The human genome is mostly transcribed, yielding a rich repository of noncoding transcripts that are involved in a myriad of biological processes including cancer. However, how many noncoding transcripts such as long noncoding RNAs (lncRNA) function in cancer is still unclear. This study identified a novel set of clinically relevant androgen-regulated lncRNAs in prostate cancer. Among this group, LINC00844 was demonstrated to be a direct androgen-regulated target that is actively transcribed in androgen receptor (AR)–dependent prostate cancer cells. The expression of LINC00844 is higher in normal prostate compared with malignant and metastatic prostate cancer clinical specimens, and patients with low expression had a poor prognosis and significantly increased biochemical recurrence, suggesting LINC00844 functions in suppressing tumor progression and metastasis. Indeed, in vitro loss-of-function studies revealed that LINC00844 prevents prostate cancer cell migration and invasion. Moreover, findings from gene expression profiling analysis indicated that LINC00844 functions in trans, affecting global androgen-regulated gene transcription. Mechanistic evidence reveals that LINC00844 is important in facilitating AR binding to the chromatin. Finally, LINC00844 mediates its phenotypic effects in part by activating the expression of NDRG1, a crucial cancer metastasis suppressor. Collectively, LINC00844 is a novel coregulator of AR that plays a central role in the androgen transcriptional network and the development and progression of prostate cancer.

Implications:

This study highlights the function of the lncRNA, LINC00844, in regulating global AR-regulated genes in prostate cancer by modulating AR binding to chromatin.

Prostate cancer is the most common noncutaneous cancer affecting men worldwide (1). Androgen receptor (AR) is a hormone-regulated transcription factor that occupies a central position in both normal prostate development and its carcinogenesis (2). AR mediates its biological effect by controlling the transcription of downstream target genes (3, 4). Therefore, antiandrogens which inhibit the transcriptional activity of AR have become a part of standard care in prostate cancer therapy. Regardless of the favorable initial response, a majority of the patients will relapse into a more lethal and incurable form called metastatic castration-resistant prostate cancer (mCRPC; refs. 5–7). Extensive research over the years has established that the reactivation or aberrant activation of the AR signaling pathway remains a critical event in the development of mCRPC (8, 9). Hence, it is imperative to elucidate the underlying molecular mechanisms of AR-mediated transcriptional regulation in order to develop better diagnostic markers and improved therapies for the management of advanced prostate cancer.

The discovery that the majority of the human genome is transcribed yielding a rich repository of noncoding RNAs (ncRNA) has generated widespread interest in understanding their functional roles (10). NcRNAs that are greater than 200 bp in length are arbitrarily classified as long noncoding RNA (lncRNA) and have been implicated in a myriad of biological processes including cancer (11, 12). With respect to prostate cancer, a number of lncRNAs have emerged as prostate cancer biomarkers such as Prostate Cancer Antigen 3 (PCA3; refs. 12, 13) and SChLAP1 (14). In gene regulation, SRA (steroid receptor RNA activator) was the first lncRNA that was identified as a transcriptional coregulator for nuclear hormone receptors including AR (15, 16). Since then, additional lncRNAs have been implicated in prostate cancer biology by directly regulating the transcriptional activity of AR. For example, Rosenfeld and colleagues showed two lncRNAs, Prostate Cancer Gene Expression Marker-1 (PCGEM1) and Prostate Cancer Noncoding RNA1 (PRNCR1), which are important for enhancing androgen-dependent transcription by promoting the chromatin looping of AR-bound enhancers to their target gene promoters (17). In addition, CTBP1-AS, an AR-stimulated antisense lncRNA, was found to repress the expression of CTBP1 in a cis-regulatory manner as well as tumor-suppressive genes in trans by recruiting the histone deacetylase corepressor complex via the PTB-associated splicing factor (18). Similarly, HOTAIR was reported to physically interact with AR in order to prevent the E3 ubiquitin ligase-mediated degradation of the receptor and thus facilitating an androgen-independent AR transcriptional program and the promotion of mCRPC transformation (19). Although we are beginning to understand the role of a few lncRNAs in prostate cancer, our current knowledge is still limited and a vast repository of lncRNAs have yet to be identified or extensively investigated with respect to AR-mediated transcriptional regulation and prostate cancer biology.

In this study, we have used a combination of bioinformatic analysis and molecular approaches to identify and functionally characterize a novel AR-regulated lncRNA in prostate cancer called Long Intergenic Long Non-Coding RNA 844 (LINC00844). We demonstrate that LINC00844 is clinically significant and potentially plays an important role in metastatic transformation of prostate cancer and provides evidence to show that LINC00844 inhibits migration and invasion of prostate cancer cells without altering the rate of cell proliferation. In addition, we establish that LINC00844 is a critical component of the AR transcriptional network and enhances the global expression of androgen-regulated genes by promoting AR recruitment to chromatin. Finally, our study also reveals that LINC00844 mediates its phenotypic effects in part by regulating the DHT-mediated activation of NDRG1, a known metastatic suppressor gene.

Reagents and antibodies

Dihydrotestosterone (DHT) was purchased from Tokyo Chemical Industry, and MDV-3100 and Biclutamimide were purchased form TCI Chemicals. The following antibodies were used for chromatin immunoprecipitation (ChIP) and Western blot analyses: anti-AR (sc-815X, sc-H280, and sc-816) from SCBT, anti-NDRG1 (ab-37897) from Abcam, and anti–β-actin and anti-GAPDH from Sigma.

Cell culture

LNCaP, VCaP, and 22Rv1 were purchased from the ATCC. LNCaP and 22Rv1 cells were maintained in RPMI medium 1640 (RPMI; Gibco) supplemented with 10% FBS (Gibco). VCaP cells were grown in DMEM (Gibco) supplemented with 10% FBS, 1 mmol/L sodium pyruvate, and 0.08% sodium bicarbonate. All cells were routinely tested for mycoplasma contamination (once in every 3 months) by MycoAlert (mycoplasma detection kit by Lonza). In addition, VCaP cells were authenticated by DiagCor (Hong Kong). Cells with early passage numbers (<30 for LNCaP and 22Rv1 and <20 for VCaP) were used for all experiments. Prior to Ethanol (ETOH) or DHT treatment, LNCaP and 22Rv1 cells were deprived of hormones for 3 days by growing them in phenol red–free RPMI containing 10% charcoal–dextran-treated FBS (HyClone), whereas VCaP cells were grown for at least a day in phenol red–free DMEM containing 10% charcoal–dextran-treated FBS.

Plasmids and cell transfection

Full-length LINC00844 was synthesized by IDT and cloned into the pcDNA3.1 vector at EcoRI and ApaI restriction sites. A pCMV vector expressing full-length human NDRG1 was purchased from Sino Biologicals. LNCaP cells were transiently transfected with plasmids using Lipofectamine 3000 (Invitrogen) according to the manufacturer's instructions.

siRNA studies

LNCaP and VCaP cells cultured for 24 hours in phenol red–free medium were transfected with 5 nmol/L of Dicer-substrate siRNAs from IDT, using Lipofectamine RNAiMAX transfection reagent (Invitrogen) according to the manufacturer's protocol. After 24 hours of incubation, the cells were transfected again in a similar manner with 5 nmol/L siRNA. Forty-eight hours after the second round of transfection, cells were treated with ETOH or 10 nmol/L DHT for another 12 hours before harvesting for real-time reverse transcription-qPCR (RT-qPCR), microarray, and Western blot analyses. For every target, 2 separate siRNAs were paired with the control nontargeting siRNA (NC). siRNA and real-time qPCR primer sequences are listed in Supplementary Table S4 in the Supplementary Material. Gene expression profiles for knockdown studies were obtained from at least three independent experiments.

ChIP-qPCR and ChIP-seq

ChIP assays were performed with minor modifications to the protocol described by Carroll and colleagues (20). For knockdown ChIP assays, 3 million LNCaP cells or 5 million VCaP cells were transfected twice with siRNA targeting LINC00844 or control. Forty-eight hours after transfection and 2 hours of 100 nmol/L DHT/ETOH treatment, cells were cross-linked with 1% formaldehyde and harvested. Cells were lysed, nuclei extracted, and DNA was sonicated using a Biorupter (Diagenode) before overnight immunoprecipitation with AR antibody conjugated to protein G magnetic beads (Dynabeads; Invitrogen). The beads were thoroughly washed and reverse cross-linked at 65°C along with 5% input before isolating DNA using a PCR-purification kit (Qiagen). AR enrichment was quantified by real-time qPCR using primer sequences listed in Supplementary Table S3 and represented as percentage of input.

For the high-throughput sequencing of the ChIP DNA (ChIP-seq), DNA libraries were constructed using the NEBNext Ultra DNA Library Prep Kit for Illumina (NEB; catalogue # E7370L) according to the manufacturer's protocol. The libraries were amplified for 12 cycles by PCR and purified using AMPure XP beads (Beckman Coulter; catalogue # A63880). The barcoded libraries were mixed in equimolar ratio and sequenced using the Illumina HiSeq-2000. Sequencing reads were mapped with BWA (21) to the indexed reference genome UCSC hg19. The obtained BAM files (as replicates) were subjected for peak calling using MAC2 (22) using default parameters. Specific and overlapping peaks between different test conditions were identified using the HOMER package (23). Heatmaps of the read density and the average binding profiles around the center of peaks (±1 kb) were produced using the R heatmap package. Genome coverage bedgraph files were converted to bigwig format using the bedGraphToBigWig tool and loaded onto the IGV browser to obtain screenshots. The raw sequencing data can be found at the GEO with the accession number GSE108704.

Migration and invasion assay

Migration and invasion assays were carried out using LNCaP cells transfected with siRNA and/or plasmids and seeded in Boyden's Transwell Chamber inserts (uncoated for migration and 1:20 diluted Matrigel coating for invasion) in RPMI media containing 5 nmol/L of siRNA in 100 μL of RPMI with 1% FBS. Note that 600 μL of RPMI with 10% FBS containing ETOH or DHT in the lower chamber acted as chemoattractant. After 48 hours, cells were fixed with 4% formalin and stained with crystal violet, and cells migrated to the other side of the insert were counted using an inverted bright-field microscope at 20X magnification.

Alamar blue cell proliferation assay

LNCaP cells transfected with siRNA for 24 hours were seeded in multiple 96-well plates (as triplicates) at a concentration of 10,000 cells/well followed by a second round of siRNA transfection. To assess for proliferation, cells were incubated with ALAMAR blue reagent for 3 hours, and the fluorescence was measured by the Spectramax with 560EX nm/590EM nm filter settings. The results from three independent experiments are presented as mean ± SEM.

Microarray expression profiling

Microarray analysis was performed as described previously (24). Briefly, purified total RNAs from three independent biological replicates were converted to cRNA using the TargetAmp-Nano Labeling Kit for Illumina Expression BeadChip (Epibio) according to the manufacturer's instructions. cRNA was hybridized onto HumanHT-12 v4 Expression BeadChips (Illumina). The BeadChips were scanned with the iScanner (Illumina), and the image data were processed using GenomeStudio. The gene expression data were analyzed using Lumi-based tools from www.arrayanalysis.org (25). All the microarray data have been deposited at the GEO database with the accession number, GSE109336.

LINC00844 is a novel androgen-upregulated prostate cancer–associated lncRNA

To identify and characterize novel clinically relevant lncRNAs that are associated with prostate carcinogenesis, we processed and analyzed RNA sequencing (RNA-seq) expression data from two independent prostate cancer studies (Fig. 1A; ref. 26, 27). We first used DESeq2 (28) to obtain a list of differentially expressed transcripts (malignant vs. benign or normal tissue) from each cohort and then used metaRNAseq (29) to perform meta-analysis on the two lists. The combined P value from multiple testing was further corrected by the Benjamini and Hochberg method, and using an FDR < 0.01, we identified 370 clinically relevant lncRNAs (Supplementary Table S1). Because our lab is interested in androgen-mediated transcription in prostate cancer cells, we decided to focus on the androgen-regulated lncRNAs within this group. To do this, we first defined a set of androgen-regulated lncRNAs with a minimum of 1.5-fold expression change from the RNA-seq data of LNCaP and VCaP cells that have been treated with or without DHT (30). Next, we overlapped this list of androgen-regulated lncRNAs with the differentially regulated list of lncRNAs from the prostate cancer studies. From this, we found 38 lncRNAs that are both clinically relevant and under androgen transcriptional regulation (Supplementary Table S2). We further ranked this list based on their abundance in LNCaP and VCaP cells and validated the top 15 candidates by qPCR (Fig. 1B). Among the top 15 candidates, we identified PCAT1 (26) and RP11-279F6.2 (DRAIC; ref. 31), two lncRNAs that have been previously shown to play an important role in the androgen signaling pathway. As for the other candidates, they are novel androgen-regulated lncRNAs and have yet to be functionally characterized in prostate cancer. We selected several of the more abundantly expressed lncRNAs in LNCaP and VCaP cells for further investigation. In this study, we present our findings for LINC00844.

Figure 1.

Identification and characterization of LINC00844 in prostate cancer cells. A, Workflow showing the steps in the discovery and identification of LINC00844. B, Heatmap representing the expression of 15 lncRNAs in hormone-starved LNCaP and VCaP cells stimulated with 10 nmol/L of DHT for the indicated amount of time. C, Genome browser view of RNA-seq and ChIP-seq events around LINC00844 with the current annotation from Gencode. CAGE peaks from normal prostate tissues were obtained from the FANTOM database. H3K27ac ChIP-seq from LNCaP cells and RNA-seq from LNCaP and VCaP cells treated with ETOH or DHT were obtained from the GEO database (GSE51621 and GSE82223). RNA-seq data in normal and prostate cancer tissues were obtained from SRA database (ERP000550).

Figure 1.

Identification and characterization of LINC00844 in prostate cancer cells. A, Workflow showing the steps in the discovery and identification of LINC00844. B, Heatmap representing the expression of 15 lncRNAs in hormone-starved LNCaP and VCaP cells stimulated with 10 nmol/L of DHT for the indicated amount of time. C, Genome browser view of RNA-seq and ChIP-seq events around LINC00844 with the current annotation from Gencode. CAGE peaks from normal prostate tissues were obtained from the FANTOM database. H3K27ac ChIP-seq from LNCaP cells and RNA-seq from LNCaP and VCaP cells treated with ETOH or DHT were obtained from the GEO database (GSE51621 and GSE82223). RNA-seq data in normal and prostate cancer tissues were obtained from SRA database (ERP000550).

Close modal

We began our characterization of LINC00844 by determining its coding potential to confirm whether it is indeed a noncoding transcript. We did this by examining the DNA sequence of LINC00844 using the Coding Potential Calculator (http://cpc.cbi.pku.edu.cn/; ref. 32). A score of −1.04818 was obtained for LINC00844 which suggests that the open reading frame of this lncRNA lacks the potential to code for a protein. Similarly, when we examined the coding potential of HOTAIR, a well-known lncRNA, we obtained a score of −1.19011. In contrast, we obtained a score of 12.246 for GAPDH.

Next, we examined the basic features of the LINC00844 transcript. LINC00844 is an annotated lncRNA (ENSG00000237949, Entrez:100507008) transcribed from an intergenic region on chromosome 10q21.1. The basic Gencode (www.gencodegenes.org) annotation maps it as a 477 bp transcript consisting of two exons transcribed from a region of 1.98 kb. To characterize LINC00844 further, we interrogated several publicly available datasets including the recently published Cap Analysis of Gene Expression (CAGE; ref. 33) on prostate tissues, H3K27 acetylation (H3K27ac) ChIP-seq in LNCaP cells (34), and the RNA-seq expression data from LNCaP and VCaP cells (35), as well as prostate cancer patient samples (27). As shown in Fig. 1C, there is a strong H3K27ac peak at the TSS of LINC00844 (marked by the CAGE peak) which increases upon DHT stimulation. Moreover, there is a concomitant increase in the transcript level of LINC00844 in both LNCaP and VCaP cells. In patient samples, LINC00844 is expressed in normal and prostate cancer. Taken together, our results suggest that LINC00844 is a novel prostate cancer–associated lncRNA that is expressed in prostate tissues and is regulated by androgen in prostate cancer cells.

LINC00844 is a direct AR target

Next, we assessed the expression and regulation of LINC00844 in different prostate cancer cell lines. LINC00844 transcript level is highest in the AR-dependent cell lines, LNCaP and VCaP, whereas it is minimally expressed in the AR-sensitive cell line, 22Rv, and undetectable in the AR-independent cell line, PC3 (Fig. 2A). With respect to its regulation, LINC00844 was stimulated by DHT in a time- and concentration-dependent manner similar to KLK3/PSA, the model AR-regulated gene (Fig. 2B and C; Supplementary Fig. S1A and S1B). In contrast, DHT did not significantly affect the expression of LINC00844 in 22Rv1 cells (Supplementary Fig. S1C). To further validate the AR-mediated regulation of LINC00844, we treated LNCaP and VCaP cells prior to DHT stimulation with either antiandrogens (Bicalutamide or MDV-3100) or siRNA against AR. In both conditions, we found DHT-mediated stimulation of LINC00844 was greatly reduced (Fig. 2D and E).

Figure 2.

LINC00844 expression is directly regulated by AR. The relative expression of LINC00844 measured by qPCR in (A) different prostate cancer cells maintained in normal media, LNCaP, and VCaP cells treated with (B) 10 nmol/L of DHT for the indicated time point, or (C) different concentration of DHT for 12 hours, or (D) pretreated with 10 μmol/L of Bicalutamide (Bic) or MDV-3100 (MDV) for 12 hours followed by 10 nmol/L DHT stimulation, or (E) transfected with siAR followed by 10 nmol/L DHT treatment for 12 hours. F, Screenshot of AR ChIP-seq peaks mapped to the promoter of LINC00844 from LNCaP and VCaP cells treated with 100 nmol/L of ETOH or DHT for 2 hours and prostate cancer patient samples from GEO dataset (GSE70079). The bottom of the screenshot shows a zoomed-in region of the AR-binding site upstream of the LINC00844 TSS containing a canonical AR-binding motif. G, AR ChIP-qPCR of the region spanning the LINC00844 promoter in LNCaP and VCaP cells pretreated with 100 μmol/L of Bicalutamide (Bic) and MDV3100 (MDV) followed by 100 nmol/L DHT stimulation.

Figure 2.

LINC00844 expression is directly regulated by AR. The relative expression of LINC00844 measured by qPCR in (A) different prostate cancer cells maintained in normal media, LNCaP, and VCaP cells treated with (B) 10 nmol/L of DHT for the indicated time point, or (C) different concentration of DHT for 12 hours, or (D) pretreated with 10 μmol/L of Bicalutamide (Bic) or MDV-3100 (MDV) for 12 hours followed by 10 nmol/L DHT stimulation, or (E) transfected with siAR followed by 10 nmol/L DHT treatment for 12 hours. F, Screenshot of AR ChIP-seq peaks mapped to the promoter of LINC00844 from LNCaP and VCaP cells treated with 100 nmol/L of ETOH or DHT for 2 hours and prostate cancer patient samples from GEO dataset (GSE70079). The bottom of the screenshot shows a zoomed-in region of the AR-binding site upstream of the LINC00844 TSS containing a canonical AR-binding motif. G, AR ChIP-qPCR of the region spanning the LINC00844 promoter in LNCaP and VCaP cells pretreated with 100 μmol/L of Bicalutamide (Bic) and MDV3100 (MDV) followed by 100 nmol/L DHT stimulation.

Close modal

To determine if AR directly regulates the expression of LINC00844 in prostate cancer by binding in the upstream or downstream regulatory region of the transcript, we interrogated AR ChIP-Seq datasets from cell lines (36, 37) and patient samples (38). As shown in Fig. 2F, there is an ARBS located at the TSS of LINC00844 in LNCaP and VCaP cells as well as in patient samples. In addition, DHT enhanced the recruitment of AR to this site similar to other ARBS associated with known AR-regulated genes (Supplementary Fig. S1D). We verified the binding of AR to the TSS of LINC00844 by ChIP-qPCR and showed that recruitment was abolished by antiandrogens in both LNCaP and VCaP cells (Fig. 2G). Upon examination of the DNA sequence of the ARBS, we identified the presence of a canonical AR response element (ARE) motif (Fig. 2F), suggesting that AR binds directly to this site. Taken together, our results suggest that LINC00844 is a direct AR-regulated lncRNA.

LINC00844 is suppressed in malignant and metastatic prostate cancer

To begin understanding the clinical significance of LINC00844 in prostate cancer, we examined the expression profile of the transcript in patient samples. The expression level of LINC00844 is significantly lower in malignant tumor samples compared with their matched controls or benign samples in four prostate cancer cohorts (26, 27, 39), including the two cohorts that we had utilized for our initial characterization of lncRNAs (Fig. 3A–D). This result suggests a potentially important tumor-suppressive functional role for LINC00844. Notably, two out of the four cohorts which also contain the expression information for metastatic prostate cancer samples showed the levels of LINC00844 transcript in these metastatic samples are further suppressed (Fig. 3B and D). Taken together, our results show LINC00844 is the first AR-regulated lncRNA to demonstrate profound association with both malignant and metastatic prostate cancer.

Figure 3.

LINC00844 is suppressed in malignant and metastatic prostate cancer. Box plot of LINC00844 expression presented as log2(FPKM) in published cancer cohorts, (A) Ren and colleagues, (B) Prensner and colleagues, (C) TCGA-PRAD, and (D) MiTranscriptome. LINC00844 expression in TCGA samples segregated based on clinical parameters including (E, left plot) Gleason score, (F, left plot) pathologic TNM staging, (G) biochemical recurrence, and (H) spread of tumor to neighboring lymph node. Linear-regression analysis was performed to calculate the correlation between LINC00844 (E, right plot) Gleason score and (F, right plot) pathologic TNM staging. I, Kaplan–Meier analysis of prostate cancer outcome. Three hundred eighty-four prostate cancer samples from the TCGA dataset were separated as upper 3 quarters and lower 1 quarter based on the expression of LINC00844 and analyzed for period of biochemical recurrence-free survival.

Figure 3.

LINC00844 is suppressed in malignant and metastatic prostate cancer. Box plot of LINC00844 expression presented as log2(FPKM) in published cancer cohorts, (A) Ren and colleagues, (B) Prensner and colleagues, (C) TCGA-PRAD, and (D) MiTranscriptome. LINC00844 expression in TCGA samples segregated based on clinical parameters including (E, left plot) Gleason score, (F, left plot) pathologic TNM staging, (G) biochemical recurrence, and (H) spread of tumor to neighboring lymph node. Linear-regression analysis was performed to calculate the correlation between LINC00844 (E, right plot) Gleason score and (F, right plot) pathologic TNM staging. I, Kaplan–Meier analysis of prostate cancer outcome. Three hundred eighty-four prostate cancer samples from the TCGA dataset were separated as upper 3 quarters and lower 1 quarter based on the expression of LINC00844 and analyzed for period of biochemical recurrence-free survival.

Close modal

We further examined the association of LINC00844 expression level with additional clinical parameters such as Gleason score, pathologic tumor–node–metastasis (TNM) stage, spread of cancer to lymph nodes, and biochemical recurrence. LINC00844 expression level exhibits a minor but significant negative correlation with Gleason score (i.e., tumors with higher Gleason score which denotes more aggressive form of the cancer show the least expression for LINC00844) and pathologic TNM stage of the tumor (Fig. 3E and F). Similarly, tumors with lower expression of LINC00844 not only have higher rate of biochemical recurrence but also have increased spread of the tumors to the neighboring or distant lymph nodes (Fig. 3G and H). Finally, we examined the expression level of LINC00844 with respect to prostate cancer patient recurrence-free survival. As expected, patients with higher expression of LINC00844 had a significantly lower recurrence-free survival (log-rank P < 0.02; Fig. 3I; Supplementary Fig. S2A and S2B). Taken together, the clinical expression data of LINC00844 suggest it may be involved in not only prostate carcinogenesis but also metastatic transformation.

LINC00844 inhibits cell migration and invasion

Because the above clinical observations indicate that LINC00844 may play an important role in metastatic transformation of malignant prostate cancer, we decided to test this out by examining the effect of LINC00844 knockdown and overexpression on LNCaP cell migration and invasion using the Boyden's chamber assay. In LNCaP cells that were cultured in normal growth media or after androgen stimulation, knockdown of LINC00844 (Fig. 4A, left) significantly increased both the migration (Fig. 4B; Supplementary Fig. S3) and invasion (Fig. 4C; Supplementary Fig. S4) of the cells, whereas ectopic expression (Fig. 4A, right) had the opposite effect on both processes (Fig. 4D and E; Supplementary Fig. S5A and S5B). Interestingly, LINC00844 also had a significant effect on basal migration and invasion levels, indicating that LINC00844 potentially may also act through AR-independent pathways. Next, we determined if the observed changes in cell migration and invasion were due to an increase in cell proliferation. As shown in Fig. 4F, we did not observe any significant increase in cell proliferation after LINC00844 depletion. Although many reported lncRNAs in prostate cancer are known to affect cell proliferation and migration (35, 40), our results show LINC00844 affects exclusively prostate cancer cell migration and invasion, without affecting cell proliferation. Collectively, our results suggest that LINC00844 (i) is critical for preventing a metastatic phenotype in prostate cancer cells and (ii) mediates its effects in part through the AR signaling pathway.

Figure 4.

LINC00844 inhibits prostate cancer cell migration and invasion. A, Bar graph showing the expression of LINC00844 in LNCaP cells transfected with either siLINC00844 or pcDNA-LINC000844. LNCaP cells were seeded in the top chamber of the Boyden's transwell support, and after 48 hours, the cells were fixed and counted. Graph representing the number of LNCaP cells per high power field (20X objective) migrated across transwell support with 8 μm pores. Transwell supports were left uncoated for (B and D) migration assay or coated with Matrigel in 1:20 dilution for (C and E) invasion assay. Error bars show the SEM of cells counted in 5 high power fields of view in 3 independent experiments. F, Relative cell proliferation measured by Alamar blue assay in LNCaP cells transfected with siLINC00844 or NC control. Normalized Alamar blue emission at 560EX nm/590EM nm measured in triplicates from 3 independent experiments. Data represent the mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Figure 4.

LINC00844 inhibits prostate cancer cell migration and invasion. A, Bar graph showing the expression of LINC00844 in LNCaP cells transfected with either siLINC00844 or pcDNA-LINC000844. LNCaP cells were seeded in the top chamber of the Boyden's transwell support, and after 48 hours, the cells were fixed and counted. Graph representing the number of LNCaP cells per high power field (20X objective) migrated across transwell support with 8 μm pores. Transwell supports were left uncoated for (B and D) migration assay or coated with Matrigel in 1:20 dilution for (C and E) invasion assay. Error bars show the SEM of cells counted in 5 high power fields of view in 3 independent experiments. F, Relative cell proliferation measured by Alamar blue assay in LNCaP cells transfected with siLINC00844 or NC control. Normalized Alamar blue emission at 560EX nm/590EM nm measured in triplicates from 3 independent experiments. Data represent the mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Close modal

LINC00844 is an integral component of the AR transcriptional network

To explore whether LINC00844 has a role in regulating AR signaling in prostate cancer, we performed microarray analysis on siNC- or siLINC00844-treated LNCaP cells before and after DHT stimulation. Overall, we detected 917 DHT-responsive genes (fold change ≥ 1.5 and P <0.01) with 529 up- and 388 downregulated genes (Fig. 5A). Among the 917 DHT-regulated genes, 520 genes (327 up and 193 down) were affected by LINC00844 knockdown (fold change ≥ 1.2). With respect to the DHT-upregulated genes that were affected, 316 genes were repressed, whereas 11 genes were increased. Moreover, this group of DHT-upregulated genes included well-known model genes such as KLK2, KLK3, FKBP5, and GREB1. As for the DHT-downregulated genes, 177 were upregulated, whereas 16 were further suppressed. We examined the expression of several model AR-regulated genes by qPCR and observed similar trends in both LNCaP and VCaP cells (Fig. 5C and D). In general, the depletion of LINC00844 suppressed the activation of androgen-regulated genes in both cell lines, although the effects were more profound in LNCaP cells. Thus, our findings suggest that LINC00844 is important regulator of androgen-dependent gene transcription.

Figure 5.

LINC00844 regulates a subset of AR target genes. A, Gene-expression profiling was performed in LNCaP cells transfected with NC or siLINC00844 with DHT or ETOH stimulation for 12 hours. Heatmap represents all the AR-regulated genes (NC-ETOH vs. NC-DHT, fold change ≥ 1.5, P < 0.01) and their corresponding expression in siLINC00844 condition. The fold change in expression is mentioned below. The AR-upregulated genes that were suppressed by ≥1.2-fold after LINC00844 knockdown are represented by orange asterisks. B, GO terms associated with biological processes enriched in LINC00844-dependent genes. The dotted line represents P < 0.05. The GO terms associated with cell migration and invasion are highlighted in green. The mRNA expression of several genes selected from the microarray list and identified to be regulated by LINC00844 was quantified by qPCR in (C) LNCaP and (D) VCaP cells. Results are shown as the mean fold change normalized to NC-ETOH ± SEM from at least three independent experiments.

Figure 5.

LINC00844 regulates a subset of AR target genes. A, Gene-expression profiling was performed in LNCaP cells transfected with NC or siLINC00844 with DHT or ETOH stimulation for 12 hours. Heatmap represents all the AR-regulated genes (NC-ETOH vs. NC-DHT, fold change ≥ 1.5, P < 0.01) and their corresponding expression in siLINC00844 condition. The fold change in expression is mentioned below. The AR-upregulated genes that were suppressed by ≥1.2-fold after LINC00844 knockdown are represented by orange asterisks. B, GO terms associated with biological processes enriched in LINC00844-dependent genes. The dotted line represents P < 0.05. The GO terms associated with cell migration and invasion are highlighted in green. The mRNA expression of several genes selected from the microarray list and identified to be regulated by LINC00844 was quantified by qPCR in (C) LNCaP and (D) VCaP cells. Results are shown as the mean fold change normalized to NC-ETOH ± SEM from at least three independent experiments.

Close modal

We next performed Gene Ontology (GO) analysis using DAVID Functional Annotation Bioinformatics Microarray Analysis (https://david.ncifcrf.gov; ref. 41) on the above androgen-regulated genes that we identified as dependent on LINC00844. In support of our functional studies, we found that androgen-regulated genes that are dependent on LINC00844 for expression are enriched in processes associated with cytoskeletal organization, extracellular matrix organization, microtubule cytoskeleton organization, cell adhesion, structural component of cytoskeleton, and anatomical structural morphogenesis, all of which are GO terms linked to cell migration or metastasis (Fig. 5B). Taken together, our results indicate that LINC00844 is an integral component of the AR transcriptional network and a decrease or loss of LINC00844 expression may promote malignant transformation of prostate cancer by suppressing the DHT-mediated activation of genes.

LINC00844 regulates AR binding to chromatin

Next, we focused on understanding the underlying mechanism on how LINC00844 might control the expression of androgen-regulated genes. Previously, our lab and others have demonstrated that AR collaborative factors such as NKX3-1 (36) and lncRNAs like PlncRNA (42, 43) act through a feed-forward mechanism to regulate AR mRNA and protein levels to influence the expression of downstream AR target genes. Thus, we examined if LINC00844 could also function in a similar fashion. As shown in Fig. 6A and B, knockdown of LINC00844 did not significantly alter the mRNA or protein level of AR in both LNCaP and VCaP cells, indicating that LINC00844 likely functions by directly or indirectly influencing the activity of AR, such as through chromatin binding.

Figure 6.

LINC00844 promotes AR association to the chromatin. LNCaP and VCaP cells treated with siLINC00844 or NC and 10 nmol/L of DHT or ETOH for 12 hours. A, The mRNA levels were quantified by qPCR, and the bar graph represents AR expression as fold change normalized to NC-ETOH ± SEM from at least three independent experiments. B, AR protein levels were analyzed by Western blot (left) and then quantified by Imagelab (Biorad). Bar graph (right plots) represents the mean normalized (to β-actin) band quantity ± SEM from at least three independent experiments. C and D, AR ChIP-qPCR was performed in LNCaP and VCaP cells transfected with siLINC00844 or NC and treated with 100 nmol/L of DHT for 2 hours. Graph represents AR enrichment at the ARBS within ± 50 kb from the TSS of the specified gene. All results represent the average of three individual experiments ± SEM. E, Venn diagram representing the common ARBS identified from 2 independent AR ChIP-seq experiments in LNCaP cells treated with siLINC00844 or NC and 100 nmol/L of DHT for 2 hours. F, A heatmap representing the sorted AR ChIP-seq signals corresponding to NC-DHT–specific regions in all the tested conditions. Signals are plotted in reference to the center of AR ChIP-seq cluster peak (−/+ 1 kb). G, A graph comparing the average ChIP-seq tag density around ± 1 kb from the AR peak center in NC-DHT compared with siLINC00844-DHT (solid lines) and NC-ETOH compared with siLINC00844-ETOH (dotted lines). H, Snapshots showing the AR ChIP-seq peak at the enhancer region of 4 model AR-regulated genes (KLK3, FKBP5, KLK2, and GREB1). Black arrow represents the ARBS with marked reduction in AR binding after LINC00844 knockdown. I, A table listing the GO terms associated with cell migration that are enriched in the ARBS specific to NC-DHT was calculated by the GREAT tool.

Figure 6.

LINC00844 promotes AR association to the chromatin. LNCaP and VCaP cells treated with siLINC00844 or NC and 10 nmol/L of DHT or ETOH for 12 hours. A, The mRNA levels were quantified by qPCR, and the bar graph represents AR expression as fold change normalized to NC-ETOH ± SEM from at least three independent experiments. B, AR protein levels were analyzed by Western blot (left) and then quantified by Imagelab (Biorad). Bar graph (right plots) represents the mean normalized (to β-actin) band quantity ± SEM from at least three independent experiments. C and D, AR ChIP-qPCR was performed in LNCaP and VCaP cells transfected with siLINC00844 or NC and treated with 100 nmol/L of DHT for 2 hours. Graph represents AR enrichment at the ARBS within ± 50 kb from the TSS of the specified gene. All results represent the average of three individual experiments ± SEM. E, Venn diagram representing the common ARBS identified from 2 independent AR ChIP-seq experiments in LNCaP cells treated with siLINC00844 or NC and 100 nmol/L of DHT for 2 hours. F, A heatmap representing the sorted AR ChIP-seq signals corresponding to NC-DHT–specific regions in all the tested conditions. Signals are plotted in reference to the center of AR ChIP-seq cluster peak (−/+ 1 kb). G, A graph comparing the average ChIP-seq tag density around ± 1 kb from the AR peak center in NC-DHT compared with siLINC00844-DHT (solid lines) and NC-ETOH compared with siLINC00844-ETOH (dotted lines). H, Snapshots showing the AR ChIP-seq peak at the enhancer region of 4 model AR-regulated genes (KLK3, FKBP5, KLK2, and GREB1). Black arrow represents the ARBS with marked reduction in AR binding after LINC00844 knockdown. I, A table listing the GO terms associated with cell migration that are enriched in the ARBS specific to NC-DHT was calculated by the GREAT tool.

Close modal

Hence, we next determined whether LINC00844 is important for the recruitment of AR to ARBS by performing ChIP assays. For this, we began by examining ARBS associated with AR-regulated genes including KLK2, KLK3, FKBP5, and GREB1, whose expression we have shown is dependent on LINC00844. We tested the recruitment of AR to these binding sites in LNCaP and VCaP cells that were treated with siLINC00844 or siNC. Interestingly, AR recruitment at these regions was markedly reduced by LINC00844 depletion (Fig. 6C and D). These results suggest that LINC00844 influences the expression of the AR-regulated transcriptome in part by facilitating the recruitment of AR to the chromatin.

To examine whether LINC00844 has a global effect on AR binding, we performed AR ChIP-seq on LNCaP cells that have been treated with or without siLINC00844. The experiment was performed in duplicates for increased confidence, and we considered only the common ARBS for downstream analysis. Overall, the depletion of LINC00844 resulted in a genome-wide reduction of a large number of ARBS (Fig. 6E; Supplementary Table S3). Specifically, 4,227 ARBS (∼30%) in the NC-DHT condition were lost after LINC00844 knockdown. Moreover, the ChIP-seq tag intensity of these 4,227 ARBS was markedly reduced after LINC00844 depletion (Fig. 6F–H; Supplementary Fig. S6A–S6F). Interestingly, AR binding at the LINC00844 promoter was also reduced after LINC00844 knockdown, suggesting a possible feed-forward regulation (Supplementary Fig. S6G). GREAT analysis of the ARBS that were affected by LINC00844 depletion showed many terms enriched for cell migration which is consistent with our initial finding that LINC00844 regulates the expression of genes associated with prostate cancer cell migration and invasion (Fig. 6I). Taken together, our results suggest that LINC00844 regulates global androgen-dependent transcription in trans by modulating the binding of AR to chromatin.

LINC00844 inhibits prostate cancer cell migration and invasion by upregulating the expression of NDRG1

Next, to identify the potential target gene(s) responsible for mediating the LINC00844-associated phenotype, we performed in-depth analysis of the androgen-regulated genes that requires LINC008844 for expression. We specifically concentrated on genes constituting the GO terms enriched for migration and invasion processes (Fig. 5B) and have been demonstrated to have tumor and/or metastasis-suppressing properties. Based on these criteria, we identified 44 LINC00844-dependent genes. After extensive literature review, we narrowed the list further to known metastasis suppressors that are suppressed after LINC00844 knockdown. We ended up with the following potential candidate genes: TUBA3C, SNAI2, TPM1, ITGAV, NKX3-1, and NDRG1. Currently, we are examining the function of several of these genes. In this work, we focus our attention on N-Myc Downstream Regulated 1 (NDRG1) as a DHT- and LINC00844-dependent gene. Our results show that knockdown of LINC00844 markedly reduced the DHT-mediated activation of NDRG1 transcript level in LNCaP and VCaP cells without any effects on its basal levels. Although NDRG1 mRNA levels were suppressed in both cell lines, protein reduction was observed only in LNCaP cells (Fig. 7A and B). In addition, we show that knockdown of LINC00844 reduced the binding of AR to an ARBS located approximately 30 kb upstream from the TSS of NDRG1 (Fig. 7C; Supplementary Fig. S7A).

Figure 7.

LINC00844 mediates prostate cancer migration and invasion in part by regulating NDRG1 expression. LNCaP and VCaP cells treated with siLINC00844 or NC and 10 nmol/L of DHT or ETOH for 12 hours. A, The mRNA levels were quantified by qPCR, and the bar graph represents NDRG1 expression as fold change normalized to NC-ETOH ± SEM from at least three independent experiments. B,NDRG1 protein levels were analyzed by Western blot (right) and were quantified by Imagelab (Biorad; left). Bar graphs represent the mean normalized (to GAPDH) band quantity ± SEM from at least three independent experiments. C, AR ChIP-qPCR was performed in LNCaP and VCaP cells treated with siLINC00844 or NC and 100 nmol/L of DHT or ETOH for 2 hours. AR binding to the ARBS at the enhancer region of the NDRG1 gene (30 kb form the TSS) is presented as the mean % of input ± SEM from at least three independent experiments. Box-plot showing the expression of NDRG1 in (D) Prensner and colleagues as log2(FPKM+1) and (E) patients in MSKCC datasets segregated based on their Gleason score and presented as log2 mRNA expression values. F, Kaplan–Meier analysis of prostate cancer outcome. One hundred forty patients from the MSKCC dataset were divided as higher or lower than median NDRG1 expression and analyzed for period of biochemical recurrence-free survival. G–L, LNCaP cells were treated with the indicated siRNA and/or plasmid and seeded in the upper chamber of the Boyden's transwell support. After 48 hours, cells were fixed and counted. Graphs representing the number of LNCaP cells per high power field (20X objective) migrated across transwell supports with 8 μm pores. Transwell supports were coated with Matrigel in 1:20 dilution for invasion assay or left uncoated for migration assay. G and H, Migration and invasion assay in LNCaP cells with knockdown of NDRG1. I and J, Migration and invasion assay in LNCaP cells with NDRG1 knockdown and LINC00844 rescue. K and L, Migration and invasion assay in overexpression of NDRG1 in LNCaP cells with knockdown of LINC00844. Error bars show the SEM of cells counted in 5 high power fields of view in 3 independent experiments. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Figure 7.

LINC00844 mediates prostate cancer migration and invasion in part by regulating NDRG1 expression. LNCaP and VCaP cells treated with siLINC00844 or NC and 10 nmol/L of DHT or ETOH for 12 hours. A, The mRNA levels were quantified by qPCR, and the bar graph represents NDRG1 expression as fold change normalized to NC-ETOH ± SEM from at least three independent experiments. B,NDRG1 protein levels were analyzed by Western blot (right) and were quantified by Imagelab (Biorad; left). Bar graphs represent the mean normalized (to GAPDH) band quantity ± SEM from at least three independent experiments. C, AR ChIP-qPCR was performed in LNCaP and VCaP cells treated with siLINC00844 or NC and 100 nmol/L of DHT or ETOH for 2 hours. AR binding to the ARBS at the enhancer region of the NDRG1 gene (30 kb form the TSS) is presented as the mean % of input ± SEM from at least three independent experiments. Box-plot showing the expression of NDRG1 in (D) Prensner and colleagues as log2(FPKM+1) and (E) patients in MSKCC datasets segregated based on their Gleason score and presented as log2 mRNA expression values. F, Kaplan–Meier analysis of prostate cancer outcome. One hundred forty patients from the MSKCC dataset were divided as higher or lower than median NDRG1 expression and analyzed for period of biochemical recurrence-free survival. G–L, LNCaP cells were treated with the indicated siRNA and/or plasmid and seeded in the upper chamber of the Boyden's transwell support. After 48 hours, cells were fixed and counted. Graphs representing the number of LNCaP cells per high power field (20X objective) migrated across transwell supports with 8 μm pores. Transwell supports were coated with Matrigel in 1:20 dilution for invasion assay or left uncoated for migration assay. G and H, Migration and invasion assay in LNCaP cells with knockdown of NDRG1. I and J, Migration and invasion assay in LNCaP cells with NDRG1 knockdown and LINC00844 rescue. K and L, Migration and invasion assay in overexpression of NDRG1 in LNCaP cells with knockdown of LINC00844. Error bars show the SEM of cells counted in 5 high power fields of view in 3 independent experiments. *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Close modal

NDRG1 is associated with a number of different types of cancers including breast, colon, pancreas, and prostate cancer (44–48). In addition, NDRG1 expression is negatively correlated with the cancer status, and overexpression studies with NDRG1 in prostate cancer cells markedly reduced in vivo metastasis (47). NDRG1 expression is also significantly reduced in prostate tumors with lymph node or bone metastasis compared with localized prostate cancer (47, 49, 50). Interrogation of data from the study by Prensner and colleagues (26) showed a reduction in NDRG1 expression in malignant and metastatic samples as compared with normal tissue (Fig. 7D). Similarly, NDRG1 expression was also negatively correlated with Gleason score (Fig. 7E), and low expressing patients showed significantly low recurrence-free survival in the MSKCC dataset (ref. 51; Fig. 7F).

From the above results, we hypothesized that LINC00844 may control the invasiveness of prostate cancer cells in part by regulating the expression of NDRG1. To test this possibility, we transfected LNCaP cells with siNDRG1 and assayed for their migration and invasion ability. If the activation of NDRG1 is essential for inhibiting cell migration and invasion, then we should observe an increase in migration and invasion upon the knockdown of NDRG1. Indeed, depletion of NDRG1 (Supplementary Fig. S7B) significantly increased both the migration (Fig. 7G; Supplementary Fig. S8A) and invasion (Fig. 7H; Supplementary Fig. S8B) activity of LNCaP cells. To provide further supporting evidence that NDRG1 is a direct downstream target of LINC00844, we carried out a set of rescue experiments. First, we overexpressed LINC00844 and demonstrated that knockdown of NDRG1 abrogated LINC00844-mediated migration and invasion (Fig. 7I and J; Supplementary Fig. S9A and S9B). Next, we depleted LINC00844 and showed that overexpression of NDRG1 (Supplementary Fig. S10A) overcame the knockdown effect of LINC00844 on migration and invasion (Fig. 7K and L; Supplementary Fig. S10B and S10C), suggesting that NDRG1 is a downstream target of LINC00844. Collectively, our results suggest that AR and LINC00844 control prostate cancer metastasis by directly regulating the transcription of NDRG1.

Recent discoveries and advancements in transcriptomics have revealed crucial roles for the so-called “junk DNA” or ncRNA in a myriad of biological processes including cancer (11). To further our understanding of ncRNAs in prostate cancer, we have systematically analyzed two well-established prostate cancer cohorts of different ethnicity to identify and characterize a novel lncRNA, LINC00844.

AR is critical for the normal development of prostate tissue and plays an important role in its carcinogenesis (3). AR-mediated transcriptional regulation is central in defining the transcriptomic landscape associated with different stages of prostate cancer like hormone-sensitive and mCRPC. The AR-regulatory complex consists of various pioneer factors, coactivators, and corepressors that are recruited to gene promoters in a coordinated and sequential pattern (4). Recent discoveries have exemplified the critical role of lncRNAs in this complex, such as mediating AR long-range chromatin interactions and altering the chromatin confirmation to regulate AR binding to the chromatin (17). Yet the current list of lncRNAs associated with AR transcriptional complex still remains small. In this study, we demonstrate LINC00844 is a novel AR-regulated and prostate cancer–associated lncRNA that is required for regulating the core AR transcriptional network.

Currently, a limited number of prostate cancer–associated lncRNAs have been identified, and only a fraction of them have been studied in detail to divulge the underlying molecular mechanisms. Interestingly, most of the lncRNAs studied so far are known to be upregulated in malignancy; however, our findings show LINC00844 is suppressed in malignancy which suggests a tumor-suppressive functional role. In this work, we showed LINC00844 is significantly suppressed not only in two test cohorts (26, 27) but also in much larger validation datasets, including TCGA (52) and MiTranscriptome (39). Interestingly, we also found the expression of LINC00844 is further suppressed in metastatic prostate cancer samples which suggest that it be may be crucial for inhibiting the metastatic transformation of malignant prostate cancer. In support of a tumor-suppressive functional role for LINC00844, we observed that prostate tumors with low expression of LINC00844 have worst prognostic outcomes and high biochemical recurrence.

Similarly, by means of loss-of-function studies, we demonstrated that the loss of LINC00844 resulted in increased cell migration and invasion without affecting the rate of proliferation. Interestingly, we also observed that the effect of LINC00844 was more pronounced on DHT-mediated cell migration and invasion compared with cells grown in normal growing media or ETOH-stimulated cells. Hence, we argue that LINC00844 mediates its phenotypic effects on prostate cancer cells by primarily targeting the AR transcriptional network.

Our current work indicates that LINC00844 is a critical component of the AR transcriptional machinery. From our microarray analysis, we showed that a loss of LINC00844 leads to a suppression of a subset of DHT-activated genes that are mainly associated with cytoskeletal organization or proliferation of prostate processes. Mechanistically, LINC00844 appears to be crucial for the binding of AR to the chromatin. To probe deeper into the mechanism through which LINC00844 regulates the binding of AR to the chromatin, we performed AR-RNA immunoprecipitation (RIP) and biotinylated LINC00844-mediated protein pulldown assays to examine if LINC00844 physically interacts with AR. Surprisingly, we did not find evidence of any direct physical interactions between LINC00844 and AR (data not shown). Hence, at this stage, we can only speculate an indirect regulation of AR by LINC00844. Plausible candidates could be coactivator(s) or collaborative factor(s) of the AR; however, this will need to be further studied.

We also provided evidence to show that LINC00844 could potentially regulate migration and invasion of prostate cancer through NDRG1. A previous study has shown that a loss of NDRG1 expression significantly increases in vivo metastasis of prostate cancer cells (47). Similar to LINC00844, NDRG1 is also significantly suppressed in malignant and metastatic prostate cancer samples, and patients with low expression have significantly increased rates of recurrence. However, significant correlation between LINC00844 and NDRG1 expression was not observed. Herein, we demonstrated in vitro DHT-mediated activation of NDRG1 suppresses prostate cancer cell migration, and invasion and LINC00844 is critical for AR recruitment to its enhancer and its expression.

In summary, our work showed LINC00844 is a novel AR-regulated lncRNA that plays an important role in the regulating prostate cancer cell migration and invasion. In addition, LINC00844 appears to regulate AR recruitment to chromatin and enhances the activation of a large number of canonical target genes.

No potential conflicts of interest were disclosed.

Conception and design: S. Lingadahalli, Y.Y. Sung, E. Cheung

Development of methodology: S. Lingadahalli, Y.Y. Sung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.Y. Sung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Lingadahalli, S. Jadhao, Y.Y. Sung, M. Chen, X. Chen, E. Cheung

Writing, review, and/or revision of the manuscript: S. Lingadahalli, E. Cheung

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Hu, E. Cheung

Study supervision: E. Cheung

This work was supported by a University of Macau Multi-Year Research Grant (MYRG2015-00196-FHS), a University of Macau Start-up Research Grant (SRG2014-00004-FHS), and the Macau Science and Technology Development Fund (FDCT/023/2014/A1 and FDCT102/2015/A3) to E. Cheung.

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.
Siegel
RL
,
Miller
KD
,
Jemal
A
. 
Cancer statistics, 2017
.
CA Cancer J Clin
2017
;
67
:
7
30
.
2.
Shen
MM
,
Abate-Shen
C
. 
Molecular genetics of prostate cancer: new prospects for old challenges
.
Genes Dev
2010
;
24
:
1967
2000
.
3.
Sung
YY
,
Cheung
E
. 
Androgen receptor co-regulatory networks in castration-resistant prostate cancer
.
Endocr Relat Cancer
2014
;
21
:
R1
11
.
4.
Chng
KR
,
Cheung
E
. 
Sequencing the transcriptional network of androgen receptor in prostate cancer
.
Cancer Lett
2013
;
340
:
254
60
.
5.
Scher
HI
,
Sawyers
CL
. 
Biology of progressive, castration-resistant prostate cancer: directed therapies targeting the androgen-receptor signaling axis
.
J Clin Oncol
2005
;
23
:
8253
61
.
6.
Harris
WP
,
Mostaghel
EA
,
Nelson
PS
,
Montgomery
B
. 
Androgen deprivation therapy: progress in understanding mechanisms of resistance and optimizing androgen depletion
.
Nat Clin Pract Urol
2009
;
6
:
76
85
.
7.
Wong
YN
,
Ferraldeschi
R
,
Attard
G
,
de Bono
J
. 
Evolution of androgen receptor targeted therapy for advanced prostate cancer
.
Nat Rev Clin Oncol
2014
;
11
:
365
76
.
8.
Andersen
RJ
,
Mawji
NR
,
Wang
J
,
Wang
G
,
Haile
S
,
Myung
JK
, et al
Regression of castrate-recurrent prostate cancer by a small-molecule inhibitor of the amino-terminus domain of the androgen receptor
.
Cancer Cell
2010
;
17
:
535
46
.
9.
Snoek
R
,
Cheng
H
,
Margiotti
K
,
Wafa
LA
,
Wong
CA
,
Wong
EC
, et al
In vivo knockdown of the androgen receptor results in growth inhibition and regression of well-established, castration-resistant prostate tumors
.
Clin Cancer Res
2009
;
15
:
39
47
.
10.
Gibb
EA
,
Brown
CJ
,
Lam
WL
. 
The functional role of long non-coding RNA in human carcinomas
.
Mol Cancer
2011
;
10
:
38
.
11.
Ponting
CP
,
Oliver
PL
,
Reik
W
. 
Evolution and functions of long noncoding RNAs
.
Cell
2009
;
136
:
629
41
.
12.
de Kok
JB
,
Verhaegh
GW
,
Roelofs
RW
,
Hessels
D
,
Kiemeney
LA
,
Aalders
TW
, et al
DD3(PCA3), a very sensitive and specific marker to detect prostate tumors
.
Cancer Res
2002
;
62
:
2695
8
.
13.
Sartori
DA
,
Chan
DW
. 
Biomarkers in prostate cancer: what's new?
Curr Opin Oncol
2014
;
26
:
259
64
.
14.
Prensner
JR
,
Iyer
MK
,
Sahu
A
,
Asangani
IA
,
Cao
Q
,
Patel
L
, et al
The long noncoding RNA SChLAP1 promotes aggressive prostate cancer and antagonizes the SWI/SNF complex
.
Nat Genet
2013
;
45
:
1392
8
.
15.
Lanz
RB
,
McKenna
NJ
,
Onate
SA
,
Albrecht
U
,
Wong
J
,
Tsai
SY
, et al
A steroid receptor coactivator, SRA, functions as an RNA and is present in an SRC-1 complex
.
Cell
1999
;
97
:
17
27
.
16.
Sun
M
,
Kraus
WL
. 
Minireview: Long noncoding RNAs: new "links" between gene expression and cellular outcomes in endocrinology
.
Mol Endocrinol
2013
;
27
:
1390
402
.
17.
Yang
L
,
Lin
C
,
Jin
C
,
Yang
JC
,
Tanasa
B
,
Li
W
, et al
lncRNA-dependent mechanisms of androgen-receptor-regulated gene activation programs
.
Nature
2013
;
500
:
598
602
.
18.
Takayama
K
,
Horie-Inoue
K
,
Katayama
S
,
Suzuki
T
,
Tsutsumi
S
,
Ikeda
K
, et al
Androgen-responsive long noncoding RNA CTBP1-AS promotes prostate cancer
.
EMBO J
2013
;
32
:
1665
80
.
19.
Zhang
A
,
Zhao
JC
,
Kim
J
,
Fong
KW
,
Yang
YA
,
Chakravarti
D
, et al
LncRNA HOTAIR enhances the androgen-receptor-mediated transcriptional program and drives castration-resistant prostate cancer
.
Cell Rep
2015
;
13
:
209
21
.
20.
Holmes
KA
,
Brown
GD
,
Carroll
JS
. 
Chromatin Immunoprecipitation-Sequencing (ChIP-seq) for mapping of estrogen receptor-chromatin interactions in breast cancer
.
Methods Mol Biol
2016
;
1366
:
79
98
.
21.
Li
H
,
Durbin
R
. 
Fast and accurate short read alignment with Burrows-Wheeler transform
.
Bioinformatics
2009
;
25
:
1754
60
.
22.
Zhang
Y
,
Liu
T
,
Meyer
CA
,
Eeckhoute
J
,
Johnson
DS
,
Bernstein
BE
, et al
Model-based analysis of ChIP-Seq (MACS)
.
Genome Biol
2008
;
9
:
R137
.
23.
Heinz
S
,
Benner
C
,
Spann
N
,
Bertolino
E
,
Lin
YC
,
Laslo
P
, et al
Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities
.
Mol Cell
2010
;
38
:
576
89
.
24.
Tan
SK
,
Lin
ZH
,
Chang
CW
,
Varang
V
,
Chng
KR
,
Pan
YF
, et al
AP-2gamma regulates oestrogen receptor-mediated long-range chromatin interaction and gene transcription
.
EMBO J
2011
;
30
:
2569
81
.
25.
Eijssen
LM
,
Goelela
VS
,
Kelder
T
,
Adriaens
ME
,
Evelo
CT
,
Radonjic
M
. 
A user-friendly workflow for analysis of Illumina gene expression bead array data available at the arrayanalysis.org portal
.
BMC Genomics
2015
;
16
:
482
.
26.
Prensner
JR
,
Iyer
MK
,
Balbin
OA
,
Dhanasekaran
SM
,
Cao
Q
,
Brenner
JC
, et al
Transcriptome sequencing across a prostate cancer cohort identifies PCAT-1, an unannotated lincRNA implicated in disease progression
.
Nat Biotechnol
2011
;
29
:
742
9
.
27.
Ren
S
,
Peng
Z
,
Mao
JH
,
Yu
Y
,
Yin
C
,
Gao
X
, et al
RNA-seq analysis of prostate cancer in the Chinese population identifies recurrent gene fusions, cancer-associated long noncoding RNAs and aberrant alternative splicings
.
Cell Res
2012
;
22
:
806
21
.
28.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
29.
Rau
A
,
Marot
G
,
Jaffrezic
F
. 
Differential meta-analysis of RNA-seq data from multiple studies
.
BMC Bioinformatics
2014
;
15
:
91
.
30.
Decker
KF
,
Zheng
D
,
He
Y
,
Bowman
T
,
Edwards
JR
,
Jia
L
. 
Persistent androgen receptor-mediated transcription in castration-resistant prostate cancer under androgen-deprived conditions
.
Nucleic Acids Res
2012
;
40
:
10765
79
.
31.
Sakurai
K
,
Reon
BJ
,
Anaya
J
,
Dutta
A
. 
The lncRNA DRAIC/PCAT29 locus constitutes a tumor-suppressive nexus
.
Mol Cancer Res
2015
;
13
:
828
38
.
32.
Kong
L
,
Zhang
Y
,
Ye
ZQ
,
Liu
XQ
,
Zhao
SQ
,
Wei
L
, et al
CPC: assess the protein-coding potential of transcripts using sequence features and support vector machine
.
Nucleic Acids Res
2007
;
35
:
W345
9
.
33.
Hon
CC
,
Ramilowski
JA
,
Harshbarger
J
,
Bertin
N
,
Rackham
OJ
,
Gough
J
, et al
An atlas of human long non-coding RNAs with accurate 5′ ends
.
Nature
2017
;
543
:
199
204
.
34.
Hazelett
DJ
,
Rhie
SK
,
Gaddis
M
,
Yan
C
,
Lakeland
DL
,
Coetzee
SG
, et al
Comprehensive functional annotation of 77 prostate cancer risk loci
.
PLoS Genet
2014
;
10
:
e1004102
.
35.
Misawa
A
,
Takayama
K
,
Urano
T
,
Inoue
S
. 
Androgen-induced long noncoding RNA (lncRNA) SOCS2-AS1 promotes cell growth and inhibits apoptosis in prostate cancer cells
.
J Biol Chem
2016
;
291
:
17861
80
.
36.
Tan
PY
,
Chang
CW
,
Chng
KR
,
Wansa
KD
,
Sung
WK
,
Cheung
E
. 
Integration of regulatory networks by NKX3-1 promotes androgen-dependent prostate cancer survival
.
Mol Cell Biol
2012
;
32
:
399
414
.
37.
Chng
KR
,
Chang
CW
,
Tan
SK
,
Yang
C
,
Hong
SZ
,
Sng
NY
, et al
A transcriptional repressor co-regulatory network governing androgen response in prostate cancers
.
EMBO J
2012
;
31
:
2810
23
.
38.
Pomerantz
MM
,
Li
F
,
Takeda
DY
,
Lenci
R
,
Chonkar
A
,
Chabot
M
, et al
The androgen receptor cistrome is extensively reprogrammed in human prostate tumorigenesis
.
Nat Genet
2015
;
47
:
1346
51
.
39.
Iyer
MK
,
Niknafs
YS
,
Malik
R
,
Singhal
U
,
Sahu
A
,
Hosono
Y
, et al
The landscape of long noncoding RNAs in the human transcriptome
.
Nat Genet
2015
;
47
:
199
208
.
40.
Li
Y
,
Luo
H
,
Xiao
N
,
Duan
J
,
Wang
Z
,
Wang
S
. 
Long noncoding RNA SChLAP1 accelerates the proliferation and metastasis of prostate cancer via targeting miR-198 and promoting the MAPK1 pathway
.
Oncol Res
2017
;
26
:
131
43
.
41.
Huang
da W
,
Sherman
BT
,
Lempicki
RA
. 
Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources
.
Nat Protoc
2009
;
4
:
44
57
.
42.
Cui
Z
,
Ren
S
,
Lu
J
,
Wang
F
,
Xu
W
,
Sun
Y
, et al
The prostate cancer-up-regulated long noncoding RNA PlncRNA-1 modulates apoptosis and proliferation through reciprocal regulation of androgen receptor
.
Urol Oncol
2013
;
31
:
1117
23
.
43.
Jin
Y
,
Cui
Z
,
Li
X
,
Jin
X
,
Peng
J
. 
Upregulation of long non-coding RNA PlncRNA-1 promotes proliferation and induces epithelial-mesenchymal transition in prostate cancer
.
Oncotarget
2017
;
8
:
26090
9
.
44.
Han
LL
,
Hou
L
,
Zhou
MJ
,
Ma
ZL
,
Lin
DL
,
Wu
L
, et al
Aberrant NDRG1 methylation associated with its decreased expression and clinicopathological significance in breast cancer
.
J Biomed Sci
2013
;
20
:
52
.
45.
Angst
E
,
Dawson
DW
,
Nguyen
A
,
Park
J
,
Go
VL
,
Reber
HA
, et al
Epigenetic regulation affects N-myc downstream-regulated gene 1 expression indirectly in pancreatic cancer cells
.
Pancreas
2010
;
39
:
675
9
.
46.
Li
Q
,
Chen
H
. 
Transcriptional silencing of N-Myc downstream-regulated gene 1 (NDRG1) in metastatic colon cancer cell line SW620
.
Clin Exp Metastasis
2011
;
28
:
127
35
.
47.
Bandyopadhyay
S
,
Pai
SK
,
Gross
SC
,
Hirota
S
,
Hosobe
S
,
Miura
K
, et al
The Drg-1 gene suppresses tumor metastasis in prostate cancer
.
Cancer Res
2003
;
63
:
1731
6
.
48.
Bandyopadhyay
S
,
Wang
Y
,
Zhan
R
,
Pai
SK
,
Watabe
M
,
Iiizumi
M
, et al
The tumor metastasis suppressor gene Drg-1 down-regulates the expression of activating transcription factor 3 in prostate cancer
.
Cancer Res
2006
;
66
:
11983
90
.
49.
Song
Y
,
Oda
Y
,
Hori
M
,
Kuroiwa
K
,
Ono
M
,
Hosoi
F
, et al
N-myc downstream regulated gene-1/Cap43 may play an important role in malignant progression of prostate cancer, in its close association with E-cadherin
.
Hum Pathol
2010
;
41
:
214
22
.
50.
Liu
W
,
Iiizumi-Gairani
M
,
Okuda
H
,
Kobayashi
A
,
Watabe
M
,
Pai
SK
, et al
KAI1 gene is engaged in NDRG1 gene-mediated metastasis suppression through the ATF3-NFkappaB complex in human prostate cancer
.
J Biol Chem
2011
;
286
:
18949
59
.
51.
Taylor
BS
,
Schultz
N
,
Hieronymus
H
,
Gopalan
A
,
Xiao
Y
,
Carver
BS
, et al
Integrative genomic profiling of human prostate cancer
.
Cancer Cell
2010
;
18
:
11
22
.
52.
Li
J
,
Han
L
,
Roebuck
P
,
Diao
L
,
Liu
L
,
Yuan
Y
, et al
TANRIC: an interactive open platform to explore the function of lncRNAs in cancer
.
Cancer Res
2015
;
75
:
3728
37
.