LncRNA OIP5-AS1 Affects the Malignant Biological Behavior of Ovarian Cancer via the miR-153-3p/KLF5 Axis


 Objectives: The purpose of this study was to investigate the expression and clinical significance of LncRNA OIP5-AS1 in ovarian cancer , as well as its effect on malignant biological behavior of ovarian cancer cells. Methods: The expression of OIP5-AS1, miR-153-3p and KLF5 in ovarian cancer (OC) tissues and cells were detected by RT-qPCR. Western Blotting was used to detect KLF5 expression. The expression patterns of OIP5-AS1, U6 and GAPDH in nuclear and cytoplasm fractions were detected using qRT-PCR. Besides, CCK-8 assay, clone formation assay, transwell, scratch test, and flow cytometry were respectively used to detect the cell activity, proliferation, invasiveness, healing of cells, and apoptosis rate of OC cells. Furthermore, The interaction between OIP5-AS1 and miR-153-3p and between miR-153-3p and KLF5 were verified by luciferase reporter assay, and the correlations among these three genes were analyzed.Results: OIP5-AS1 expression was up-regulated in ovarian cancer cell lines and tissues. Si-OIP5-AS1 inhibited cell proliferation, invasion and migration, and induced the apoptosis to a certain extent. Subcellular fraction assay revealed the location of OIP5-AS1 was mainly situated in the cytoplasm. In addition, miR-153-3p was a target of OIP5-AS1, and KLF5 was directly targeted by miR-153-3p. Si-OIP5-AS1 inhibited KLF5 expression, miR-153-3p inhibitor promoted KLF5 expression, and si-KLF5 inhibited OIP5-AS1 expression. Interestingly, expression of OIP5-AS1 and miR-153-3p, and expression of miR-153-3p and KLF5 were negatively correlated, while expression of OIP5-AS1 and KLF5 was positively correlated. In addition, si-KLF5 inhibited the malignant biological behavior of ovarian cancer cells, while miR-153-3p inhibitor had the opposite effect. Most importantly, the addition of si-OIP5-AS1 to mir-153-3p silenced cells could reverse the promotion effect of miR-153-3p inhibitor on the malignant biological behavior of ovarian cancer cells.Conclusions: OIP5-AS1 can be used as an effective prognostic indicator of ovarian cancer, which has the potential to be a new drug target.


Introduction
Ovarian cancer is one of the three malignancies of the female reproductive system, which is the leading cause of female death [25,19] . The ovary is located deep in the pelvic cavity, so early ovarian cancer is not easy to nd. Many patients go to the clinic with symptoms at an advanced stage. Ovarian cancer is easy to metastasize and widely spread. The 5-year survival rate still lingers at 30%~40%, and the mortality rate ranks the rst among gynecological malignant tumors [32] . Therefore, it is of great signi cance to explore the pathogenesis of ovarian cancer.
It has been reported that long-non-coding RNAs (lncRNAs) can present different types of high or low expression states in cancer. Moreover, lncRNAs play their roles through a variety of complex molecular mechanisms [20] , such as affecting the epigenetic activity of tumor cells in the nucleus, gene transcription and post-transcriptional modi cation of mRNA [2] . LncRNA OIP5-AS1 has been reported to inhibit the proliferation of multiple malignant tumors [29,28] . It can regulate a variety of biological activities of cells in some diseases and tumors. For example, lncRNA OIP5-AS1 regulates proliferation and apoptosis in multiple myeloma. LncRNA OIP5-AS1 can inhibit GAK expression [16] , thus controlling mitosis at the cytoskeleton level. It also binds to the RNA-binding protein HuR [15] , thereby inhibiting the proliferation of tumor cells [7] . Although OIP5-AS1 plays an important role in cancer, its role in ovarian cancer is still unknown. Through bioinformatics analysis, we found that OIP5-AS1 may have a potential binding relationship with miR-153-3p.
Mir-153-3p is a miRNA that plays a key role in disease progression. It has been reported that miR-153-3p plays an anticancer role in human cancers including melanoma [18] , gastric cancer [12] and ovarian cancer [3] . Through bioinformatics analysis, we found that KLF5 is the potential target gene of miR-153-3p. KLF5 expression is up-regulated in many cancers [10,4] , and our previous studies have shown that KLF5 expression is also up-regulated in ovarian cancer tissues. However, its regulatory mechanism is still unclear. We speculated that the OIP5-AS1/miR-153-3p/KLF5 axis may exist in ovarian cancer and play an important regulatory role in the development of ovarian cancer.
In this study, the expression level of OIP5-AS1/miR-153-3p/KLF5 in ovarian cancer tissues were detected, and the effect of si-OIP5-AS1 on the proliferation, invasion and apoptosis ability of ovarian cancer cells by regulating miR-153-3p/KLF5 axis was also detected, which laid the experimental foundation for the later in-depth study.

Cell culture
The human EOC cell lines HO-8910, SKOV3, A2780 and normal human ovary surface epithelial cells (HOSE) were cultured in DMEM containing 10% fetal bovine serum. The above cell lines were inoculated at the atmosphere of 37 °C and 5% CO 2 .

CCK-8 assay
The CCK-8 assay was performed to detect the proliferation of SKOV3 and A2780 cells. The speci c determination method was as follows: the cells were inoculated into a 96-well plate, each well was added with 10 µl CCK8 reagent and cultured for 24 h, and the absorbance value of 450 nm was determined by a microplate microscope.

Colony formation assay
After SKOV3 and A2780 cells was transfected for 48 h, cells of each group were seeded in a 60 mm cell culture dish with 1 × 10 3 cells/well. After 14 days, the cells were stained with crystal violet staining solution, and colonies containing ≥ 30 cells were counted. Representative colonies were photographed and counted with Olympus inverted uorescence microscope.

Flow cytometry assay
The semi-fused SKOV3 and A2780 cells were transfected and cultured for 48 h. After these ovarian cancer cells were washed with PBS solution, 100uL PBS solution was added and resuspended.
Subsequently, 5uL Annexin V-FITC and 5uL PI were added. After mixing evenly, the mixture was incubated at room temperature and away from light for 15 min. Then added 400uL binding buffer, mixed well and thus placed on ice for 1 h for up ow cytometry detection. The apoptotic cells were measured by BD Caliber ow cytometry (Becton Dickinson).
2.8 Transwell invasion assay 50 ul matrix glue was added to Transwell chamber to solidify at 37℃ for 3 h. Serum-free medium was added to the upper chamber and complete medium was added to the lower chamber. SKOV3 and A2780 cells of each group were seeded in the upper chamber of Transwell chamber and cultured for 48 h. Erased all cells that had not crossed the membrane. Paraformaldehyde was used to xed the residual cells, and then 0.5% crystal violet (Solarbio, Beijing, China) was used to stain and take photos. Assays were independently repeated three times.
2.9 Wound Healing Assay 3 × 10 5 cells/ml SKOV3 and A2780 cells were inoculated in a 6-well plates. The medium lance was used to line the cells vertically and horizontally when the cells coverd 80% of the plates. After the cells were washed with PBS solution, the cells were cultured for 2 h and 24 h in a humidi ed incubator at 37℃ and 5% CO 2 , and the scratch healing was observed under a microscope.

Nuclear and cytoplasmic RNA fraction isolation
Adherent SKOV3 and A2780 cells were placed in an ice bath and cell lysis buffer was added for 10 min. The whole experiment process was carried out according to the Ambion PARISTM kit. The lysate was collected and centrifuged at 3000r/min for 10 min (centrifugation radius was 10 cm). The supernatant was the cytoplasm and the precipitation part was mainly the nucleus. The supernatant was placed in the ice bath. The precipitate was further lysed and placed in an ice bath. Then RNA was extracted and RT-qPCR was performed.

Western blot
Total protein of each group was extracted with RIPA solution and quanti ed with BAC kit. The samples were sampled at 30ug per well and the proteins were separated by 10% SDS-PAGE. The protein was imprinted on the PVDF membrane by semi-dry transfer method and sealed with 5% skim milk for 2 h.
Then incubated with rabbit anti-KLF5 (ab137676; 1:1,000; abcam, USA). β-actin was used as inner loading control. The protein expression bands on the membrane were detected by chemiluminescence.
The second antibody was added to incubate for 2 h, and the ECL color developing solution was added drop for development.

Statistical analysis
SPSS19.0 statistical software was used for Student's t-test and one-way ANVOA. GraphPad Prism 6.0 software was used for graphing. Measurement data were expressed as mean ± standard deviation.

LncRNA OIP5-AS1 was upregulated in ovarian cancer tissues and cells
Primarily, 30 pairs of OC tissue samples and normal para-tumor tissue samples were collected. The results showed that OIP5-AS1 was signi cantly up-regulated in OC tissues compared with normal paratumor tissues (Fig. 1A). RT-qPCR results also showed that OIP5-AS1 was signi cantly up-regulated in OC cell lines (HO-8910, SKOV3 and A2780) compared with normal ovarian epithelial cell line (HOSE) (Fig. 1B).

Knockdown of OIP5-AS1 inhibited the proliferation, invasion, migration and promoted the apoptosis of ovarian cancer cells
To investigate the biological roles of OIP5-AS1 in the behavior of ovarian cancer cells, we used small interfering RNA (siRNA) to perform the following functional de ciency tests. OIP5-AS1 expression was silenced when si-OIP5-AS1 (knockdown of OIP5-AS1) was transfected into SKOV3 and A2780 cells ( Fig. 2A). Furthermore, comparing with si-NC group, CCK-8 assay revealed that si-OIP5-AS1 in the si-OIP5-AS1 group signi cantly reduced the proliferative ability of OC cells (Fig. 2B). Colony formation assay showed that si-OIP5-AS1 in the si-OIP5-AS1 group signi cantly reduced the mumber of cloning in the si-OIP5-AS1 group ( Fig. 2C and D). Flow cytometry analysis revealed that si-OIP5-AS1 in the si-OIP5-AS1 group signi cantly increased apoptotic rate of SKOV3 and A2780 cells ( Fig. 2E and F). Transwell invasive assay illustrated that si-OIP5-AS1 in the si-OIP5-AS1 group signi cantly inhibited the invasive ability of SKOV3 and A2780 cells ( Fig. 2G and H). The scratch test revealed that si-OIP5-AS1 in the si-OIP5-AS1 group signi cantly inhibited the migration ability of SKOV3 and A2780 cells (Fig. 2I, and H). Thus, we conclude that knockdown of OIP5-AS1 inhibited the proliferation, invasion, migration and promoted the apoptosis of ovarian cancer cells.
3.3 MiR-153-3p was a target of LncRNA OIP5-AS1 To determine the subcellular location of OIP5-AS1, the locations of OIP5-AS1 in SKOV3 and A2780 cells were analyzed by the subcellular of RNA fractions analysis. The RT-qPCR results showed that, the location of OIP5-AS1 was mainly situated in the cytoplasm. This nding sent out the idea that OIP5-AS1 might function as the 'sponge' of miRNAs. To indentify the interaction between OIP5-AS1 and miR-153-3p, in the previous stage, we predicted that OIP5-AS1 may have a potential binding relationship with miR-153-3p through bioinformatics online tools analysis (TargetScan, http://www.targetscan.org/vert_71/).

Discussion
Targeted therapy is a new method for the treatment of malignant tumors, which can inhibit the proliferation of cancer cells through speci c factors or even directly kill cancer cells. Therefore, it is of great importance to study the molecular mechanism of the occurrence and development of ovarian cancer for the molecular targeted therapy of ovarian cancer [22] . In recent years, many studies have pointed out the abnormal expression of lncRNA in various types of tumors [21,6,11] . As a new regulatory factor, lncRNA is currently a hot research eld in oncology, and its biological functions are mainly expressed as the regulation of gene activation or gene expression inhibition [27,13] . LncRNA OIP5-AS1 is a highly conserved LncRNA in mammals, which is mainly expressed in the cytoplasm [26] . It has been reported that OIP5-AS1 may act as a biological sponge, acting as competing endogenous RNAs (ceRNA) to inhibit the binding mRNA of RNA-binding protein HuR (HuR was a member of the human embryo lethal abnormal visual RBP family), and then inhibited the expression of HuR, thus participating in the regulation of phenotypes [15] . In addition, miRNAs play an important role in regulating gene expression and cell function, and there are many reports on their involvement in tumor genesis and development.
Regulatory networks of non-coding RNA are involved in the development of many types of tumors [34] . Studies have shown that lncRNAs and other transcriptional substances can bind to miRNAs through miRNA recognition elements and play the role of ceRNA [14,17] . For instance, studies have reported that, OIP5-AS1 is a ceRNA in Hepatoblastoma cells through modulating miR-186a-5p/ZEB1 [35] . For another example, OIP5-AS1 as ceRNA positively regulates PAPPA expression through the spongization of miRNA-152-3p [31] . Our research team reported that LncRNA OIP5-AS1 acted as a ceRNA to drive proliferation, invasion, migration and apoptosis of ovarian cancer cells via the spongization of miR-153-3p/KLF5 axis.
In this study, we found that OIP5-AS1 expression was up-regulated in ovarian cancer cell lines and tissues. The abnormal expression of OIP5-AS1 in ovarian cancer cells indicates that oIP5-AS1 may trigger a series of pathophysiological processes as an oncogene. The knockdown of OIP5-AS1 inhibited the proliferation, invasion and migration of ovarian cancer cells, and promoted the apoptosis of ovarian cancer cells to a certain extent. OIP5-AS1 has been identi ed as an oncogene of human diseases in a variety of cancer cells, such as breast cancer cells [30] , non-small cell lung cancer cells [5] , pancreatic ductal carcinoma cells [29] , and multiple myeloma cells [28] .
In order to further study the deep mechanism of OIP5-AS1 regulating the occurrence of ovarian cancer tumors, the locations of OIP5-AS1 in SKOV3 and A2780 cells were analyzed by the subcellular of RNA fractions analysis. We found that OIP5-AS1 was mainly localized in the cytoplasm, suggesting its posttranscriptional regulation. This nding stimulated the idea that lncRNA OIP5-AS1 might function as the 'sponge' of miRNAs. To indentify the interaction between OIP5-AS1 and miR-153-3p, in the previous stage, we predicted that OIP5-AS1 may have a potential binding relationship with miR-153-3p through bioinformatics online tools analysis. Luciferase gene reporter assay con rmed that OIP5-AS1 was a sponge of miR-153-3p. Our hypothesis was con rmed that lncRNA CASC15 could act as the "sponge" of miRNA. Subsequently, on the basis of previous studies, we further veri ed that KLF5 mRNA and its product proteins were targets of miR-153-3p through luciferase gene reporter assay, which was consistent with previous research results [33] , thus constructed the OIP5-AS1/ miR-153-3p/KLF5 axis. KLF5 is one of the most studied members of the KLFs family. It belongs to the family of DNA-binding transcriptional regulators and has various biological functions. In recent years, studies have found that KLF5 is closely related to the occurrence and development of tumors. KLF5 is abnormally expressed in different types of tumors and plays different roles in different tumors [36,23,1] . Thus, on the other hand, RT-qPCR was also used to detect the expressions of miR-153-3p and KLF5 in ovarian cancer cell lines and tissues, and we found that a signi cant down-regulation of miR-153-3p and a signi cant up-regulation of KLF5 were shown in ovarian cancer. Interestingly, a negative correlation between OIP5-AS1 and miR-153-3p expression, a positive correlation between OIP5-AS1 and KLF5 expression, and a negative correlation between miR-153-3p and KLF5 expression were found in ovarian cancer tissues.
Through the above luciferase reporter gene validation, we con rmed that miR-153-3p was a target of LncRNA OIP5-AS1, and KLF5 was directly targeted by miR-153-3p. Next, we found that the interaction of OIP5-AS1/miR-153-3p/KLF5 axis related genes played an important role in the malignant biological behavior of ovarian cancer cells. Si-KLF5 inhibited the malignant biological behavior of ovarian cancer cells, as shown in our study. Our ndings are further supported by the fact that KLF5 is an important transcription factor involved in the survival, migration, invasion and apoptosis of osteosarcoma cells [8] . What's more, our study also found that miR-153-3p inhibitor promoted the malignant biological behavior of ovarian cancer cells, while the addition of si-OIP5-AS1 to mir-153-3p silenced cells could reverse the promotion effect of miR-153-3p inhibitor on the malignant biological behavior of ovarian cancer cells, suggesting that OIP5-AS1 acted as a ceRNA to drive proliferation, invasion, migration and apoptosis of ovarian cancer cells via the spongization of miR-153-3p/KLF5 axis. A new study reports that, OIP5-AS1 promoted the progression of gastric cancer cells through miR-153-3p/ZBTB2 axis, and its research target between OIP5-AS1 and miR-153-3p was surprisingly consistent with our results [9] . Moreover, OIP5-AS1 knockdown inhibits breast cancer progression by competitively binding miR-216a-5p [30] . Down-regulation of OIP5-AS1 can inhibit the proliferation and migration of hepatocellular carcinoma cells and promote apoptosis via regulating miR-3163/VEGFA [24] .
As mention above, our data suggested that lncRNA OIP5-AS1, by sponging miR-153-3p/KLF5 axis, depressed cell proliferation, invasion and migration, and induced the apoptosis to a certain extent, suggesting an important regulation of OIP5-AS1 on the ovarian cancer tumorigenesis.

Conclusions
OIP5-AS1 can be used as an effective prognostic indicator of ovarian cancer. By elucidating the mechanism of OIP5-AS1 in the malignant biological behavior of ovarian cancer, we demonstrate that OIP5-AS1 has the potential to be a new drug target.

Declarations
Authors' contributions SLW, CCL, ARZ, YCL, and YXX conceived and designed the experiments. SLW, CCL, JWQ and XLC performed the experiments. JB and RL analyzed the data and contributed to the acquisition of reagents and materials. SLW and YXX wrote the manuscript.        collected. RT-qPCR assay was used to measure KLF5 expression in tissues. ***p<0.001. (D) RT-qPCR assay was used to measure KLF5 expression in OC cell lines (HO-8910, SKOV3 and A2780) and normal ovarian epithelial cell line (HOSE). *p<0.05, compared with HOSE group. (E) RT-qPCR assay for the correlation between miR-153-3p level and KLF5. (F) RT-qPCR assay for the correlation between OIP5-AS1 level and KLF5.   (E) RT-qPCR assay revealed that KLF5 siRNA signi cantly decreased the expression of OIP5-AS1 in SKOV3 and A2780 cells. *p<0.05, compared with si-NC group. (E) RT-qPCR assay revealed that KLF5 siRNA signi cantly decreased the expression of OIP5-AS1 in SKOV3 and A2780 cells. *p<0.05, compared with si-NC group.