Estrogen receptor-related receptors (ERR) are orphan nuclear receptors, which are constitutively activated without estrogen binding. Recent evidence indicates that the ligand-independent ERRs may be involved in similar ER-mediated regulatory pathways and modulate estrogen responsiveness in certain target cells. We recently showed that an ERR subtype, ERRγ, is coexpressed with ERβ in normal human prostatic epithelial cells and exhibits reduced expression in many prostate cancer cell lines and clinical neoplastic prostate tissues. Based on this, we hypothesize that ERRγ may have growth regulatory roles in prostate and prostate cancer. We showed in this study that ERRγ was expressed in epithelial cell nuclei in fetal and pubertal human prostates, whereas its nuclear expression became reduced in advanced prostate cancer lesions. Stable ERRγ expression by retroviral transduction suppressed significantly both in vitro cell growth and in vivo tumorigenicity of two prostate cancer cell lines, LNCaP and DU145, as evidenced by a cell-cycle arrest at G1-S transition and also induction of two cyclin-dependent kinase inhibitors p21WAF1/CIP1 and p27KIP1. We further showed by reporter assay that induction of p21 and p27 by ERRγ was mediated through direct transactivation of their gene promoters. Moreover, we also showed that a selective ERRγ-agonist, DY131, could potentiate the ERRγ-induced growth inhibition in LNCaP-ERRγ and DU145-ERRγ cells in a dose-dependent manner compared with respective parental cells. Taken together, our results show that ERRγ may perform an antiproliferative or tumor-suppressing function in prostate cancer cells. More importantly, our results suggest that ERRγ could be a novel therapeutic target for prostate cancer treatment. [Cancer Res 2007;67(10):4904–14]
- Estrogen receptor–related receptor
- orphan nuclear receptor
- estrogen receptor
- tumor suppression
- prostate cancer
Besides androgens, which play an important role in normal and neoplastic growths of prostate gland, estrogens have been suggested for a long time to play synergetic or distinct roles in the same processes ( 1, 2). It is generally believed that direct estrogenic effects in prostate or their possible roles in prostate cancer and benign prostatic hyperplasia are partially mediated through their specific receptors, estrogen receptor (ER) α and ERβ, which bind directly to estrogen response elements (ERE) or other response sites via indirect interactions with other heterologous transcription factors [e.g., Sp1 and activator protein 1 (AP-1)] in the target gene promoters to mediate transcriptional regulation. Both ERα and ERβ are differentially expressed in adult prostatic tissues, with the epithelial cells expressing mainly ERβ whereas the stromal cells expressing ERα at low levels ( 3). However, the exact roles of ERs and estrogen-signaling pathways in prostate still remain unclear. Previous studies of ER knock-out mice show that functional inactivation of ERs induces no apparent abnormality in both gross and microscopic structure of prostate ( 4– 6), whereas estrogen-imprinting effects in neonatal prostate are mediated primarily through the stromal ERα and not epithelial ERβ ( 7). However, some recent studies suggest that ERβ may play an antiproliferative role in the regulation of prostatic epithelial differentiation, probably through activation by 5α-androstane-3β,17β-diol, a newly identified ERβ ligand derived metabolically from dihydrotestosterone ( 8– 10).
Recent advances in nuclear receptors show that members of the ligand-independent orphan nuclear receptors are important regulators of development and many cellular functions via their transcriptional regulations of gene expression. Among these orphan receptors, estrogen receptor-related receptors [estrogen related-receptors (ERR)], which are closely related to ERs and sharing high homology in their DNA-binding domains (DBD), are constitutively active without binding to natural estrogen ( 11, 12). Three different ERR subtypes (α, β, and γ) have been characterized thus far. All ERR members show considerable homology, especially in their DBD and the highly conserved activation function-2 (AF-2) domain, with the highest similarity between ERRβ and ERRγ ( 13, 14). Like ERs, all ERRs bind to the classic ERE (AGGTCAn3TGACCT), suggesting that ERRs may be involved in similar ER-mediated signaling pathways via regulation of similar target genes and competition for coregulatory proteins in same target cells. ERRs also bind to some ERE-related response elements with extended half-site core sequences (TNAAGGTCA; ERRE/SF-1RE; ref. 15), suggesting that ERRs may also have their own independent regulatory pathways or functions distinct from ERs. In addition, ERRs can regulate transcription in a similar manner to ERα via a tethered pathway by interacting with the Sp1 and AP-1 transcription factors at their respective DNA binding sites ( 16– 18). Therefore, it is likely that ERRs may regulate a broad spectrum of genes in their target cells.
Human ERRγ (ERR3/NR3B3/ESRRG) was first cloned by rapid amplification of cDNA ends (RACE)–PCR from fetal brain ( 19) and subsequently from other organ cDNA libraries ( 13, 20), whereas mouse ERRγ was isolated by yeast two-hybrid screening of an embryo cDNA library using the transcriptional coactivator glucocorticoid receptor interacting protein 1 (GRIP1) as bait ( 21). Rat ERRγ has recently been cloned by yeast two-hybrid screening of a rat liver cDNA library with an orphan nuclear receptor, small heterodimer partner (SHP) as bait ( 15) and RACE-PCR from rat prostate ( 14). Multiple spliced ERRγ transcripts (ERRγ1 and ERRγ2), probably derived from alternative splicing at the 5′-end, have been identified from different cDNA libraries and show tissue-specific expression patterns ( 13, 20). Recently, a novel spliced variant ERRγ3, which lacks the exon F encoding the second zinc finger motif in DBD, has been cloned, and interestingly, its transcript expression is confined to the prostate and adipocytes ( 22). However, their protein expressions in cells and tissues are not confirmed yet. Besides ERE and ERRE, ERRγ can also bind to a variety of ERRE-related sequences ( 15), suggesting that ERRγ may have its own distinct target genes, including the recently characterized ERRα, DAX-1, and SHP ( 23– 26). A recent structural study indicates that ERRγ is a true orphan receptor as its ligand-binding domain (LBD) adopts a transcriptionally active conformation, which can recruit transcriptional coactivators in the absence of ligand ( 27). The ligand-independent transactivation of ERRγ depends on its binding via the AF-2 domain to some p160 transcriptional coactivators, including GRIP1, steroid receptor coactivator-1, amplified in breast cancer 2/ASC-2, and peroxisome proliferator-activated receptor γ coactivator-1α/PGC-1α, while repressed by receptor interacting protein 140/RIP140, which recruits histone deacetylases ( 21, 23, 24, 27, 28). However, it is shown that some synthetic ER agonist (diethylstilbestrol) and selective ER modulators (SERMs; tamoxifen, 4-hydroxytamoxifen) can bind to ERRβ/γ as antagonists by disrupting the ERR-coactivator interactions ( 29, 30), whereas some synthetic acyl hydrazones can bind to ERRβ/γ as agonists ( 31, 32). By Northern blot and reverse transcription-PCR (RT-PCR), ERRγ transcripts are detected in diverse fetal and adult human tissues and organs, including prostate, with particularly high levels in brain, gastrointestinal tract, heart, kidney, skeletal muscle, and placenta ( 13, 19, 20, 33). Similar expression patterns are also shown in mouse and rat adult tissues ( 14, 21), suggesting that ERRγ may play important roles in the differentiation and functions in these tissues and organs. However, in vitro or in vivo functional study of ERRγ is absent thus far.
Recently, we showed that ERR members (α, β, and γ) are coexpressed with ERs in normal prostatic epithelial cells, whereas they exhibit an apparent down-regulation in many prostate cancer cell lines and clinical neoplastic tissues ( 33), implying that the ligand-independent ERRs and ligand-dependent ERs may coregulate the growth and functions of prostatic cells. Among ERRs, ERRγ shows a down-regulation pattern in some androgen-independent prostate cancer lines and also clinical neoplastic tissues, whereas its transient ectopic expression inhibits the in vitro proliferation of PC-3 prostate cancer cells. To show whether ERRγ could be a useful marker for prostate cancer development, in this study, we investigated the roles of ERRγ in the growth regulation of prostate cancer cells by retroviral transducing ERRγ in both androgen-sensitive (LNCaP) and androgen-insensitive (DU145) prostate cancer cells. Moreover, the role of cell-cycle regulatory proteins, p21WAF1/CIP1 and p27KIP1 (hereafter referred to as p21 and p27), in the ERRγ-induced cell cycle arrest was also investigated. Finally, we also examined the efficacy of a selective ERRγ agonist on the ERRγ-positive prostate cancer cell growth.
Materials and Methods
Human Prostatic Tissues and Immunohistochemistry
Archival blocks of formalin-fixed fetal, pubertal, and neoplastic human prostates were used for immunohistochemistry ( 34). Immunohistochemistry was done using a glucose-oxidase-diaminobenzidine-nickel procedure ( 33). All human tissues were obtained with informed consent and approval from Clinical Research Ethics Committee, the Chinese University of Hong Kong.
Cell Lines and Cell Culture
Human prostate cancer lines, LNCaP and DU145, cervical carcinoma line HeLa, and a mouse packaging cell line PA317 were obtained from the American Type Culture Collection. All cell lines and derived clones were cultured according to the American Type Culture Collection recommendations. Total RNA and proteins were extracted from cells grown to 70% to 80% confluence.
Mammalian expression vectors. FLAG-tagged or untagged full-length human ERRγ cDNAs were generated by PCR using pSG5-ERRγ (33; NM_001438) as template and subcloned into pBabe-puro for retroviral transduction and pcDNA3.1 for transfections, at BamHI and EcoRI sites, yielding pBabe-ERRγ, pBabe-FLAG-ERRγ, pcDNA3.1-ERRγ, and pcDNA3.1-FLAG-ERRγ, respectively. FLAG-tagged truncated mutant ERRγ-ΔZFI with deletion of first zinc finger in DBD was generated by fusion-PCR method and subcloned into pcDNA3.1 and pBabe.
Reporter constructs. Luciferase reporter plasmids, pGL3-ERE3-Luc and pGL3-ERRE3-Luc containing three copies of ERE and ERRE, were constructed as described ( 33, 35). Human gene promoter sequences, p21 (−2,324 to +11) and p27 (−3,568 to +1), and their serial deletion constructs were amplified by PCR from human genomic DNA extracted from DU145 cells and cloned into pGL3 at XhoI and HindIII sites as pGL3-p21-Luc and pGL3-p27-Luc, respectively. All plasmid constructs were confirmed by DNA sequencing.
Transient Transfection and Luciferase Reporter Assay
HeLa cells were seeded at 5 × 104 cells/well in 24-well plates and grown for 24 h in MEM with 10% fetal bovine serum (FBS) before transfection. Cells were cotransfected with 0.2 μg reporter plasmid (pGL3 or pGL3-ERE3/ERRE3/p21/p27), 0.2 μg expression plasmid (pcDNA3.1, pcDNA3.1-FLAG-ERRγ/FLAG-ERRγ-ΔZFI), and 0.04 μg pRL-CMV using FuGENE6 transfection reagent. After 48 h post-transfection, cells were harvested for luciferase reporter assay using the Dual-Luciferase Reporter Assay System (Promega). The luciferase activation activity was normalized to that of the internal control Renilla luciferase activity by pRL-CMV as relative luciferase unit and fold of activation of relative luciferase unit as compared with controls (transfection with pcDNA3.1) was determined. All assays were done in three separate experiments done in triplicate, and data were presented as mean ± SD.
Retroviral Transduction and Generation of ERRγ-Stable Clones
pBabe-FLAG-ERRγ, pBabe-FLAG-ERRγ-ΔZFI, or pBabe were transfected, respectively, into PA317 cells growing in DMEM and 10% FBS using FuGENE 6 reagent. Transfected cells were incubated in fresh medium for another 24 h. After 48 and 72 h, culture medium containing infectious virions was harvested for retroviral transduction. LNCaP and DU145 cells were incubated with the virus-containing medium mixed with 8 μg/mL polybrene. After 48 h incubation, puromycin was added to eliminate the noninfected cells, whereas stable clones were isolated after 8-day drug selection. Vector control clones were generated by infecting cells with empty vector pBabe-puro.
In vitro Cell Growth Assays
Cell counting. Stable ERRγ clones and their parental cells were plated at 5 × 103 per well in 24-well plates and cultured in growth media for 10 days, with fresh media replaced every 3 days. Viable cells were counted daily by trypan blue exclusion assay. All counts were done on triplicate wells and repeated in three independent experiments, and mean ± SD of cell number was plotted against time.
5-Bromodeoxyuridine incorporation assay. Cells at same density were plated on poly-lysine–coated coverslips placed in 24-well plates and cultured for 24 h in growth media. After 24 h culture, cells were incubated with 10 μmol/L 5-bromodeoxyuridine (BrdUrd) for 1 h and fixed in cold methanol-acetone (1:1) for 15 min. Fixed cells were stained with a mouse monoclonal anti-BrdUrd antibody (Becton Dickinson) followed by an anti-mouse immunoglobulin G–FITC. Stained cells were detected and counted under a fluorescence microscope and analyzed with morphometry software (Neurolucida-2000, Vermont). Percentages of BrdUrd-positive cells were determined after counting at least 500 cells in both parental cells and stable clones. Data were presented as mean ± SD obtained from three independent experiments.
Soft agar assay for anchorage-independent growth. Cells were suspended in culture media with 0.3% agar and plated at 5 × 103 cells per well in six-well plates, which were precoated with 0.5% base agar in the same media. After culture for 28 days, cells were fixed in 4% formaldehyde, stained with 0.1% crystal violet for colony visualization, and counted under microscope. Colony-formation efficiency was determined as the number of colonies, with sizes more than 100 μm in diameter, formed per total number of cells seeded. All experiments were done in triplicates.
Cell Cycle and Apoptosis Analyses
Parental cells and stable clones were cultured to 80% confluence and harvested. Trypsinized cells (1 × 106), suspended in PBS, were fixed in ice-cold 70% ethanol overnight at −20°C. Fixed cells were incubated in PBS containing 50 μg/mL propidium iodide (PI) and 10 μg/mL DNase-free RNase for 15 min. DNA flow cytometry was done in a flow cytometer (ALTRA Cell Sorting System, Coulter) using a total of 3 × 105 cells. Cell-cycle distribution was analyzed with commercial software (MoFit LT2). Apoptotic or necrotic cells were detected by an FITC-Annexin V-PI double staining method (ApopNexin Apoptosis detection Kit, Chemicon) and analyzed by bivariate DNA flow cytometry. By this staining method, the AV+/PI− cell population represents the apoptotic cells, whereas AV+/PI+ and AV−/PI+ as the necrotic or late apoptotic cells.
In vivo Tumor Growth Assay
To evaluate the impact of ectopic ERRγ expression on in vivo tumorigenicity of prostate cancer cells, stable ERRγ clones or their parental cells (suspended in 100 μL 1:1 growth medium–Matrigel mixture) were s.c. injected into the flanks of male non-obese diabetic/severe combined immunodeficient (SCID) mice (3 × 106 LNCaP cells per mouse) or athymic nude mice (5 × 106 DU145 cells per mouse) and allowed to grow for 4 to 8 weeks. All mice were housed in an air-filtered pathogen-free condition. Tumor growth was monitored weekly. Tumor volumes (mm3) were measured using electronic calipers and calculated using the formula [(width) × (length) × (height)]/2. Growth of tumor volumes (mean ± SD obtained from three independent animals) was plotted against time. All animal experiments were done with approval from the Animal Experimentation Ethics Committee, The Chinese University of Hong Kong, and in accordance with the LASEC Guidelines for laboratory animals.
RNA and Protein Analyses
RT-PCR analysis. Total RNA was extracted from cultured cells using TRIzol reagent. Approximately 2 μg DNase I-treated RNA samples were reverse transcribed to cDNAs by M-MLV reverse transcriptase using oligo(dT) primer, and 1 μL cDNA samples was used for PCR. Primers used are as follows: ERRγ: 5′-primer: 5′-ACCATGAATGGCCATCAGAA-3′, 3′-primer: 5′-ACCAGCTGAGGGTTCAGGTAT-3′; and β-actin: 5′-primer: 5′-ATGGATGATGATATCGCCGCG-3′, 3′-primer: 5′-CTCCATGTCGTCCCAGTTGGT-3′.
Western blot analysis. Total cellular proteins were extracted from cultured cells grown to 70% to 80% confluence. Cells were lysed in 100 μL ice-cold lysis buffer ( 33) supplemented with a protease inhibitor cocktail for 30 min at 4°C. Supernatants were collected after centrifugation at 11,625 × g at 4°C. Protein concentration was determined by a modified Lowry method using bovine serum albumin as standard. Protein samples, loaded at 40 to 80 μg per lane, separated by 10% to 15% SDS-PAGE gels and transblotted onto polyvinylidene diflouride membranes, were probed with primary antibodies at optimal dilutions, followed by an alkaline phosphatase–conjugated secondary antibody for colorimetric visualization. Primary antibodies used are as follows: (a) an affinity-purified rabbit polyclonal ERRγ antibody ( 14, 33), (b) a mouse monoclonal ERRγ antibody (R&D Systems), (c) p21 (Lab Vision), (d) p27 (Lab Vision), (e) p53 (Cell Signaling), (f) cyclin A (Santa Cruz), (g) cyclin B1 (Cell Signaling), (h) cyclin D1 (Lab Vision), (i) cyclin D3 (Lab Vision), and (j) anti-FLAG M2 (Sigma).
In vitro Treatments with ERRγ Agonist
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay and BrdUrd incorporation. The efficacy of an ERRγ-specific agonist DY131 ( 31) on prostate cancer cell growth was evaluated by cell viability 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay and BrdUrd incorporation. DY131 (Tocris) was dissolved in DMSO at 0.1 to 30 mmol/L as stock solutions and diluted (1:103) in culture medium at different concentrations for in vitro treatments. For cell growth studies, ERRγ clones and parental cells were seeded at 2 × 103 per well onto 96-well plates in their growth media and allowed to adhere for 24 h before culturing in the same medium containing DY131 at doses ranging from 0.1 to 30 μmol/L for 1 to 5 days. For controls, cells were treated with DMSO diluted (1:103) in the same medium. Media with or without DY131 were renewed twice for treatments longer than 3 days. After treatments, the number of viable cells was determined by MTT assay. Experiments on each drug concentration were done in quadruplicate, and data as mean ± SD were obtained from three independent experiments. DNA synthesis in DY131-treated cells was determined by BrdUrd incorporation. After 24-h culture, cells were treated with DY131 at various doses in the same medium and cultured for 3 to 4 days. After treatments, 10 μmol/L BrdUrd was added to DY131-treated cells followed by BrdUrd staining procedure as mentioned above.
Luciferase assay. Human embryonic kidney epithelial cell line 293 was used for the study of DY131 on ERRγ-driven ERE/ERRE-Luc reporter activities. Cells were seeded at 7 × 104 cells per well in 24-well plates and cultured in DMEM medium with 10% FBS for 24 h before transfection. Cells were cotransfected with the reporter and ERRγ-expression plasmids as mentioned above. After 12 h post-transfection, cells were treated with DY131 at various doses. After treatments for 48 h, cells were harvested for luciferase reporter assay as mentioned above.
All results are expressed as the mean ± SD. Statistical analyses of data were done using two-tailed Student's t test, and differences were considered significant where P < 0.01.
ERRγ expression is reduced in prostatic intraepithelial neoplasia and carcinoma. The specificities of newly generated ERRγ antibodies were verified and evaluated by blotting with FLAG-ERRγ expressed in transient transfected 293 cells and bacterial expressed recombinant FLAG-ERRγ fusion proteins ( Fig. 1A ). Immunohistochemistry of ERRγ revealed that in fetal prostate, the developing epithelial cells in proximal ducts extending from the urethra expressed moderate to intense ERRγ immunoreactivity in their nuclei, whereas epithelial cells in the developing acini and budding outgrowths of epithelial nests in distal region showed variable nuclear staining ( Fig. 1B1). Mesenchymal cells in the stroma also showed positive nuclear staining. In pubertal prostate, both the basal and luminal epithelial cells exhibited moderate to intense nuclear immunoreactivity ( Fig. 1B2). The stromal cells were variably stained at their nuclei. The aged normal prostate also showed similar staining pattern as in pubertal prostate but weaker nuclear immunoreactivities in the epithelial cells ( Fig. 1B3), whereas the ERRγ immunoreactivity of stromal cells in aged prostate became markedly reduced or negative. The dysplastic epithelial cells in prostatic intraepithelial neoplasia (PIN) lesions showed reduced staining in their nuclei ( Fig. 1B4). In well-differentiated adenocarcinoma lesions, the malignant cells showed significantly reduced or negative immunoreactivity in their nuclei ( Fig. 1B5). The controls without primary antibody showed negativity in both epithelial and stromal cells ( Fig. 1B6).
Ectopic ERRγ expression suppresses in vitro growth of prostate cancer cells, which is mediated through its intact DBD. To further elucidate the functional significance of ERRγ in prostate cancer cell growth, we generated ERRγ-stable clones in LNCaP and DU145 cells (LNCaP expresses low endogenous level of ERRγ, whereas DU145 expresses no ERRγ) by retroviral transduction and selected more than 10 stable clones for in vitro and in vivo growth studies ( Fig. 2A ). We did not observe any significant morphologic alterations in DU145-ERRγ clones, as compared with its parental line. However, we observed an induction of long dendritic cellular processes in LNCaP-ERRγ clones, as compared with their empty vector pBabe-infected clones and noninfected parental cells (data not shown). All generated LNCaP-ERRγ and DU145-ERRγ clones proliferated significantly slower than their corresponding pBabe clones and parental cells ( Fig. 2A). Soft agar assay showed that colony formation efficiencies and colony sizes of LNCaP-ERRγ and DU145-ERRγ clones were significantly suppressed by stable expression of ERRγ, as compared with their pBabe clones and parental cells that formed colonies in soft agar ( Fig. 2B).
To show whether the ERRγ-induced growth inhibition in prostate cancer cells were mediated through its intact DBD, we generated ERRγ-ΔZFI mutant with deletion of the first zinc finger in its DBD ( Fig. 2C). We found that the mutant was unable to transactivate the ERE/ERRE-Luc reporters ( Fig. 2C) and did not show any inhibitory effect on the proliferation of LNCaP and DU145 cells ( Fig. 2D).
ERRγ expression prevents in vivo tumor growth of prostate cancer cells. The effect of stable ERRγ expression on the tumorigenicity of LNCaP and DU145 cells was evaluated in immunodeficient mice. Significant suppression of tumor formation was observed in mice bearing the inoculated LNCaP-ERRγ and DU145-ERRγ clones, as compared with mice bearing the empty vector clones or parental cells, in which tumors were formed within 4 to 7 weeks after injection ( Fig. 3 ).
Inhibition of prostate cancer cell proliferation by ERRγ is mediated through suppression of S-phase progression but not apoptosis. Because ectopic ERRγ expression suppressed both in vitro cell proliferation and tumor growth of prostate cancer cells, we next investigated the status of cell-cycle progression and cell proliferation in ERRγ clones by DNA flow cytometry and BrdUrd incorporation. The results of flow cytometry revealed that the isolated ERRγ clones and the pooled ERRγ infectants of LNCaP and DU145 cells exhibited a significant decrease of cell populations at S phase (22–24% in parental cells versus 7–8% in infectants) and a concomitant moderate increase of cell fractions at G0-G1 phase, as compared with their respective pBabe infectants (isolated or pooled infectants) and parental cells ( Fig. 4A and B ). Similar to pBabe infectants and parental cells, the mutant ERRγ-ΔZFI infectants showed no reduction of S-phase populations ( Fig. 4B). Fluorescence-activated cell sorting (FACS) analysis of FITC-Annexin V-PI–labeled cells showed that there was no significant induction of apoptotic and necrotic cells in the LNCaP-ERRγ and DU145-ERRγ clones ( Fig. 4C). The pattern of decreased cell populations at S phase was also consistent to the decreased BrdUrd incorporation in the ERRγ clones ( Fig. 2D), suggesting that DNA replication rate was reduced in these clones. These results, together with the cell count study, suggest that the inhibition of cell proliferation and tumor growth by stable ERRγ expression in LNCaP and DU145 cells was mediated mainly through a blockage of S-phase entry and suppression of DNA replication.
Cell-cycle arrest in prostate cancer cells by ERRγ is accompanied with up-regulation of p21 or p27. We next investigated whether the arrest of cell-cycle progression at G1-S transition in ERRγ-transduced prostate cancer cells could be associated with altered expression of cell-cycle regulators. Western blot analysis showed that there was an increased protein expression of cyclin-dependent kinase inhibitors p21 and p27 in DU145-ERRγ and p21 in LNCaP-ERRγ clones ( Fig. 5A ). A slightly decreased expression of cyclin A but no significant changes in cyclin B1, D1, and D3 were observed in the LNCaP-ERRγ and DU145-ERRγ clones.
Transactivation of p21 and p27 gene promoters by ERRγ. We next sought to find out whether the elevated protein expressions of p21 and p27 in ERRγ infectants was a direct transcription activation of their gene promoters by ectopic ERRγ or an indirect result mediated by other inhibitory growth–related signaling pathways or DNA damage ( 36). Luciferase reporter assay showed that the wild-type ERRγ but not ERRγ-ΔZFI mutant was transcriptionally active because it could activate both ERE3-Luc and ERRE3-Luc reporters in transient transfected ER-negative HeLa cells in the absence of ligand, showing that the exogenously expressed ERRγ was a functional protein ( Fig. 2C). We further manifested by luciferase assay that ERRγ could activate the p21 gene promoter reporter (5-fold increase) and moderately on the p27 gene promoter reporter (2-fold increase) in HeLa cells ( Fig. 5B). However, the ERRγ-ΔZFI mutant was unable to activate any reporter constructs ERE3/ERRE3/p21/p27-Luc ( Figs. 2C and Fig. 5B). Our results suggest that both p21 and p27 could be the target genes of ERRγ, whereas the increased p21 and p27 protein expressions in LNCaP-ERRγ and DU145-ERRγ infectants could be partially mediated through the stimulation of p21 and p27 gene promoter activities.
DY131 inhibits in vitro proliferation of prostate cancer cells. The effect of an ERRγ agonist DY131 on ERRγ-driven ERE3- and ERRE3-Luc reporter activities were evaluated. Administration of DY131 further enhanced ERRγ-activated ERE- and ERRE-driven luciferase transcription activities, with higher response on ERRE3-Luc than ERE3-Luc reporter, at concentrations >1 μmol/L ( Fig. 6A ). DY131 at high concentrations also induced weak luciferase activities in 293 cells transfected with the mutant ERRγ-ΔZFI (data not shown). We next investigated whether DY131 could potentiate further inhibition in ERRγ clones than their parental cells. Cultured cells were exposed to DY131 at concentrations ranging from 0.1 to 30 μmol/L. Under these conditions, DY131 inhibited cell proliferation and BrdUrd incorporation in LNCaP-ERRγ and LNCaP in a dose-dependent manner, with higher inhibition in LNCaP-ERRγ than LNCaP at lower concentrations of DY131 (IC50 ≅ 5 μmol/L for LNCaP; IC50 ≅ 2 μmol/L for LNCaP-ERRγ; Fig. 6B). Similarly, DY131 also inhibited cell proliferation in DU145-ERRγ cells in a dose-dependent manner but only slight inhibition in DU145 cells (IC50 ≅ 26 μmol/L for DU145; IC50 ≅ 14 μmol/L for DU145-ERRγ; Fig. 6C). No significant change in cellular morphology or induction of apoptotic cells was seen in DY131-treated cells (data not shown).
In the present study, we reconfirmed our previous report that ERRγ is mainly expressed as a nuclear protein in epithelial cells in normal adult prostate, whereas its protein expression exhibits a negative pattern with the prostate cancer progression ( 33). By immunohistochemistry, we also observed that besides epithelial localization, ERRγ was also localized to mesenchymal or stromal cells in fetal and pubertal prostates, whereas the stromal cell expression of ERRγ became markedly reduced in adult and aged prostates. The significance of stromal cell localization of ERRγ with reference to epithelial-stromal interaction is unclear. Interestingly, this expression pattern seems to be similar to ERβ in developing and adult prostates ( 37). The colocalization of the ligand-independent ERRγ and ligand-dependent ERβ in prostatic epithelial cells suggests that the two nuclear receptors may cross-talk with one another via the regulation of the same target genes or function independently.
We further investigated the functional role of ERRγ in cell growth regulation in prostate cancer cells and showed for the first time that ERRγ could suppress cell proliferation and tumorigenicity in both androgen-sensitive (LNCaP) and androgen-insensitive (DU145) prostate cancer cells, and the ERRγ-induced cell growth inhibitory effect was linked to a perturbation in cell-cycle progression with a significant suppression of S-phase fractions and decreased DNA replication, but with no apparent induction of apoptosis in ERRγ infectants. The results shown in isolated and pooled LNCaP-ERRγ and DU145-ERRγ infectants are consistent with our recent findings on rat ERRγ that the rat homologue also exerts an antiproliferative action in immortalized rat prostatic epithelial cells and cancer cells stably transfected with ERRγ ( 14), suggesting that the inhibitory function on cell proliferation or arrest of cell-cycle progression by ectopic ERRγ is similar in both human and rat prostatic cells. We also asked whether this growth inhibitory effect was related to cell senescence or differentiation. Senescence-associated β-galactosidase staining revealed no significant difference between the ERRγ clones and their parental cells, suggesting that the ERRγ-induced cell growth inhibition in prostate cancer cells may not be associated with cell senescence (data not shown).
By immunoblot analysis, we further showed that the ERRγ-induced suppression of cell cycle S phase (together with G1 and G2-M arrest) in LNCaP and DU145 cells was accompanied with an increased protein expression of p21 and p27. Because p21 and p27 are key inhibitors of cell-cycle progression and play crucial role in the G1-S phase transition, it is believed that the blockage of G1-S transition in LNCaP-ERRγ and DU145-ERRγ cells may be partially exerted through the elevated levels of p21 or p27. In addition, the inhibition of DNA replication (BrdUrd incorporation) in LNCaP-ERRγ and DU145-ERRγ cells could also be a result of increased levels of p21 because it is well characterized that p21 can directly inhibit DNA replication by its interaction with proliferating cell nuclear antigen (PCNA) and blocking the PCNA-dependent DNA polymerase δ activity ( 38). On the other hand, it is also known that both p21 and p27 genes are estrogen responsive ( 39, 40) and ERα/β inducible in endometrial and breast cancer cells ( 41– 43), whereas their transcriptional regulation by ERα is mediated through interaction with transcription factor Sp1 and histone deacetylase via the Sp1-binding sites in their promoters ( 44, 45). However, it is unclear whether p21 is cotargeted by ERs and ERRγ in prostate cancer cells. In this study, we also showed by luciferase reporter assay that upon transient transfection, ERRγ could stimulate the gene promoter activities of p21 and p27, suggesting that the elevation of p21 and p27 in ERRγ-transduced prostate cancer cells might be attributed to direct transactivation of their gene promoters, in addition to other possible inhibitory cell growth signals ( 36). To map its binding sites, we have done a serial of 5′-deletion analysis on the p21 gene promoter and its activation by transiently expressed ERRγ in HeLa cells. Our result showed that deletion of a region located between nucleotides −2,324 and −2,132 upstream of the p21 transcription start site resulted in a significant loss of activation (6-fold reduction). Chromatin immunoprecipitation assay also confirmed that ERRγ bound directly to p21 promoter (data not shown). However, sequence information revealed that no canonical ERE or 1/2 ERE site was found within this region. Further detailed mapping and binding analysis is required to identify and confirm the putative ERRγ-binding sites required for the positive regulation of p21 transcription by ERRγ or whether its activation at these sites also involves indirect interactions with other transcription factors. In a recent study on the role of RIP140 on transcriptional regulation by ERRs, Castet et al. ( 18) show that besides ERα, ERRα, and ERRβ, ERRγ can also weakly transactivate the p21 gene promoter through two proximal Sp1 response elements located at nucleotides between −82 and −69 in p21 promoter, whereas its transcriptional activity can be further enhanced by RIP140.
In this study, we also evaluated the in vitro efficacy of a selective ERRβ/γ-agonist DY131 ( 31) on prostate cancer cell growth. Our cell-based assay results showed that DY131 at micromolar levels could potentiate the ERRγ-mediated growth inhibition in LNCaP-ERRγ and DU145-ERRγ clones compared with their parental cells (without significant induced apoptosis). It is also believed that the inhibitory effect of DY131 on the LNCaP and DU145 parental cells may be mediated through their endogenous low expression of ERRγ and ERRβ. By comparison, LNCaP/LNCaP-ERRγ cells showed higher sensitivities to DY131 than DU145/DU145-ERRγ cells, which may be attributed to the endogenous expression of ERRγ in LNCaP but not DU145. Our encouraging results suggest that the orphan nuclear receptor ERRγ, which may have an antiproliferative function in prostate cancer, could be a novel therapeutic target for treatment of prostate cancer. However, further experiments are required to elucidate its pharmacology, including molecular mechanism of action besides targeting on ERRγ, in vivo tumor growth inhibition, efficacy in androgen-independent or metastatic prostate cancer, and systemic cytotoxicity. It is worthy to further explore and develop selective ligands targeting to ERRs as new therapeutic drugs or anticancer agents supplementing the conventional hormonal therapies for prostate cancer. Previous studies in prostate cancer cell lines and animal models show that ER antagonists or SERMs, targeting on the ligand-dependent ERα and ERβ, can inhibit growth of prostate cancer and its metastasis ( 46, 47), whereas it is claimed that these drugs may be useful as chemopreventive agents for prostate cancer ( 48). Characterization by binding affinity, reporter gene assays, and structural studies also reveal that some SERMs (tamoxifen, 4-hydroxytamoxifen) can act as ligands or antagonists to ERRs besides ERs ( 27, 29, 30, 49, 50). Thus, these SERMs, if they have any effects on the prostate cancer cells, may target on both ERRs and ERs.
In summary, we showed in this study that ERRγ inhibited the cell proliferation of androgen-sensitive and androgen-insensitive prostate cancer cells by arresting cell-cycle progression at the G1-S phase transition, which was accompanied with an increased expression of two cyclin-dependent kinase inhibitors p21 and p27, suggesting an antiproliferative role of ERRγ in prostate cancer. We further showed by reporter assay that the induction of p21 and p27 by ERRγ was mediated at transcriptional and gene promoter levels. Considering that ERRγ is down-regulated in clinical prostate cancer, our findings suggest that ERRγ may play a negative role in transformed prostatic cell growth and a tumor suppressor function in prostate cancer cells. Finally, we also showed that ERRγ could be useful as a potential therapeutic target for treating prostate cancer because a selective ERRγ agonist, DY131, could potentiate the growth inhibition of ERRγ-positive prostate cancer cells.
Grant support: Direct Grants for Research (2004 and 2005) and Research Grants Council Competitive Earmarked Research Grant CUHK4411/06M.
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.
We thank Prof. Y.C. Wong for critical reading of the manuscript.
- Received October 18, 2006.
- Revision received January 26, 2007.
- Accepted March 7, 2007.
- ©2007 American Association for Cancer Research.