The PCA3 ncRNA is one of the most prostate-specific genes described to date, is highly overexpressed in PCa tumors, and has been extensively characterized as a tumor biomarker [1, 2, 4, 7, 23]. However, no function has been attributed to this transcript in PCa cells [1, 2, 4, 7, 23]. Our primary aim was to elucidate the putative roles of this ncRNA in PCa cell biology. Previous data have partially supported the concept that PCA3 is a functional transcript, as argued for other ncRNAs [1, 12]. The limited expression of PCA3 in prostate tissues suggests that it may function specifically in this organ. This also suggests that its expression may be tightly regulated, as would be expected for a functional transcript. The results obtained here support the notion that PCA3 is involved in PCa survival pathways by controlling cell growth and viability, at least in part through controlling the AR pro-survival signaling. In addition, our results accord with the hypothesis that PCA3 is involved in transcriptional modulation of AR target genes, although it may act through a still-unknown mechanism.
Other ncRNAs have also been described as being involved in cancer cell survival, including PlncRNA-1, GAS5, HOTAIR, and several miRNA genes [24–27], which act by controlling apoptosis, the cell cycle, cell proliferation, or viability, through their interactions with intracellular signaling networks. Other ncRNAs involved in PCa cell survival and proliferation have also been described, including PCAT-1 , PRNCR1 , PCGEM1 , and PlncRNA-1 . Similarly to PCA3, the AR signaling pathway also modulates the expression of PRNCR1, PlncRNA-1, and PCGEM1. Additionally, several androgen-responsive intronic ncRNAs have been described, indicating that intronic ncRNAs, such as PCA3 , may have control mechanisms that are common to protein-coding transcripts, such as those involving hormonal control of gene-promoter activation . The facts that PCA3 is expressed in higher levels in the androgen-responsive cell line (LNCaP) and that some ncRNA expressed in PCa cells are involved in the AR signaling [24, 29, 30] led us to hypothesize that PCA3 ncRNA expression may also be modulated by this pathway, and that this transcript may also be involved in the control of genes related to this cell signaling.
Uncontrolled cell growth is a result of a progression of changes at the cellular and genetic levels, which ultimately reprogram a cell to undergo uncontrolled division, and is one of the first steps in carcinogenesis. In order to investigate the putative involvement of PCA3 in controlling this early step of tumorigenesis, we investigated the behavior of LNCaP cell growth after PCA3 knockdown by RNA interference. PCA3-silenced LNCaP cells showed a significant attenuation of cell growth, with a corresponding increase in the number of cells undergoing apoptosis. Transfected cells with siScr also showed a slight decrease in cell growth after 20 h post-transfection, which could be due to a cytotoxicity effect promoted by the lipofectamin transfection reagent and acting specifically on LNCaP cells, which has not been observed in other cell lines. Notably, siPCA3-LNCaP transfected cells show a greater decrease in cell growth rates than do siScr-transfected cells, indicating that PCA3 ncRNA is involved in PCa survival and that siPCA3 specifically targeted PCA3 transcripts, since growth rates were not modified in cells that do not express PCA3. Considering that PCA3 is specifically expressed in prostate tissues , it is possible that PCA3 silencing would be an interesting therapeutic approach, especially to inhibit PCa growth and progression, as has been proposed for other genes involved in PCa survival [33, 34]. The evidence presented in this study, for a potential modulating role of PCA3 in the AR signaling, highlights the importance of this transcript as a potential target for treatment. This may also be an interesting approach during PCa progression, especially when androgen resistance is developed . In addition, PCA3 expression at different PCa stages  further reinforces the notion that this ncRNA plays an essential role during PCa tumorigenesis and progression.
In order to further emphasize the role of PCA3 in modulating PCa cell survival, we also investigated the effect of PCA3 knockdown in an androgen-independent PCa cell line, which simulates a more aggressive PCa disease. However, conflicting results have been reported regarding the androgen independence of these cells and AR protein expression in the PC3 and DU145 cell lines, as well in castration-resistant prostate cancer tumors [36, 37]. Although the majority of human prostate-cancer cell lines are reported to be AR-negative [38, 39], several studies have indicated that the DU-145 and PC-3 prostate-cancer cell lines express detectable levels of the AR mRNA [40–44]. For this reason, the exact role of AR in PC3 cells is still controversial . Although PC3 cells showed lower PCA3 expression than did the LNCaP cells, we asked whether PCA3 also regulates cell survival in these cells. Considering the heterogeneity of PCa tumors regarding gene expression profiling and tumor progression behavior, we then wanted to evaluate survival features after PCA3 knockdown even in cell lines that express low PCA3 levels, and that in principle are androgen-independent. Since PC3 is a well-established cell-line model representing a more aggressive stage of prostate cancer tumors, our data provide additional evidence for a role of PCA3 in modulating PCa cell survival. We also wanted to investigate how these androgen-independent cells would respond to PCA3 knockdown. Our results demonstrated that PCA3 knockdown, as demonstrated by the significant decrease in PCA3 levels after transfecting PC3 cells with the siPCA3 molecule, increased the proportion of cells with pyknotic nuclei (compared to PC3 transfected with siScr control), which was indicative of a larger number of cells undergoing apoptosis, further evidencing the role of PCA3 in modulating PCa cell survival, even in an androgen-independent cell-line model that expresses lower PCA3 levels compared to LNCaP cells. However, how PCA3 can modulate PCa cell survival in these cells and whether it is mediated by different mechanisms from those observed in siPCA3-LNCaP transfected cells, should be further investigated.
Due to the increased PCA3 expression in androgen-responsive cells compared with androgen-insensitive cells , and because AR signaling is an important pathway controlling PCa survival, we tested whether PCA3 expression was modulated by the androgen-active metabolite DHT and whether this expression pattern involved the activated AR. Upregulation of PCA3 expression in response to LNCaP stimulation with DHT was significantly counteracted by the AR antagonist flutamide, indicating that PCA3 expression was induced by the activated AR. AR activation was further confirmed by the observation that LNCaP cells stimulated with DHT also showed AR transcriptional activity. Consistently, all of the AR target genes tested that contain canonical AR response elements (AREs) in their promoter sequences, were upregulated upon DHT treatment. Although eight of the genes showed at least a 1.5-fold increase after AR activation, only two of them showed a significant increase in their expression levels. Interestingly, PCA3 upregulation upon DHT treatment has been observed previously [11, 20], but no study has demonstrated the involvement of activated AR in PCA3 expression by using AR antagonists. Although our data also suggest that PCA3 is an androgen-responsive gene, the precise molecular mechanism by which PCA3 expression responds to this activation is still unknown. One hypothesis is that activated AR can directly activate the PCA3 promoter, as has been demonstrated for the miR-101 and miR-21 regulatory regions [45, 46], which are also modulated by the activated AR. However, no consensus AREs have been identified in the 500-bp PCA3 promoter region . We further screened for consensus ARE elements in the entire PCA3 genomic region at the 5 Kb region upstream from the PCA3 transcription start site, and have so far identified no canonical element (data not shown). Nevertheless, we cannot exclude the possibility that other, non-canonical ARE elements could also promote AR binding and directly activate PCA3 expression, as has been previously described for other genes modulated by the AR activation [48, 49]. PCA3-upregulated expression in response to DHT treatment could also be a result of activated AR binding to the regulatory regions of other AR-responsive genes, which in turn could induce PCA3 expression. Further experiments should investigate direct AR binding to different PCA3 genomic regions, in order to answer these open questions.
Our data support a pro-survival role for PCA3, since its downregulation, in addition to inhibiting PCa survival, decreased the expression of AR target genes, most of them typically involved in androgen-dependent cell growth. The close association between the involvement of PCA3 in PCa cell survival and the modulation of the expression of AR should be further investigated, in an attempt to elucidate how PCA3 expression levels can regulate AR signaling and target genes. Furthermore, PCA3 knockdown inhibited the expression of all AR target genes tested, even under DHT treatment, indicating that the final effect of PCA3 downregulation may be stronger than the effect of DHT stimulation in modulating the expression of AR target genes. On the other hand, Akt and ERK phosphorylation levels remained unchanged in siPCA3-transfected cells, indicating that alternative pathways able to activate AR irrespective of ligand activation  were not altered after PCA3 knockdown. Taken together, these data suggest that the role of PCA3 in modulating the expression of these AR target genes may function downstream from AR activation.
As an approach to investigate the signal by which PCA3 controls PCa cell survival, we analyzed the transcript expression of PSA, AR, TMPRSS2, NDRG1, GREB1, FGF8, CDK1, CDK2, and PMEPA1 genes, all of which have key roles in PCa growth and progression, and are classical AR target genes . Also highly regulated by androgens, fibroblast growth factor 8 (FGF8) , cyclin-dependent kinase 1 (CDK1), cyclin-dependent kinase 2 (CDK2) , and the gene regulated in breast cancer 1 (GREB1)  gene products have classical stimulating roles in prostate growth and proliferation. Conversely, the PMEPA1 gene, although a direct transcriptional target of the AR, has been described as a negative regulator of cell growth in the prostate epithelium, as well as negatively regulating AR protein levels in different cell-culture models [55, 56]. We also observed that the AR transcription level was downregulated after PCA3 knockdown. These results accord with previously published data, which demonstrated that the AR gene is transcriptionally regulated by AR through binding to AR regulatory elements (autoregulation). However, differently from the other AR-responsive genes tested here, the ARE elements required for this process have not been found in the AR promoter or in the 5'-flanking region, but rather in AR coding sequences . Since DHT treatment also upregulated AR transcript expression, it is possible that PCA3 could also modulate the transcriptional activity of AR, as has been postulated for other ncRNAs [58, 59]. Although it shows both oncogenic and tumor-suppressor roles, PSA also has key roles in promoting tumor progression and metastasis . NDRG1-ERG fusions, encoding a chimeric protein, are also androgen-regulated and correspond to one of the recurrent erythroblast transformation-specific rearrangements found in PCa samples. Presumably, NDRG1 participates in angiogenesis, metastasis, and mechanisms leading to anti-cancer drug resistance . TMPRSS2, another component of these typical androgen-regulated PCa translocations, is highly expressed in PCa cells, contributing to prostate tumorigenesis [62, 63]. The observed downregulation of some of these tested AR target genes after PCA3 silencing could be part of the events related to inhibition of PCa cell survival, especially because some of these genes are classical positive modulators of PCa progression. Considering that some of the AR target genes tested were downregulated by PCA3 knockdown, we hypothesize that PCA3, similarly to other ncRNAs, could be a key modulator of the AR signaling pathway, as has been observed for other ncRNA gene products in controlling key pathways . However, highly aggressive metastatic PCa cell lines that are termed AR-insensitive, such as PC3 and DU145, only express PCA3 at very low levels, indicating that PCA3 may play a role (in combination with other factors) by promoting the transition from an AR-dependent to a hormone-refractory disease. Data concerning the expression of AR in PC3 and DU145 cells are contradictory. Although they are classically termed androgen non-responsive and AR-negative cells, the expression of AR transcript and protein has been clearly demonstrated in PC3 and DU145 cells, as has been AR nuclear translocation, but not transcriptional activity . Because PCA3 silencing in PC3 cells also inhibited PC3 viability, it is possible that PCA3 may be involved in AR signaling at alternative steps that are able to control the expression of AR-responsive genes in an androgen-independent manner, as described for other gene targets . Notably, PC3 cells, although they expressed lower PCA3 levels than LNCaP cells, showed greater inhibition of cell survival after PCA3 silencing. Possibly, due to lower PCA3 levels in PC3 cells, silencing of this transcript was more effective, with a stronger negative effect on cell survival.
Based on these data, we suggest that PCA3 probably behaves as a modulator of the expression of the AR target gene, although the underlying mechanism by which PCA3 regulates the expression of these AR target genes remains elusive. One hypothesis is that PCA3 long ncRNA or its putative processed RNA products could directly control the transcription of AR-regulated genes, as has been reported for other ncRNAs [59, 66]. It is also possible that PCA3 transcripts could modulate the transcriptional activity of AR-regulated genes by controlling the interaction and/or expression of multiple AR co-regulatory proteins, which facilitate the formation of an active transcription complex to activate transcription of AR target genes, similarly to what has been described for the Steroid Receptor Coactivator (SRA), a non-coding RNA that confers functional specificity during transcriptional activation . Inactive or under-represented co-activators could not evoke AR binding to their cognate binding sites on AR-responsive genes, promoting the downregulation of these genes . This possibility accords with our data regarding the downregulation of the tested AR target genes after PCA3 knockdown, indicating that a key positive AR signal was disrupted. Furthermore, the downregulation of these AR target genes occurred even in LNCaP cells that were not DHT-activated, indicating that the role of PCA3 in modulating the transcription of these genes is independent of AR ligand activation. Our current studies highlight the need for further investigation to elucidate the exact mechanisms by which PCA3 controls PCa cell survival and how it modulates AR signaling and cell growth.
The subcellular location of ncRNAs has also provided significant insights into their functions. We found in the present study that PCA3 is mainly expressed in the nuclear and microsomal subcellular fractions in LNCaP prostate epithelial cells, but not in prostate-tumor stromal cells. Contradictory findings regarding PCA3 expression have indicated that this transcript was restricted to the nuclear  or cytoplasmic  cell compartments. In order to clarify this, we used a cell-fractionated PCA3 transcript expression analysis approach, in which most PCA3 transcripts were located in nuclear fractions, and to a lesser extent in microsomal compartments, which classically contain ribosomal and vesicle particles. Other long non-coding RNAs have also been found in nuclear fractions, especially accumulated in specific nuclear bodies . Other reports have also identified ncRNAs at nuclear speckles, where these molecules could regulate alternative splicing by modulating the activity of splicing regulatory proteins . Other roles described for ncRNAs located in nuclear compartments, including their roles in modulating the activity of transcription factors and in chromatin remodeling, have been reviewed elsewhere . Co-purification with ribosomes has also been described for other long ncRNAs, such as HULC, implying that they play a role in translational control  and in nonsense-mediated mRNA decay (NMD) . The presence of PCA3 transcripts in both nuclear and microsomal compartments may indicate that PCA3 could perform its main roles in both cell niches. These data may help to direct our future approaches in attempting to identify the exact roles of PCA3 in modulating the transcription of AR target genes. Because PCA3 is not expressed in stromal cells, it seems that its role in controlling PCa cell survival may be restricted to prostate epithelial cells and possibly their AR signaling pathway.