Loss of expression of context-specific tumor suppressors is a critical event that facilitates the development of prostate cancer. Zinc finger and BTB domain containing transcriptional repressors, such as ZBTB7A and ZBTB16, have been recently identified as tumor suppressors that play important roles in preventing prostate cancer progression. In this study, we used combined ChIP-seq and RNA-seq analyses of prostate cancer cells to identify direct ZBTB7A-repressed genes, which are enriched for transcriptional targets of E2F, and identified that the androgen receptor (AR) played a critical role in the transcriptional suppression of these E2F targets. AR recruitment of the retinoblastoma protein (Rb) was required to strengthen the E2F–Rb transcriptional repression complex. In addition, ZBTB7A was rapidly recruited to the E2F–Rb binding sites by AR and negatively regulated the transcriptional activity of E2F1 on DNA replication genes. Finally, ZBTB7A suppressed the growth of castration-resistant prostate cancer (CRPC) in vitro and in vivo, and overexpression of ZBTB7A acted in synergy with high-dose testosterone treatment to effectively prevent the recurrence of CRPC. Overall, this study provides novel molecular insights of the role of ZBTB7A in CRPC cells and demonstrates globally its critical role in mediating the transcriptional repression activity of AR.

Significance:

ZBTB7A is recruited to the E2F–Rb binding sites by AR and negatively regulates the transcriptional activity of E2F1 on DNA replication genes.

Prostate cancer is one of the most common cancers in men. The development of primary prostate cancer depends on the activity of androgen receptor (AR), a ligand-dependent nuclear receptor transcription factor. The standard treatment of prostate cancer is surgical or medical castration (known as androgen deprivation therapy, ADT) to reduce circulating androgens. However, patients invariably relapse into more aggressive castration-resistant prostate cancer (CRPC) with increased expression and restored activity of AR (1). Although CRPC can be further treated with more aggressive ADT using agents such as abiraterone and enzalutamide (2, 3), the tumors generally relapse within 1 year and a large portion of these relapsed tumors still express AR and AR-regulated genes. Although AR is known for its transcriptional activator function, it can also act as a transcriptional repressor to suppress the expression of a subset of genes, including AR itself, androgen synthetic genes, and genes mediating DNA replication and repair (4, 5). Mechanistically, we have shown that AR globally recruits hypophosphorylated retinoblastoma protein (Rb) to the promoters/enhancers of DNA replication gene loci and strengthens the activity of E2F–Rb suppressor complex (6). Therefore, loss of this tumor suppressor activity of AR after ADT is likely to be one mechanism contributing to the progression to CRPC (1). This transcriptional repressor activity of AR also provides one of the mechanisms for the high-dose testosterone therapy in patients with CRPC (7–10).

In addition to AR, the recently reported zinc finger and BTB domain containing transcription repressors, such as ZBTB7A and ZBTB16, can also function as tumor suppressors in preventing the progression of prostate cancer (11–14). ZBTB7A, also known as LRF/POKEMON, consists of a protein–interacting BTB domain at the N-terminus and DNA binding zinc fingers at the C-terminus (13). Even though ZBTB7A has been identified as a proto-oncogene in other cancer types, a recent study using a transgenic mouse model indicates that it functions as a tumor suppressor in prostate cancer and that loss of its expression can drive the development of aggressive invasive tumor in Pten-null prostate epithelial cells by bypassing the Pten-loss induced cellular senescence (15). Mechanistically, ZBTB7A was shown to repress the activity of SOX9, a proto-oncogene in prostate cancer (16), and to impair the SOX9 regulation of an RB-targeting miRNA, thus allowing cells to bypass the Pten-loss induced senescence (15). Although its tumor suppressor activity has been demonstrated in mouse prostate cancer cells, the activities of ZBTB7A at the chromatin level in human prostate cancer cells remain to be characterized. Using a combined analysis of the ZBTB7A cistrome and transcriptome, we have mapped the binding sites of ZBTB7A and identified direct ZBTB7A-regulated genes. Significantly, the direct ZBTB7A-repressed genes were enriched for the activation function of E2Fs, suggesting that ZBTB7A may function to repress their oncogenic activities in prostate cancer cells.

Because our previous studies indicated AR can function as a transcriptional repressor to suppress DNA replication genes through enhancing the chromatin binding of Rb that reinforces the suppressor activity of E2F–Rb complex (6), we next determined how ZBTB7A chromatin binding in prostate cancer cells globally impacts the transcriptional activity of AR. By co-analyzing the previous reported AR cistrome database in prostate cancer cells (4, 16), we show that a significant portion of ZBTB7A binding sites overlap with AR binding sites and that these ZBTB7A and AR overlapping sites are significantly associated with the repression activity of AR on gene transcription. More importantly, we also show that ZBTB7A binding at those AR repression sites is rapidly increased upon androgen stimulation and the increased binding is highly associated with E2F–Rb binding, indicating that AR-recruited ZBTB7A may cooperate with Rb in regulating the transcriptional activity of E2Fs. By co-immunoprecipitation assays, we demonstrated that ZBTB7A can physically interact with AR, Rb, and E2F1, further indicating that ZBTB7A may be an additional component of the AR–Rb repressor complex. Furthermore, we carried out in vitro and in vivo studies to examine the effects of overexpression of ZBTB7A on CRPC tumor growth and results show that overexpression of ZBTB7A in CRPC cells significantly reduced the cancer development, and that overexpression of ZBTB7A can synergize with high-dose testosterone therapy in treating CRPC. Overall, this study has provided novel insights into the tumor suppressor activity of ZBTB7A in prostate cancer cells and identified ZBTB7A as a critical mediator required for AR-dependent transcriptional repression activity.

Cell lines and cell culture

The VCaP and C4-2 cell lines were purchased from ATCC. All the cell lines were recently authenticated using short tandem repeat (STR) profiling by DDC Medical and tested for Mycoplasma contamination (negative result) by using MycoAlert Mycoplasma Detection Kit (Lonza). VCaP cells were cultured in DMEM medium with 10% FBS (Gibco). VCaP-tet-shZBTB7A (tetracycline-inducible ZBTB7A silencing) cells were maintained in DMEM medium with 10% tetracycline-free FBS. C4-2 and C4-2-shZBTB7A (stable ZBTB7A silencing) cells were cultured in RPMI1640 medium supplemented with 2% FBS plus 8% CSS (charcoal-dextran stripped FBS; Gibco). C4-2-tet-ZBTB7A (tetracycline-inducible ZBTB7A overexpressing) cells were maintained in RPMI1640 medium with 2% tetracycline-free FBS plus 8% CSS. For androgen stimulation assays, cells were generally grown to 50% to 60% confluence in medium containing 5% CSS for 2 to 3 days and then treated with DHT or inhibitors for indicated time.

Chromatin immunoprecipitation

For preparation of chromatin immunoprecipitation (ChIP), dispensed cells were formalin fixed, lysed, and sonicated to break the chromatin into 500 to 800 bp fragments, followed by immunoprecipitation. The qPCR analysis was carried out using the SYBR Green method on the QuantStudio 3 real-time PCR system (Thermo Fisher Scientific). The primers for MCM7-pro, BLM-pro, FANCI-pro, and TK1-pro were previously listed (6).

RT-PCR and immunoblotting

The expression of genes was measured using real-time RT-PCR analyses with Taqman one-step RT-PCR reagents (Thermo Fisher Scientific) and results were normalized to coamplified GAPDH. The primer and probe sets for the following genes: MCM2, MCM7, FANCI, BLM, TK1, PCAT1, and GAPDH were purchased as inventoried mix from Applied Biosystems at Thermo Fisher Scientific. For immunoblotting, cells were lysed with RIPA buffer with protease inhibitors (Thermo Fisher Scientific) and anti-ZBTB7A (Bethyl), anti-AR PG21 (Millipore), anti-Rb, anti-E2F1 (Cell Signaling), anti-V5, anti-HA (Sigma), anti-HDAC1, anti-GAPDH, or anti-β-actin (Abcam) antibodies were used. Immunoblotting results shown are representative of at least 3 independent experiments.

RNAi and transfection

siRNAs against ZBTB7A and nontarget control (NTC) were purchased from Dharmacon and transfected into cells using lipofectamine 2000 (Thermo Fisher Scientific). C4-2-shZBTB7A cells were generated using lentiviral shRNA against ZBTB7A or NTC (Dharmacon). VCaP-tet-shZBTB7A cells were generated using tetracycline-inducible lentiviral shRNA against ZBTB7A or NTC (Dharmacon). C4-2-tet-ZBTB7A cells were generated by stable infection with lenti-virus containing tetracycline-inducible V5-tagged ZBTB7A cDNA in C4-2 cells. COS-7 cells cultured in DMEM with 10% FBS were transiently transfected with plasmids expressing HA-tagged AR 1-539aa, AR 1-628aa, AR 538-919aa, or AR 662-919aa for overnight, followed by immunoprecipitation assay.

ChIP-seq and RNA-seq

For ChIP-seq, VCaP cells were fixed and then lysed, and the chromatin was sheared to 300 to 500 bp in size using the Bioruptor sonicator (Diagenode). The chromatin was incubated with ZBTB7A antibody bound to protein G beads (Thermo Fisher Scientific) for overnight. After immunoprecipitation, samples were reversely crosslinked in 65°C water bath, and DNA was extracted with QIAquick PCR Purification Kit (Qiagen). The ChIP-seq libraries were prepared using ThruPlex DNA-seq 48D Kit (Rubicon Genomics) and then sequenced using Illumina HiSeq 2500. For RNA-seq, VCaP cells transfected with siZBTB7A or siNTC were harvested for RNA extraction using RNeasy mini kits (Qiagen), followed by RNA-Seq library preparation with TruSeq Stranded RNA LT Kit (Illumina).

For data analysis, ChIP-seq raw reads were aligned to hg19 using bwa (version 0.7.2-r351). The resulted sam files are converted to bam with samtools [version 0.1.18 (r982:295)]. MACS2 (version 2.0.10.20131216) was used to call peak on the bam files. bedGraph files containing signal per million reads produced from MACS2 was converted to bigwig files with ucsctool kit (315). The R package ChIPpeakAnno (version 3.10.1) was used for analyzing peak intervals. deepTools (version 2.4.1) was used to extract and visualize signal from bigwig files. The RNA-seq differential gene expression analysis was performed using TopHat pipeline on Galaxy. The GEO accession for ChIP-seq and RNA-seq is GSE123091.

Cell proliferation assay

C4-2-shNTC and C4-2-shZBTB7A cells were maintained in RPMI1640 supplemented with 2% FBS plus 8% CSS. After DHT treatments, cells were stained with Muse Count & Viability Assay Kit for 5 minutes and then counted by Muse Cell Analyzer (EMD Millipore).

Luciferase reporter assay

HEK293 cells were transfected with a Firefly luciferase reporter construct containing ∼800bp promoter of MCM7 gene together with a Renilla luciferase reporter construct for 24 hours prior to the treatments. The activities of Firefly luciferase and Renilla luciferase were measured using the dual-luciferase reporter assay (Promega) and the results were normalized for Renilla activities.

Mouse xenografts

C4-2-tet-ZBTB7A xenografts were established in the flanks of castrated male SCID mice by injecting ∼2 million cells mixed with 50% Matrigel. Doxycycline-supplemented food was introduced at ∼6 weeks postinjection and tumor volume was measured by manual caliper using the formula V = (W2 × L)/2. Frozen sections were examined to confirm that the samples used for RNA and protein extraction contain predominantly nonnecrotic tumor. All animal experiments were approved by the UMass Boston Institutional Animal Care and Use Committee and were performed in accordance with institutional and national guidelines.

BETA analysis

As previously described (6), binding and expression target analysis (BETA) was performed to assess the association of AR and ZBTB7A binding sites with the expression of AR-activated versus AR-repressed genes. BETA software package was used (with default parameters) to integrate ChIP-seq of ZBTB7A and AR with androgen-regulated gene expression profiling. The red and the purple lines represent the AR-activated and AR-repressed genes, respectively. The black dashed line indicates the nondifferentially expressed genes as background. Genes are cumulated by the rank on the basis of the regulatory potential score from high to low. P-value represent the significance of difference in the AR-activated or AR-repressed group compared with the nondifferentially expressed group by the Kolmogorov–Smirnov test.

Statistical analysis

Data in bar graphs represent mean ± SD of at least 3 biological repeats. Statistical analysis was performed using Student t test by comparing treatment versus vehicle control or otherwise as indicated. P value <0.05 (*) was considered to be statistically significant. For animal studies, 1-way ANOVA was performed for the tumor volume data measured at the final day of the treatments.

Characterization of ZBTB7A transcriptional program in prostate cancer cells

The zinc finger and BTB domain containing transcription repressors, including ZBTB7A and ZBTB16, were recently identified as critical tumor suppressors in prostate cancer cells. Using TCGA prostate cancer dataset (17), we found that over 6% of prostate cancer samples have deep deletion of ZBTB7A (1.2%) or ZBTB16 (5%; Fig. 1A), indicating the critical role of this gene family in prostate cancer development. We then compared the expression levels of ZBTB7A and ZBTB16 in clinical cohorts of primary prostate cancer versus metastatic CRPC. As shown in Fig. 1B, expression levels of ZBTB7A and ZBTB16 were higher in the primary prostate cancer cohort (TCGA) than in the CRPC cohort (SU2C; ref. 18), suggesting that their expression is decreased during CRPC development. Although the transcriptional repression activity of ZBTB16 in prostate cancer cells has been described (14), the activity of ZBTB7A on chromatin has not yet been determined. Therefore, we next examined its chromatin binding in prostate cancer cells. ChIP-seq analysis of ZBTB7A in VCaP cells (an AR-amplified CRPC cell line) cultured in full serum condition identified 17,691 high confidence peaks as potential ZBTB7A binding sites. Interestingly, these binding sites were significantly enriched for promoter region (13.8% vs. 1.1% background) and 5′ UTR region (8.9% vs. 0.4% background; Fig. 1C), suggesting that ZBTB7A may preferentially bind to the upstream or downstream sites near transcription start sites (TSS). Using a motif enrichment analysis, we found that the previously known ZBTB7A binding motif was significantly associated with ZBTB7A binding sites (Fig. 1D). We then clustered and ranked the transcription factors that may potentially co-occupy the ZBTB7A sites based on their motif enrichment scores. As shown in Fig. 1E, the top-ranked clusters were identified, and the possible factors include POLR3A (a subunit of RNA polymerase III), TRIM28 (a Tripartite motif containing transcription cofactor), PDX1 (a homeodomain transcription factor), NFKB1 (the DNA binding subunit of NF-κB), ING4 (a PHD-finger containing chromatin remodeling protein), and E2F2 (an E2F family transcription factor). The POLR3A cluster (cluster 1) also includes a number of nuclear hormone receptors.

Figure 1.

Functional characterization of ZBTB7A transcriptional activity in prostate cancer cells. A, Genetic alterations of ZBTB7A and ZBTB16 in TCGA prostate cancer cohort (retrieved in cBioPortal). B, The expression levels of ZBTB7A and ZBTB16 (normalized to GAPDH) in SU2C mCRPC cohort versus TCGA primary prostate cancer cohort. C, ZBTB7A ChIP-seq was done in VCaP cells cultured in full FBS and the genomic distribution of ZBTB7A binding is shown. D, A Bayesian DNA motif comparison method, BLiC (33), was applied on ZBTB7A ChIP-seq peaks (cutoff: P = 10−15). The ZBTB7A binding motif was found with z-score = −50.87. E, Potential ZBTB7A co-occupied factors were identified in 6 clusters with high confidence z-score (< −40). Other identified factors with high similarity-score (>2.85) in each cluster are also listed. F, BETA was performed to assess the direct regulation of ZBTB7A target genes using the ChIP-seq analysis of ZBTB7A and the differential gene expression analysis (from RNA-seq) in VCaP cells transfected with siZBTB7A and nontarget control (confirmed by immunoblotting). G, The direct ZBTB7A-regulated genes identified from BETA analysis were subjected to KEGG pathway analysis by DAVID 6.8 to identify pathways associated with ZBTB7A activation or repression function. H, GSEA was carried out on ZBTB7A-regulated genes.

Figure 1.

Functional characterization of ZBTB7A transcriptional activity in prostate cancer cells. A, Genetic alterations of ZBTB7A and ZBTB16 in TCGA prostate cancer cohort (retrieved in cBioPortal). B, The expression levels of ZBTB7A and ZBTB16 (normalized to GAPDH) in SU2C mCRPC cohort versus TCGA primary prostate cancer cohort. C, ZBTB7A ChIP-seq was done in VCaP cells cultured in full FBS and the genomic distribution of ZBTB7A binding is shown. D, A Bayesian DNA motif comparison method, BLiC (33), was applied on ZBTB7A ChIP-seq peaks (cutoff: P = 10−15). The ZBTB7A binding motif was found with z-score = −50.87. E, Potential ZBTB7A co-occupied factors were identified in 6 clusters with high confidence z-score (< −40). Other identified factors with high similarity-score (>2.85) in each cluster are also listed. F, BETA was performed to assess the direct regulation of ZBTB7A target genes using the ChIP-seq analysis of ZBTB7A and the differential gene expression analysis (from RNA-seq) in VCaP cells transfected with siZBTB7A and nontarget control (confirmed by immunoblotting). G, The direct ZBTB7A-regulated genes identified from BETA analysis were subjected to KEGG pathway analysis by DAVID 6.8 to identify pathways associated with ZBTB7A activation or repression function. H, GSEA was carried out on ZBTB7A-regulated genes.

Close modal

To study the effect of ZBTB7A binding on gene transcription in prostate cancer cells, we performed an RNA-seq analysis in VCaP cells (cultured in full serum condition) transfected with siRNA targeting ZBTB7A or nontarget control (NTC) and then carried out BETA (15) to identify potential ZBTB7A directly regulated genes. As shown in Fig. 1F, chromatin binding of ZBTB7A was associated with both ZBTB7A-activated (P = 10−37) and ZBTB7A-repressed genes (P = 10−21), suggesting that ZBTB7A may have both transcriptional activator and repressor functions in prostate cancer cells. As the BETA analysis also ranked the potential for genes that may be directly regulated by ZBTB7A (Supplementary Table S1), we next performed Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (using DAVID 6.8) on these potential direct targets of ZBTB7A. Interestingly, the activator activity of ZBTB7A was associated with specific cancer types, including prostate and non–small cell lung cancers (Fig. 1G), suggesting that ZBTB7A may retain some tumor promoting activities in prostate cancer cells through gene activation. The repressor function of ZBTB7A was enriched for transcription and translation regulation, chemokine signaling pathway, and DNA replication, supporting its tumor suppressor role in prostate cancer cells. Interestingly, the suppression of chemokine signaling was also supported by a recent study showing that ZBTB7A regulates the infiltration and composition of immune cells within prostate cancer tumor through repressing the expression of a chemokine, CXCL5 (19). To further determine which signaling pathways are most impacted by ZBTB7A, we performed a gene set enrichment analysis (GSEA) on ZBTB7A-regulated genes. As seen in Fig. 1H, although ZBTB7A-activated genes were not significantly associated with any known pathways, the expression of ZBTB7A-repressed genes were highly enriched for the activation function of E2Fs, an oncogenic transcription factor family that plays critical role in activating DNA replication and cell-cycle progression, indicating ZBTB7A may function to suppress the activity of E2Fs. Collectively, these global cistrome and transcriptome studies of ZBTB7A revealed its transcription factor activities in prostate cancer cells and linked its tumor suppressor function to the possible downregulation of E2F activity.

ZBTB7A chromatin binding is associated with transcriptional repression activity of AR

Our previous studies in multiple prostate cancer models and human prostate cancer samples have indicated that repressing E2F-activated DNA replication is the major transcriptional repression activity of AR (4, 6). Therefore, we next assessed the global impact of ZBTB7A on AR transcriptional activity. Because the above global studies were done in VCaP cells under the full serum condition, which contains substantial levels of testosterone and DHT (dihydrotestosterone, a potent form of testosterone) as well as other steroid hormones (20), we were able to perform combined analyses with the previously published AR ChIP-seq and gene expression profiling datasets based on VCaP cells treated with 10 nmol/L DHT (4, 16), we found that ∼30% of ZBTB7A binding peaks (5,875 of 17,691) were overlapped with AR binding sites (Fig. 2A) and these sites were highly enriched for promoter binding (Supplementary Fig. S1). The binding intensity of ZBTB7A was also highly correlated with the intensity of AR binding (Fig. 2B), indicating a possible interaction of ZBTB7A and AR to coregulate gene transcription. We then performed BETA to determine the global association of ZBTB7A binding with the AR regulation of gene transcription. As seen in Fig. 2C, although AR repressive function appeared to associate with ZBTB7A unique sites (P = 10−16), it was more strongly correlated with AR and ZBTB7A overlapping sites (P = 10−39) and not with AR unique sites (P = 0.839). In contrast, AR activation function was associated more strongly with ZBTB7A-absent AR binding sites (P = 10−17), and was less significantly associated with ZBTB7A and AR overlapping sites (P = 10−15) or ZBTB7A unique sites (P = 10−7). Collectively, these global analyses of ZBTB7A and AR binding sites strongly indicate that ZBTB7A may be involved in AR-mediated transcriptional repression in prostate cancer cells.

Figure 2.

ZBTB7A chromatin binding is associated with transcriptional repression activity of AR. A, The Venn diagram shows overlap between previously identified AR binding sites in VCaP cells and ZBTB7A binding sites. B, The heatmap view for the signal intensities of the aligned reads from ChIP-ZBTB7A and ChIP-AR across all ZBTB7A binding sites (±1 kb). C, BETA analysis on AR-regulated genes with nearby AR and ZBTB7A unique or overlapping binding sites. D, ZBTB7A-repressed genes with overlapping AR and ZBTB7A binding sites were analyzed by KEGG pathway analysis. E, Motif enrichment analysis was carried out at the AR binding sites (3kb up/downstream of TSS) within AR-repressed gene (1.5-fold cutoff) loci. The potential factors within the top-ranked clusters are listed (z-score < −10, similarity score >2.85). AR motif was the most enriched motif within 600bp around TSS. F, ZBTB7A was immunoprecipitated in VCaP cells treated with or without 10 nmol/L DHT for 4 hours, followed by immunoblotting for AR and ZBTB7A. G, COS7 cells were cotransfected with V5-tagged ZBTB7A and HA-tagged AR fragments (NTD: 1-539 aa, NTD+DBD: 1-628 aa, DBD+LBD: 538-919 aa, and LBD: 662-919 aa). AR fragments were immunoprecipitated by anti-HA beads, followed by immunoblotting for V5 (V5-ZBTB7A).

Figure 2.

ZBTB7A chromatin binding is associated with transcriptional repression activity of AR. A, The Venn diagram shows overlap between previously identified AR binding sites in VCaP cells and ZBTB7A binding sites. B, The heatmap view for the signal intensities of the aligned reads from ChIP-ZBTB7A and ChIP-AR across all ZBTB7A binding sites (±1 kb). C, BETA analysis on AR-regulated genes with nearby AR and ZBTB7A unique or overlapping binding sites. D, ZBTB7A-repressed genes with overlapping AR and ZBTB7A binding sites were analyzed by KEGG pathway analysis. E, Motif enrichment analysis was carried out at the AR binding sites (3kb up/downstream of TSS) within AR-repressed gene (1.5-fold cutoff) loci. The potential factors within the top-ranked clusters are listed (z-score < −10, similarity score >2.85). AR motif was the most enriched motif within 600bp around TSS. F, ZBTB7A was immunoprecipitated in VCaP cells treated with or without 10 nmol/L DHT for 4 hours, followed by immunoblotting for AR and ZBTB7A. G, COS7 cells were cotransfected with V5-tagged ZBTB7A and HA-tagged AR fragments (NTD: 1-539 aa, NTD+DBD: 1-628 aa, DBD+LBD: 538-919 aa, and LBD: 662-919 aa). AR fragments were immunoprecipitated by anti-HA beads, followed by immunoblotting for V5 (V5-ZBTB7A).

Close modal

We next examined the functions for the ZBTB7A-repressed genes that have nearby AR and ZBTB7A overlapping sites (Supplementary Table S2, ranking of genes based on BETA). As shown in Fig. 2D, this subset of genes were enriched for pathways mediating translation and DNA replication and damage repair, the latter of which was consistent with the previously reported major function of direct AR-repressed genes (4, 6). To further confirm that ZBTB7A can bind to AR occupied repression sites, we performed a motif enrichment analysis on the AR binding sites adjacent to the androgen-repressed gene loci. Consistently, we found that ZBTB7A binding motif was among the top-ranked motifs enriched at those AR repression sites (Fig. 2E). The E2F binding motif was also top ranked, consistent with our previous studies showing AR interaction with E2F–Rb complex at the repression sites (4, 6). A previous study found that ZBTB7A can directly interact with AR through the POZ domain of ZBTB7A and ligand-binding domain (LBD) of AR, and this interaction impairs AR activation of PSA (21). Using a coimmunoprecipitation assay in VCaP cells, we confirmed that the endogenous ZBTB7A can interact with AR and the interaction appeared to be enhanced by androgen stimulation (Fig. 2F). We next sought to determine whether the LBD of AR is the domain responsible for AR interaction with ZBTB7A. Experiments in cells transiently overexpressing AR fragments and ZBTB7A were then carried out and the result indicated that the entire C-terminal domains, including both LBD and DNA-binding domain (DBD), may be required for the full interaction with ZBTB7A (Fig. 2G). Overall, these data support the function of ZBTB7A in mediating AR-dependent transcriptional repression activity in prostate cancer cells through a possible direct interaction with AR protein.

AR recruitment of ZBTB7A is required for the transcriptional repression function of AR on E2F-regulated genes

The current model on AR transcriptional repression activity based on our previous studies is that the rapid AR recruitment of Rb can reinforce the E2F–Rb repressor complex to suppress the transcription of genes mediating DNA replication/repair and cell cycle (6). Because the recruitment of Rb is one major event that mediates this specific activity of AR, we next determined whether AR may also recruit ZBTB7A to those suppression sites. To test this hypothesis, we performed additional ChIP-seq of ZBTB7A in VCaP cells treated with or without 10 nmol/L DHT for 4 hours. As seen in Fig. 3A and B, the binding intensity of ZBTB7A at those ZBTB7A and AR overlapping sites was rapidly increased by the short-term androgen treatment and correlated with the binding intensity of AR and Rb, suggesting that AR may directly recruit ZBTB7A to those sites. Supportively, the binding intensity of ZBTB7A at those AR and Rb overlapping sites was also similarly androgen-induced (Supplementary Fig. S2A). We then conducted BETA to determine whether AR recruited ZBTB7A and Rb are associated with AR repression function. As shown in Fig. 3C and D, AR repression function was more significantly associated with AR+/ZBTB7A+ sites (P = 10−43) and AR+/ZBTB7A+/Rb+ sites (P = 10−31) than AR activation function (P = 10−23 and P = 10−13, respectively). Moreover, the absence of either Rb (AR+/ZBTB7A+/Rb) or ZBTB7A binding (AR+/ZBTB7A/Rb+) at those AR repression sites significantly decreased their association with AR repression function (P = 10−5 and P = 10−9, respectively), suggesting that AR recruitment of ZBTB7A and Rb may be both required for the full repression activity of AR.

Figure 3.

AR recruitment of ZBTB7A is required for the transcriptional repression activity of AR on E2F-regulated genes. A, ZBTB7A ChIP-seq was done in VCaP cells treated with or without 10 nmol/L DHT for 4 hours in hormone-depleted media. The heatmap view for the signal intensities of aligned reads across all ZBTB7A binding sites (±3 kb) is shown and the ZBTB7A bindings were compared with Rb and AR bindings in DHT-treated VCaP cells. B, The mean signal density of AR, ZBTB7A, and Rb bindings centered at ZBTB7A binding sites (±3 kb). C, The Venn diagram illustrating overlap among AR, ZBTB7A, and Rb binding sites (DHT-stimulated). D, BETA analysis was used to assess the correlation of AR, ZBTB7A, and Rb binding sites with the expression of AR-activated genes (red) and AR-repressed genes (blue) over static background (black). E, ChIP-qPCR for ZBTB7A binding at indicated promoters in VCaP cells treated with vehicle or 10 nmol/L DHT for 1 hour. F and G, MCM7 promoter (∼800bp)–driven luciferase report activity was measured in HEK293 cells transfected with E2F1, Rb, and/or ZBTB7A (F) in full serum, or E2F1 and/or AR in presence or absence of 10 nmol/L DHT (G). H, VCaP cells stably expressing tetracycline-regulated lentiviral shRNA against ZBTB7A (VCaP-tet-shZBTB7A) were established and ZBTB7A expression was examined by immunoblotting in cells pretreated with 0.1 μg/mL doxycycline for 2 days and then with 10 nmol/L DHT for 24 hours. The mRNA expressions of a panel of Rb mediated AR-repressed genes were measured. *, P < 0.05.

Figure 3.

AR recruitment of ZBTB7A is required for the transcriptional repression activity of AR on E2F-regulated genes. A, ZBTB7A ChIP-seq was done in VCaP cells treated with or without 10 nmol/L DHT for 4 hours in hormone-depleted media. The heatmap view for the signal intensities of aligned reads across all ZBTB7A binding sites (±3 kb) is shown and the ZBTB7A bindings were compared with Rb and AR bindings in DHT-treated VCaP cells. B, The mean signal density of AR, ZBTB7A, and Rb bindings centered at ZBTB7A binding sites (±3 kb). C, The Venn diagram illustrating overlap among AR, ZBTB7A, and Rb binding sites (DHT-stimulated). D, BETA analysis was used to assess the correlation of AR, ZBTB7A, and Rb binding sites with the expression of AR-activated genes (red) and AR-repressed genes (blue) over static background (black). E, ChIP-qPCR for ZBTB7A binding at indicated promoters in VCaP cells treated with vehicle or 10 nmol/L DHT for 1 hour. F and G, MCM7 promoter (∼800bp)–driven luciferase report activity was measured in HEK293 cells transfected with E2F1, Rb, and/or ZBTB7A (F) in full serum, or E2F1 and/or AR in presence or absence of 10 nmol/L DHT (G). H, VCaP cells stably expressing tetracycline-regulated lentiviral shRNA against ZBTB7A (VCaP-tet-shZBTB7A) were established and ZBTB7A expression was examined by immunoblotting in cells pretreated with 0.1 μg/mL doxycycline for 2 days and then with 10 nmol/L DHT for 24 hours. The mRNA expressions of a panel of Rb mediated AR-repressed genes were measured. *, P < 0.05.

Close modal

Because the previous study indicates that ZBTB7A may regulate Rb expression in transgenic mouse model (12), we first examined whether ZBTB7A upregulates Rb in prostate cancer cells. As shown in Supplementary Fig. S2B, Rb protein expression was not affected by knocking down ZBTB7A in VCaP cells, suggesting that ZBTB7A does not regulate Rb in human CRPC cells. To further determine whether ZBTB7A contributes to the AR-mediated repression activity on DNA replication, we selected a panel of previously identified Rb-dependent AR-repressed DNA replication/repair genes, including MCMs (minichromosome maintenance complex genes), BLM (a Bloom syndrome RecQ like helicase gene), FANCI (a Fanconi anemia complementation group gene), and TK1 (a thymidine kinase gene; ref. 6), for the subsequent studies. The rapid increase of ZBTB7A binding by androgen-stimulation at those previously identified AR-repressed promoters (6) was confirmed using ChIP-qPCR (Fig. 3E). We next cloned the MCM7 promoter (∼800bp fragment, containing AR, ZBTB7A, and E2F1 binding sites) into a luciferase reporter to examine the effect of ZBTB7A on E2F1 activity. As seen in Fig. 3F, E2F1 significantly induced MCM7-promoter activity, which can be suppressed by Rb. Importantly, the expression of ZBTB7A alone markedly downregulated the E2F1 activity and coexpression of ZBTB7A and Rb can further repress the E2F1 activity. These result suggest that ZBTB7A and Rb may function in parallel to collaboratively suppress E2F1-mediate transcriptional activation. Interestingly, overexpressing AR alone was not sufficient to suppress E2F1 activity (Fig. 3G), further suggesting that AR may indirectly repress E2F1 activity in PCa cells through recruitment of ZBTB7A and Rb. Furthermore, we also demonstrated that doxycycline-induced silencing of ZBTB7A in VCaP-tet-shZBTB7A stable cells (expressing tetracycline-regulated lentiviral shRNA against ZBTB7A) impaired the androgen-induced repression on these DNA replication genes (Fig. 3H). Collectively, these genomic and molecular studies indicated an important function of AR recruited ZBTB7A by collaboration with Rb to repress E2F-regulated transcriptional activation of DNA replication.

ZBTB7A mediates AR repression of PCAT-1 lncRNA

The gene profiling database used to study androgen regulation in VCaP cells was based on Affymetrix gene microarrays (4), which primarily detect protein coding genes. However, recent studies have revealed important functions of noncoding RNAs, particularly lncRNAs, in regulating gene transcription in tumor cells (22). Therefore, we sought to identify the androgen-repressed lncRNAs in androgen-treated prostate cancer cells by using RNA-seq. In VCaP cells treated with 10 nmol/L DHT for 24 hours, 111 lncRNAs were identified as androgen-induced genes and 245 lncRNAs were identified as androgen-repressed genes (Supplementary Table S3). Among the identified androgen-induced lncRNA subset, PCAT-14 and PCAT-18 have been previously suggested as biomarkers for predicting prostate cancer outcomes (23, 24). PRCAT47 (also called ARLNC1) was another recently identified androgen-upregulated gene that functions to stabilize the AR transcript and enhances AR signaling and prostate cancer tumor growth (25). Among the androgen-repressed lncRNA subset, PCAT-29 has been reported previously as an androgen-repressed tumor suppressor gene in prostate cancer cells (26). Importantly, we identified PCAT-1 (prostate cancer–associated transcript 1) as a novel androgen-repressed gene, which promotes prostate cancer by mechanisms such as stabilizing MYC protein, repressing BRCA2, and activating AKT (27–29). Therefore, suppressing the expression of this oncogenic lncRNA may be an important activity for AR to act as a tumor suppressor in prostate cancer.

We next examined androgen-regulation of PCAT-1 expression in prostate cancer cells. As seen in Supplementary Fig. S3A, PCAT-1 expression was significantly decreased by DHT treatment in VCaP cells and this suppression activity was abolished when cells were treated with an HDAC1 inhibitor (mocetinostat), indicating a critical function of HDAC1 that may be required for AR repression on PCAT-1. Interestingly, although PCAT-1 was not clearly androgen-repressed in androgen-dependent LNCaP prostate cancer cells, which express less AR than VCaP cells (4), it was significantly repressed by androgen in AR-overexpressing LNCaP (LN-AR) cells (Supplementary Fig. S3B), suggesting that the repression on PCAT-1 requires high levels of AR expression. To assess the androgen regulation of PCAT-1 in vivo, we examined its expression in a previously established VCaP xenograft tumor progression model (30). In this model, we have clearly demonstrated the restoration of the expression of PSA and TMPRSS2-ERG due to regained AR activation activity and the increased expression of AR, AKR1C3 (an androgen synthesis gene), and DNA replication genes due to loss of AR repression activity in the castration-resistant stage of the tumor (4, 6, 30). As seen in Supplementary Fig. S3C, the expression of PCAT-1 was rapidly increased in at least a portion of xenograft tumors upon castration and the increased level was retained in the relapsed tumors, suggesting ADT may alleviate the AR repression on PCAT-1 and hence result in the increased expression of PCAT-1 in CRPC.

We next determined whether ZBTB7A is involved in the AR-mediated repression of PCAT-1. Through examining the AR and ZBTB7A ChIP-seq results, we identified AR and ZBTB7A chromatin binding sites in the PCAT-1 locus. Interestingly, although the identified AR binding site (named S1) was very close to the ZBTB7A binding site (named S2), these 2 peaks were not exactly overlapping. Using ChIP-qPCR analysis, we showed that AR chromatin binding was significantly stimulated by DHT treatment on S1 and weakly increased on S2 site (Supplementary Fig. S3D, left). However, ZBTB7A binding was increased by DHT on the S2 site (Supplementary Fig. S3D, middle), suggesting that AR binding at S1 site may increase the recruitment of ZBTB7A at S2 site. Moreover, we also found that silencing ZBTB7A impaired the AR repression activity on PCAT-1 transcript in VCaP cells (Supplementary Fig. S3E), indicating that ZBTB7A contributes to the full repression activity of AR on PCAT-1.

ZBTB7A suppresses CRPC tumor growth in vitro

Although loss of Zbtb7a expression in conjunction with loss of Pten in mouse prostate epithelial cells promote prostate cancer development (12), it is not clear whether ZBTB7A can similarly function as a tumor suppressor in human CRPC cells. Although VCaP cells, which were derived from CRPC bone metastases, were used as the primary model for the mechanistic studies, these cells cannot directly form xenograft tumors in castrated mice and it is difficult to use this model for the subsequent functional studies. Therefore, to further study the tumor suppressor role of ZBTB7A in vitro and in vivo, we selected a well-known LNCaP-derived CRPC model, C4-2 cells, to generate the stable cell line that can inducibly overexpress ZBTB7A. Similar to VCaP cells, AR is overexpressed and androgen can suppress cell growth through an Rb-dependent mechanism in C4-2 cells (6). More importantly, in these cells AR also similarly suppresses the expression of the majority of DNA-replication genes that are androgen-repressed in VCaP cells and this repression activity of AR is partially impaired by silencing of Rb (Supplementary Fig. S4). C4-2 cells express comparable level of ZBTB7A as in VCaP cells and high-passage LNCaP cells, which are also resistant to ADT (31), although the AR-negative PC-3 cells express higher level of ZBTB7A (Fig. 4A). Importantly, ZBTB7A was also recruited by AR to the promoter regions of the DNA replication genes and this chromatin recruitment can be impaired by AR antagonist treatment (enzalutamide) or AR-targeted siRNA (Supplementary Fig. S5A and S5B). Silencing ZBTB7A in C4-2 cells increased cell proliferation in hormone-reduced condition and compromised the antiproliferative effect of DHT in C4-2 cells (Fig. 4B–D), indicating that the endogenous ZBTB7A is required for tumor suppressor function of AR in CRPC cells.

Figure 4.

ZBTB7A suppresses CRPC tumor growth in vitro. A, Immunoblotting for ZBTB7A in four CRPC cell lines. B, Immunoblotting for ZBTB7A expression in C4-2 cells stably expressing lentiviral shRNA against ZBTB7A or nontarget-control (shNTC). C, Cell proliferation (by counting live cells) was measured in the ZBTB7A-silencing cells versus control C4-2 cells. D, Cell proliferation was also measured in these cell lines treated with 10 nmol/L DHT for 2 days. E, Immunoblotting for V5 expression in C4-2-tet-ZBTB7A cells stimulated with 0 to 1 μg/mL doxycycline for 2 days. F and G, The expressions of a panel of AR-repressed DNA replication genes were examined in C4-2-tet-ZBTB7A cells treated with 0.25 μg/mL doxycycline and 10 nmol/L DHT for 2 days. The mRNA expression was measured by qRT-PCR (F) and the protein expression was measured by immunoblotting (G). H, V5-ZBTB7A was immunoprecipitated using V5-antibody-condugated beads in C4-2-tet-ZBTB7A cells stimulated by 0.25 μg/mL doxycycline for 2 days, followed by immunoblotting for AR, Rb, HDAC1, and ZBTB7A. I, E2F1 or Rb was immunoprecipitated in C4-2-tet-ZBTB7A cells stimulated by 0.25 μg/mL doxycycline for 2 days, followed by immunoblotting for ZBTB7A, Rb, and E2F1. J, Cell proliferation was measured in C4-2-tet-ZBTB7A cells treated with 0.1 μg/mL doxycycline and with 0 to 100 nmol/L DHT for 0 to 6 days. K, The expressions of a panel of AR-repressed DNA replication genes were examined in C4-2-tet-ZBTB7A cells treated with 0.1 μg/mL doxycycline and 1 nmol/L DHT for 2 days. *, P < 0.05.

Figure 4.

ZBTB7A suppresses CRPC tumor growth in vitro. A, Immunoblotting for ZBTB7A in four CRPC cell lines. B, Immunoblotting for ZBTB7A expression in C4-2 cells stably expressing lentiviral shRNA against ZBTB7A or nontarget-control (shNTC). C, Cell proliferation (by counting live cells) was measured in the ZBTB7A-silencing cells versus control C4-2 cells. D, Cell proliferation was also measured in these cell lines treated with 10 nmol/L DHT for 2 days. E, Immunoblotting for V5 expression in C4-2-tet-ZBTB7A cells stimulated with 0 to 1 μg/mL doxycycline for 2 days. F and G, The expressions of a panel of AR-repressed DNA replication genes were examined in C4-2-tet-ZBTB7A cells treated with 0.25 μg/mL doxycycline and 10 nmol/L DHT for 2 days. The mRNA expression was measured by qRT-PCR (F) and the protein expression was measured by immunoblotting (G). H, V5-ZBTB7A was immunoprecipitated using V5-antibody-condugated beads in C4-2-tet-ZBTB7A cells stimulated by 0.25 μg/mL doxycycline for 2 days, followed by immunoblotting for AR, Rb, HDAC1, and ZBTB7A. I, E2F1 or Rb was immunoprecipitated in C4-2-tet-ZBTB7A cells stimulated by 0.25 μg/mL doxycycline for 2 days, followed by immunoblotting for ZBTB7A, Rb, and E2F1. J, Cell proliferation was measured in C4-2-tet-ZBTB7A cells treated with 0.1 μg/mL doxycycline and with 0 to 100 nmol/L DHT for 0 to 6 days. K, The expressions of a panel of AR-repressed DNA replication genes were examined in C4-2-tet-ZBTB7A cells treated with 0.1 μg/mL doxycycline and 1 nmol/L DHT for 2 days. *, P < 0.05.

Close modal

To further study the effect of restoring ZBTB7A expression in CRPC cells, we established a C4-2 stable cell line that express tetracycline-regulated V5-tagged ZBTB7A (Fig. 4E). First, induced overexpression of ZBTB7A further decreased the mRNA and protein expression levels of the DNA replication genes that were suppressed by DHT treatment (Fig. 4F and G), consistent with the effect of silencing ZBTB7A in VCaP cells. Second, using a V5 pull-down assay we also demonstrated that AR, Rb, and HDAC1 can all interact with ZBTB7A (Fig. 4H), suggesting that ZBTB7A may be a component of the AR–Rb suppressor complex. Conversely, we also immunoprecipitated endogenous E2F1 or Rb in these cells and found that the V5-tagged ZBTB7A can be coimmunoprecipitated with these proteins (Fig. 4I), indicating that ZBTB7A can interact with E2F1. However, the interaction of E2F1 with Rb was not significantly affected by overexpression of ZBTB7A, suggesting that the repression activity of ZBTB7A on E2F1 may not be mediated through Rb. Consistently, silencing or overexpressing ZBTB7A had little effect on Rb binding to the promoters of DNA replication genes (Supplementary Fig. S6A–S6C), further suggesting that ZBTB7A can suppress E2F activity without increasing chromatin binding of Rb–E2F repressor complex. Nonetheless, induced expression of ZBTB7A by using lower dose of doxycycline to minimize any toxicity effect (Supplementary Fig. S7) also led to decreased cell growth, and more importantly could act in synergy with DHT treatment, particularly with lower doses of DHT (1 nmol/L; Fig. 4J). Consistent with the effect on growth, overexpression of ZBTB7A also enhanced the suppression effect of 1 nmol/L DHT on DNA replication genes (Fig. 4K).

Overexpression of ZBTB7A delays the recurrence of CRPC tumor treated by high-dose testosterone

Finally, we assessed the effect of ZBTB7A overexpression on CRPC tumor growth in vivo. As seen in Fig. 5A, induced overexpression of ZBTB7A (prior to the establishment of xenograft tumor) significantly delayed the development of C4-2 CRPC xenograft tumor, suggesting a critical and potent tumor suppressor function of ZBTB7A in CRPC cells. The doxycycline supplemented food had no toxic effect on the health of mice and the growth of C4-2 xenograft tumors (Supplementary Fig. S8A and S8B). We then determined whether overexpression of ZBTB7A can enhance the efficacy of high-dose testosterone treatment in this CRPC model. For this experiment, we allowed the CRPC xenograft tumor to establish (∼50 mm3) prior to the treatments. As seen in Fig. 5B, although the induction of ZBTB7A expression or the treatment of testosterone markedly suppressed the CRPC tumor growth, the tumors began to relapse after ∼1 month of treatments. Significantly, the combination treatment was able to delay the recurrence of the CRPC tumor, suggesting that overexpression of ZBTB7A can synergize with testosterone to suppress CRPC progression. Despite the limited tissue materials in treatment groups, we were able to extract a small amount of RNA to examine the effect of the combination treatment on a few DNA replication genes. As shown in Supplementary Fig. S9, although the expression levels of the DNA replication genes may be restored in the doxycycline- or testosterone-treated group, they appeared to remain repressed in the combination treatment group. Overall, these in vitro and in vivo studies clearly demonstrated the critical tumor suppressor function of ZBTB7A in CRPC cells and suggested a potential therapeutic strategy by enhancing ZBTB7A expression or activity to improve the efficacy of the high-dose testosterone therapy in CRPC patients. The model for ZBTB7A-mediated transcriptional repression activity of AR on a subset of E2F-regulated DNA replication genes was summarized in Fig. 5C.

Figure 5.

Overexpression of ZBTB7A delays the recurrence of CRPC tumor treated by high-dose testosterone. A, C4-2-tet-ZBTB7A xenograft tumors were established and passaged in castrated male SCID mice. A cohort of 12 mice (at ∼6 weeks post-injection) was randomly divided into two arms fed with regular diet or doxycycline-supplemented diet and the development of xenograft tumors was monitored for over 4 weeks. B, The C4-2-tet-ZBTB7A CRPC xenograft tumors were established (∼50 mm3) prior to the treatments. A cohort of 24 mice was randomly divided into four arms: (i) on regular diet with vehicle injection (cotton oil); (ii) on regular diet with testosterone treatment (10 mg/kg via i.p. injection every day); (iii) on doxycycline-supplemented diet with vehicle injection (cotton oil); and (iv) on doxycycline-supplemented diet with the same testosterone treatment, and the development of xenograft tumors was monitored for over 5 weeks. C, The current model for the ZBTB7A-mediated AR transcriptional repression activity.

Figure 5.

Overexpression of ZBTB7A delays the recurrence of CRPC tumor treated by high-dose testosterone. A, C4-2-tet-ZBTB7A xenograft tumors were established and passaged in castrated male SCID mice. A cohort of 12 mice (at ∼6 weeks post-injection) was randomly divided into two arms fed with regular diet or doxycycline-supplemented diet and the development of xenograft tumors was monitored for over 4 weeks. B, The C4-2-tet-ZBTB7A CRPC xenograft tumors were established (∼50 mm3) prior to the treatments. A cohort of 24 mice was randomly divided into four arms: (i) on regular diet with vehicle injection (cotton oil); (ii) on regular diet with testosterone treatment (10 mg/kg via i.p. injection every day); (iii) on doxycycline-supplemented diet with vehicle injection (cotton oil); and (iv) on doxycycline-supplemented diet with the same testosterone treatment, and the development of xenograft tumors was monitored for over 5 weeks. C, The current model for the ZBTB7A-mediated AR transcriptional repression activity.

Close modal

Loss of expression of zinc finger and BTB domain containing transcription factors, such as ZBTB7A and ZBTB16, are commonly seen in prostate cancer tumors, and their expression is decreased in more aggressive CRPC (see Fig. 1). However, the biological functions of these genes in prostate cancer cells remain to be identified. ZBTB16 is a classic AR regulated gene and it can function as a tumor suppressor in prostate cancer cells through inhibiting the MAPK pathways (14). In contrast, ZBTB7A is not regulated by AR and was previously known as an oncogene in many cancer types but recently reported to play a tumor suppressor role in prostate cancer. The loss of Zbtb7a expression in conjunction with loss of Pten in mouse prostate epithelial cells was previously shown to drive the development of aggressive prostate cancer (12). This tumor suppressive activity of ZBTB7A was largely attributed to maintaining the expression of Rb, a critical cell cycle regulator that mediates G1–S transition, through the negative regulation of SOX9 activity on activating an Rb-targeting miRNA. In this study, our integrated ZBTB7A cistrome and transcriptome analyses have indicated that ZBTB7A directly represses the E2F-regulated genes, consistent with those previous findings. More importantly, our mechanistic study of ZBTB7A has demonstrated that ZBTB7A can repress DNA replication and cell-cycle progression through a very distinct mechanism by which ZBTB7A was recruited by AR to the E2F binding sites and suppresses its transcriptional activation function. We have previously reported this transcriptional repressor activity of AR and revealed its critical biological function on repressing DNA replication and cell cycle (4). We have further shown in another study that this transcriptional repressor activity of AR was through enhancing Rb chromatin binding and thus strengthening the Rb–E2F repressor complex (6). Importantly, a recent study using castration-resistant LuCaP PDX models also indicates that the most robust molecular phenotype for the high-dose testosterone treatment is the suppression of E2F transcriptional output (32). In this study, we report that AR can rapidly recruit ZBTB7A to those AR and E2F–Rb overlapping sites within DNA replication genes and this recruitment can further enhance the transcriptional repression of the target genes. Mechanistically, we have also demonstrated that ZBTB7A can directly interact with E2F1 and negatively regulate its transcriptional activity (see Figs. 3F and 4I). Because ZBTB7A was known to recruit HDACs (also see Fig. 4H), this transcriptional corepressor function of ZBTB7A on AR may be through strengthening the recruitment of HDACs that can deacetylate histone 3 lysine 27 and thus represses gene transcription activated by E2F.

In addition to DNA replication genes, we have also identified the lncRNA PCAT-1 as a novel AR-repressed gene in prostate cancer cells. One of the major functions of this oncogenic lncRNA in prostate cancer cells is to regulate MYC oncoprotein (27). Therefore, ZBTB7A may suppress MYC activity through transcriptionally repressing PCAT-1 expression. Overall, this study identified a biologically important lncRNA as a novel repression target of AR and demonstrated the role of ZBTB7A in mediating this repression process.

Furthermore, we have demonstrated in vitro and in vivo that ZBTB7A is a potent tumor suppressor, which can markedly repress CRPC tumor growth. More importantly, we have demonstrated that overexpression of ZBTB7A can enhance the growth suppressive activity of high-dose androgen treatment on CRPC cells in vitro and in vivo. Based on our results (see Fig. 4J), even lower concentration of androgen treatment can suppress prostate cancer cell proliferation when ZBTB7A is expressed in high levels. This finding may provide an explanation of why a subset of prostate cancer or CRPC tumors have to select for the decrease or loss of ZBTB7A expression in order to escape from AR-mediated growth-suppression activity. Significantly, our study also provides a rationale to therapeutically enhance the high-dose testosterone treatment in CRPC (currently in phase II clinical trials; ref. 10) through elevating the expression or activity of ZBTB7A. Future study is clearly needed to identify the actionable targets that are involved in regulating ZBTB7A expression or activity in CRPC cells.

S.P. Balk is a consultant at Sanofi, Kronos, Constellation Pharamceuticals, and Radius, and has provided expert testimony for Atellas. No potential conflicts of interest were disclosed by the other authors.

Conception and design: D. Han, S. Chen, S. Gao, S.P. Balk, H.H. He, C. Cai

Development of methodology: D. Han, J.N. Owiredu, H.H. He, C. Cai

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D. Han, W. Han, J.N. Owiredu, M. Li, H.H. He

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D. Han, S. Chen, W. Han, S. Gao, J.N. Owiredu, M. Li, S.P. Balk, H.H. He, C. Cai

Writing, review, and/or revision of the manuscript: D. Han, S. Chen, W. Han, S. Gao, S.P. Balk, H.H. He, C. Cai

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Han, M. Li, H.H. He

Study supervision: H.H. He, C. Cai

This work was supported by grants from NIH (R00 CA166507 and R01 CA211350 to C. Cai and P01 CA163227 to S.P. Balk), DOD (W81XWH-15-1-0554 to S. Gao and W81XWH-16-1-0445 to C. Cai), CIHR (142246, 152863, 152864, and 159567 to H.H. He), Prostate Cancer Canada (RS2016-1022 and TAG2018-2061 to H.H. He), NSERC (498706 to H.H. He), Terry Fox New Investigator Award (1069 to H.H. He), and Princess Margaret Cancer Foundation (to H.H. He). We thank Dr. Jill A. Macoska and Susan C. Patalano (Genomics Core, University of Massachusetts Boston) for assistance and guidance of high-throughput sequencing.

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.
Yuan
X
,
Cai
C
,
Chen
S
,
Chen
S
,
Yu
Z
,
Balk
SP
. 
Androgen receptor functions in castration-resistant prostate cancer and mechanisms of resistance to new agents targeting the androgen axis
.
Oncogene
2014
;
33
:
2815
25
.
2.
de Bono
JS
,
Logothetis
CJ
,
Molina
A
,
Fizazi
K
,
North
S
,
Chu
L
, et al
Abiraterone and increased survival in metastatic prostate cancer
.
N Engl J Med
2011
;
364
:
1995
2005
.
3.
Green
SM
,
Mostaghel
EA
,
Nelson
PS
. 
Androgen action and metabolism in prostate cancer
.
Mol Cell Endocrinol
2012
;
360
:
3
13
.
4.
Cai
C
,
He
HH
,
Chen
S
,
Coleman
I
,
Wang
H
,
Fang
Z
, et al
Androgen receptor gene expression in prostate cancer is directly suppressed by the androgen receptor through recruitment of lysine-specific demethylase 1
.
Cancer Cell
2011
;
20
:
457
71
.
5.
Zhao
JC
,
Yu
J
,
Runkle
C
,
Wu
L
,
Hu
M
,
Wu
D
, et al
Cooperation between Polycomb and androgen receptor during oncogenic transformation
.
Genome Res
2012
;
22
:
322
31
.
6.
Gao
S
,
Gao
Y
,
He
HH
,
Han
D
,
Han
W
,
Avery
A
, et al
Androgen receptor tumor suppressor function is mediated by recruitment of retinoblastoma protein
.
Cell Rep
2016
;
17
:
966
76
.
7.
Schweizer
MT
,
Antonarakis
ES
,
Wang
H
,
Ajiboye
AS
,
Spitz
A
,
Cao
H
, et al
Effect of bipolar androgen therapy for asymptomatic men with castration-resistant prostate cancer: Results from a pilot clinical study
.
Sci Transl Med
2015
;
7
:
269ra2
.
8.
Schweizer
MT
,
Wang
H
,
Luber
B
,
Nadal
R
,
Spitz
A
,
Rosen
DM
, et al
Bipolar androgen therapy for men with androgen ablation naive prostate cancer: results from the phase II BATMAN study
.
Prostate
2016
;
76
:
1218
26
.
9.
Lam
HM
,
Corey
E
. 
Supraphysiological testosterone therapy as treatment for castration-resistant prostate cancer
.
Front Oncol
2018
;
8
:
167
.
10.
Teply
BA
,
Wang
H
,
Luber
B
,
Sullivan
R
,
Rifkind
I
,
Bruns
A
, et al
Bipolar androgen therapy in men with metastatic castration-resistant prostate cancer after progression on enzalutamide: an open-label, phase 2, multicohort study
.
Lancet Oncol
2018
;
19
:
76
86
.
11.
Lunardi
A
,
Ala
U
,
Epping
MT
,
Salmena
L
,
Clohessy
JG
,
Webster
KA
, et al
A co-clinical approach identifies mechanisms and potential therapies for androgen deprivation resistance in prostate cancer
.
Nat Genet
2013
;
45
:
747
55
.
12.
Wang
G
,
Lunardi
A
,
Zhang
J
,
Chen
Z
,
Ala
U
,
Webster
KA
, et al
Zbtb7a suppresses prostate cancer through repression of a Sox9-dependent pathway for cellular senescence bypass and tumor invasion
.
Nat Genet
2013
;
45
:
739
46
.
13.
Liu
XS
,
Haines
JE
,
Mehanna
EK
,
Genet
MD
,
Ben-Sahra
I
,
Asara
JM
, et al
ZBTB7A acts as a tumor suppressor through the transcriptional repression of glycolysis
.
Genes Dev
2014
;
28
:
1917
28
.
14.
Hsieh
CL
,
Botta
G
,
Gao
S
,
Li
T
,
Van Allen
EM
,
Treacy
DJ
, et al
PLZF, a tumor suppressor genetically lost in metastatic castration-resistant prostate cancer, is a mediator of resistance to androgen deprivation therapy
.
Cancer Res
2015
;
75
:
1944
8
.
15.
Wang
S
,
Sun
H
,
Ma
J
,
Zang
C
,
Wang
C
,
Wang
J
, et al
Target analysis by integration of transcriptome and ChIP-seq data with BETA
.
Nat Protoc
2013
;
8
:
2502
15
.
16.
Cai
C
,
Wang
H
,
He
HH
,
Chen
S
,
He
L
,
Ma
F
, et al
ERG induces androgen receptor-mediated regulation of SOX9 in prostate cancer
.
J Clin Invest
2013
;
123
:
1109
22
.
17.
Cancer Genome Atlas Research N
. 
The molecular taxonomy of primary prostate cancer
.
Cell
2015
;
163
:
1011
25
.
18.
Robinson
D
,
Van Allen
EM
,
Wu
YM
,
Schultz
N
,
Lonigro
RJ
,
Mosquera
JM
, et al
Integrative clinical genomics of advanced prostate cancer
.
Cell
2015
;
161
:
1215
28
.
19.
Bezzi
M
,
Seitzer
N
,
Ishikawa
T
,
Reschke
M
,
Chen
M
,
Wang
G
, et al
Diverse genetic-driven immune landscapes dictate tumor progression through distinct mechanisms
.
Nat Med
2018
;
24
:
165
75
.
20.
Song
W
,
Khera
M
. 
Physiological normal levels of androgen inhibit proliferation of prostate cancer cells in vitro
.
Asian J Androl
2014
;
16
:
864
8
.
21.
Cui
J
,
Yang
Y
,
Zhang
C
,
Hu
P
,
Kan
W
,
Bai
X
, et al
FBI-1 functions as a novel AR co-repressor in prostate cancer cells
.
Cell Mol Life Sci
2011
;
68
:
1091
103
.
22.
Mouraviev
V
,
Lee
B
,
Patel
V
,
Albala
D
,
Johansen
TE
,
Partin
A
, et al
Clinical prospects of long noncoding RNAs as novel biomarkers and therapeutic targets in prostate cancer
.
Prostate Cancer Prostatic Dis
2016
;
19
:
14
20
.
23.
Crea
F
,
Watahiki
A
,
Quagliata
L
,
Xue
H
,
Pikor
L
,
Parolia
A
, et al
Identification of a long non-coding RNA as a novel biomarker and potential therapeutic target for metastatic prostate cancer
.
Oncotarget
2014
;
5
:
764
74
.
24.
White
NM
,
Zhao
SG
,
Zhang
J
,
Rozycki
EB
,
Dang
HX
,
McFadden
SD
, et al
Multi-institutional analysis shows that low PCAT-14 expression associates with poor outcomes in prostate cancer
.
Eur Urol
2017
;
71
:
257
66
.
25.
Zhang
Y
,
Pitchiaya
S
,
Cieslik
M
,
Niknafs
YS
,
Tien
JC
,
Hosono
Y
, et al
Analysis of the androgen receptor-regulated lncRNA landscape identifies a role for ARLNC1 in prostate cancer progression
.
Nat Genet
2018
;
50
:
814
24
.
26.
Malik
R
,
Patel
L
,
Prensner
JR
,
Shi
Y
,
Iyer
MK
,
Subramaniyan
S
, et al
The lncRNA PCAT29 inhibits oncogenic phenotypes in prostate cancer
.
Mol Cancer Res
2014
;
12
:
1081
7
.
27.
Prensner
JR
,
Chen
W
,
Han
S
,
Iyer
MK
,
Cao
Q
,
Kothari
V
, et al
The long non-coding RNA PCAT-1 promotes prostate cancer cell proliferation through cMyc
.
Neoplasia
2014
;
16
:
900
8
.
28.
Prensner
JR
,
Chen
W
,
Iyer
MK
,
Cao
Q
,
Ma
T
,
Han
S
, et al
PCAT-1, a long noncoding RNA, regulates BRCA2 and controls homologous recombination in cancer
.
Cancer Res
2014
;
74
:
1651
60
.
29.
Shang
Z
,
Yu
J
,
Sun
L
,
Tian
J
,
Zhu
S
,
Zhang
B
, et al
LncRNA PCAT1 activates AKT and NF-kappaB signaling in castration-resistant prostate cancer by regulating the PHLPP/FKBP51/IKKalpha complex
.
Nucleic Acids Res
2019
;
47
:
4211
25
.
30.
Cai
C
,
Wang
H
,
Xu
Y
,
Chen
S
,
Balk
SP
. 
Reactivation of androgen receptor-regulated TMPRSS2:ERG gene expression in castration-resistant prostate cancer
.
Cancer Res
2009
;
69
:
6027
32
.
31.
Igawa
T
,
Lin
FF
,
Lee
MS
,
Karan
D
,
Batra
SK
,
Lin
MF
. 
Establishment and characterization of androgen-independent human prostate cancer LNCaP cell model
.
Prostate
2002
;
50
:
222
35
.
32.
Lam
HM
,
Nguyen
HM
,
Labrecque
MP
,
Brown
LG
,
Coleman
IM
,
Gulati
R
, et al
Durable response of enzalutamide-resistant prostate cancer to supraphysiological testosterone is associated with a multifaceted growth suppression and impaired DNA damage response transcriptomic program in patient-derived xenografts
.
Eur Urol
2019
.
doi: 10.1016/j.eururo.2019.05.042. [Epub ahead of print].
33.
Habib
N
,
Kaplan
T
,
Margalit
H
,
Friedman
N
. 
A novel Bayesian DNA motif comparison method for clustering and retrieval
.
PLoS Comput Biol
2008
;
4
:
e1000010
.