, immunohistochemistry of Cyclin D1+/+ mice ventral prostates, comparing vehicle to DHT-treated animals, show an induction in Ki-67 staining and quantitated in D. In cyclin D1−/− animals, the Ki-67 signal in response to DHT is blunted. E, TUNEL staining as a marker of apoptosis with data quantitated in F as mean ± SEM for N = 3 separate mice in each category.

Figure 1.

Cyclin D1 is required for DHT-induced prostate cellular proliferation in vivo. A, immunohistochemical staining for cyclin D1 in cyclin D1+/+ mice showed a robust induction of cyclin D1 by DHT in the epithelial compartment. In cyclin D1−/− mice, no cyclin D1 positive cells were observed. B, histologic evaluation of H&E-stained ventral prostates from castrated cyclin D1+/+ versus cyclin D1−/− mice treated with either vehicle or DHT. C, immunohistochemistry of Cyclin D1+/+ mice ventral prostates, comparing vehicle to DHT-treated animals, show an induction in Ki-67 staining and quantitated in D. In cyclin D1−/− animals, the Ki-67 signal in response to DHT is blunted. E, TUNEL staining as a marker of apoptosis with data quantitated in F as mean ± SEM for N = 3 separate mice in each category.

Close modal

Endogenous cyclin D1 determines prostate cancer cell proliferation

To determine the role of endogenous cyclin D1 in DHT-mediated cellular proliferation, 3 AR expressing lines were examined (LNCaP, NeuT-PEC, c-Myc-PEC). The c-Myc and NeuT lines are isogenic lines we derived through retroviral transduction of primary murine (FVB) prostate epithelium (27). Western blot analysis demonstrated the reduction of cyclin D1 abundance in the cyclin D1 siRNA-transduced lines (Fig. 2A–C). In the LNCaP cell line, cyclin D1 siRNA reduced the relative abundance of the AR approximately 55% (Fig. 2A). In the NeuT-PEC line, cyclin D1 siRNA reduced cyclin D1 abundance approximately 90% and reduced DHT-induced AR abundance approximately 80% (Fig. 2B). Cyclin D1 siRNA reduced cyclin D1 abundance approximately 50% in c-Myc-PEC lines, but did not affect AR induction by DHT (Fig. 2C). Cell counting demonstrated a significant reduction in DHT-induced cellular proliferation in each of the 3 prostate cancer cell lines (Fig. 2D–F). Together, these studies indicate endogenous cyclin D1 enhances DHT-induced prostate cancer cellular proliferation and that the role of cyclin D1 in the regulation of AR abundance is prostate cancer cell line specific.

Figure 2.

Endogenous cyclin D1 maintains prostate cancer cell proliferation. Three AR positive prostate cancer lines LNCaP (A), NeuT-PEC (B), and c-Myc-PEC (C) were stably transduced with either control-siRNA or cyclin D1–siRNA (80 nmol/L). Control-siRNA and D1 siRNA lines were treated with DHT (10 nmol/L) or vehicle, protein lysates recovered, and Western blot analysis was conducted for abundance of cyclin D1 and AR. Proliferation assays were conducted on control-siRNA and cyclin D1 siRNA lines from LNCaP (D), NeuT-PEC (E), and c-Myc-PEC treated with DHT or vehicle (F). Cell numbers were collected daily over a period of 6 days following cyclin D1 knockdown.

Figure 2.

Endogenous cyclin D1 maintains prostate cancer cell proliferation. Three AR positive prostate cancer lines LNCaP (A), NeuT-PEC (B), and c-Myc-PEC (C) were stably transduced with either control-siRNA or cyclin D1–siRNA (80 nmol/L). Control-siRNA and D1 siRNA lines were treated with DHT (10 nmol/L) or vehicle, protein lysates recovered, and Western blot analysis was conducted for abundance of cyclin D1 and AR. Proliferation assays were conducted on control-siRNA and cyclin D1 siRNA lines from LNCaP (D), NeuT-PEC (E), and c-Myc-PEC treated with DHT or vehicle (F). Cell numbers were collected daily over a period of 6 days following cyclin D1 knockdown.

Close modal

Cyclin D1 determines basal and androgen-dependent gene expression in vivo

To characterize the genes regulated by DHT in a cyclin D1–dependent manner, genome-wide microarray analysis was conducted of the mRNA derived from mice subjected to castration and androgen replacement (Fig. 3A and Supplementary Fig. S2). The log 2-fold change versus Log2 mean intensity was well distributed. A total of 328 genes were differentially expressed between the cyclin D1−/− versus cyclin D1+/+ mice (Fig. 3A and Supplementary Table S1). DHT regulated the expression of 2,253 genes in the cyclin D1+/+ mice prostate (Supplementary Fig. S2 and Supplementary Table S1). The expression of cyclin D1 also determined DHT-mediated gene expression, both induction and repression (Supplementary Fig. S2A), as the cyclin D1−/− mice prostate epithelium showed DHT-mediated alteration in expression of only 186 genes (Supplementary Table S1). Thus, approximately 2,126 genes were regulated by DHT in a manner that was dependent upon endogenous cyclin D1. The function of genes regulated by cyclin D1 examined using KEGG analysis, demonstrated the induction of ECM and cell adhesion, cell cycle and Wnt pathway signaling, and the induction of cancer pathways (bladder, glioma, pancreatic, colorectal, endometrial, chronic myeloid leukemia; Fig. 3C). Cyclin D1 inhibited pathways involved in metabolism, dependent upon mitochondria (Fig. 3C), consistent with the known ability of cyclin D1 to inhibit mitochondrial metabolism in other tissues, including fibroblasts and the mammary gland (11, 12). DHT induced and repressed pathways known to be regulated by DHT (Supplementary Fig. S3A and S3B), however, the vast majority of these pathways were not regulated by DHT in the cyclin D1−/− mice (Supplementary Fig. S3C and S3D). The genes associated with the pathways induced by cyclin D1 in vivo [including ECM, cell adhesion receptors and genes involved in promoting Wnt/β-catenin signaling, and stem cell function (EphA2, Wnt7A, SOX4)], are shown in Fig. 3D.

Figure 3.

Microarray of mRNA expression from cyclin D1+/+ and cyclin D1−/− prostate mice. A, pie chart of 328 genes differentially regulated between cyclin D1+/+ mouse prostates (n = 3) and cyclin D1−/− mouse prostates (n = 3; P < 0.05, fold change ≥2.0). Red signifies upregulated and green downregulated expression. B, unsupervised hierarchical clustering of 328 genes differentially regulated between cyclin D1+/+ mouse prostates and cyclin D1−/− mouse prostates. C, molecular pathways regulated by cyclin D1 (KEGG); pathways depicted are graphically represented using the enrichment score (ES score). D, functional annotation clustering of genes induced by cyclin D1 in vivo. The clusters are labeled with suitable terms that describe the function; clusters 1 and 8 contain gene pathway terms also identified by KEGG analysis.

Figure 3.

Microarray of mRNA expression from cyclin D1+/+ and cyclin D1−/− prostate mice. A, pie chart of 328 genes differentially regulated between cyclin D1+/+ mouse prostates (n = 3) and cyclin D1−/− mouse prostates (n = 3; P < 0.05, fold change ≥2.0). Red signifies upregulated and green downregulated expression. B, unsupervised hierarchical clustering of 328 genes differentially regulated between cyclin D1+/+ mouse prostates and cyclin D1−/− mouse prostates. C, molecular pathways regulated by cyclin D1 (KEGG); pathways depicted are graphically represented using the enrichment score (ES score). D, functional annotation clustering of genes induced by cyclin D1 in vivo. The clusters are labeled with suitable terms that describe the function; clusters 1 and 8 contain gene pathway terms also identified by KEGG analysis.

Close modal

Cyclin D1 downregulation is required for EMT induction; however, the microarray gene expression analysis identified genes that either promote (Sox4) or inhibit (DKK) EMT (26). Therefore, to determine the effect of cyclin D1 genetic deletion on EMT in the prostate in vivo, we conducted analysis of key markers of EMT phenotype (Fig. 4). Initially, we compared the genes differentially regulated between the cyclin D1−/− and cyclin D1+/+ prostate to those known to regulate EMT. Taube and colleagues (30) defined a gene signature shared by upregulation of GSC (Goosecoid), Snail, and Twist. Forty genes regulated by cyclin D1 are concordantly regulated in the EMT core signature (Fig. 4A). Quantitative real-time PCR (qRT-PCR) of 3 separate mice of each genotype was conducted. In the cyclin D1−/− prostate, the mRNA abundance of Snail was induced 3.5-fold, Slug was increased 2-fold (Fig. 4B). Immunohistochemical staining demonstrated increased abundance of Snail (Fig. 4C) and N-cadherin (Fig. 4D).

Figure 4.

Cyclin D1–induced gene expression signatures are inversely correlated with EMT-induced expression. A, genes concordant in directionality to genes downregulated in EMT and upregulated in EMT based on a core EMT signature published by Taube and colleagues (fold >1.2 and P < 0.05; t test; ref. 30). B, relative mRNA abundance for genes governing EMT, including Snail, Slug, with data shown as mean ± SEM for 3 separate mice of each genotype. C and D, immunohistochemical staining of ventral prostate from cyclin D1+/+ mouse versus cyclin D1−/− mouse for Snail (C) or N-cadherin (D). Data are quantitated as mean ± SEM for three separate mice of each genotype. E and G, mean expression (E) and individual arrays (G) of cyclin D1 induced gene signature in EMT-inducing conditions versus control. Cyclin D1–induced gene signature is significantly downregulated in mesenchymal relative to epithelial cells in Gsc, shEcad, Snail, TGF-β, and Twist-induced EMT. F and H, mean expression (F) and individual arrays (H) of cyclin D1 repressed signature in EMT-inducing conditions versus control. Cyclin D1–repressed signature is significantly upregulated in mesenchymal cells in the induced EMT signature.

Figure 4.

Cyclin D1–induced gene expression signatures are inversely correlated with EMT-induced expression. A, genes concordant in directionality to genes downregulated in EMT and upregulated in EMT based on a core EMT signature published by Taube and colleagues (fold >1.2 and P < 0.05; t test; ref. 30). B, relative mRNA abundance for genes governing EMT, including Snail, Slug, with data shown as mean ± SEM for 3 separate mice of each genotype. C and D, immunohistochemical staining of ventral prostate from cyclin D1+/+ mouse versus cyclin D1−/− mouse for Snail (C) or N-cadherin (D). Data are quantitated as mean ± SEM for three separate mice of each genotype. E and G, mean expression (E) and individual arrays (G) of cyclin D1 induced gene signature in EMT-inducing conditions versus control. Cyclin D1–induced gene signature is significantly downregulated in mesenchymal relative to epithelial cells in Gsc, shEcad, Snail, TGF-β, and Twist-induced EMT. F and H, mean expression (F) and individual arrays (H) of cyclin D1 repressed signature in EMT-inducing conditions versus control. Cyclin D1–repressed signature is significantly upregulated in mesenchymal cells in the induced EMT signature.

Close modal

Cyclin D1–induced genes correlate with gene expression signatures of EMT restraint

Induction of the EMT-inducing TGF-β/BMT-Smad pathway is a restraining force in prostate cancer in mice (9). To examine the role the cyclin D1 gene expression network, we compared gene expression signatures governed by the expression of known EMT-inducing agents, including goosecoid (GSC), shRNA to E-cadherin, Snail, Twist, or TGF-β with the genes induced by cyclin D1 (Fig. 4E; ref. 30). The gene network induced by cyclin D1 was downregulated in EMT-producing conditions relative to control (Fig. 4E and G; ANOVA, P = 1.38 × 10−7; Gsc vs. control, P = 8.68 × 10−8; shEcad vs. control, P = 9.5 × 10−4; Snail vs. control, P = 6.15 × 10−7; TGF-β vs. control, P = 4.5 × 10−5; Twist vs. control, P = 1.7 × 10−5). Similarly, the expression signature of cyclin D1 repressed genes was upregulated in EMT produced by expression of Gsc (Fig. 4F and H; ANOVA, P = 1.01 × 10−6; Gsc vs. control, P = 1.7 × 10−6). Collectively, these studies demonstrate the gene expression profile induced by cyclin D1 expression correlated with gene expression that was reduced during EMT. This finding is consistent with the role of cyclin D1–dependent gene expression in restraining EMT gene expression in the prostate in vivo.

Cyclin D1–induced gene signature predicts BCR of human prostate cancer

Evidence suggests that cyclin D1 is a driver of prostate cancer progression in murine models (9, 17). We examined the hypothesis that cyclin D1–regulated gene expression may, therefore, predict risk of BCR in human prostate cancer samples. Genes up- or downregulated by cyclin D1 were used to generate a signature of cyclin D1 activity. Investigation of gene expression in human prostate tumors indicated that the majority of genes induced by cyclin D1 (cluster 2 in red) are expressed at a higher level in the high-risk cohort (Fig. 5A). Cyclin D1–repressed genes also correlated with poor outcome by BCR (cluster 2 in red; Fig. 5C and D). We compared the cyclin D1 signature defined herein against a 4-gene signature capable of predicting BCR or metastatic lethal outcome (Fig. 5E; ref. 9) using Cox regression, Kaplan–Meier analysis, and the concordance index. K-means clustering of samples according to expression of cyclin D1–induced genes classifies tumors into 2 groups associated with a significant difference in risk of BCR (K-means HR = 3.75, P = 0.002; Fig. 5B and E). Similar analysis with the signature based on cyclin D1–repressed genes identified tumors with a difference in risk that approached significance (K-means HR = 2.02, P = 0.07; Fig. 5D and E). The signature of cyclin D1–induced genes compares favorably to this 4-gene signature in Kaplan–Meier analysis (4-gene: HR = 2.6, P = 0.012; cyclin D1–induced signature: HR = 3.75, P = 0.002) and concordance index (4-gene: C-statistic = 0.75; cyclin D1–induced gene signature: C-index = 0.77).

Figure 5.

Expression signatures of cyclin D1 activity are associated with poor prognosis. A, heat map of expression of cyclin D1–induced gene signature in human prostate cancer patients. Rows, genes; columns, tumors. Green points, downregulation; red points, upregulation. K-means clustering identifies 2 patient clusters. The sample color bar indicates the classification of tumors into each cluster (black, cluster 1; red, cluster 2). The cyclin D1–induced gene signature is upregulated in cluster 2 tumors. B, the Kaplan–Meier plot of the risk of BCR for tumors clustered in Fig. 5A. Cluster 1 patients have significantly less risk of BCR. C, heat map of expression of cyclin D1-repressed gene signature in human prostate cancer patients. D, the Kaplan–Meier plot of the risk of BCR for tumors clustered in C. E, comparison of cyclin D1 signatures with a 4-gene signature, recently described in ref. 9.

Figure 5.

Expression signatures of cyclin D1 activity are associated with poor prognosis. A, heat map of expression of cyclin D1–induced gene signature in human prostate cancer patients. Rows, genes; columns, tumors. Green points, downregulation; red points, upregulation. K-means clustering identifies 2 patient clusters. The sample color bar indicates the classification of tumors into each cluster (black, cluster 1; red, cluster 2). The cyclin D1–induced gene signature is upregulated in cluster 2 tumors. B, the Kaplan–Meier plot of the risk of BCR for tumors clustered in Fig. 5A. Cluster 1 patients have significantly less risk of BCR. C, heat map of expression of cyclin D1-repressed gene signature in human prostate cancer patients. D, the Kaplan–Meier plot of the risk of BCR for tumors clustered in C. E, comparison of cyclin D1 signatures with a 4-gene signature, recently described in ref. 9.

Close modal

A similar analysis was used to evaluate the association of genes regulated by Cyclin D1 and AR signaling. Genes up- or downregulated by cyclin D1 or DHT were identified and evaluated for association with BCR as described above. A majority of the genes induced by cyclin D1 are also repressed by DHT. As such, a similar relationship between K-means clusters and risk of BCR is observed (HR = 3.75, P = 0.002). Conversely, an expression signature defined from genes repressed by cyclin D1 and induced by DHT is also associated with risk of BCR (HR, 2.21, P = 0.03). Collectively, these findings suggest that the signature of genes regulated by endogenous cyclin D1 correlates with poor patient prognosis and enhances the HR associated with tumors classified by a cyclin D1 expression signature in comparison to a 4-gene signature from 2.6 to 3.75 (P < 0.002).

Cyclin D1 expression network is enriched in stem cells and prostatospheres

Studies of prostate cancer cells have suggested EMT induction is associated with suppression of TIC properties based on gene expression (24). We examined the gene expression profile of prostatospheres and compared those to the gene signature induced by cyclin D1 in the prostate in vivo (Fig. 6A and B; ref. 31). In this data set, the expression signature of cyclin D1–repressed genes is downregulated in prostatospheres relative to parental cell lines (Fig. 6A). Thirty-six genes were upregulated by cyclin D1 and expressed at a higher level in prostatospheres relative to parental cell lines. Conversely, 118 genes are downregulated by cyclin D1 and expressed at a lower level in prostatospheres relative to parental cell lines (Fig. 6B). Cyclin D1 may promote stem cell behavior by repressing the expression of these genes.

Figure 6.

Cyclin D1 enhances prostate cancer stem cell populations in vitro and cyclin D1 DNA-bound form induces prostate stem cell targets. A, mean expression of cyclin D1–repressed signature is downregulated in prostatospheres relative to parental cell lines (31). B, 154 genes with correlated expression patterns either upregulated in prostatospheres and induced by cyclin D1 (red) or downregulated in prostatospheres and repressed by cyclin D1 (green). C, prostatosphere formation was decreased by cyclin D1. Note that 4,000 cells/mL of LNCaP cells infected with lentivirus encoded either cyclin D1 shRNA (cyclin D1 shRNA #3 and #4) or scrambled control shRNA were seeded onto ultralow contact plates with DMEM/F12 media containing N4 and B27 supplements, cultured for 10 days and counted in 96-well plate. D, Chip-Seq analysis of cyclin D1–bound genomic regions revealed association between cyclin D1 and the genes coding for Epha2, Dkkl, Sox4, and Nanog. The depicted tag density profiles are represented on a chromosomal chart showing distance to transcriptional start site and peak tag height (*). E and H, comparison of SOX2 and Nanog positivity between cyclin D1+/+ and cyclin D1−/− ventral prostates. F and I, comparative quantitative PCR for relative abundance of SOX2 and Nanog mRNA transcripts between ventral prostates of cyclin D1+/+ and cyclin D1−/− mice. G and J, HEK-293T cells were transfected with promoters driving luciferase reporter genes for SOX2 and Nanog, together with control vector or cyclin D1 expression vector (50 or 100 ng plasmid DNA). Data, mean ± SEM for N > 5 separate experiments.

Figure 6.

Cyclin D1 enhances prostate cancer stem cell populations in vitro and cyclin D1 DNA-bound form induces prostate stem cell targets. A, mean expression of cyclin D1–repressed signature is downregulated in prostatospheres relative to parental cell lines (31). B, 154 genes with correlated expression patterns either upregulated in prostatospheres and induced by cyclin D1 (red) or downregulated in prostatospheres and repressed by cyclin D1 (green). C, prostatosphere formation was decreased by cyclin D1. Note that 4,000 cells/mL of LNCaP cells infected with lentivirus encoded either cyclin D1 shRNA (cyclin D1 shRNA #3 and #4) or scrambled control shRNA were seeded onto ultralow contact plates with DMEM/F12 media containing N4 and B27 supplements, cultured for 10 days and counted in 96-well plate. D, Chip-Seq analysis of cyclin D1–bound genomic regions revealed association between cyclin D1 and the genes coding for Epha2, Dkkl, Sox4, and Nanog. The depicted tag density profiles are represented on a chromosomal chart showing distance to transcriptional start site and peak tag height (*). E and H, comparison of SOX2 and Nanog positivity between cyclin D1+/+ and cyclin D1−/− ventral prostates. F and I, comparative quantitative PCR for relative abundance of SOX2 and Nanog mRNA transcripts between ventral prostates of cyclin D1+/+ and cyclin D1−/− mice. G and J, HEK-293T cells were transfected with promoters driving luciferase reporter genes for SOX2 and Nanog, together with control vector or cyclin D1 expression vector (50 or 100 ng plasmid DNA). Data, mean ± SEM for N > 5 separate experiments.

Close modal

To consider possible mechanisms promoting cyclin D1–mediated prostatosphere expansion, we compared the cyclin D1 signature with stem cell signatures compiled from the literature (32, 33; Supplementary Fig. S4A). These studies demonstrate an enrichment of the cyclin D1 upregulated gene signature in the ES-like signature, which represents genes upregulated in ES cells relative to adult tissue stem cells (ES-like.Segal, P = 0.017; ref. 33) and a set of c-Myc target genes (Myc.targets.1, P = 0.007). The individual genes within the cyclin D1–upregulated network in the prostate mapped to the Wnt/β-catenin signaling pathway (Supplementary Fig. S4B individual genes upregulated by cyclin D1 are shown in yellow). In addition, functional annotation clustering of the cyclin D1–induced genes revealed the most enriched set corresponded to Wnt signaling pathway (Supplementary Table S2).

Endogenous cyclin D1 enhances prostatosphere expansion

Activation of the Wnt/β-catenin pathway is known to enhance prostate stem cell expansion (Supplementary Fig. S4B; refs. 23, 29). We therefore investigated the possibility that cyclin D1 may contribute to the propagation of stem-like prostate epithelial cells. Stem/progenitor cells can be expanded using in vitro sphere assays in which sphere number over multiple passages measures self-renewal capacity and sphere size reflect progenitor proliferation capacity (23). A surrogate assay was conducted using in vitro sphere formation comparing LNCaP cells that were transduced either with shRNA to cyclin D1 or controlled scrambled shRNA (Fig. 6C). These studies demonstrated an approximately 40% reduction in the proportion of prostatospheres formed in the second-generation LNCaP prostatospheres transduced with cyclin D1 shRNA (Fig. 6C). Trop2+/CD45f+ immunophenotyping by fluorescence-activated cell sorting analysis has been used to define prostate stem cells (23, 34, 35). To determine whether endogenous cyclin D1 maintains the Trop2+/CD45f population of prostate stem cells, LNCaP cells were transduced with a cyclin D1 shRNA tomato red fluorescent vector. Tetracycline-mediated induction of cyclin D1 shRNA reduced cyclin D1 abundance (Supplementary Fig. S5A and S5B) and reduced the proportion of Trop2+/CD45f cells by >50% (Fig. 6D; Supplementary Fig. S5C).

Cyclin D1 chromatin immunoprecipitation sequencing identifies occupancy of ES-inducing target genes

Cyclin D1 has been shown to regulate gene expression and directly binds to promoter regulatory regions of target genes in the context of local chromatin using chromatin immunoprecipitation (ChIP) assays (36–39). To examine the possibility that cyclin D1 binds in the context of local chromatin to genes governing Wnt/ES cell signaling, we interrogated our previously published chromatin immunoprecipitation sequencing (ChIP-Seq) analysis (39). Cyclin D1−/− fibroblasts were transduced with an expression vector encoding an amino terminal Flag-tagged cyclin D1 and ChIP-Seq analysis conducted (39). These studies demonstrated cyclin D1 occupancy of EphA2, Dkk1, SOX4, and Nanog genes (Fig. 6D). SOX2 abundance was increased in the cyclin D1+/+ compared with cyclin D1−/− VDL prostate by immunohistochemical staining (Fig. 6E) and by qRT-PCR of ventral/dorsolateral (VDL) prostate mRNA (Fig. 6F). Cotransfection of a mammalian expression vector encoding cyclin D1 enhanced transcriptional activity of the SOX2 promoter linked to a luciferase reporter gene >2-fold (Fig. 6G). We also determined the expression of Nanog in the VDL prostate of cyclin D1+/+ versus cyclin D1−/−. Nanog protein and mRNA abundance were reduced in cyclin D1−/− prostate epithelial cells (Fig. 6H and I). To determine whether cyclin D1 was capable of directly inducing the activity of the Nanog promoter luciferase reporter gene, experiments were conducted in LNCaP cells. Cyclin D1 expression induced the Nanog promoter in a dose-dependent manner (Fig. 6J), and bound directly to the Nanog promoter in ChIP analysis (data not shown).

Herein, endogenous cyclin D1 promoted prostate epithelial cellular proliferation in vivo and silencing of cyclin D1 reduced basal and DHT-induced proliferation of prostate cancer cell lines (LNCaP, NeuT-PEC, c-Myc-PEC). The role of cyclin D1 in androgen-induced prostate cell proliferation in vivo was previously unknown. The current findings are consistent with the previously described role for cyclin D1 in other tissues. Enforced cyclin D1 expression enhanced PC3 cell growth in xenograft tumors (16). Cyclin D1 mRNA levels are induced by growth factors in human prostate cancer cell lines and are increased in a subset of prostate cancer samples (40, 41). Reduction of cyclin D1 mRNA and protein levels via specific drugs, including flavonoids, inhibited cell-cycle progression in prostate cancer lines and in vivo (10, 14). Cyclin D1 siRNA reduced ErbB2-induced cell-cycle progression in LNCaP cells, suggesting an important role for endogenous cyclin D1 in prostate cancer cellular proliferation (18). These findings are in contrast with studies of forced cyclin D1 overexpression in which cyclin D1 inhibited cellular proliferation (42). The difference in findings may result from the different experimental systems used comparing the function of endogenous cyclin D1 herein versus study of forced overexpression in tissue culture (19).

An analysis of gene expression profiling demonstrated that cyclin D1 governs basal and DHT-mediated expression of gene clusters in vivo. In the current studies, endogenous cyclin D1 determined DHT-depending gene expression for approximately 90% of all AR-responsive genes in the prostate in vivo. The molecular mechanisms by which cyclin D1 coordinates such substantial changes in AR-mediated gene expression is likely to be multifactorial, with direct and indirect effects. First, direct effects of cyclin D1 evidenced herein, include findings that cyclin D1 bound to and regulated transcription of, specific target genes. The DNA-bound form of cyclin D1 promotes chromosomal instability (CIN) and expression of genes that promote CIN (39). In the current studies, cyclin D1 was identified in the context of local chromatin by ChIP Seq analysis of several genes governing stem cell function (EphA2, c-Myc, SOX4, Nanog). Second, cyclin D1 regulates recruitment in the context of local chromatin of several chromatin-modifying enzymes, including p300, HDAC, SUV39, HP1α, and modulates histone acetylation and methylation (36, 37). Cyclin D1 binds AR coactivators (p300, P/CAF, HDAC1, HDAC3, BRCA1), which may contribute indirectly to alterations in gene expression. Third, cyclin D1 binds to and regulates the AR and other nuclear receptors (36, 37, 43, 44). A physical interaction between cyclin D1 and the AR was previously identified in immunoprecipitation–Western blot analysis and forced cyclin D1 overexpression inhibited a synthetic AR reporter gene activity in cultured cells (43, 44). The AR, however, is known to convey context-specific pro- versus antiproliferative effects and pro- versus antimigratory and growth effects (45). Fourth, cyclin D1 abundance in vivo determined the expression of growth factors and other cell-cycle control proteins, which may also indirectly affect AR signaling. Collectively, the current studies establish the essential function of cyclin D1 in AR signaling in vivo.

The current studies extend prior studies in other cell types that demonstrate cyclin D1 promotes cellular proliferation and inhibits EMT in other cell types. Herein, the cyclin D1–regulated gene expression signature was anticorrelated with EMT gene expression induced by a variety of mechanisms (Snail, GSC, Twist, TGF-β, shE-cadherin) in the prostate in vivo. Cyclin D1 signaling was anticorrelated with EMT and correlated with prostatosphere gene expression. In recent studies, the knockdown of EMT factors in mesenchymal-like prostate cancer cells caused a gain of TIC properties (24). TIC constitutes a subpopulation of cells capable of initiating sustained growth of tumors in immunodeficient mice. The hallmarks of TIC include expression of specific embryonic stem cell genes (e.g., SOX2, OCT3/4, Nanog), altered expression of stem cell markers (Trop2, CD49f), increased potential for anchorage-independent growth, and capacity to form spheroids in vitro (21, 23, 35). The current studies provide several lines of evidence that cyclin D1 enhances prostate stem cell expansion. Cyclin D1 shRNA transduction of prostate cancer cell lines reduced second-generation prostatosphere formation. Second, in vivo gene expression studies demonstrated cyclin D1 induced the expression of a cluster of genes involved in promoting stem cell expansion, which are enriched in prostatospheres (31). Several genes maintained by endogenous cyclin D1 in the prostate in vivo, including Oct4 and Nanog, were also enriched in prostatospheres (31). Third, cyclin D1 shRNA reduced the abundance of the Trop2/CD49f+ population that is thought to define a population of prostate stem-like cells (34). Fourth, cyclin D1 was shown to bind the regulatory region of key genes governing stem cell function (EphA2, SOX4, Nanog, c-Myc), and to induce transcription, mRNA and protein levels of SOX2 and Nanog. The maintenance of Nanog abundance by cyclin D1 may involve direct transcriptional induction consistent with prior findings that a DNA-bound form of cyclin D1 contributes to transcriptional regulation (36, 39). SOX9 is known to maintain the committed stem cell compartment and, herein, cyclin D1 induced mRNA for SOX2, SOX9, and SOX11.

The current studies suggest cyclin D1 may drive prostate stem cell expansion by enhancing Wnt/β-catenin signaling thereby promoting direct transcriptional induction of genes promoting prostate stem cell expansion. Herein, endogenous cyclin D1 enhanced Wnt gene expression and signaling in the prostate in vivo (induced genes shown in red; Supplementary Fig. S3D). Cyclin D1 maintained expression of TCF4, which was induced 2.5-fold by cyclin D1 and as a Wnt/β-catenin responsive gene itself, serves as useful reporter of endogenous Wnt activity (Supplementary Table S1). Cyclin D1 is itself a TCF/β-catenin responsive gene (46), which may provide a positive feed-forward loop to amplify Wnt signaling in the prostate.

Herein, endogenous cyclin D1 maintained a prostate gene expression signature in vivo, which predicted BCR of patients with prostate cancer. The cyclin D1 expression signature identified herein was associated with BCR and poor prognosis in human prostate cancer. This signature enhances the predictive value of a gene signature, which included cyclin D1 itself, together with SPP1, SMAD4, and PTEN, that was shown to enhance the prognostic accuracy of the Gleason score in predicting metastatic lethal outcome (9). How might the gene expression profile regulated by endogenous cyclin D1 contribute to poor prognosis in human prostate cancer? One possibility is that cyclin D1–regulated gene expression may contribute to the formation of prostate cells with self-renewing capacity or TICs. Within the population of genes induced by cyclin D1, we identified a cluster of genes known to contribute to the expression of “TICs,” including EphA2, Wnt7A, and SOX4. TIC and cancer stem cell share gene networks with ES and adult stem cells that are essential for self-renewal and pluripotency. The finding that cyclin D1 enhances prostate cancer stem cell expansion provides the rational basis for targeting this novel node to reduce therapeutic resistance and recurrence.

No potential conflicts of interest were disclosed.

Conception and design: X. Ju, S. Katiyar, M.P. Lisanti, R.G. Pestell

Development of methodology: X. Ju, M.C. Casimiro, H. Meng, X. Jiao, S. Katiyar, K. Chen, R.G. Pestell

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Ju, M.C. Casimiro, X. Jiao, S. Katiyar, M. Crosariol, R.G. Pestell

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Ju, M. Casimiro, M. Gormley, X. Jiao, S. Katiyar, M. Wang, A.A. Quong, A. Ertel, R.G. Pestell

Writing, review, and/or revision of the manuscript: X. Ju, M.C. Casimiro, M. Gormley, S. Katiyar, M. Wang, A.A. Quong, M.P. Lisanti, A. Ertel, R.G. Pestell

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Ju, H. Meng, M. Crosariol, R.G. Pestell

Study supervision: X. Ju

This work was supported in part by awards from R01CA70896, R01CA75503, R01CA86072, R01CA137494 (R.G. Pestell), and the Kimmel Cancer Center (1P30CA56036-08; R.G. Pestell). This project is funded in part from the Dr. Ralph and Marian C. Falk Medical Research Trust and a grant from the Pennsylvania Department of Health (R.G. Pestell). The Department specifically disclaims responsibility for an analysis, interpretation, or conclusions. M.P. Lisanti and his laboratory were supported via the resources of Thomas Jefferson University.

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.

1.
Jemal
A
,
Siegel
R
,
Xu
J
,
Ward
E
. 
Cancer statistics, 2010
.
CA Cancer J Clin
2010
;
60
:
277
300
.
2.
Martin
NE
,
Mucci
LA
,
Loda
M
,
Depinho
RA
. 
Prognostic determinants in prostate cancer
.
Cancer J
2011
;
17
:
429
37
.
3.
Glinsky
GV
,
Glinskii
AB
,
Stephenson
AJ
,
Hoffman
RM
,
Gerald
WL
. 
Gene expression profiling predicts clinical outcome of prostate cancer
.
J Clin Invest
2004
;
113
:
913
23
.
4.
Taylor
BS
,
Schultz
N
,
Hieronymus
H
,
Gopalan
A
,
Xiao
Y
,
Carver
BS
, et al
Integrative genomic profiling of human prostate cancer
.
Cancer Cell
2010
;
18
:
11
22
.
5.
Markert
EK
,
Mizuno
H
,
Vazquez
A
,
Levine
AJ
. 
Molecular classification of prostate cancer using curated expression signatures
.
Proc Natl Acad Sci USA
2011
;
108
:
21276
81
.
6.
Eisenberger
MA
. 
The current knowledge of hormonal therapy in the treatment of prostate cancer
. In:
Pestell
RG
,
Nevalainen
MT
,
editors
. 
Prostate cancer—signaling networks, genetics, and new treatment strategies
.
Santa Monica, CA
:
Humana Press
; 
2008
. p.
339
54
.
7.
Clark
J
,
Attard
G
,
Jhavar
S
,
Flohr
P
,
Reid
A
,
De-Bono
J
, et al
Complex patterns of ETS gene alteration arise during cancer development in the human prostate
.
Oncogene
2008
;
27
:
1993
2003
.
8.
Mosquera
JM
,
Perner
S
,
Genega
EM
,
Sanda
M
,
Hofer
MD
,
Mertz
KD
, et al
Characterization of TMPRSS2-ERG fusion high-grade prostatic intraepithelial neoplasia and potential clinical implications
.
Clin Cancer Res
2008
;
14
:
3380
85
.
9.
Ding
Z
,
Wu
CJ
,
Chu
GC
,
Xiao
Y
,
Ho
D
,
Zhang
J
, et al
SMAD4-dependent barrier constrains prostate cancer growth and metastatic progression
.
Nature
2011
;
410
:
269
73
.
10.
Fu
M
,
Wang
C
,
Li
Z
,
Sakamaki
T
,
Pestell
RG
. 
Minireview: cyclin D1: normal and abnormal functions
.
Endocrinology
2004
;
145
:
5439
47
.
11.
Wang
C
,
Li
Z
,
Lu
Y
,
Du
R
,
Katiyar
S
,
Yang
J
, et al
Cyclin D1 repression of nuclear respiratory factor 1 integrates nuclear DNA synthesis and mitochondrial function
.
Proc Natl Acad Sci USA
2006
;
103
:
11567
72
.
12.
Sakamaki
T
,
Casimiro
MC
,
Ju
X
,
Quong
AA
,
Katiyar
S
,
Liu
M
, et al
Cyclin D1 determines mitochondrial function in vivo
.
Mol Cell Biol
2006
;
26
:
5449
69
.
13.
Holnthoner
W
,
Pillinger
M
,
Groger
M
,
Wolff
K
,
Ashton
AW
,
Albanese
C
, et al
Fibroblast growth factor-2 induces Lef/Tcf-dependent transcription in human endothelial cells
.
J Biol Chem
2002
;
277
:
45847
53
.
14.
Hulit
J
,
Wang
C
,
Li
Z
,
Albanese
C
,
Rao
M
,
Di Vizio
D
, et al
Cyclin D1 genetic heterozygosity regulates colonic epithelial cell differentiation and tumor number in ApcMin mice
.
Mol Cell Biol
2004
;
24
:
7598
611
.
15.
Perry
JE
,
Grossmann
ME
,
Tindall
DJ
. 
Epidermal growth factor induces cyclin D1 in a human prostate cancer cell line
.
Prostate
1998
;
35
:
117
24
.
16.
Chen
Y
,
Martinez
LA
,
LaCava
M
,
Coghlan
L
,
Conti
CJ
. 
Increased cell growth and tumorigenicity in human prostate LNCaP cells by overexpression to cyclin D1
.
Oncogene
1998
;
16
:
1913
20
.
17.
Casimiro
M
,
Rodriguez
O
,
Pootrakul
L
,
Aventian
M
,
Lushina
N
,
Cromelin
C
, et al
ErbB-2 induces the cyclin D1 gene in prostate epithelial cells in vitro and in vivo
.
Cancer Res
2007
;
67
:
4364
72
.
18.
Wang
C
,
Pattabiraman
N
,
Zhou
JN
,
Fu
M
,
Sakamaki
T
,
Albanese
C
, et al
Cyclin D1 repression of peroxisome proliferator-activated receptor gamma expression and transactivation
.
Mol Cell Biol
2003
;
23
:
6159
73
.
19.
Schiewer
MJ
,
Morey
LM
,
Burd
CJ
,
Liu
Y
,
Merry
DE
,
Ho
SM
, et al
Cyclin D1 repressor domain mediates proliferation and survival in prostate cancer
.
Oncogene
2009
;
28
:
1016
27
.
20.
Mulholland
DJ
,
Xin
L
,
Morim
A
,
Lawson
D
,
Witte
O
,
Wu
H
. 
Lin-Sca-1+CD49fhigh stem/progenitors are tumor-initiating cells in the Pten-null prostate cancer model
.
Cancer Res
2009
;
69
:
8555
62
.
21.
Rajasekhar
VK
,
Studer
L
,
Gerald
W
,
Socci
ND
,
Scher
HI
. 
Tumour-initiating stem-like cells in human prostate cancer exhibit increased NF-kappaB signalling
.
Nat Commun
2011
;
2
:
162
.
22.
Wang
XD
,
Leow
CC
,
Zha
J
,
Tang
Z
,
Modrusan
Z
,
Radtke
F
, et al
Notch signaling is required for normal prostatic epithelial cell proliferation and differentiation
.
Dev Biol
2006
;
290
:
66
80
.
23.
Stoyanova
T
,
Goldstein
AS
,
Cai
H
,
Drake
JM
,
Huang
J
,
Witte
ON
. 
Regulated proteolysis of Trop2 drives epithelial hyperplasia and stem cell self-renewal via beta-catenin signaling
.
Genes Dev
2012
;
26
:
2271
85
.
24.
Celia-Terrassa
T
,
Meca-Cortes
O
,
Mateo
F
,
de Paz
AM
,
Rubio
N
,
Arnal-Estape
A
, et al
Epithelial-mesenchymal transition can suppress major attributes of human epithelial tumor-initiating cells
.
J Clin Invest
2012
;
122
:
1849
68
.
25.
Vega
S
,
Morales
AV
,
Ocana
OH
,
Valdes
F
,
Fabregat
I
,
Nieto
MA
. 
Snail blocks the cell cycle and confers resistance to cell death
.
Genes Dev
2004
;
18
:
1131
43
.
26.
Thiery
JP
,
Acloque
H
,
Huang
RY
,
Nieto
MA
. 
Epithelial-mesenchymal transitions in development and disease
.
Cell
2009
;
139
:
871
90
.
27.
Ju
X
,
Ertel
A
,
Casimiro
MC
,
Yu
Z
,
Meng
H
,
McCue
PA
, et al
Novel oncogene-induced metastatic prostate cancer cell lines define human prostate cancer progression signatures
.
Cancer Res
2013
;
73
:
978
89
.
28.
Wu
K
,
Jiao
X
,
Li
Z
,
Katiyar
S
,
Casimiro
MC
,
Yang
W
, et al
Cell fate determination factor Dachshund reprograms breast cancer stem cell function
.
J Biol Chem
2011
;
286
:
2132
42
.
29.
Bisson
I
,
Prowse
DM
. 
WNT signaling regulates self-renewal and differentiation of prostate cancer cells with stem cell characteristics
.
Cell Res
2009
;
19
:
683
97
.
30.
Taube
JH
,
Herschkowitz
JI
,
Komurov
K
,
Zhou
AY
,
Gupta
S
,
Yang
J
, et al
Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes
.
Proc Natl Acad Sci USA
2010
;
107
:
15449
54
.
31.
Duhagon
MA
,
Hurt
EM
,
Sotelo-Silveira
JR
,
Zhang
X
,
Farrar
WL
. 
Genomic profiling of tumor initiating prostatospheres
.
BMC Genomics
2010
;
11
:
324
.
32.
Ben-Porath
I
,
Thomson
MW
,
Carey
VJ
,
Ge
R
,
Bell
GW
,
Regev
A
, et al
An embryonic stem cell-like gene expression signature in poorly differentiated aggressive human tumors
.
Nat Genet
2008
;
40
:
499
507
.
33.
Wong
DJ
,
Liu
H
,
Ridky
TW
,
Cassarino
D
,
Segal
E
,
Chang
HY
. 
Module map of stem cell genes guides creation of epithelial cancer stem cells
.
Cell Stem Cell
2008
;
2
:
333
44
.
34.
Goldstein
AS
,
Huang
J
,
Guo
C
,
Garraway
IP
,
Witte
ON
. 
Identification of a cell of origin for human prostate cancer
.
Science
2010
;
329
:
568
71
.
35.
Goldstein
AS
,
Lawson
DA
,
Cheng
D
,
Sun
W
,
Garraway
IP
,
Witte
ON
. 
Trop2 identifies a subpopulation of murine and human prostate basal cells with stem cell characteristics
.
Proc Natl Acad Sci USA
2008
;
105
:
20882
7
.
36.
Fu
M
,
Rao
M
,
Bouras
T
,
Wang
C
,
Wu
K
,
Zhang
X
, et al
Cyclin D1 inhibits peroxisome proliferator-activated receptor gamma-mediated adipogenesis through histone deacetylase recruitment
.
J Biol Chem
2005
;
280
:
16934
41
.
37.
Wang
C
,
Fan
S
,
Li
Z
,
Fu
M
,
Rao
M
,
Ma
Y
, et al
Cyclin D1 antagonizes BRCA1 repression of estrogen receptor alpha activity
.
Cancer Res
2005
;
65
:
6557
67
.
38.
Bienvenu
F
,
Jirawatnotai
S
,
Elias
JE
,
Meyer
CA
,
Mizeracka
K
,
Marson
A
, et al
Transcriptional role of cyclin D1 in development revealed by a genetic-proteomic screen
.
Nature
2010
;
463
:
374
8
.
39.
Casimiro
MC
,
Crosariol
M
,
Loro
E
,
Ertel
A
,
Yu
Z
,
Dampier
W
, et al
ChIP sequencing of cyclin D1 reveals a transcriptional role in chromosomal instability in mice
.
J Clin Invest
2012
;
122
:
833
43
.
40.
Albanese
C
,
D'Amico
M
,
Reutens
AT
,
Fu
M
,
Watanabe
G
,
Lee
RJ
, et al
Activation of the cyclin D1 gene by the E1A-associated protein p300 through AP-1 inhibits cellular apoptosis
.
J Biol Chem
1999
;
274
:
34186
95
.
41.
Fu
GF
,
Lin
XH
,
Han
QW
,
Fan
YR
,
Xu
YF
,
Guo
D
, et al
RNA interference remarkably suppresses bcl-2 gene expression in cancer cells in vitro and in vivo
.
Cancer Biol Ther
2005
;
4
:
822
9
.
42.
Petre-Draviam
CE
,
Williams
EB
,
Burd
CJ
,
Gladden
A
,
Moghadam
H
,
Meller
J
, et al
A central domain of cyclin D1 mediates nuclear receptor corepressor activity
.
Oncogene
2005
;
24
:
431
44
.
43.
Knudsen
KE
,
Cavenee
WK
,
Arden
KC
. 
D-type cyclins complex with the androgen receptor and inhibit its transcriptional transactivation ability
.
Cancer Res
1999
;
59
:
2297
301
.
44.
Reutens
AT
,
Fu
M
,
Wang
C
,
Albanese
C
,
McPhaul
MJ
,
Sun
Z
, et al
Cyclin D1 binds the androgen receptor and regulates hormone-dependent signaling in a p300/CBP-associated factor (P/CAF)-dependent manner
.
Mol Endocrinol
2001
;
15
:
797
11
.
45.
Niu
Y
,
Altuwaijri
S
,
Lai
KP
,
Wu
CT
,
Ricke
WA
,
Messing
EM
, et al
Androgen receptor is a tumor suppressor and proliferator in prostate cancer
.
Proc Natl Acad Sci U S A
2008
;
105
:
12182
7
.
46.
Shtutman
M
,
Zhurinsky
J
,
Simcha
I
,
Albanese
C
,
D'Amico
M
,
Pestell
R
, et al
The cyclin D1 gene is a target of the beta-catenin/LEF-1 pathway
.
Proc Natl Acad Sci USA
1999
;
96
:
5522
7
.