LINC01088 inhibits tumorigenesis of ovarian epithelial cells by targeting miR-24-1-5p

The roles of long non-coding RNAs (lncRNAs), a class of long non-protein-coding RNAs, in the tumorigenesis of ovarian epithelial cells remain unknown. In this study, we discovered that the expression of long intergenic non-coding RNA 1088 (LINC01088) was clearly reduced in benign epithelial ovarian tumor tissues compared to matched normal ovarian tissues. This was shown by global cDNA gene chip scanning and real-time qPCR, and validated in 42 clinical specimens. Furthermore, we found that LINC01088 inhibited the growth of ovarian cancer xenografts in nude mice. Correlation analysis between LINC01088 and mircoRNAs (miRNAs) conducted using primary clinical samples and RNA co-precipitation experiments revealed that miR-24-1-5p was one of the targets of LINC01088. Overexpression of miR-24-1-5p facilitated cell proliferation both in vitro and in vivo, however, LINC01088 could partially reverse the cell proliferation induced by miR-24-1-5p. Finally, we demonstrated that p21 activated kinase 4 (PAK4) was one of the downstream key targets of miR-24-1-5p by luciferase reporter assay and Western blotting; and our results showed a remarkable decrease in cell proliferation after overexpression of PAK4. We conclude that LINC01088 might function as a tumor suppressor, inhibiting the tumorigenesis of ovarian epithelial cells through LINC01088/ miR-24-1-5p/ PAK4 axis.


Results
LINC01088 is down-regulated in benign epithelial ovarian tumors. To evaluate the main lincR-NAs involved in tumorigenesis of ovarian epithelial cells, we analysed lincRNA expression in 3 benign epithelial ovarian tumor tissue specimens and 3 normal ovarian epithelial tissue specimens using gene chip scanning. Compared with the normal group, the expression levels of the most significant four lincRNAs (LINC01088, LINC01018, LINC01436 and LINC00243) were remarkably down-regulated in the tumor group (Fig. 1A), and the down-regulation of LINC01088 was the most marked. Also, RT-qPCR on the same tissue specimens gave similar results with the gene chip analyses (Fig. 1B). Next, we compared the expression levels of these four lincRNAs in normal ovarian epithelial tissue specimens, and found that the expression level of LINC01088 was the highest (Fig. 1C). Thus, we chose LINC01088 for further study. Moreover, in an experiment with increased sample size, we discovered once again that the expression of LINC01088 was prominently decreased in benign epithelial ovarian tumor tissues compared to the normal tissues. Additionally, LINC01088 expression declined even more in borderline and malignant epithelial ovarian tumors (Fig. 1D). The above results suggested that LINC01088 was a molecule that was negatively associated with tumorigenesis and might be a tumor suppressor gene.

miR-24-1-5p is a target of LINC01088.
To determine the role of LINC01088 in tumorigenesis of ovarian epithelial cells, we constructed LINC01088-lentiviral vectors and transfected them into A2780 cells. RT-qPCR verified that we have establish a cell line highly expressing LINC01088 (Fig. 2A). The tumor growth assay showed that LINC01088 significantly inhibited tumor growth (Fig. 2B), implying that LINC01088 was a tumor suppressor gene, although its mechanism remained unclear. Based on the role of lncRNAs as molecular "sponges", we speculated that LINC01088 had target miRNAs and served as a molecular "sponge". Thus, we used the BLAST function on www.ncbi.nlm.nih.gov and found out that miR-24-1-5p had two binding sites for LINC01088 (Fig. 2C). To verify the interaction between LINC01088 and miR-24-1-5p, we conducted RT-qPCR analysis using a total of 22 RNA was isolated from the clinical specimens using Trizol. After reverse transcription, real-time qPCR was conducted to determine the LINC01088 level in 3 benign epithelial ovarian tumor tissue specimens and the matched normal tissues. Data are represented as mean +/− SD, *means P < 0.05 vs normal group (ANOVA). (C) The expressive abundance of four lincRNAs. Normalized signals of LINC01088, LINC00243, LINC01436 and LINC01018 in normal ovarian epithelial tissues showed by gene chip above. (D) LINC01088 in primary epithelial ovarian tumors. RNA was isolated from the clinical specimens using Trizol. After reverse transcription, real-time qPCR was conducted to determine the LINC01088 level in 12 benign epithelial ovarian tumor tissue specimens, 8 borderline epithelial ovarian tumor tissue specimens, 12 malignant epithelial ovarian tumor tissue specimens and the matched normal tissues. Data are represented as mean +/− SD, **means P < 0.01 vs normal group (ANOVA). clinical tissue samples, and found a negative correlation between the expression levels of LINC01088 and miR-24-1-5p (spearman rank correlation was −0.568, P < 0.01) (Fig. 2D,E). Furthermore, we transfected A2780 cells with miR-24-1-5p expressing plasmids labeled with biotin and tested LINC01088 by PCR using RNA sediments "pulled-down" by co-precipitation, and the results proved the existence of LINC01088 (Fig. 2F). For further verification, pMIR-LINC01088 luciferase reporter plasmids were constructed and transfected into A2780 cells. The luciferase activity assay showed that miR-24-1-5p could significantly inhibited firefly luciferase activity (Fig. 2G), indicating that LINC01088 could target miR-24-1-5p. The above experiments demonstrated that LINC01088 might suppress cell growth by targeting miR-24-1-5p.
LINC01088 reduces cell proliferation promoted by miR-24-1-5p. The experiments described above demonstrated the interaction between LINC01088 and miR-24-1-5p. Next, we went on to explore the effect of miR-24-1-5p on the growth of ovarian epithelial cells. Firstly, we assessed the impact of miR-24-1-5p on cell proliferation. Plasmids expressing miR-24-1-5p were transfected into A2780 cells to establish stable cell line expressing miR-24-1-5p. The MTS assay demonstrated that miR-24-1-5p cloud remarkably facilitated cell proliferation, while LINC01088 distinctly inhibited cell proliferation promoted by miR-24-1-5p (Fig. 3A). Secondly, we investigated the role of miR-24-1-5p in cell migration, and the scratch assay showed that miR-24-1-5p had no regulatory effect on cell migration (Fig. 3B). At last, we explored the influence of miR-24-1-5p on tumorigenicity in animals. Balb/c nude mice were subcutaneously implanted with cells stably overexpressing miR-24-1-5p. Four weeks after implantation, it was evident that not only the tumor cell growth was markedly accelerated, but also the tumor volume and weight were significantly increased in the miR-24-1-5p group, compared with those in the control group (Fig. 3C,D). Moreover, we performed in vivo experiment with cells stably overexpressing miR-24-1-5p + LINC01088, the result showed that LINC01088 inhibited tumor growth facilitated by miR-24-1-5p (Fig. 3E). These findings suggested that miR-24-1-5p promoted cell proliferation but did not affect cell migration, hinting that miR-24-1-5p was an oncogene and LINC01088 could reduce cell proliferation promoted by miR-24-1-5p. Female Balb/c nude mice were subcutaneously implanted with A2780 cells transfected with LINC01088-lentiviral vector or blank-lentiviral vector, respectively. 28 days after implantation, the nude mice in each group were sacrificed. The tumor volume was calculated. Data are represented as mean +/−SD, **means P < 0.01 vs control (Repeated Measure ANOVA). (C) Prediction of binding sites between LINC01088 and miR-24-1-5p. Gene Blasting was performed and found two potential binding sites between LINC01088 and miR-24-1-5p. (D,E) Level of miR-24-1-5p in primary epithelial ovarian tumors. RNA was isolated from the clinical specimens using Trizol. After reverse transcription, real-time qPCR was conducted to determine the miR-24-1-5p level in 12 benign epithelial ovarian tumor tissue specimens and the matched normal tissues. Data are represented as mean +/− SD, **means P < 0.01 vs normal group (ANOVA) and showed a negative relation between the expression LINC01088 and miR-24-1-5p (spearman rank correlation was −0.568, P < 0.01). (F) RNA co-precipitation. A2780 cells were transiently transfected with miR-24-1-5p expression plasmids labeled with biotin, PCR was performed to detect LINC01088 after reverse transcription through Oligo dT15. Results showed that there was an interaction between LINC01088 and miR-24-1-5p. (G) Validating the interaction between LINC01088 and miR-24-1-5p. The relative luciferase activity was notably decreased in A2780 cells co-transfected with pcDNA6.2-GW/EmGFP-miR-24-1-5p, PRL-TK and pMIR-LINC01088 reporter plasmids. miR-LacZ was taken as the control. Data are represented as mean +/−SD, **means P < 0.01 vs control (ANOVA).
PAK4 is the potential target of miR-24-1-5p. It had previously been demonstrated that miR-24-1-5p increased in benign epithelial ovarian tumors and promoted cell proliferation, and was also involved in the development of epithelial ovarian tumors. Hence, we studied the target genes downstream of miR-24-1-5p further. Gene 'BLAST' on NCBI suggested that PAK4 was the possible target of miR-24-1-5p (Fig. 4A). To test this assumption, firstly we performed a IHC assay on normal ovarian epithelial tissue sections and benign epithelial ovarian tumor tissue sections. The results showed that PAK4 was expressed in both normal ovarian epithelial tissues and benign epithelial ovarian tumor tissues, but PAK4 expression seemed to be lower in benign epithelial ovarian tumor (Fig. 4B). Next, pMIR-PAK4 3′UTR luciferase reporter plasmids were constructed and transfected into A2780 cells. The luciferase activity assay showed that miR-24-1-5p inhibited firefly luciferase activity dramatically (Fig. 4C). Lastly, we detected PAK4 expression in miR-24-1-5p-expressing cells. The western blotting data indicated that PAK4 expression in A2780 cells stably transfected with miR-24-1-5p recombinant plasmids was indeed lessened (Fig. 4D), while LINC01088 had no direct effect on the expression of PAK4 (data not shown). Taken together, these findings demonstrated that PAK4 was the target of miR-24-1-5p, indicating that miR-24-1-5p might participate in the development and progression of epithelial ovarian tumors by targeting PAK4.

PAK4 inhibits cell proliferation.
To further study the biological function of PAK4 in ovarian epithelial cells, we transfected A2780 cells with plasmids expressing PAK4 (Fig. 5A), and evaluated their proliferation and migration abilities respectively. The data showed that overexpressed PAK4 significantly decreased cell proliferation (Fig. 5B), but promoted cell migration (Fig. 5C), which implied that PAK4 was a cell proliferation-inhibiting protein.

Discussion
This study demonstrates that the expression of LINC01088 is lowered in benign epithelial ovarian tumors compared to normal ovarian epithelial tissues, and LINC01088 expression has a negative correlation with that of miR-24-1-5p. The study also proves that LINC01088 can target miR-24-1-5p to regulate the expression of PAK4, and influence cell proliferation and migration. The brief diagrammatic representation is made in the Fig. 6. This implies that LINC01088 might inhibit the development of epithelial ovarian tumors through its interaction with miR-24-1-5p and the downstream effector protein PAK4.
LncRNAs perform their biological function through regulation of gene expression. Their interaction with miRNAs is very important in tumorigenesis and tumor progression [14][15][16][17] , and lncRNA-mediated sponge regulatory network in miRNAs and pri-miRNAs is one of the modes of action. For example, Kumar et al. 18 proved that Hmga2 promoted lung cancer progression by competing for let-7 occupancy with TGF-beta type III receptor (Tgfbr3t) through transforming growth factor-β (TGF-β) signaling pathway. Also, lncRNAGAS5 acted as a tumor suppressor in hepatocellular carcinoma through negative regulation of miR-21 to up-regulate its targets programmed cell death 4 (PDCD4) and phosphatase and tensin homologue (PTEN), resulting in inhibition of cancer cell migration and invasion 19 . In our study, we demonstrated that LINC01088 targeted miR-24-1-5p to regulate the expression of PAK4, but the exact mechanism of the interaction between LINC01088 and miR-24-1-5p required further exploration.
MiR-24 was initially discovered through a research involving invertebrates and vertebrates by Lagos-Quintana et al. 20 in 2001. Has-mir-24 has two forms, mir-24-1 and mir-24-2, located on Chromosome 9 and Chromosome 19 respectively 21 , which includes three mature sequences, miR-24-1-5p, miR-24-3p and miR-24-2-5p. To date, the action mechanism of miR-24-1-5p in carcinomas remains unclear. As reported in Braoudaki's retrospective study in patients with ependymoma (EP) 22 , miR-24-1-5p was up-regulated significantly in relapse and progression cases compared to clinical remission and survival cases. Moreover, miR-24-1-5p was considered as an oncogene associated with multiple endocrine neoplasia type1 (MEN1) 23 . While Goto et al. 24 indicated that miR-24-1-5p could clearly inhibit cell proliferation, migration and metastasis in prostate cancer. Inoguchi et al. 25 also verified that miR-24-1-5p inhibited bladder cancer cell proliferation by targeting forkhead box protein M1 (FOXM1). However, the possible roles of miR-24-1-5p in epithelial ovarian tumors haven't been reported yet. This study confirmed that miR-24-1-5p could promote epithelial ovarian tumors by negatively regulating PAK4 expression. In fact, miRNAs have innumerable target molecules, and regulate every aspect of cellular activity, such as cell proliferation, differentiation, apoptosis and so on [26][27][28][29] . Furthermore, miRNAs play multiple roles in development and progression of tumors, including angiogenesis, metastasis, and also exosome-mediated regulation [30][31][32][33][34] . Hence, it can be concluded that the results of our study elucidate only one of the countless signaling pathways involved in various biological processes associated with tumors, therefore further studies are required.
During the course of this study, it was demonstrated that PAK4 was the target of miR-24-1-5p. PAK is a class of evolutionarily conserved serine/threonine protein kinases 35 , and PAK4 is the earliest and most profoundly studied protein of the group II PAKs. The PAK4 gene is located on chromosome 19 at the 19q13 locus; it is 3064 bp in length and encodes 591 amino acids. Recent studies on PAK4 reveal its significant role in controlling cellular activities. It is considered to be a signaling integrator, regulating numerous fundamental cellular processes, such as actin cytoskeletal dynamics, cell morphology and motility, cell survival, embryonic development, immune defense and oncogenic transformation 36,37 . Besides, it also plays key roles in gallbladder carcinoma 38 , pancreatic cancer 39,40 , colorectal cancer 41 , lung cancer 42 and so on.
Studies show that the activation and expression levels of PAK4 plays significant roles in the development and progression of tumors. It is reported that among the 100 types of cancers affecting humans, 78% have increased  Supplementary Fig. S1).

(D) Detection of PAK4 in miR-24-1-5pexpressing cells. Western blot was conducted to examine the protein expression of PAK4 in A2780 cells stably transfected with pcDNA6.2-GW/ EmGFP-miR-24-1-5p or control plasmids (cropped; full length blots can be found in
expression of PAK4 43 . For example, PAK4 expression is significantly increased in breast cancer and is positively correlated with tumor progression 44 . This is inconsistent with our results which suggest that cells with increased expression of miR-24-1-5p and decreased expression of PAK4 demonstrate enhanced proliferation. Coarfa et al. 45 characterized the proteomic footprint of a panel of 8 miRNAs by using reverse phase protein arrays (RPPA) in prostate cancer; and illustrated that for at most 12% of the proteins, the expression level was determined by direct interaction between miRNAs and target mRNAs, but for the majority of them, various factors were involved. Thus, it can be summarized that the proliferation of ovarian cancer cells is regulated by numerous factors, miR-24-1-5p and PAK4 are only the tip of the iceberg.  Cell lines. The human ovarian cancer cell line A2780 was purchased from the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in RPMI-1640 (Gibco, USA) with 10% fetal bovine serum (FBS) (Invitrogen, USA) and 100 IU/ml gentamycin at 37 °C in a humidified atmosphere with 5% CO 2 .
The gene chip scanning. 3 normal ovarian epithelial tissue and 3 benign epithelial ovarian tumor tissue were used for RNA isolation through AMBION kit (Invitrogen, USA). After passing quality inspection, the total RNA was transcribed into double stranded cDNA, synthesized into cRNA and labeled with Cyanine-3-CTP. The labeled cRNAs were hybridized onto the microarray (Agilent SurePrint G3 Human Gene Expression v3, 8*60 K, Design ID: 072363) and scanned by the Agilent Scanner G2505C (Agilent Technologies, USA). The raw data was normalized using Extraction and GeneSpring. Differentially expressed genes were identified through fold change as well as P values calculated using Student's-test. The threshold set for up-and down-regulated genes was a fold change ≥ 2.0 and a P value ≤ 0.01. The gene chip scanning was conducted in the laboratory of the OE Biotech Company (Shanghai, China).
Real-time qPCR assay. RNA was isolated from the clinical specimens or cells using Trizol (Invitrogen, USA) and reverse transcription was performed using RNA and primers specific for mature miR-24-1-5p or LINC01088. The reverse transcription of LINC01088 was conducted at 42 °C for 60 min and then 95 °C for 5 min (PrimeScript 1st Strand cDNA Synthesis Kit, TaKaRa, Japan). Real-time qPCR of the reverse transcription products of LINC01088 was performed using Premix Ex Taq (TaKaRa, Japan). The reverse transcription of miR-24-1-5p was conducted at 16 °C for 30 min and then 42 °C for 30 min, inactivated by incubating at 85 °C for 5 min (TaqMan, USA). The reverse transcription products of miR-24-1-5p were mixed with TaqMan universal PCR master mix II (ABI, USA) and the mixture was incubated at 95 °C for 10 min, followed by 40 cycles, with an extension time of 15 s each at 95 °C and 60 s each at 60 °C. The relative LINC01088 expression was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the relative miR-24-1-5p expression was normalized to miR-484, and analysed by the 2 −ΔΔCt method. RNA co-precipitation. A2780 cells were transiently transfected with 5ug miR-24-1-5p expression plasmids labeled with biotin using Lipofectamine 3000 (Invitrogen, USA) for 48 h. The cells were then lysed in RIPA buffer and centrifuged at 14 800 rpm to remove the precipitate. A total of 30 µl biotin-avidin conjugated agarose was added to the supernatant and mixed via vortexing for 2 h at 4 °C. The mixture was then centrifuged and the supernatant was carefully discarded. 500 µl chloroform was added to the precipitate for nucleic acids extraction after it was washed thrice with RIPA buffer. PCR was performed to detect LINC01088 after reverse transcription using Oligo dT15. miR-155 labeled with biotin was used as negative control and LINC01088 as positive control.
Immunohistochemistry. Formalin-fixed, paraffin-embedded tissue sections were deparaffinized, hydrated and soaked in 3% H 2 O 2 at room temperature for 10 min. Then the slides were incubated with rabbit anti-human PAK4 antibody (10 µg/ml, Abcam, UK) overnight at 4 °C in a humidified chamber. The negative control was obtained by omitting the primary antibody. The next day, biotin-labeled sheep anti-rabbit IgG antibody was added to sections and incubated at 37 °C for 10 to 30 min. Then, the HRP streptavidin solution was added and the slides were incubated at 37 °C for 10 to 30 min. Finally, the slides were added with DAB solution, incubated for 5-10 min and counterstained with hematoxylin for 2 min and mounted.
Scratch assay. Cells in two groups were implanted into 3.5 cm 2 culture plates with approximately 90% confluence. Wounds were created with a 200 µl pipette tip and washed with PBS, then serum-free medium was reintroduced at the experiment start time point of 0 hour. The diameter of the scratch was recorded under light microscopy at 24 h.
In vivo tumor growth assay. Female Balb/c nude mice (5-6 weeks old) were randomly divided into groups (control, miR-24-1-5p, LINC01088 and miR-24-1-5p + LINC01088), and each group contained five mice. Cells transfected with LINC01088-, miR-24-1-5p-or LINC01088 + miR-24-1-5p-overexpressing plasmids were injected subcutaneously in the dorsal flank of the nude mice. The mice were observed for tumor formation by measuring the tumor major axis (a) and minor axis (b) every 7 days. The tumor volume was calculated: v = ab 2 /2. Afterwards the mice were sacrificed by cervical dislocation on the 28th day after injection, the tumors were subsequently recovered and the weight of each tumor was determined.
Data availability. All data generated or analysed during this study are included in this published article.
Ethical statement. We solemnly stated that all methods were performed in accordance with the relevant guidelines and regulations. All human and animal studies have been approved by the Ethics Committee of the Second Affiliated Hospital, College of Medicine, Zhejiang University. Written informed consent was obtained from each patient prior to inclusion in the study.