Antithetic hTERT Regulation by Androgens in Prostate Cancer Cells: hTERT Inhibition Is Mediated by the ING1 and ING2 Tumor Suppressors

Simple Summary The expression of the catalytic subunit of the human telomerase reverse transcriptase subunit (hTERT) is hormonally controlled. Androgen treatment suppresses the hTERT expression at a transcriptional level in prostate cancer cells. Here, we identified the responsive promoter element that mediates the androgen receptor induced transrepression of hTERT. The negative androgen response element (nARE) is identified as 62 bp located in the core promoter of hTERT. Chromatin immunoprecipitations indicate an androgen-dependent recruitment of the androgen receptor (AR) ING1 and ING2 to the hTERT promoter. Interestingly, the androgen-induced transrepression is mediated by the class II tumor suppressors inhibitor of growth 1 and 2, namely ING1 and ING2, respectively. Abstract The human telomerase is a key factor during tumorigenesis in prostate cancer (PCa). The androgen receptor (AR) is a key drug target controlling PCa growth and regulates hTERT expression, but is described to either inhibit or to activate. Here, we reveal that androgens repress and activate hTERT expression in a concentration-dependent manner. Physiological low androgen levels activate, while, notably, supraphysiological androgen levels (SAL), used in bipolar androgen therapy (BAT), repress hTERT expression. We confirmed the SAL-mediated gene repression of hTERT in PCa cell lines, native human PCa samples derived from patients treated ex vivo, as well as in cancer spheroids derived from androgen-dependent or castration resistant PCa (CRPC) cells. Interestingly, chromatin immuno-precipitation (ChIP) combined with functional assays revealed a positive (pARE) and a negative androgen response element (nARE). The nARE was narrowed down to 63 bp in the hTERT core promoter region. AR and tumor suppressors, inhibitor of growth 1 and 2 (ING1 and ING2, respectively), are androgen-dependently recruited. Mechanistically, knockdown indicates that ING1 and ING2 mediate AR-regulated transrepression. Thus, our data suggest an oppositional, biphasic function of AR to control the hTERT expression, while the inhibition of hTERT by androgens is mediated by the AR co-repressors ING1 and ING2.


Introduction
Prostate cancer (PCa) is the most commonly diagnosed cancer and the second leading cause of cancer mortality for males in the USA [1]. The androgen receptor (AR) regulates the development and physiological function of the normal prostate, as well as the proliferation of cancerous prostate tissue. PCa is initially an androgen-sensitively growing tumor that eventually becomes castration-resistant (CRPC) and drug resistant [2]. While age is a major risk factor for this disease, androgen levels also contribute to the growth of the tumor. Importantly, AR represents a major drug target in the treatment of PCa. An important hormone therapy is androgen deprivation therapy (ADT) to reduce androgen production. This is often combined with AR-antagonist treatment for full blockade of the AR.
Interestingly however, there is a paradox in the androgen response of PCa cells. In addition to androgen deprivation, supraphysiological levels of androgens (SAL) also inhibit PCa growth [3][4][5]. This observation led to clinical trials using bipolar androgen therapy (BAT) composed of applying SAL within repeated cycles in combination with androgen deprivation [6]. Supraphysiologic levels of testosterone used for BAT have been shown to be effective for inhibiting PCa progression. It is suggested that the sudden switch in the androgen levels provide less time for adaptive responses of AR-signaling [7]. Mechanistically, we found that SAL induces cellular senescence in human PCa tumor cell lines, as well as in PCa tissues obtained from patients with prostatectomy [8]. This suggests that SAL may induce a tumor suppressive program in PCa cells. In line with this, SAL represses hTERT expression, the catalytical subunit of the telomerase [9], which might be a factor as a tumor-suppressive function of AR.
AR is a ligand-controlled transcription factor and a member of the steroid and nuclear hormone receptor superfamily. Androgen-activated AR translocates to the nucleus, where it binds to chromatin at androgen response elements (ARE) to control the gene expression. AR can interact with co-activators and co-repressors to either activate or repress the expression of target genes [10,11].
The regulation of hTERT expression by androgen is controversial. On the one hand, the induction of hTERT by androgens was reported [12]. Mechanistically, an indirect effect of androgens on hTERT activation was postulated [13]. The use of AR antagonists indicate a down-regulation of hTERT leading also to the assumption that androgens activate hTERT [14]. However, on the other hand, the first evidence of androgen regulation of telomerase activity was obtained from castrated animals [15]. In line with the role of androgens as a protective agent, androgen ablation in animal model systems led to the induction of telomerase activity [15,16]. These data suggest that the telomerase activity is repressed by androgens, and reducing the androgen levels leads to its activation. The underlying mechanism of how AR activates or represses hTERT expression remains elusive. It has further been reported that hTERT is repressed by androgen treatment in PCa cells [9,12,17].
As we have recently identified novel AR co-repressors that act as tumor suppressors [18,19], we aimed to investigate the underlying mechanism of hTERT repression by androgens. The inhibitor of growth (ING) family of proteins represents a tumor suppressor family [20]. Unlike the other ING family members, human ING1 and ING2 are closely related proteins exhibiting a high homology in the amino acid sequence [21,22]. Interestingly, ING1 and ING2 have been found to physically interact with AR and to act as AR co-repressors [18,19]. It remains unclear whether ING1 and ING2 are also involved in ARmediated transrepression. We also included an analysis of the first-and second-generation AR-antagonists Bicalutamide, Enzalutamide (Enz), and Darolutamide.
Here, we show that both the androgen-mediated activation and inhibition of hTERT expression are androgen dose-dependent. Furthermore, the data suggest that the androgenregulation of hTERT expression is time-dependent. In native human PCa patient samples treated ex vivo, the expression of hTERT is repressed by SAL. The hTERT repression is profound in SAL, whereas the physiological low androgen levels (LAL) instead activate the hTERT expression. In line with this, we identified in the distal hTERT promoter a positive androgen response element and in the proximal promoter region a negative androgen Cancers 2021, 13,4025 3 of 17 response element, nARE, spanning 63 bp. Interestingly, the androgen-regulated hTERT repression was mediated through ING1 and ING2. Notably, AR antagonists also repressed the hTERT expression, albeit less efficiently than SAL.

Cell Culture and Retroviral Transduction
The cultivation of both LNCaP and C4-2 cells were described previously [8]. As defined earlier based on androgen-dependent growth, we used 1nM R1881 or 10nM DHT as SAL, 1pM R1881 or 10pM DHT as LAL, and DMSO as the solvent control [8]. For the knockdown of ING1 and ING2, retroviral vectors pLMP-shING1b and pSR-shING2a were used with pLMP-shluc or pSR-shluc as the controls, respectively. Transient transfection and retroviral transduction experiments were performed as described earlier [9,18,19]. Transient transfections were performed with the indicated hTERT promoter fragments fused to the luciferase gene. The hTERT-3996 fragment harbors the entire promoter sequence fused to the luciferase gene, while the sequence between the numbers indicates the hTERT promoter regions fused to the TATA-Box-luciferase gene fusion.

RNA-Sequencing and Transcriptome Analysis
The total RNA was isolated from LNCaP and C4-2 cells treated for 72 h with solvent control and androgens, in triplicate, using peqGOLD TriFast (Peqlab, Erlangen, Germany), according to the manufacturer's protocol. The sequencing library was prepared using SMARTer Stranded Total RNA Sample Prep Kit-HI Mammalian (TAKARA, Kusatu, Japan). The paired-end sequencing was done using Illumina NextSeq 500/550 High-Output v2 Kit (150 cycles, Illumina, San Diego, CA, USA). Fastq files' quality check was performed using FastQC (v0.11.5). The first four nucleotides were removed from the sequenced reads using fastx_trimmer (FASTX Toolkit 0.0.13). The low-quality nucleotides from the end of each read were removed using fastq_quality_trimmer (FASTX Toolkit 0.0.13) with -Q 33 -t 20 -l 25 parameters. Only the first sequenced read from the pair end (R1) was used for the downstream analysis. The fastq files were mapped to the hg19 genome using tophat (v2.1.0) with the following parameters -bowtie1 -no-coverage-search -a 5. The number of reads covered by each gene was calculated by htseq-count (0.11.2) with -s no -a 0 -t exon -m intersection-nonempty parameters and hg19 gencode.v19 annotation. Before further analysis, all of the rRNA genes (5srRNA, rRNA, and mt-rRNA) were removed from the count data. To calculate the p-value and normalized count (based on the geometric library size factors), DESeq2 (1.20.0) R package with the default parameters and paired test was used. For the gene ontology analysis, the deferentially expressed genes were used in gprofiler function from gProfileR (v0.7.0) R package, searching in the GO:BP, GO:MF, GO:CC, and REAC data bases.
For the gene set enrichment analysis (GSEA), normalized counts (for each gene in all of the samples) were scaled using the scale function in R (with center = TRUE, scale = TRUE parameters). The average of the z-scores (scaled normalized counts) was calculated for each group (one value for each gene per group) and was used for plotting and statistical analysis. Wilcoxon paired test was performed for p-value calculation. Reactome analysis was performed according to [23]. The RNA-sequencing data are available in the gene expression omnibus (GEO) database under the accession number GSE151492.

Generation of Spheroids and Immunostaining
Spheroids were generated, as described earlier [24], by forced floating methods in 96-well ULA plates with seeding of 1000 cells/well. Then, 24 h after seeding, AR ligands or DMSO were added with each six technical replicates.

Senescence-Associated Beta-Galactosidase (SA β-Gal) Staining of Spheroids
After eight days of treatment, the spheroids were fixed in 4% paraformaldehyde and washed three times with 1× PBS pH 6.0. For the SA β-Gal activity staining, the spheroids were incubated in the staining solution according to [8].

Reporter Gene Assay
The reporter gene assays were performed as described earlier [9,18,19].

Reverse Transcription Quantitative Real-Time PCR (RT-qPCR)
RNA was isolated from the cells using peqGOLD TriFast (Peqlab, Erlangen, Germany), according to the manufacturer's protocol. Two-step RT-qPCR was conducted using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) and the SsoFast EvaGreen Supermix (Bio-Rad, Munich, Germany), gene specific primers, and Bio-Rad CFX96 Real Time PCR detection system. The RT-qPCR results were analyzed via the ∆∆Ct method [25] using CFX manager software from Bio-Rad. The primer sequences are listed as 5 →3 :

Ex Vivo Treatment of Prostate Cancer Samples
The ex vivo treatment of native PCa samples from patients with radical prostatectomies was described previously [26]. All of the patients gave informed consent and all were informed about the purpose of the study. The study was approved by the Ethics Committee of the Friedrich Schiller University (ethical approvals 3286-11/11 and 2019-1502), and it conformed to the Declaration of Helsinki.

Chromatin Immunoprecipitation (ChIP)
ChIP was essentially performed as described previously [9], with the modification that the beads were washed with RIPA buffer and suspended in a 100 µL RIPA buffer. The antibodies were used according to Table 1. For the hTERT genomic proximal promoter locus, nested PCR was used.

Statistical Analysis
For the statistical analyses, except otherwise stated, two-tailed unpaired Student's t-tests were performed using the GraphPad Prism 8.0. Within each experiment, treatments were performed in technical duplicates. For the statistical analysis, data from three independent experiments were analyzed. Western Blot analysis and RT-qPCR experiments were performed for at least two independent experiments. Error bars show the mean ± SEM values, calculated from the mean, SD, and number of independent replicates. For statistical analysis of spheroid growth repeated measure two-way ANOVA was performed.

Molecular Profiling on Transcriptomics Level Reveals Repression of hTERT Expression by SAL
To identify the mRNA profile affected by androgen treatment at SAL, we performed transcriptome analyses with multiple comparisons. RNA sequencing (RNA-Seq) was performed with both the androgen-sensitive human LNCaP and the CRPC C4-2 cell lines treated with and without the androgen R1881 (methyltrienolone) at SAL (1 nM) or with DMSO as the solvent control. As dihydrotestosterone (DHT) can be rapidly metabolized and the metabolites can activate the estrogen receptor beta [27], the more stable and ARspecific compound R1881 was used as an AR agonist to treat the cells for 72 h. RNA-seq has been performed three times independently for each condition in both cell lines with all treatments. The number of significantly differentially expressed genes (DEGs), including the number of upregulated genes and the number of downregulated genes, indicates that LNCaP responds more sensitively to androgen treatment compared with the CRPC cell line C4-2 ( Figure 1A). This may be due to the fact that the AR is more active in CRPC cells, and thus additional added ligands lead to a lower number of DEGs compared with the LNCaP cells.
has been performed three times independently for each condition in both cell lines with all treatments. The number of significantly differentially expressed genes (DEGs), including the number of upregulated genes and the number of downregulated genes, indicates that LNCaP responds more sensitively to androgen treatment compared with the CRPC cell line C4-2 ( Figure 1A). This may be due to the fact that the AR is more active in CRPC cells, and thus additional added ligands lead to a lower number of DEGs compared with the LNCaP cells. (H) Native PCa tissue pieces obtained from patients with radical prostatectomy were treated ex vivo with R1881 at SAL as described previously (8). RT-qPCR was used to detect the regulation of hTERT expression by SAL. n = 4, one-tailed student t-test, * p-value: 0.028, n.s.: not significant.
Significantly down-regulated genes after androgen treatment were significantly enriched in telomere-related gene ontology (GO) terms in both the LNCaP and C4-2 cell lines ( Figure 1B,D). Interestingly, upon androgen treatment at SAL, enrichment of the down-regulated genes was observed in terms such as telomere organization, telomere capping, and telomerase activity in both cell lines ( Figure 1B,D). Furthermore, the reactome terms indicate that the significantly down-regulated genes were also enriched in telomeric terms by SAL treatment in both cell lines ( Figure 1C,E). No significantly enriched GO term was observed from the up-regulated genes in both cell lines. The number of enriched terms in LNCaP was higher compared with the C4-2 cells, indicating that telomeric function is less androgen-regulated in CPPC C4-2 cells, and telomere-related genes in LNCaP show a stronger response to androgen treatment ( Figure 1F). This observation was confirmed by analyzing the shelterin complex gene set ( Figure 1F). The shelterin complex is a multi-subunit protein complex responsible for protecting telomeric DNA from unwanted degradation and end-to-end fusion events [28]. These data suggest an androgen regulation of telomere protecting factors more significantly in LNCaP compared with C4-2 cells ( Figure 1F). This observation may support the findings that AR plays a role in telomere stability and in the replication of telomere DNA in PCa cells [29,30]. Moreover, the upregulation of telomere protecting factors, the expression of the catalytic subunit of telomerase, hTERT, is repressed by SAL ( Figure 1G). This suggests that androgen treatment at SAL regulates the expression of telomeric genes.

Repression of hTERT by SAL in PCa Tumor Samples Ex Vivo and in Spheroids
To verify the androgen-mediated repression of hTERT in PCa tissues, native human PCa specimens from radical prostatectomy were obtained and treated ex vivo for 48 h with or without SAL [8]. Reverse transcription qRT-PCR data confirmed the down-regulation of hTERT by SAL in the ex vivo treated patient tumor samples ( Figure 1H). This indicates that the AR transrepresses the hTERT expression at SAL.
To further verify that androgens repress hTERT expression in a 3D tumor cell entity, PCa spheroids from both LNCaP and C4-2 cell lines were generated (Figure 2A,C). 3Dspheroids are suggested to mimic a tumor better in terms of complexity and drug delivery compared with monolayer cultures [24,31,32]. The spheroids were treated with DHT at SAL or R1881 at LAL or SAL, and the growth was compared with the solvent control for the indicated days. Treatment of spheroids with the androgens DHT or R1881 at SAL conditions significantly inhibited the growth of the LNCaP (Figure 2A,B) and C4-2 ( Figure 2C,D) spheroids. The SAL-mediated growth repression was slightly more pronounced in LNCaP compared with the C4-2 spheroids. Although LAL promotes the growth of LNCaP cells in adherent cell cultures [8], LAL treatment did not show a significant effect on the growth of LNCaP spheroids. The growth of spheroids CRPC C4-2 cells was unaffected by LAL treatment. were analyzed every second day. The mean ± SEM values were calculated from three independent experiments, whereas for each biological experiment, six technical replicates were performed. The significance of the data from the treatments versus DMSO was analyzed with two-way repeated measure ANOVA with Dunnett's multiple comparisons test (** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001). Both LNCaP (E,F) and C4-2 (G,H) spheroids were cultured for 8 days with the indicated ligands, and were sliced (4 µM thick) and subsequently immunostained for Ki-67 as a proliferation marker (E,G), and finally were analyzed for SA β-Gal activity as a hallmark for senescent cells (F,H). The scale bar is indicated for 100 µm. Spheroids generated from LNCaP (I) or C4-2 (J) cells were analyzed for hTERT mRNA by qRT-PCR. Values were calculated relative to DMSO, which was set arbitrarily as 1. Error bars indicate SEM. *** p ≤ 0.001.
To confirm growth reduction of spheroids by SAL treatment, LNCaP ( Figure 2E,F) and C4-2 ( Figure 2G,H) were sliced. Immunostaining of the proliferation marker Ki-67 was performed and co-stained with DAPI. Ki-67 staining was found in the outer layer of both the LNCaP and C4-2 spheroids ( Figure 2E,G). The data further suggest that SAL treatment strongly reduces Ki-67 staining, indicating that the level of proliferating cells is reduced by SAL in both LNCaP and C4-2 spheroids.
To analyze whether SAL induces cellular senescence, the spheroid slices were analyzed for the senescence-associated β-galactosidase (SA β-Gal) activity ( Figure 2F,H). We observed a robust induction of SA β-Gal activity by SAL treatment in the spheroids of LNCaP, which was less pronounced in those derived from C4-2 cells. An increase of SA β-Gal activity was not observed with the LAL treatment. These data suggest that PCa spheroids respond to androgen treatment and that high doses of androgens inhibit growth and induce cellular senescence.
To confirm the inhibition of hTERT mRNA levels by SAL also in spheroids, the RNA from both the LNCaP and C4-2 spheroids were extracted. The obtained data suggest that the expression of hTERT is reduced by SAL treatment in the 3D spheroid model system of both cell lines ( Figure 2I,J). Notably, the hTERT expression is repressed more potently by SAL in LNCaP cells compared with the castration-resistant C4-2 cells. This confirms the RNA-seq transcriptome analysis and is consistent with the hTERT mRNA expression data of the adherent monolayer. While DHT treatment leads to the inhibition of the hTERT expression in LNCaP spheroids, we observed only mild inhibition of hTERT in the C4-2 spheroids.
Thus, SAL treatment leads to a repression hTERT expression in adherent PCa monolayer cells, in PCa spheroids, and in patient tumor samples.

Androgen-Mediated Repression and Activation of hTERT Expression Is Time-and Concentration-Dependent
To analyze the concentration-dependent androgen regulation of the hTERT expression, RT-qPCR was employed. We compared the androgen response of the known positively regulated AR target gene KLK3 in both human PCa cell lines LNCaP (Figure 3A,B) and C4-2 ( Figure 3C,D). As expected, by increasing the androgen concentrations, the KLK3 expression was enhanced in both cell lines, with a stronger effect in the LNCaP compared with the C4-2 cells (Figure 3A,C). An upregulated KLK3 expression remained induced when higher androgen concentrations were used. However, interestingly, the hTERT expression was up-regulated at lower androgen levels, whereas it was inhibited at higher levels of androgens ( Figure 3B,D). This antithetic up-and down-regulation was observed in both cell lines. Notably, the androgen-mediated induction, as well as the repression of hTERT, was less prominent in the castration-resistant C4-2 cells. This bidirectional response of hTERT expression depends on the androgen concentrations, which may explain, at least in part, some of the controversial discussions about its regulation by AR. Thus, the data indicate a dual role of AR in influencing the hTERT expression dependent on androgen levels.  Next, we analyzed the time-dependency of the androgen response of hTERT mRN The results suggest that the level of hTERT mRNA was reduced at 48 and 72 h by SA treatment (Figure 3E). At a protein level, hTERT was slightly repressed at 48 h and mo pronounced at 72 h with the SAL treatment ( Figure 3F, the uncropped Western blots a shown in Figure S5.). Surprisingly, treatment with the second-generation AR antagon Enz resulted also in a reduction of the hTERT protein level at 72 h ( Figure 3F). This obs vation is consistent with reduced hTERT mRNA levels by different AR antagonists in bo cell lines ( Figure S1). These results indicate that unexpectedly, the AR antagonists Bica tamide, Enz, and Darolutamide did not fully act in an opposite manner compared w the agonists in order to regulate hTERT.

Identification of a Positive Androgen Response element in the Distal hTERT Promoter
To obtain additional molecular insights into the dual regulation of hTERT by andr gens, we used reporter assays in CV1 cells that lack AR, and the related receptors for g cocorticoids, progesterone, and estrogen that may interfere with the assay system. Wi out the co-expression of AR, no androgen regulation of hTERT was observed [9]. Bas Next, we analyzed the time-dependency of the androgen response of hTERT mRNA. The results suggest that the level of hTERT mRNA was reduced at 48 and 72 h by SAL treatment (Figure 3E). At a protein level, hTERT was slightly repressed at 48 h and more pronounced at 72 h with the SAL treatment ( Figure 3F, the uncropped Western blots are shown in Figure S5). Surprisingly, treatment with the second-generation AR antagonist Enz resulted also in a reduction of the hTERT protein level at 72 h ( Figure 3F). This observation is consistent with reduced hTERT mRNA levels by different AR antagonists in both cell lines ( Figure S1). These results indicate that unexpectedly, the AR antagonists Bicalutamide, Enz, and Darolutamide did not fully act in an opposite manner compared with the agonists in order to regulate hTERT.

Identification of a Positive Androgen Response Element in the Distal hTERT Promoter
To obtain additional molecular insights into the dual regulation of hTERT by androgens, we used reporter assays in CV1 cells that lack AR, and the related receptors for glucocorticoids, progesterone, and estrogen that may interfere with the assay system. Without the co-expression of AR, no androgen regulation of hTERT was observed [9]. Based on chromatin immunoprecipitation (ChIP) experiments, AR is recruited androgen-dependently to both the −4 kb and −0.1 kb regions of the hTERT promoter [9] (Figure S2). Therefore, the promoter fragments spanning these regions were analyzed functionally in more detail using promoter fragments fused to a TATA-box-luciferase construct. The hTERT promoter activity spanning the entire region from −4 kb to the start codon is repressed by the androgen treatment ( Figure 4A,B). Focusing on the distal −4 kb hTERT, various hTERT fragments were analyzed for their transcriptional activity. Using hTERT −3988 to −3276 and smaller fragments, the androgen treatment led to potent up-regulation of the reporters, suggesting a positive ARE (pARE) in the distal promoter region of hTERT ( Figure 4A,C,D) at −4 kb. Thus, the data indicate the existence of a pARE that was narrowed down to −3985 to −3956 spanning only 29 bp.

Identification of a Negative Androgen Response Element in the Proximal hTERT Promoter
Focusing on the −0.1 kb region to which AR is also recruited in an androgen-dependent manner, various promoter deletions were generated ( Figure 5A). In the absence of AR (empty vector-e.v.) no androgen-mediated transrepression was observed ( Figure 5B). Co-expressing AR indicates androgen-mediated repression by fragments spanning the proximal promoter region of −120 bp to −58 bp upstream of the transcription start site as the nARE ( Figure 5C). The promoter fragment lacking this sequence lacked androgenmediated transrepression. In addition, the AR mutant AR-T877A mediated transrepression via this nARE ( Figure 5D). The AR-T877A mutant mediated resistance to the AR antagonist flutamide, and was expressed in both LNCaP and C4-2 cells. Using the −240 in both orientations, it was suggested that the transrepressive element was located within this fragment, but acted in an orientation-dependent sense manner. Further deletions of the −120 to −58 bp sequence led to loss of transrepression ( Figure 5E). These data suggest the minimal fragment for AR, and the androgen-dependent transrepression of hTERT was the 63bp spanning negative response element nARE at −120 to −58 bp.

Identification of a Negative Androgen Response Element in the Proximal hTERT Promoter
Focusing on the −0.1 kb region to which AR is also recruited in an androgen-dependent manner, various promoter deletions were generated ( Figure 5A). In the absence of AR (empty vector-e.v.) no androgen-mediated transrepression was observed ( Figure 5B). Co-expressing AR indicates androgen-mediated repression by fragments spanning the proximal promoter region of −120 bp to −58 bp upstream of the transcription start site as the nARE ( Figure 5C). The promoter fragment lacking this sequence lacked androgenmediated transrepression. In addition, the AR mutant AR-T877A mediated transrepression via this nARE ( Figure 5D). The AR-T877A mutant mediated resistance to the AR antagonist flutamide, and was expressed in both LNCaP and C4-2 cells. Using the −240 in both orientations, it was suggested that the transrepressive element was located within this fragment, but acted in an orientation-dependent sense manner. Further deletions of the −120 to −58 bp sequence led to loss of transrepression ( Figure 5E). These data suggest the minimal fragment for AR, and the androgen-dependent transrepression of hTERT was the 63bp spanning negative response element nARE at −120 to −58 bp.

The Tumor Suppressor ING1 and ING2 Mediate AR-Regulated Transrepression of hTERT
To address the underlying mechanism of AR-mediated gene repression, various corepressors were over-expressed in CV1 cells, including NCoR, SMRT, Alien, CoREST, and TSGA, without showing a response to the hTERT proximal promoter activity (Supplemental Figure S3). However, the over-expression of ING1 reduced the activity of the proximal −120 to −58bp hTERT sequence ( Figure S3). The overexpression of ING1 may squelch repressive factors, and the data suggest that ING1 is associated with the AR-mediated repression of hTERT.
Therefore, we hypothesized that the tumor suppressor ING1 might be recruited to the hTERT proximal promoter. In addition, we also analyzed ING2 recruitment, as both ING1 and ING2 were identified as AR co-repressors [18,19]. For this purpose, ChIP experiments were performed with LNCaP cells to detect whether ING1 and ING2 were recruited to the hTERT proximal promoter ( Figure 6A). The data suggest that compared with the solvent control (DMSO), AR, ING1, and ING2 were recruited to the hTERT promoter by SAL treatment of the LNCaP cells ( Figure 6A). Interestingly, the other ING family members ING3, ING4, and ING5 did not show an androgen-dependent recruitment ( Figure S4). Based on these data, we further hypothesized that the ING1 and ING2 tumor suppressors are involved in the AR-mediated transrepression of hTERT.

The Tumor Suppressor ING1 and ING2 Mediate AR-Regulated Transrepression of hTERT
To address the underlying mechanism of AR-mediated gene repression, various corepressors were over-expressed in CV1 cells, including NCoR, SMRT, Alien, CoREST, and TSGA, without showing a response to the hTERT proximal promoter activity (Supplemental Figure S3). However, the over-expression of ING1 reduced the activity of the proximal −120 to −58bp hTERT sequence ( Figure S3). The overexpression of ING1 may squelch repressive factors, and the data suggest that ING1 is associated with the AR-mediated repression of hTERT.
Therefore, we hypothesized that the tumor suppressor ING1 might be recruited to the hTERT proximal promoter. In addition, we also analyzed ING2 recruitment, as both ING1 and ING2 were identified as AR co-repressors [18,19]. For this purpose, ChIP experiments were performed with LNCaP cells to detect whether ING1 and ING2 were recruited to the hTERT proximal promoter ( Figure 6A). The data suggest that compared with the solvent control (DMSO), AR, ING1, and ING2 were recruited to the hTERT promoter by SAL treatment of the LNCaP cells ( Figure 6A). Interestingly, the other ING family members ING3, ING4, and ING5 did not show an androgen-dependent recruitment ( Figure S4). Based on these data, we further hypothesized that the ING1 and ING2 tumor suppressors are involved in the AR-mediated transrepression of hTERT.  Figure 6D,E). This suggests that AR-regulated transrepression is mediated by ING1 and ING2. Hence, the data confirm our hypothesis that the tumor suppressors and AR co-repressors ING1 and ING2 functionally mediate the transrepression of androgen-activated AR on hTERT. * p ≤ 0.05

Discussion
Pieces of evidence suggest that supraphysiological androgens inhibit PCa growth. In line with this, pharmacological SAL is currently being used in clinical trials within the BAT to treat metastatic CRPC patients. BAT includes ADT with intermittent cycles of ADT in combination with SAL [3,4]. It is suggested that BAT might prevent adaptive changes in AR expression, leading to a prolonged therapy response of patients [6]. Here, we show that SAL treatment of human PCa cells leads to the repression of hTERT expression. The repression of hTERT by androgen was observed with the treatment of 48 h in PCa tumor samples ex vivo, as well as in spheroids. Telomerase function is, on the one hand, to protect telomere erosion by DNA replication, but also has a non-telomeric function in promoting cell growth and mediating more independence towards growth factors in tumorigenesis [33,34]. Inhibition of the hTERT activity in tumor cells might therefore be benefi- Therefore, to confirm and functionally analyze the role of ING1 and ING2 in the hTERT expression, knockdown experiments were performed in LNCaP cells using a retroviral expression system. Western blot data indicate a knockdown of ING1 and ING2 ( Figure 6B,C, the uncropped Western blots are shown in Figure S5). The vector sh-luc served as a knockdown control. qRT-PCR was performed to analyze the effect on hTERT expression upon ING1 or ING2 knockdown. The data suggest that in either ING1 or ING2 knockdown cells, the androgen-mediated hTERT repression is alleviated ( Figure 6D,E). This suggests that AR-regulated transrepression is mediated by ING1 and ING2. Hence, the data confirm our hypothesis that the tumor suppressors and AR co-repressors ING1 and ING2 functionally mediate the transrepression of androgen-activated AR on hTERT. * p ≤ 0.05.

Discussion
Pieces of evidence suggest that supraphysiological androgens inhibit PCa growth. In line with this, pharmacological SAL is currently being used in clinical trials within the BAT to treat metastatic CRPC patients. BAT includes ADT with intermittent cycles of ADT in combination with SAL [3,4]. It is suggested that BAT might prevent adaptive changes in AR expression, leading to a prolonged therapy response of patients [6]. Here, we show that SAL treatment of human PCa cells leads to the repression of hTERT expression. The repression of hTERT by androgen was observed with the treatment of 48 h in PCa tumor samples ex vivo, as well as in spheroids. Telomerase function is, on the one hand, to protect telomere erosion by DNA replication, but also has a non-telomeric function in promoting cell growth and mediating more independence towards growth factors in tumorigenesis [33,34]. Inhibition of the hTERT activity in tumor cells might therefore be beneficial in tumor growth suppression. In the long term, this might be one beneficial aspect of SAL treatment in BAT, as shown by the SAL-suppressed growth of PCa spheroids. Moreover, it is interesting that androgens regulate hTERT expression at least in hormone responsive cancers, which might be an important point as an interesting option of drug therapy.
Interestingly, the androgen-mediated transrepression of hTERT is androgen-concentration, AR-dependent, and treatment-time dependent. Interestingly, on the one hand, lower androgen concentrations enhance hTERT expression. On the other hand, higher doses of androgens at supraphysiological, pharmacological concentrations repress hTERT expression. This opposite regulation of the hTERT response dependent on the androgen concentration may account for the discrepancies in the findings of how hTERT is regulated by androgens. Activation of the hTERT expression by androgens was observed [12,13], whereas repression of hTERT by androgen was also shown [9,15,17]. The androgen-mediated up-regulation of hTERT at LAL may be mediated by the −4 kb pARE, whereas at SAL, the nARE might be overriding the pARE, leading in sum to a SAL-dependent hTERT repression. Interestingly, the identified nARE sequence has a very high GC content and overlaps the recently identified putative quadruplex forming sequence in the hTERT promoter [35].
In addition, the treatment time of SAL also accounts for the repression. Androgen treatment of one day or less did not lead to an inhibition of hTERT mRNA levels, which is in line with previous findings [9]. Longer treatment periods revealed a downregulation of both the hTERT mRNA and hTERT protein by SAL.
However, additional factors may also account for the discrepancies in the androgen regulation of hTERT expression. It may play an important role whether synthetic androgens or the metabolizable DHT were used for experiments. The metabolites of DHT activate the estrogen receptor beta [27]. It is known that the hTERT promoter recruits the estrogen receptor, leading to the activation of the hTERT expression [36,37]. Thus, in cells with a higher DHT metabolism, there might be a balance between the repression of hTERT by DHT and activation by DHT metabolites through the estrogen receptor. A further point that may lead to the different observations is that androgens are produced in PCa cells [38]. Thus, dependent on the particular PCa tumor or PCa subline, the androgen level could be different and contribute specifically to the regulation of the hTERT expression. Previously, we reported that the wild-type but not AR mutant T877A mediates the repression of hTERT mRNA levels [9]. This effect was observed through analysis within 24 h of the hormone treatment. It is nevertheless possible that the AR-T877A mutant mediates a different response on hTERT as the wild-type AR observed. This might also be reflected by the hTERT deletion constructs −1301 to −238 and −58 to −1. Both non-overlapping constructs are enhanced more potently by the AR-T877A mutant compared with wild-type AR, and may therefore account for a difference between wild-type and mutant AR.
Notably, the down-regulation of hTERT by androgens is not necessarily coupled to an up-regulation of hTERT by AR antagonists. In fact, as LAL up-regulates the hTERT expression, we expect a down-regulation from treatment with AR antagonists. As AR antagonists are shown to repress hTERT, the conclusion was drawn that androgens would show the opposite activity and may activate hTERT expression. However, this seemed to not be the case. We tested various AR antagonists and indeed observed a down-regulation of hTERT expression in the same experimental setup using both cell lines ( Figure S1), being in line with previous observations [12,39]. The underlying mechanism for antagonistmediated hTERT repression is, however, unclear. Interestingly, the sexual hormones estrogens and androgens regulate the hTERT expression at least in hormone responsive cancers, which might be an interesting option for drug therapy.
Thus, the data suggest that when analyzing the regulation of hTERT expression, it is critical whether DHT or other synthetic agonists are used, as well as the androgen concentration and the treatment period. In addition, AR antagonists do not act in the opposite manner as agonists when analyzing the hTERT expression. These particularities seem to be important factors in analyzing the hormonal regulation of hTERT.

Conclusions
Knowledge about negative hormone response elements, including nAREs, is very limited. The identification of ING1 and ING2 tumor suppressors in the AR-mediated transrepression of hTERT provides some first insights into the link between tumor suppressors and the AR (graphical abstract). ING1 and ING2 belong to the ING family of tumor suppressors, with five family members. Both ING1 and ING2 are known as transcriptional repressors interacting with the mSIN3A/HDAC complex [40]. In addition, ING1 and ING2 were shown to act as AR co-repressors that act on the androgen-activated AR [18,19]. This study reveals that the recruitment of both ING1 and ING2 to the proximal promoter of hTERT is enhanced by androgen treatment. Moreover, knockdown experiments confirm functionally that ING1 and ING2 are mediated with SAL-induced hTERT repression. Thus, the androgen-dependent recruitment of these ING factors could mechanistically explain the androgen-dependent repression of hTERT. Taken together, here, we show an antithetic response of hTERT expression dependent on the androgen concentrations and the interplay of the ING1 and ING2 tumor suppressors on hTERT, indicating a novel pathway.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/cancers13164025/s1, Figure S1: AR antagonists repress hTERT expression, Figure S2: Genome Browser depicts the recruitment of AR to the hTERT promoter, Figure S3: Overexpression of selected corepressors do not change the AR-mediated transrepression of hTERT, Figure S4: AR is recruited in an androgen-dependent manner to both the proximal and distal hTERT sites, Figure S5: The uncropped Western blots,

Conflicts of Interest:
The authors declare no conflict of interest.