Human PCa cell lines DU145, PC3, 22Rv1 and LNCaP were obtained from the Shanghai Institute of Biochemistry and Cell Biology (CAS). Lenti-X™293T cells were obtained from Clontech Laboratories, Inc. (Mountainview, CA, USA). Cell lines were cultured routinely in RPMI-1640 medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) or DMEM (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% fetal bovine serum (FBS, Gibco; Thermo Fisher Scientific, Inc.), penicillin and streptomycin (×100). All cells were maintained in 5% CO₂ at 37°C. PCR detection method was used to check mycoplasma contamination. All cell lines were verified using short tandem repeat assays by Genetic Testing Biotechnology, Suzhou, China.

5′/3′ Rapid amplification of cDNA ends (5′/3′ RACE). The complementary DNA (cDNA) was synthesized using RNA (1.0 µg) extracted from PC3 cells. The 5′/3′ RACE was carried out using the SMARTer™ RACE cDNA Amplification kit (Clontech Laboratories, Inc.) according to the manufacturer's instructions. The gene-specific primers used for the PCR of the RACE analysis are given in Table I.

The full-length of MYU (Fig. 1) and VPS9D1, which is fused with a FLAG epitope tag with the C terminus, Fig. 2) were subcloned into pLVX-IRES-Neo vector. The short hairpin RNAs (shRNAs) that target different sites of MYU, VPS9D1 and c-Myc CDS region were inserted into the pSIH1-H1 vector. The shRNA primers are listed in Table II. MYU full-length and 3′ untranslated regions (UTR) of c-Myc containing the miR-184 binding sites were synthesized in the dual luciferase reporter vector pmirGLO (Promega Corporation, Madison, WI, USA).

Total RNA was isolated from PCa cells using RNAiso Plus reagent (Takara, Shiga, Japan) and RNA (1 µg) was reverse transcribed using the PrimerScript RT-PCR kit (Takara) according to the manufacturer's instructions. The expression levels of MYU were determined by qRT-PCR using the SYBR^® Green dye (Takara) detection method. We used the comparative Ct method to quantify transcripts. The relative gene expression was defined using the equation: ∆Ct=Ct _target - Ct _reference, and the fold-change for target genes normalized by GAPDH was determined by the formula 2^−∆∆Ct.

Total cell lysates were prepared using RIPA reagent (Beyotime, Shanghai, China) and sodium dodecyl sulfate (SDS) loading buffer (Beyotime). The concentration of total proteins was examined using BCA protein kit (Beyotime). Equivalent quantities of proteins were loaded to each lane. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate proteins and then the proteins were transferred from the gel onto NC or PVDF membranes. Membranes were incubated with primary antibodies against GAPDH (AG019, 1:5,000), H3 (AH433, 1:5,000) and β-actin (AA128, 1:5,000) all from Beyotime, Flag (F3165, 1:1,000) from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany), ZO-1 (8193S, 1:1,000), caludin-1 (13995S, 1:1,000), E-cadherin (3195S, 1:1,000), and c-Myc (13987S, 1:1,000) from Cell Signaling Technology, Inc. (Danvers, MA, USA), N-cadherin (ab76011, 1:1,000), vimentin (ab92547, 1:1,000) from Abcam (Cambridge, MA, USA). After incubation with suitable HRP-conjugated secondary antibodies, immunoreactive proteins were finally visualized using ECL chemiluminescence system (CLiNX, Shanghai, China). All qRT-PCR primers are listed in Table III.

LNCaP cells were captured with 1 ml PBS and 200 µl were extracted as the positive control, the reminder were snap-frozen with liquid nitrogen and then quickly dissolved in a water bath, repeating this operation twice. Then, 200 µl buffer A1 (10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.1% NP-40 and 1% protease inhibitor) was added to the cells, and ground 15–20 times with a grinding pestle. After incubation on ice for 5 min, the lysate was centrifuged at 2,000 rpm for 5 min at 4°C. The supernatant was used to separate the cytoplasmic fraction. The remaining pellet was resuspended in 200 µl of buffer A2 (10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, and 1% protease inhibitor) and ground 5–10 times with a grinding pestle. Then, the lysate was resuspended in another 300 µl of buffer A2. After incubating on ice for 5 min, the lysate was centrifuged at 2000 rpm for 5 min at 4°C. The remaining pellet was resuspended in 500 µl of buffer S1 (0.25 mM sucrose, 10 mM MgCl₂, and 1% protease inhibitor), and then added 500 µl of buffer S2 (0.35 mM sucrose, 0.5 mM MgCl₂, and 1% protease inhibitor) slowly from the top of the liquid. The lysate was centrifuged at 3,500 rpm for 5 min at 4°C. The remaining pellet was resuspended in 200 µl RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1% NP-40, and 0.5% deoxycholate) and incubated on ice for 30 min. The supernatant containing the nuclear fraction was collected by centsrifugation at 12,000 rpm for 15 min at 4°C. qRT-PCR and western blot analysis were used to assess the total RNA and protein isolated from the subcellular fractions. GAPDH served as the cytoplasmic control, and U1 as nuclear control. GAPDH, U1 and MYU expression levels were measured by qRT-PCR.

The packaging plasmids, recombined plasmids and LNAs for overexpression and knockdown were transfected into Lenti-X™ 293T cells. Two LNAs (LNA-1 and LNA-2) and LNA-Ctrl (300610, Negative control A Antisense LNA GapmeR) were purchased from Exiqon A/S (Vedbaek, Denmark). The hsa-miR-184 (miR-184) mimics and the negative control miR-NC were purchased from RiboBio (Guangzhou, China). Cells were harvested 48 h after transfection. These corresponding plasmids or miRNA mimics were transfected with PEI or Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). The final concentrations of miRNAs/LNAs employed in this study were as follows: miR-184 mimic/negative control mimic 50 nM/ml and LNA-1/LNA-2/LNA-Ctrl 50 nM/ml.

Cells were seeded into 96-well plates at a density of 1.0×10³ cells/well. Cell viability was assessed using CellTiter-Glo (CTG) reagent (Promega Corporation) every 48 h following the manufacturer's instructions. In addition, 1.0×10³ cells were seeded per well of 6-well plates. After 2 weeks, the number of clones was counted with staining using ImageJ (https://imagej.nih.gov/ij/). Wound healing and Transwell chamber assays were performed to determine cell migration ability. For wound healing assay, cells were seeded into two well silicones (ibidi, GmbH, Munich, Germany) in 24-well plates. When the cell density reached 90–100% confluence, the inserts were removed and washed with PBS, and then cultured with serum-free medium. Images were collected at 0 and 24 h under an inverted microscope (Ziess, Jena, Germany). Wound healing images were analyzed using ImageJ. For the Transwell chamber assay, the cell suspension, 3.5×10⁴ cells/well in 200 µl, was added to a 8.0-µm pore-size insert of a 24-well plate (Corning Costar, Corning, NY, USA). The lower well was filled with 500 µl medium with 10% FBS. The Transwell plate was then incubated for 24 h at 37°C with 5% CO₂. Hereafter, the cells on the top surface of the insert were wiped off and the cells on the lower surface of the insert were fixed with 95% ethanol for 10–15 min, and then dyed with 0.5% crystal violet. Images were collected under an inverted microscope. Finally, the cells were decolorized with 95% ethanol and the same amount of eluent was used to measure the absorbance at 580 nm wavelength. These experiments were conducted in triplicate.

PC3 cells were grown to a 70–80% confluent monolayer, washed three times with PBS, and grown in serum-free culture medium for 3 days. Exosomes were collected by gradient centrifugation, cleaned by PBS, and stored at −80°C.

Exosomes were isolated from the MYU-overexpressing or control PC3 supernatant. An appropriate amount of serum-free 1640 medium was added and filtered with a 0.22-µm filter. For cell viability assay, PC3 cells were plated in a 96-well plate on the first day, and then treated with exosomes from MYU-overexpressing or control PC3 cells, which were assessed by CTG assays after incubation for 5 days. For Transwell assays, the PC3 cell suspension, 3.5×10⁴ cells/well in 200 µl, was added to an 8.0-µm pore-size insert of a 24-well plate. The lower well contained 500 µl exosomes and 10% FBS. The Transwell plate was then incubated for 24 h at 37°C with 5% CO₂.

Lenti-X™ 293T cells grown in a 24-well plate were co-transfected with 200 ng of dual luciferase reporter vector comprising MYU and its mutant or 3′UTR of c-Myc, 50 nM miRNA mimics using Lipofectamie 3000. Cells were harvested 48 h after transfection and analyzed using the dual luciferase reporter assay kit (Promega Corporation) with a microplate reader (BioTek Instruments, Inc., Winooski, VT, USA) according to the manufacturer's protocol.

The expression level of lncRNA MYU in prostate adenocarcinoma (PRAD) based on the clinical information was obtained from The Cancer Genome Atlas (TCGA) (17) and the matched expression profiles of MYU in Mitranscriptome (18). TCGA PRAD patient miRNA expression (Level 3 data, illuminahiseq mirnaseq) and mRNA expression (level 3 data, RNA-seq Version 2) was download from the Broad GDAC Firehose (19).

Sequencing data from GSE96019, GSE96399 and GSE96418 were downloaded from GEO. Reads from the PC3 H3K4me3, H3K27ac and H3K36me3 ChIP-Seq samples were mapped to human genome version hg19 using HISAT2. Peak calling was performed using MACS (20) according to the published protocols (21). Data were visualized using the UCSC Genome Browser (22).

The data were collected and expressed as means ± SEM. Statistical differences (P-values) between two groups were obtained by using a two-tailed Student's t-test. Statistical differences (P-values) in multiple groups were obtained by using a two-way ANOVA analysis. All experiments were carried out in triplicate. A two-tailed P<0.05 was considered to indicate statistical significance. All statistical analyses were undertaken using GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA).

In order to search for the differentially expressed lncRNAs associated with PCa, we analyzed clinical patient data in TCGA and matched expression profiles of lncRNAs from Mitranscriptome. We noted that lncRNA MYU was significantly upregulated in multiple cancers compared with adjacent normal tissues, and focused on PCa (P<0.0001, Fig. 3A). To investigate the potential cause of higher MYU expression in PCa, we examined the methylation levels of the MYU promoter region, which is a major cause of gene deregulation in tumorigenesis (23). Apparently, the methylation levels of histone H3 of the MYU promoter region in PCa tissues was significantly lower than those of adjacent normal tissues (P=0.0163, Fig. 3B). Moreover, receiver operating characteristic (ROC) curve was used to evaluate the sensitivity and specificity of MYU expression in predicting PCa tissues from normal tissues. MYU displayed considerable predictive significance, with an area under the curve (AUC) of 0.906 (Fig. 3C). In summary, MYU is downregulated in PCa and may be a potentially diagnostic indicator.

To explore the function of MYU more precisely, we performed 5′/3′-RACE to obtain the accurate transcript of MYU, which is a 1,461 bp transcript with four exons (Fig. 3D and Fig. 1). Published PCa chromatin immunoprecipitation and sequencing (ChIP-seq) data confirmed that the transcriptional start site (TSS) of MYU was marked by H3-lysine-27-acetylation (H3K27ac) (24), H3-lysine-4-trimethylation (H3K4me3) (25) and that its gene body harbored H3-lysine-36-trimethylation (H3K36me3) (26), an epigenetic signature consistent with lncRNAs. In addition, the UCSC Genome Browser revealed that MYU sequences across the different species were extremely unconserved (Fig. 3E). The PhyloCSF analysis showed that MYU has no potential protein-coding ability (Fig. 3F). Moreover, subcellular localization of MYU showed about 58% MYU in the nucleus in LNCaP cells (Fig. 3G). Taken together, these data revealed that MYU is an active transcribed lncRNA, which exists both in the nucleus and cytoplasm.

We found that MYU has a corresponding sense-cognate coding gene VPS9D1 from the UCSC Genome Browser. Many studies have shown that antisense transcripts regulate the expression of sense genes, particular for ncRNAs as antisense transcripts (27,28). We investigated the potential relationship between MYU and VPS9D1. There was no difference in expression levels of VPS9D1 between PCa tissues and adjacent normal tissues (Fig. 4A). The ROC analysis showed that the AUC of VPS9D1 was 0.621 (Fig. 2B). Moreover, there was no statistically significant correlation in RNA level between MYU and VPS9D1 (r=0.0525, P=0.2418) (Fig. 4C). Furthermore, we examined the potential regulation between MYU and VPS9D1 using qRT-PCR and western blotting assays. The qRT-PCR results revealed that knockdown of either gene did not affect the mRNA level of another gene (Fig. 4D), and the result of overexpression was the same (Fig. 4E). In order to determine whether MYU expression affected the protein level of VPS9D1, we overexpressed VPS9D1 in DU145 cells (Fig. 4F). Then we knocked down MYU in the VPS9D1-overexpressed cells, and observed the same protein levels of VPS9D1 in either MYU-knockdown or control cells (Fig. 4G). These results suggest that MYU did not regulate VPS9D1 at either the mRNA or protein level. Thus, we focused on studying lncRNA MYU.

To evaluate the biological effects of MYU on PCa, we firstly performed qRT-PCR to examine MYU expression level in four PCa cell lines (Fig. 5A). To clarify the function of MYU, we established cell lines with stably overexpressed MYU and MYU-specific shRNAs (Fig. 5B). CTG assays indicated that PCa cell proliferation was repressed by MYU shRNAs compared to that of cells stably expressing the scrambled shRNA, whereas enhancement of cell viability occurred after overexpression of MYU (Fig. 5C). In colony formation assay, MYU-upregulated PC3 cells exhibited increased colony growth, while knockdown of MYU in DU145 cells reduced colony growth compared to the control cells (Fig. 5D). Additionally, wound healing (Fig. 5E), Transwell chamber (Fig. 5F) and western blotting (Fig. 5G) assays indicated that knockdown of MYU reduced the DU145 cell migration rate and the protein level of N-cadherin and vimentin, but increased ZO-1, claudin-1 and E-cadherin. In contrast, overexpression of MYU in the PC3 cells resulted in a higher migration rate, accompanied by downregulated ZO-1, claudin-1 and E-cadherin and upregulated N-cadherin and vimentin. To further confirm these results, we also used locked nucleic acids (LNAs) to knock down the expression of MYU. Similar results for cell viability, clonogenicity and motility were obtained in the DU145 cells (Fig. 6A-F). These results suggest that MYU may act as a non-coding oncogene in PCa cells.

We aimed to ascertain the potential underlying mechanism of MYU, which may mediate PCa progression. lncRNAs are known to work as sponges to recruit miRNAs, resulting in the activation of miRNA target genes (13,29). We analyzed the correlation of the expression level between MYU and mature miRNAs and found that MYU showed a significantly negative correlation with 15 miRNAs (r< −0.3, P<0.0001, Table V). We then tested if three miRNAs including miR-187-3p, miR-143-3p and miR-184 (r< −0.35, P<0.0001) can bind to MYU. The dual luciferase reporter assays revealed that only miR-184 downregulated the luciferase activity of lncRNA MYU. Then, we searched the reported target coding genes of miR-184 in PubMed. Noteworthy, miR-184 has been reported as a tumor-suppressor gene in multiple cancers (30,31), and downregulated c-Myc expression (32,33). It has been reported that MYU is a direct target of c-Myc (14). We hypothesized that MYU may promote the proliferation of PCa as a ceRNA to sponge miR-184, thus upregulating c-Myc.

TCGA data revealed that miR-184 was significantly downregulated in PCa tissues (Fig. 7A) and negatively correlated with MYU (r=−0.1966, P=0.0037) (Fig. 7B). RNAhybrid (34) was used to predict the binding sites of miR-184 /MYU and miR-184/3′ UTR of c-Myc (Fig. 7C and D). We constructed luciferase reporters containing full-length MYU or mutated miR-184 binding sites or 3′ UTR of c-Myc. We found that overexpression of miR-184 mimic (Table IV) reduced luciferase activity of MYU and c-Myc 3′UTR, but not of the empty vector or the mutated reporter vector (Fig. 7E). Next, we explored the effect of MYU on the expression of c-Myc. The expression level of c-Myc was positively correlated with MYU (r=0.3598, P<0.0001) in PCa tissues (Fig. 7F). Furthermore, we detected that MYU knockdown in DU145 cells decreased the mRNA and protein level of c-Myc. Inversely, MYU overexpression in PC3 cells increased c-Myc expression (Fig. 7G). In addition, c-Myc knockdown suppressed the proliferation of PC3 cells and inhibited the effect of MYU overexpression on promoting cell proliferation. Additionally, the protein level of c-Myc was consistent with the cell proliferation (Fig. 7H), indicating that c-Myc can partly influence MYU. To further clarify the relationship of MYU, miR-184 and c-Myc, we transfected miR-184 mimics into MYU-overexpressing PC3 cells. We found that overexpression of miR-184 significantly abrogated the effect of MYU overexpression on promoting cell proliferation and suppressed the protein level of c-Myc (Fig. 7I). On the base of the above findings, MYU likely regulates c-Myc by functioning as ceRNA to sponge miR-184.

Recent studies suggest that exosomes, which contain proteins, mRNAs, miRNAs and lncRNAs, may participate in the biological effects in cancer cells (35,36). To examine whether MYU is present in exosomes, we isolated exosomes from the culture medium of MYU-overexpressing or control PC3 cells. Nanosight particle tracking analysis demonstrated the size and number of exosomes (Fig. 8A). Western blot analysis further confirmed their identity by the exosomal markers TSG-101 and CD9 (Fig. 8B). We detected the expression level of MYU by qRT-PCR in the exosomes of both PC3 cell lines. There was a significant upregulation of MYU in the exosomes isolated from the MYU-overexpressing cells, which was consistent with the cellular expression (Fig. 8C). We next investigated the existing pattern of MYU from exosomes. The levels of MYU were unchanged upon RNase A treatment but significantly decreased after being treated with RNase A and Triton X-100 (Fig. 8D), indicating that MYU was mainly wrapped by the membrane of exosomes. These findings suggest that MYU can be delivered into the extracellular milieu by exosomes. In order to explore whether exosome-contained MYU can function in neighboring cells, we purified exosomes and then fed PC3 cells. The CTG (Fig. 8E) and Transwell chamber assays (Fig. 8F) demonstrated that exosomes from the MYU-overexpressing PC3 cells significantly increased proliferation and migration ability of recipient cells, compared to those from control cells. Correspondingly, the c-Myc protein level was significantly increased in the recipient PC3 cells fed with exosomes from the MYU-overexpressing cells (Fig. 8G). These findings suggest that MYU can be delivered into the extracellular milieu by exosomes, and then function in other PCa cells.