Silencing of androgen-regulated genes using a fusion of AR with the PLZF transcriptional repressor

Article metrics

Abstract

The androgen receptor (AR) is a member of the nuclear receptor superfamily of ligand-activated transcription factors and plays a key role in the development and progression of prostate cancer. Current therapies include the use of antiandrogens aimed at inhibiting the transcriptional activation of AR-regulated genes by AR. Here, we explore a strategy aimed at obtaining silencing of AR-regulated genes, based on the properties of the transcriptional repressor promyelocytic leukamia zinc-finger protein (PLZF). In order to do this, we have made a fusion protein between PLZF and AR, named PLZF-AR, and show that PLZF-AR is able to bring about silencing of genomically encoded AR-regulated genes and inhibit the androgen-regulated growth of LNCaP prostate cancer cells. Together, our results show that this strategy is able to bring about potent repression of AR-regulated responses and, therefore, could be of value in the development of new therapies for prostate cancer.

Introduction

The human prostate is a prototypical androgen-dependent organ. Androgens are required not only for normal growth and development but also for structural and functional integrity (Cunha et al., 1987; Davies and Eaton, 1991). Prostate adenocarcinoma is the most prevalent cancer of men in the West, affecting approximately 25 000 men in the UK each year, and is the second commonest cause of cancer-related deaths in men in industrialized nations. The probability of invasive prostate cancer incidence increases with age, with a lifetime risk of one in six in the USA (Greenlee et al., 2000). The finding by Huggins and Hodges, in the 1940s, that prostate cancer is an androgen-dependent neoplasm led to the demonstration that androgen ablation by orchiectomy (removal of testes) induces regression of prostate cancer (Huggins, 1967). Although orchiectomy is effective, orchiectomy by androgen depletion can now be achieved pharmacologically, through the use of gonadotropin-releasing hormone (GnRH) receptor agonists that act on the pituitary to block luteinizing hormone (LH) release, leading to inhibition of testosterone secretion from the testis, while total androgen ablation involves combination therapy using an LHRH agonist, together with antiandrogens that compete with androgens for binding to the androgen receptor (Labrie et al., 1986, 1993). Current therapies therefore rely on initial prostatectomy or radiotherapy, and androgen ablation.

Androgen actions are mediated by the androgen receptor (AR), a member of the nuclear hormone receptor superfamily of ligand-activated transcription factors, the AR being most closely related to the glucocorticoid, mineralocorticoid and progesterone receptors. The nuclear receptor superfamily also includes receptors for oestrogen, vitamin D3, thyroid hormone and retinoic acid, as well as receptors activated by peroxisome proliferators, bile acids and prostaglandins (Chawla et al., 2001). These receptors are characterized by a DNA binding domain (DBD) comprised of two zinc-fingers. AR binds as a homodimer to palindromic sequences known as androgen response elements in promoters of androgen-regulated genes. The hormone or ligand binding domain (LBD) is encoded within a region of about 300 amino acids and is bound by androgens and antiandrogens. The LBD also contains a hormone-activated transcription activation function, AF2, as well as sequences required for ligand-dependent dimerization. N-terminal to the DBD is transcription activation function AF1, which in the AR is characterized by the presence of a polyglutamine repeat, expansion of which reduces transactivation. Expansion in the polyglutamine repeat has been implicated in spinobulbar muscular atrophy or Kennedy's disease, where affected men often display mild androgen insensitivity (Mhatre et al., 1993). In the absence of ligand, the AR exists in the cytoplasm, maintained in an inactive state in complex with heat shock proteins. Ligand binding causes a conformational change, release from the heat shock protein complex, dimerization and translocation to the nucleus and regulation of gene expression (Montgomery et al., 2001). Stimulation of the expression of androgen-regulated genes requires the ligand-regulated recruitment of transcriptional coactivator complexes that facilitate transcription by stabilization of the preinitiation complex and/or through chromatin remodelling and modification (Glass and Rosenfeld, 2000; Dilworth and Chambon, 2001; McKenna and O’Malley, 2002).

The AR is expressed in virtually all prostate cancers and the AR gene is frequently amplified and subject to mutation in this disease (Sadi et al., 1991; van der Kwast et al., 1991). In addition to androgen ablation, AR antagonists are used to oppose the action of androgens. However, a major problem with current therapies is that a large number of tumours that initially respond to endocrine and other therapies eventually develop resistance, as the surviving tumour cells lose their dependency on androgens for growth and proliferate in the absence of serum androgens. The majority of resistant prostate cancers continue to express AR and although the molecular mechanisms underlying the acquisition of androgen-independent growth in the continuing presence of AR have not been defined, several potential mechanisms have been proposed. These include AR gene amplification, mutations of the AR gene, resulting in promiscuous ligand binding and activation of AR by other endocrine or paracrine growth factors, independent of ligand binding (Feldman and Feldman, 2001). These findings argue that AR continues to play an important role in the acquisition of resistance to endocrine therapies.

Taken together, current evidence strongly suggests that AR regulation is important, even in recurrent prostate cancers that have become androgen independent and indicate that androgen-regulated genes are important in progression of androgen-independent prostate cancer. In considering this, we have explored the possibility of inhibiting the expression of such genes by facilitating the recruitment of transcriptional corepressor complexes that silence gene expression by chromatin remodelling and histone modification. Such a strategy is exemplified by the mechanisms underlying acute promyelocytic leukaemia, where chromosomal translocations lead to the formation of fusion proteins between the nuclear hormone receptor retinoic acid α (RARα) and transcriptional repressor proteins (Zelent et al., 2001). Such fusion proteins act to bind to retinoic acid receptor regulatory elements encoded in retinoic acid-regulated genes and result in the recruitment of histone deacetylase complexes, thereby bringing about gene repression. In this respect, the fusion of the DBD and LBD of RARα with the promyelocytic leukamia zinc-finger protein, PLZF, produces a potent transcriptional repressor that specifically acts to inhibit retinoic acid-regulated genes, thereby blocking myelocyte differentiation in acute promyelocytic leukaemia (David et al., 1998; Grignani et al., 1998; Lin et al., 1998).

Here, we have explored the possibility that fusion of the PLZF sequences found in PLZF-RARα, to AR sequences homologous to the RARα sequences present in PLZF-RARα, would generate a specific and potent inhibitor of androgen-regulated gene expression. We show that PLZF-AR is indeed a potent repressor of androgen-regulated reporter genes and genomically encoded androgen-regulated genes. Further, we show that PLZF-AR can be virally transduced into the AR-positive and androgen-regulated prostate cancer cell line, LNCaP (Horoszewicz et al., 1980), to bring about repression of androgen-regulated genes and inhibition of androgen-regulated growth. Together, these findings suggest that PLZF-AR may be useful as the basis of a gene therapy for prostate cancer.

Results

PLZF-AR inhibits the expression of an androgen-regulated reporter gene

The PLZF-RARα fusion protein contains amino acids 1–456 of PLZF fused to sequences encoded within exons 4–8 of the RARα gene, comprising the DBD and LBD (Chen et al., 1993a). The analogous AR sequences encoding amino acids 530–919, which include the AR DBD and LBD, were cloned in frame with amino acids 1–456 of PLZF to generate PLZF-AR (Figure 1a). Immunoblotting using antibodies specific for the LBD of AR (AR-C) showed the presence of a 100 kDa polypeptide, consistent with the predicted molecular weight of PLZF-AR (Figure 1b, lane 4). As expected, this polypeptide was not detectable by using an antibody specific for the N-terminal transcription activation domain AF1 (AR-N). Immunoblotting with anti-PLZF revealed PLZF-AR and PLZF, but did not detect AR (lanes 2, 4, 6).

Figure 1
figure1

Schematic representation of the fusion proteins used in this study. (a) The positions of the DNA binding domain (DBD), the ligand binding domain (LBD) and transcription activation functions AF1 and AF2 of RARα, ERα and AR are shown. PLZF is also shown, with the POZ domain, required for mediating recruitment of the Sin3A HDAC complex and the nine zinc-fingers (ovals) present in PLZF being highlighted. PLZF-RARα is the protein product resulting from a t(11; 17) translocation in APL. Replacement of the RARα sequences by the analogous portion of AR resulted in PLZF-AR. In PANER, the AR LBD was replaced by the LBD from ERα. (b) Whole-cell extracts from transiently transfected COS-1 cells were immunoblotted using AR antibodies specific for AF1 of AR (AR-N), absent from PLZF-AR and PANER, the C-terminus of AR (AR-C) or PLZF

In transiently transfected COS-1 cells, AR activated an ARE-containing CAT reporter gene in the presence of the agonist R1881, but not in the absence of ligand (Figure 2a, lanes 3, 4). Cotransfection of PLZF-AR with AR reduced reporter gene activity to basal levels (lane 6), whereas PLZF did not repress AR activity (lane 8). PLZF and PLZF-AR did not activate the reporter gene in the presence or absence of R1881. Indeed, the basal activity of the reporter was reduced by PLZF-AR (lanes 9–12).

Figure 2
figure2

PLZF-AR inhibits an androgen-responsive reporter gene in COS-1 cells. (a) COS-1 cells were transiently transfected with an androgen-responsive CAT reporter gene, together with full-length AR, PLZF-AR or PLZF, in the presence of R1881 (1 nM) or vehicle, as appropriate. Results represent three independent experiments. AR activity in the presence of R1881 was taken as 100%. All other activities are shown relative to this. (b) In order to determine whether the transcriptional repression by PLZF-AR is ligand-regulated, COS-1 cells were transiently transfected with ARΔLBD, together with PLZF-AR or PLZF. ARΔLBD activity in the presence of R1881 was taken as 100%. (c) Responsiveness of PLZF-AR to androgens and antiandrogens was investigated following transient transfection of COS-1 cells with full-length AR or ARΔLBD, together with PLZF-AR. AR activity in the presence of R1881 was taken as 100%. All other activities are shown relative to this. (d) COS-1 cells were transiently transfected with full-length AR or ARΔLBD, together with PLZF-AR or PANER. The androgen R1881 (1 nM), 17β-oestradiol (E2, 10 nM) and 4-hydroxytamoxifen (OHT; 100 nM) were added as shown. AR activity in the presence of R1881 was taken as 100%. In all cases, (ad), the results of three independent experiments are shown

As expected, AR lacking the LBD (ARΔLBD) stimulated the reporter gene in the absence, as well as in the presence of ligand (Figure 2b, lanes 3, 4). Interestingly, PLZF-AR inhibited ARΔLBD activity in the presence of R1881, but not in its absence (Figure 2b, lanes 5, 6), indicating that the repression by PLZF-AR is ligand dependent. Moreover, the ability of PLZF-AR to inhibit androgen-regulated reporter gene activity reflected the agonist/antagonist activity of the ligand. Thus, PLZF-AR inhibited reporter gene activity in the presence of the agonists R1881, testosterone and dihydrotestosterone (Figure 2c, lanes 13–15), but not in the absence of ligand (lane 12), or in the presence of the antiandrogens casodex or hydroxyflutamide, that block AR activation (compare lanes 7, 8 and lanes 16, 17), whereas, cyproterone acetate (CPA), which is a mixed agonist/antagonist, activated AR, albeit to a lower extent than the full agonists (see lane 9) and was also able to stimulate repression by PLZF-AR (Figure 2c. lane 18). These results indicate that PLZF-AR is a potent ligand-regulated inhibitor of androgen-responsive gene expression.

To further explore the ligand dependence of repression of androgen-regulated genes by PLZF-AR fusion proteins, we generated PANER, in which the AR LBD in PLZF-AR was substituted by the oestrogen receptor α LBD (Figure 1a). PANER did not stimulate reporter gene activity (Figure 2d lanes 11–16) and inhibited reporter gene activation by ARΔLBD in a ligand-independent manner, although the inhibition was greater when the ERα ligands, oestrogen or the anti-oestrogen 4-hydroxytamoxifen (OHT) were added (lanes 2–6). PANER also inhibited the reporter when cotransfected with AR in the presence of R1881 (lane 18). Interestingly, addition of oestrogen or OHT increased reporter gene inhibition (lanes 20, 22), indicating that transcriptional repression by PANER is ligand-regulated, although considerable ligand-independent repression is observed for PANER, in contrast to PLZF-AR.

Previous studies have established that AR is cytoplasmic in the absence of ligand and that ligand binding enables localization to the nucleus. Furthermore, intracellular localization of AR has been shown to be mediated by sequences located close to, or within the LBD (Zhou et al., 1994; Tyagi et al., 2000; Tomura et al., 2001). The lack of transcriptional repression by unliganded PLZF-AR may be due to its exclusion from the nucleus. In order to investigate this possibility, COS-1 cells were transiently transfected with AR or PLZF-AR, in the presence of androgen agonists and antagonists. Indirect immunofluorescence showed that AR was cytoplasmic in the absence of ligand and localized to the nucleus in the presence of the ligands, R1881, dihydrotesterosterone (DHT) or the antiandrogens CPA and casodex (Figure 3). Similarly, unliganded AR was cytoplasmic. However, while agonist bound PLZF-AR was nuclear, only partial localization to the nucleus was observed in the presence of CPA, while PLZF-AR was exclusively cytoplasmic in the presence of casodex.

Figure 3
figure3

Intracellular localization of PLZF-AR is regulated by ligand binding. COS-1 cells transiently transfected with AR or PLZF-AR were treated with R1881 (1 nM), DHT (100 nM), CPA (100 nM) or casodex (100 nM) for 48 h prior to methanol fixation and incubation with an AR antibody that recognizes sequences within AF1 (Zegers et al., 1991) or a PLZF antibody, as appropriate. The cells were subsequently incubated with FITC-conjugated goat anti-mouse immunoglobulins and DAPI for visualization of nuclei. Immunofluorescence staining was visualized using a Zeiss LS510 confocal microscope

Stable expression of PLZF-AR inhibits the expression of androgen-regulated genes

In order to determine if PLZF-AR can silence the expression of genomically encoded androgen-regulated genes, we stably transfected the T47D breast cancer cell line. T47D cells express AR and have been used for investigation of AR function (Cleutjens et al., 1996; Takane and McPhaul, 1996; Cleutjens et al., 1997; Blankvoort et al., 2001). Moreover, several well-characterized androgen-regulated genes are expressed in an androgen-dependent manner in T47D cells (Monne et al., 1994; Hsieh et al., 1997; Heemers et al., 2000), but unlike the AR-positive LNCaP prostate cancer cells, T47D cells are not dependent on androgens for growth, thereby allowing investigation of the silencing of androgen-regulated gene expression in a growth-independent manner. Stable lines were generated following transfection of pcDNA3.1/hygromycin, encoding PLZF or PLZF-AR. After several weeks of selection with hygromycin B, antibiotic-resistant colonies were picked and used to derive independent lines for evaluation. RT–PCR was used initially to identify PLZF-AR- and PLZF-expressing lines (data not shown; and see Figure 4a). Three lines expressing PLZF-AR, one line expressing PLZF and a line transfected with the control vector were further characterized for expression and regulation of androgen-responsive genes. Immunoblotting for AR and PLZF confirmed that PLZF- and PLZF-AR are expressed in the appropriate lines (Figure 4b). Moreover, immunoblotting with AR-C, which recognizes AR and PLZF-AR, showed that levels of PLZF-AR in the PLZF-AR lines are similar to that of endogenous AR (Figure 4b), while immunblotting with AR-N, which recognizes an epitope present in AR, but absent in PLZF-AR, indicated that AR levels in these lines were similar to the levels of AR in the parental T47D cells.

Figure 4
figure4

Characterization of T47D cell lines stably expressing PLZF-AR. (a) Total RNA prepared following treatment of T47D, or T47D-transfected stable cell lines with R1881 or vehicle, for 48 h was subjected to RT–PCR for DRG-1, PLZF-AR, AR and GAPDH using 35 cycles of amplification. (b) Whole-cell lysates were subjected to SDS–PAGE and immunoblotting using antibodies to PLZF or AR (AR-C, AR-N). Positions of PLZF, PLZF-AR and AR are indicated on the right of each panel. (c, d) Selected lines were transiently transfected with ARE-2-tk-CAT or the oestrogen-regulated reporter, ERE-G-CAT. The results of three independent experiments are shown. Reporter gene activity in the parental T47D cells in the absence of ligand (lane 1) was taken as 1 and all other activities are shown relative to this

To investigate the functionality of stably expressed PLZF-AR, the stable lines were transiently transfected with ARE-2-tk-CAT. As expected, CAT activity was stimulated by R1881 in T47D cells and to a lower extent by the partial antagonist CPA, but not by casodex (CSX) (Figure 4c, lanes 1–4). Similar results were obtained for lines stably transfected with control vector or PLZF (Figure 4c, lanes 5–12). For each of the PLZF-AR clones tested, no significant reporter gene activity above background was observed in the presence of R1881 or CPA (lanes 13–24), indicating that PLZF-AR represses AR activity in these lines, while the lack of agonist activity for CSX prevents any conclusions regarding PLZF-AR activity in the presence of CSX.

The androgen receptor binds as a homodimer to palindromic response elements with each half-site conforming to the consensus sequence AGAACA. The related oestrogen receptors bind to a related but distinct palindromic sequence, each half-site conforming to the sequences AGGTCA (Schwabe et al., 1993). In order to determine whether inhibition of androgen-regulated genes by PLZF-AR in these lines is due to direct binding to promoter sequences and not to nonspecific effects such as sequestration of limiting transcription intermediary factors, the activity of an oestrogen-responsive CAT reporter gene was investigated in the T47D PLZF-AR lines. In T47D cells, ERE-globin-CAT (ERE-G-CAT) activity was stimulated by 17β-oestradiol (E2), but not by R1881, which is not a ligand for oestrogen receptors (Figure 4d). Similar activities were obtained in the pcDNA3.1 and PLZF-transformed lines and in the PLZF-AR lines. Although reporter gene activity was lower in the PLZF-AR03 line, the level of induction by E2, compared to the no ligand control, was similar to that observed in the other lines and in the parental T47D cells. Moreover, no repression by PLZF-AR was evident in the presence of R1881, suggesting that the lower CAT activity in this line is not due to PLZF-AR, but is more likely to be due to other effects, such as reduced ER levels in these cells.

RT–PCR analysis was also used to investigate the effects of PLZF-AR on the expression of androgen-regulated genes in T47D cells (Figure 4a). DRG-1 is a well-defined androgen-regulated gene whose expression is greatly stimulated by androgens in the androgen-responsive prostate cancer-derived LNCaP cells (Ulrix et al., 1999). Low level of DRG-1 expression was strongly stimulated by R1881 in T47D cells, as well as in the pcDNA3.1 and PLZF lines. By contrast, expression of DRG-1 was not stimulated by R1881 in PLZF-AR lines, demonstrating that PLZF-AR represses the expression of genomically encoded androgen-regulated genes.

Adenoviral gene transduction of PLZF-AR can be used to inhibit androgen-regulated gene expression and growth in prostate cancer cells

In order to ascertain if PLZF-AR can lead to silencing of androgen-regulated gene expression following viral transduction, we utilized the pADEasy adenoviral system (He et al., 1998b). PLZF-AR was cloned into the pADTrack-CMV shuttle vector, which encodes the green fluorescent protein (GFP) gene, allowing infected cells to be visualized, as well as enabling infected cells to be separated from the uninfected population by Fluorescence Activated Cell Sorting. The androgen-responsive LNCaP human prostate cancer cell line was transiently transfected with pADTrack-CMV encoding PLZF-AR (PLZF-AR Adeno) or the control vector (Control Adeno). Transfected cells were purified by cell sorting for GFP expression and cultured in the presence or absence of the AR agonist R1881 for 72 h, at which time culture media were collected and the amount of prostatic serum antigen (PSA) was determined by immunoassay. Expression of the PSA gene is androgen-regulated and its levels are elevated in the serum of patients with prostate cancer. Thus, serum PSA level determination is used to aid diagnosis for prostate cancer and its levels are used for monitoring response to endocrine therapies. In the presence of androgen, expression of PSA is stimulated in prostate cells, including LNCaP cells (Horoszewicz et al., 1980; Chung et al., 2001). Measurement of PSA levels in medium from LNCaP cells transfected and purified by cell sorting showed that addition of R1881 resulted in a 25-fold increase in levels of secreted PSA in cells transfected with control Adeno (Figure 5a, lanes 3, 4). By contrast, PSA levels in PLZF-AR Adeno-transfected cells were raised fivefold, demonstrating a fivefold decrease in PSA levels, relative to the control (compare lanes 2 and 4). At the time of assaying for PSA, approximately 70% of the sorted cells remained GFP positive, in both PLZF-AR and in control adenovirus-transfected cultures. This suggests that the R1881-induced PSA activity in PLZF-AR-transfected cultures arises from cells that have lost construct DNA, or GFP-negative cells that may have inadvertently been purified through the cell sorting and have expanded, although the GFP negativity may be due to misfolding of GFP in these cells.

Figure 5
figure5

PLZF-AR gene transduction in LNCaP cells. (a) The human prostate LNCaP cell line was transiently transfected with the adenoviral shuttle vector pADTrack-CMV encoding PLZF-AR (PLZF-AR Adeno) or empty vector (Control Adeno). Transfected cells were sorted for GFP expression and cultured in the presence or absence of the AR agonist R1881. After 72 h, culture media were collected and the amount of PSA was determined by immunoassay. Cells transfected with the parent vector were treated in the same way. PSA levels were also determined for the medium used to culture the cells (lane 5). (b) LNCaP cells were virally transduced with control adenovirus or adenovirus encoding PLZF-AR. Infected cells were purified by cell sorting for GFP expression and cultured in the presence or absence of the AR agonist R1881 for 48 h. Expression of the androgen-regulated genes PSA, DRG-1 and GAPDH (control) was determined by RT–PCR. (−) refers to uninfected control cells. (c) Equal numbers of adenovirally infected LNCaP cells purified by cell sorting were seeded in complete medium lacking androgen. R1881 (0.1 nM) was added after 48 h and cell growth followed by counting. Counts for cells infected with control virus treated with R1881 (□) or vehicle () and PLZF-AR-infected cells treated with R1881 () or vehicle () are shown and represent at least three replicates. The error bars are too small to be visualized in the graphs shown

Following the encouraging results of the LNCaP transfections, PLZF-AR and control adenoviruses were generated and used to infect LNCaP cells. Infected cells were purified by cell sorting for GFP and cultured in the presence of R1881, and the expression of the PSA, DRG-1 and GAPDH (control) genes was determined by RT–PCR. Expression of PSA and DRG-1 was similar in uninfected LNCaP and in cells infected with the control adenovirus, whereas expression of both genes was significantly reduced following infection with the PLZF-AR adenovirus (Figure 5b).

The growth of adenovirally infected LNCaP cells was clearly stimulated by R1881, a greater than 10-fold increase in cell number being obtained after 9 days in culture, when compared with similarly infected cells cultured in the absence of androgen (Figure 5c). By contrast, little androgen-stimulated growth was observed for cells infected with Ad-PLZF-AR.

Discussion

Regulation of eukaryotic gene expression and Gene ICE

The last decade has seen major advances in understanding the mechanisms by which DNA binding transcription factors regulate gene expression, through the recruitment of transcriptional coregulators (Workman and Kingston, 1998; Narlikar et al., 2002). Of particular importance has been the demonstration that many coregulator complexes remodel chromatin through ATP-dependent mechanisms and the description of coregulator complexes that facilitate gene regulation by post-translational modification of the amino-terminal tails of core histones in chromatin. Interplay between different types of histone tail modification result in dynamic as well as long-term transitions between transcriptionally active or silent chromatin states. Particular attention has been given to histone acetylation, where it has been recognized that hyperacetylation of histone tails in nucleosomes around gene promoters generally correlates with transcription activation and the identification of several transcriptional coactivators that are histone acetyl transferases and which are recruited to the chromatin by interaction with DNA binding transcription factors (Roth et al., 2001). Conversely, transcriptional repressors can recruit histone deacetylase (HDAC) complexes that deacetylate histone tails to bring about gene silencing (Ayer, 1999; Cheung et al., 2000). The importance of HDACs in gene silencing has been further emphasized by the discovery of HDACs in complexes involved in binding to methylated CpG-islands (Bird and Wolffe, 1999). CpG methylation has long been known to be associated with transcriptional repression (Razin and Riggs, 1980) and these discoveries provide the basis for novel ideas of how chromatin remodelling, DNA methylation and histone deacetylation might be linked together, to achieve long-term gene silencing (Zhang et al., 1998, 1999).

By fusing a protein possessing, or capable of recruiting, a repressive activity to a specific DNA binding protein provides a method for specifically silencing a gene(s) of interest. This could be achieved by direct targeting of histone deacetylases, histone methylases or other chromatin-modifying activities to the target gene promoter, or alternatively, by targeting transcriptional repressors, which can recruit chromatin-modifying activities. The PLZF-RARα fusion protein, resulting from the chromosomal translocation t(11,17), which underlies the pathogenesis of retinoic-acid-resistant acute promyelocytic leukaemia (APL) (Chen et al., 1993b) provides a naturally occurring example of gene repression by fusion of a DNA binding transcriptional regulator (RARα) to a transcriptional repressor (PLZF). PLZF possesses a strong transcriptional repressor activity that is mediated, at least in part, by interaction with components of the mSin3A–corepressor complex, including mSin3A, NCoR, SMRT and HDAC1/2, to recruit histone deacetylase activity (David et al., 1998; Grignani et al., 1998; Guidez et al., 1998; He et al., 1998a; Lin et al., 1998; Wong and Privalsky, 1998). Fusion of the PLZF transcriptional repressor with RARα, a member of the NR family of transcription factors (Chawla et al., 2001) leads to targeted inactivation of retinoic acid (RA)-regulated genes (Chen et al., 1993b). Capitalizing on this, we decided to test whether an analogous construct, expressing PLZF-AR could be used for the targeted repression of androgen-regulated genes. Further, we have named this approach to gene silencing, Gene Inactivation by Chromatin Engineering, or ‘Gene ICE’.

Directed silencing of the expression of androgen-regulated genes

In agreement with previous reports, immunofluorescent staining demonstrated that unliganded AR is cytoplasmic, whereas agonist- and antagonist-bound AR is localized to the nucleus. PLZF-AR was also cytoplasmic in the absence of ligand but nuclear in the presence of the agonists R1881 and DHT. However, while AR was nuclear in the presence of the mixed agonist/antagonist CPA and the complete antagonist casodex, only partial nuclear localization was observed for PLZF-AR in the presence of CPA and PLZF-AR was excluded from the nucleus in the presence of casodex. This implies a role for the N-terminal AF-1 of AR mediating nuclear localization. Alternatively, the PLZF sequences are responsible for the cytoplasmic localization of PLZF-AR, although PLZF is itself a nuclear protein (Reid et al., 1995).

In transient transfection studies, PLZF-AR inhibited the activation of an androgen-responsive reporter gene, when cotransfected with AR. Cotransfection with constitutively active AR that lacks the LBD (ARΔLBD) showed that repression by PLZF-AR is ligand dependent. Thus, no repression was observed in the absence of ligand, while AR agonists R1881, testosterone and DHT, as well as the partial antagonist CPA, enabled repression by PLZF-AR. However, the pure antiandrogens hydroxyflutamide and casodex prevented repression by PLZF-AR. The simplest explanation for this ligand regulation is that ligand binding is required for nuclear localization of PLZF-AR, as the unliganded PLZF-AR and the casodex-bound PLZF-AR were excluded from the nucleus, whereas in the presence of R1881, DHT and CPA, PLZF-AR was nuclear and inhibited AR activity.

T47D cells stably expressing PLZF-AR similarly repressed the activity of an androgen-regulated reporter gene, but not that of an oestrogen-regulated reporter gene, while stable lines expressing PLZF did not affect androgen-regulated reporter gene activity, indicating that PLZF-AR specifically targets androgen-responsive genes. Note, however, that in T47D-PLZF cells reporter gene activity in the presence of CPA was similar to that observed in the presence of R1881. It has been proposed, at least in the case of ERα, that the activity of mixed agonists/antagonists is due to coactivator/corepressor ratios in cells (Shang and Brown, 2002). The increased agonist activity of CPA observed in this line may reflect the action of PLZF in sequestering away corepressors, and thereby enabling greater recruitment of coactivators by the CPA-bound AR. A similar mechanism is unlikely to be at work in the case of a pure antagonist such as CSX, perhaps due to the fact that CSX-bound PLZF-AR is not nuclear.

Importantly, PLZF-AR expression was seen to inhibit stimulation of the expression of the androgen-regulated DRG-1 gene in the T47D lines. Moreover, expression of PSA and DRG-1 genes following transfection or transduction of the LNCaP prostate cancer cell with an adenovirus encoding PLZF-AR was also inhibited. Moreover, proliferation of infected LNCaP cells in response to androgen was almost totally inhibited, when compared with cells infected with a control adenovirus.

Although repressors designed to obtain inhibition of androgen-regulated gene expression using AR fusion proteins have been described before (Bramlett et al., 2001), these have not been shown to inhibit endogenous gene expression, or affect androgen-regulated growth of prostate cancer cells. Our results show that PLZF-AR is able to bind to and silence endogenous genes such as PSA and DRG-1, targeted by the DNA binding domain portion of PLZF-AR. Further, we demonstrate that Gene ICE proteins, such as PLZF-AR, are able to be transduced using adenovirus, a vector system that has already seen application in gene therapy, including prostate cancer gene therapies (Mabjeesh et al., 2002). In summary, our results suggest that a strategy directed at chromatin remodelling in prostate cancer potentially offers strong silencing of androgen-regulated genes in the tumour.

Materials and methods

Plasmids

The expression vector pSG5 (Green et al., 1988), pSG5-PLZF (Chen et al., 1993a) and pSVAR0, encoding the human AR coding sequence (Bevan et al., 1999), have previously been described. PLZF-AR was generated by site-directed mutagenesis of pSG5-PLZF to create an XhoI site following sequences encoding amino acid 456 of PLZF. Sequences encoding amino acids 530–919 of human AR were cloned into the XhoI site following mutagenesis, thereby replacing amino acids 457–673 of PLZF. ARΔLBD was generated by site-directed mutagenesis of pSVAR0, by converting valine 659 to the stop codon TGA. PLZF-AR coding sequences were excised from pSG5-PLZF-AR by digestion with BamHI/EcoRI and cloned into pcDNA3.1/Hygromycin B (+) (Invitrogen, UK). NarI sites were generated in PLZF-AR encompassing the codons encoding amino acids 625/626 of AR and in ERα, (HEG0; Tora et al., 1989) by changing codons encoding amino acids 261/262. The NarI/BamHI fragment from PLZF-AR was then replaced by the NarI/BamHI fragment from HEG0 to generate PANER. All constructs were verified by automated DNA sequencing.

Cell transfection, extract preparation and immunoblotting

COS-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco-BRL), supplemented with 5% fetal calf serum (FCS). Cells were plated in 9 cm dishes, 16–24 h prior to transfection with 5 μg of each expression plasmid by the calcium phosphate co-precipitation technique, followed by extract preparation, as described (Chen et al., 1999). Immunoblotting was performed by fractionation of 10 μg of WCE by SDS–PAGE and immunodetection using antibodies specific for the N- or C-terminal portions of AR (SP16 (Zegers et al., 1991) and C-19 (Santa Cruz, UK), respectively) and PLZF (Reid et al., 1995), as described (Chen et al., 1999).

Reporter gene assays

COS-1 cells were seeded into 24-well plates containing DMEM lacking phenol red and supplemented with 5% dextran-coated charcoal-treated FCS (DSS) (Berthois et al., 1986) and transfected as above, using 400 ng of ARE-2-tk-CAT (kind gift of Dr C Bevan), and 50 ng of the β-galactosidase reference plasmid, pCH110 (Promega, UK), together with 50 ng of pSG5, AR, ARΔLBD, PANER, PLZF and/or PLZF-AR expression plasmids, as appropriate. Bluescript M13+ (BSM) was included as a carrier to a total of 2 μg of DNA. T47D and derived cell lines were similarly transfected with 1 μg of ARE-2-tk-CAT, 125 ng of pCH110 and BSM to a total of 5 μg of DNA. Ligands or an equal volume of vehicle (ethanol) were added 30 min after transfection. Precipitates were removed by washing cells 16 h after transfection and fresh ligands added, as appropriate. After a further 24 h (COS-1 cells) or 48 h (T47D and derived cell lines), the cells were harvested in 200 μl of 1 × reporter gene assay SDS-lysis buffer and β-galactosidase and CAT assays were performed using ELISA kits (Roche, UK).

Indirect immunofluorescence

COS-1 cells seeded in DMEM supplemented with 5% DSS onto sterile round glass coverslips in 24-well plates were transiently transfected with 300 ng of AR or PLZF-AR using Fugene-6 (Roche, UK). Ligands were added 30 min following transfection. Cells were washed with ice-cold PBS 48 h after transfection, fixed in ice-cold 100% methanol for 10 min, and incubated with 10% goat serum in PBS for 30 min, followed by addition of AR (SP16; Zegers et al., 1991) or PLZF antibodies for 1 h at room temperature. Following three washes with PBS and blocking with 10% goat serum in PBS, the cells were incubated with FITC-conjugated goat anti-mouse immunoglobulins (Sigma, UK) for 1 h at room temperture. The cells were washed with PBS and mounted onto microscope slides using Vectashield mountant containing DAPI nuclear stain (Vector Laboratories, UK) and the cells visualized using a Zeiss LS510 confocal microscope.

Generation of T47D cells stably transfected with PLZF-AR

T47D cells, cultured in DMEM supplemented with 10% FCS, were transfected with pcDNA3.1, encoding PLZF or PLZF-AR, using Effectene (Qiagen, UK), according to the manufacturer's protocol. The cells were washed 16 h following transfection and after a further 24 h hygromycin B (80 μg/ml) was added to the culture medium. Hygromycin B-resistant clones were screened for PLZF and PLZF-AR expression by RT–PCR, immunofluorescence staining and immunoblot analysis.

Recombinant adenovirus construction and infection of LNCaP cells

Recombinant adenovirus genomes were constructed using the system of He et al. (1998b). PLZF-AR was cloned into the EcoRV restriction enzyme site of the adenoviral shuttle vector pAdTrack-CMV to generate pAdTrack-CMV-PLZF-AR, which was cotransformed with the adenoviral genome contained in the plasmid pAdEasy into Escherichia coli BJ5183 cells, where homologous recombination was used to generate the recombinant adenoviral genome Ad-PLZF-AR. A control adenoviral genome, in which pAdTrack-CMV had been recombined with pAdEasy was also generated. Recombinant adenoviral genomes were packaged in HEK293 cells by transfection followed by rounds of infection and lysis in these cells to obtain packaged viral stocks. Viral stocks were used to infect LNCaP cells and transduced cells were recovered by cell sorting for GFP expression (FACS Vantage; Becton Dickinson, UK). Sorted cells were used in cell growth and gene expression analysis.

LNCaP cell growth assays

Cells (2 × 104) were seeded in 24-well dishes in DMEM supplemented with 10% FCS and left overnight to settle. The media were changed to phenol red-free DMEM supplemented with 10% DSS and stimulated with 100 pM R1881 or equivalent volume of ethanol. The cells were trypsinized and cell numbers determined using a Coulter counter (Coulter Electronics, UK) at varying time intervals, following R1881 addition.

RT–PCR analysis of gene expression

RT–PCR analysis of gene expression was carried out as described previously (Liu et al., 1996). RT–PCR was carried out for the androgen-regulated genes encoding PSA(5′-IndexTermCTTGTGGCCTCTCGTGGCAG-3′ and 5′-IndexTermCTGAGGGTGAACTTGCGCAC-3′) and DRG-1 (5′-IndexTermCGAGAGCTTTACATGGCTCTG-3′ and 5′-IndexTermTCATTGATGAACAGGTGCAG-3′). Control RT–PCR analysis for mRNA dosage was carried out for glyceraldheyde-3-phosphate dehydrogenase (GAPDH; 5′-IndexTermCCACCCATGGCAAATTCCATGGCA-3′ and 5′-IndexTermTCTAGACGGCAGGTCAGGTCCAC-3′).

PSA assay

PSA levels in medium harvested from LNCaP cell cultures were determined using the Architect chemiluminescence total PSA kit (Abbott, UK).

References

  1. Ayer DE . (1999). Trends Cell Biol., 9, 193–198.

  2. Berthois Y, Katzenellenbogen JA and Katzenellenbogen BS . (1986). Proc. Natl. Acad. Sci. USA, 83, 2496–2500.

  3. Bevan CL, Hoare S, Claessens F, Heery DM and Parker MG . (1999). Mol. Cell. Biol., 19, 8383–8392.

  4. Bird AP and Wolffe AP . (1999). Cell, 99, 451–454.

  5. Blankvoort BM, de Groene EM, van Meeteren-Kreikamp AP, Witkamp RF, Rodenburg RJ and Aarts JM . (2001). Anal. Biochem., 298, 93–102.

  6. Bramlett KS, Dits NF, Sui X, Jorge MC, Zhu X and Jenster G . (2001). Mol. Cell. Endocrinol., 183, 19–28.

  7. Chawla A, Repa JJ, Evans RM and Mangelsdorf DJ . (2001). Science, 294, 1866–1870.

  8. Chen D, Pace PE, Coombes RC and Ali S . (1999). Mol. Cell. Biol., 19, 1002–1015.

  9. Chen SJ, Zelent A, Tong JH, Yu HQ, Wang ZY, Derre J, Berger R, Waxman S and Chen Z . (1993a). J. Clin. Invest., 91, 2260–2267.

  10. Chen Z, Brand NJ, Chen A, Chen SJ, Tong JH, Wang ZY, Waxman S and Zelent A . (1993b). EMBO J., 12, 1161–1167.

  11. Cheung WL, Briggs SD and Allis CD . (2000). Curr. Opin. Cell. Biol., 12, 326–333.

  12. Chung BH, Mitchell SH, Zhang JS and Young CY . (2001). Carcinogenesis, 22, 1201–1206.

  13. Cleutjens KB, van der Korput HA, van Eekelen CC, van Rooij HC, Faber PW and Trapman J . (1997). Mol. Endocrinol., 11, 148–161.

  14. Cleutjens KB, van Eekelen CC, van der Korput HA, Brinkmann AO and Trapman J . (1996). J. Biol. Chem., 271, 6379–6388.

  15. Cunha GR, Donjacour AA, Cooke PS, Mee S, Bigsby RM, Higgins SJ and Sugimura Y . (1987). Endocr. Rev., 8, 338–362.

  16. David G, Alland L, Hong SH, Wong CW, DePinho RA and Dejean A . (1998). Oncogene, 16, 2549–2556.

  17. Davies P and Eaton CL . (1991). J. Endocrinol., 131, 5–17.

  18. Dilworth FJ and Chambon P . (2001). Oncogene, 20, 3047–3054.

  19. Feldman BJ and Feldman D . (2001). Nat. Rev. Cancer, 1, 34–45.

  20. Glass CK and Rosenfeld MG . (2000). Genes Dev., 14, 121–141.

  21. Green S, Issemann I and Sheer E . (1988). Nucleic Acids Res., 16, 369.

  22. Greenlee RT, Murray T, Bolden S and Wingo PA . (2000). CA Cancer J. Clin., 50, 7–33.

  23. Grignani F, De Matteis S, Nervi C, Tomassoni L, Gelmetti V, Cioce M, Fanelli M, Ruthardt M, Ferrara FF, Zamir I, Seiser C, Lazar MA, Minucci S and Pelicci PG . (1998). Nature, 391, 815–818.

  24. Guidez F, Ivins S, Zhu J, Soderstrom M, Waxman S and Zelent A . (1998). Blood, 91, 2634–2642.

  25. He LZ, Guidez F, Tribioli C, Peruzzi D, Ruthardt M, Zelent A and Pandolfi PP . (1998a). Nat. Genet., 18, 126–135.

  26. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW and Vogelstein B . (1998b). Proc. Natl. Acad. Sci. USA, 95, 2509–2514.

  27. Heemers H, Vanderhoydonc F, Heyns W, Verhoeven G and Swinnen JV . (2000). Biochem. Biophys. Res. Commun., 269, 209–212.

  28. Horoszewicz JS, Leong SS, Chu TM, Wajsman ZL, Friedman M, Papsidero L, Kim U, Chai LS, Kakati S, Arya SK and Sandberg AA . (1980). Prog. Clin. Biol. Res., 37, 115–132.

  29. Hsieh ML, Charlesworth MC, Goodmanson M, Zhang S, Seay T, Klee GG, Tindall DJ and Young CY . (1997). Cancer Res., 57, 2651–2656.

  30. Huggins C . (1967). Cancer Res., 27, 1925–1930.

  31. Labrie F, Belanger A, Dupont A, Luu-The V, Simard J and Labrie C . (1993). Clin. Invest. Med., 16, 475–492.

  32. Labrie F, Dupont A, Belanger A, St-Arnaud R, Giguere M, Lacourciere Y, Emond J and Monfette G . (1986). Endocr. Rev., 7, 67–74.

  33. Lin RJ, Nagy L, Inoue S, Shao W, Miller Jr WH and Evans RM . (1998). Nature, 391, 811–814.

  34. Liu QY, Niranjan B, Gomes P, Gomm JJ, Davies D, Coombes RC and Buluwela L . (1996). Cancer Res., 56, 1155–1163.

  35. Mabjeesh NJ, Zhong H and Simons JW . (2002). Endocr. Relat. Cancer, 9, 115–139.

  36. McKenna NJ and O’Malley BW . (2002). Cell, 108, 465–474.

  37. Mhatre AN, Trifiro MA, Kaufman M, Kazemi-Esfarjani P, Figlewicz D, Rouleau G and Pinsky L . (1993). Nat. Genet., 5, 184–188.

  38. Monne M, Croce CM, Yu H and Diamandis EP . (1994). Cancer Res., 54, 6344–6347.

  39. Montgomery JS, Price DK and Figg WD . (2001). J. Pathol., 195, 138–146.

  40. Narlikar GJ, Fan HY and Kingston RE . (2002). Cell, 108, 475–487.

  41. Razin A and Riggs AD . (1980). Science, 210, 604–610.

  42. Reid A, Gould A, Brand N, Cook M, Strutt P, Li J, Licht J, Waxman S, Krumlauf R and Zelent A . (1995). Blood, 86, 4544–4552.

  43. Roth SY, Denu JM and Allis CD . (2001). Annu. Rev. Biochem., 70, 81–120.

  44. Sadi MV, Walsh PC and Barrack ER . (1991). Cancer, 67, 3057–3064.

  45. Schwabe JW, Chapman L, Finch JT and Rhodes D . (1993). Cell, 75, 567–578.

  46. Shang Y and Brown M . (2002). Science, 295, 2465–2468.

  47. Takane KK and McPhaul MJ . (1996). Mol. Cell. Endocrinol., 119, 83–93.

  48. Tomura A, Goto K, Morinaga H, Nomura M, Okabe T, Yanase T, Takayanagi R and Nawata H . (2001). J. Biol. Chem., 276, 28395–28401.

  49. Tora L, Mullick A, Metzger D, Ponglikitmongkol M, Park I and Chambon P . (1989). EMBO J., 8, 1981–1986.

  50. Tyagi RK, Lavrovsky Y, Ahn SC, Song CS, Chatterjee B and Roy AK . (2000). Mol. Endocrinol., 14, 1162–1174.

  51. Ulrix W, Swinnen JV, Heyns W and Verhoeven G . (1999). FEBS Lett, 455, 23–26.

  52. van der Kwast TH, Schalken J, Ruizeveld de Winter JA, van Vroonhoven CC, Mulder E, Boersma W and Trapman J . (1991). Int. J. Cancer, 48, 189–193.

  53. Wong CW and Privalsky ML . (1998). J. Biol. Chem., 273, 27695–27702.

  54. Workman JL and Kingston RE . (1998). Annu. Rev. Biochem., 67, 545–579.

  55. Zegers ND, Claassen E, Neelen C, Mulder E, van Laar JH, Voorhorst MM, Berrevoets CA, Brinkmann AO, van der Kwast TH, Ruizeveld de Winter JA, Trapman J and Boersma WJA . (1991). Biochim. Biophys. Acta, 1073, 23–32.

  56. Zelent A, Guidez F, Melnick A, Waxman S and Licht JD . (2001). Oncogene, 20, 7186–7203.

  57. Zhang Y, LeRoy G, Seelig HP, Lane WS and Reinberg D . (1998). Cell, 95, 279–289.

  58. Zhang Y, Ng HH, Erdjument-Bromage H, Tempst P, Bird A and Reinberg D . (1999). Genes Dev., 13, 1924–1935.

  59. Zhou ZX, Sar M, Simental JA, Lane MV and Wilson EM . (1994). J. Biol. Chem., 269, 13115–13123.

Download references

Acknowledgements

We are grateful to members of the group for continual advice and support. This work was carried out with support from the Charing Cross & Hammersmith Hospitals Trustees, the Prostate Cancer Charity, Cancer Research UK and the Association for International Cancer Research.

Author information

Correspondence to Simak Ali or Laki Buluwela.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Pike, J., Holmes, D., Kamalati, T. et al. Silencing of androgen-regulated genes using a fusion of AR with the PLZF transcriptional repressor. Oncogene 23, 7561–7570 (2004) doi:10.1038/sj.onc.1208030

Download citation

Keywords

  • prostate cancer
  • androgen receptor
  • PLZF
  • gene expression

Further reading