Tumor (n = 25) and normal (n = 17) tissue were obtained from women undergoing surgery for EOC between 2015 and 2019 at our institution. None had received preoperative chemoradiotherapy. Histopathology was used to confirm specimens prior to storage in a liquid nitrogen tank at –196 °C until the extraction of RNA. All patients gave informed consent. The major clinicopathological features of cohort are indicated below (Table 1). For the relationship between MYB, miR-150-5p and clinicopathological features, see Supplementary Tables 1,2. Ethical approval for the study was given by the host institute (No. 2017AE02133, approval date December 3, 2017) and the Declaration of Helsinki guidelines were followed.

The human cell lines studied here were provided by the Chinese Science Academy (Shanghai, China) and Zhong Qiao Xin Zhou (Shanghai, China). These consisted of normal ovarian surface epithelial cells (IOSE80, ZQ0721, Zhong Qiao Xin Zhou, Shanghai, China) and three EOC cell lines ((A2780, ZQ0486, Zhong Qiao Xin Zhou, Shanghai, China), (SKOV3, TCHu185, Chinese Science Academy, Shanghai, China) and (OVCAR3, TCHu228, Chinese Science Academy, Shanghai, China)). SKOV3 was grown in McCoy’s 5A medium (Gibco, New York, USA) with 10% fetal bovine serum (FBS), while remaining cell lines were grown in Dulbecco’s modified Eagle medium (DMEM) with high-glucose and 10% FBS. Cells were kept at 37 °C in a humidified incubator with 5% CO${}_2$ concentration. Mycoplasma testing has been done for the cell lines used. The cell lines used have been authenticated and Short Tandem Repeat (STR) was used for the authentication.

RNA was extracted with Trizol (Invitrogen, Carlsbad, USA) and cDNA for qPCR was prepared using HiScript II SuperMix (Vazyme, Nanjing, China). qRT-PCR (Q6, Life) was carried out with SYBR® Green PCR Master Mix (Applied Biosystems, Waltham, USA) and primer sequences as follow: PART1 forward 5${}^{\prime }$CTCCTGCGGTTTCCCATT3${}^{\prime }$, reverse 5${}^{\prime }$ATCTCACCAGACACCTGCCTAC3${}^{\prime }$; miR-150-5p forward 5${}^{\prime }$CGGGCTCTCCCAACCCTTGT3${}^{\prime }$, reverse 5${}^{\prime }$CAGCCACAAAAGAGCACAAT3${}^{\prime }$; MYB forward 5${}^{\prime }$GGCACACAAGAGACTGGGGA3${}^{\prime }$, reverse 5${}^{\prime }$CGACCTTCCGACGCATTGTA3${}^{\prime }$; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward 5${}^{\prime }$ACCCACTCCTCCACCTTTGAC3${}^{\prime }$, reverse 5${}^{\prime }$TGTTGCTGTAGCCAAATTCGTT3${}^{\prime }$; U6 forward 5${}^{\prime }$CTCGCTTCGGCAGCACA3${}^{\prime }$, reverse 5${}^{\prime }$AACGCTTCACGAATTTGCGT3${}^{\prime }$. The cycling conditions were: stage 1 (pre-denaturation): 95 °C, 10 min; stage 2 (cyclic reaction, 40 times): 95 °C for 10 sec; 60 °C for 30 sec; stage 3 (dissolution curve): 95 °C for 15 sec; 60 °C for 60 sec; 95 °C for 15 sec. Expression levels for lncRNA, miRNA and mRNA were detected quantitatively by the 2${}^{-\mathrm{\Delta }\mathrm{\Delta }\text{Ct}}$ method [22], with GAPDH the internal references.

StarBase v3.0 (https://starbase.sysu.edu.cn/index.php) and miRcode (http://www.mircode.org/index.php) were employed to predict potential miRNAs that bind to PART1 [23, 24]. StarBase v3.0 was also used to identify candidate mRNAs that bind to miR-150-5p. This software integrates 5 different miRNA prediction softwares (TargetScan, PicTar, microT, miRmap, miRanda/mirSVR). The GEPIA database was used to analyze the expression levels of the target genes for miR-150-5p [25].

PART1-shRNA-lentiviruses (sequence: sh1:GAAC TCAATTACGACTACATA; sh2:CCAGAUGAGACUAC GAUAATT; sh3:GAACAGAGUUGACUUUGUGTT; sh4:GCAAAGUAUCCAAGACCAATT) used for the in vivo experiments were provided by GenePharma (Shanghai, China), as were the siRNAs that target PART1. Transfection of cells was performed with Lipofectamine 3000 (Invitrogen, Carlsbad, USA). MiR-150-5p inhibitors (sequence: CACUGGUACAAGGGUUGGGAGA) were also from GenePharma. Triplicate experiments were performed throughout.

CCK-8 assay was used to quantify the proliferation rate of cells. Following inoculation into 96-well culture plates (1.5 $\times$ 10${}^3$ cells/well), cells were grown in 100 µL DMEM containing 10% FBS. CCK-8 solution (10 µL) was added 0, 24, 48, 72 and 96 h after cellular attachment. The 450 nm absorbance was subsequently read following 2 h of incubation by Enzyme label (Multiskan FC, Thermo, Waltham, USA).

Transwell chambers (Millipore, Darmstadt, Germany) used for cell invasion assays were first pretreated with 50 µL of a 1:9 Matrigel/DMEM solution (BD, New Jersey, USA). Subsequently, 1 $\times$ 10${}^5$ cells were dispersed in DMEM without FBS (1 mL) and 200 µL of cellular solution was placed in the upper chamber. Following this, DMEM containing 10% FBS (600 µL) was placed into the lower chamber to act as a chemotactic agent. 48 h later, residual cells in the upper chamber were scraped off, while invading cells were immobilized in 4% paraformaldehyde then dyed in 2% crystal violet. The invading cells were counted by light microscopy (D-35578, Leica, Weztlar, Germany). The experimental procedure used to quantify the migration of cells was identical to the above, but with no Matrigel pretreatment.

Cells were dispersed into a single-cell suspension at 48 h after transfected. The colony forming assay was carried out by incubating 1 $\times$ 10${}^3$ cells at 37 °C for two weeks in a culture dish containing 10% FBS medium. The cells were then stabilized, dyed using 0.1% crystal violet, and the colonies counted manually. Experiments were repeated three times for all groups.

Five $\times$ 10${}^5$ cells per well were inoculated in 6-well plates, grown for 48 h, then rinsed in PBS and subsequently immersed in ice-cold lysis buffer. The concentration of protein in the lysate was quantified with the bicinchoninic acid (BCA) method (P0012S, Beyotime, Shanghai, China). Protein from each sample was separated with 10% SDS-PAGE then transferred to polyvinylidene fluoride film (162017, Bio-Rad, Hercules, USA). The film was subsequently incubated for 2 h with a solution of 5% non-fat dried milk in TBS with 0.1% Tween-20 to block non-specific proteins. MYB- (1:2000, ab109127, abcam, Cambridgeshire, UK) or GAPDH-specific antibody (1:5000, 60004-1-Ig, Proteintech, Wuhan, China) was then incubated with the polyvinylidene fluoride film for 24 h at 4 °C. The film was subsequently rinsed before incubation for 2 h with horseradish peroxidase (HRP)-conjugated secondary antibody (1:5000, Proteintech, 60004-1-Ig, Wuhan, China) at 37 °C. Immunoblot analysis was performed using an enhanced chemiluminescence reagent (34579, Invitrogen, Carlsbad, USA) and the film was subsequently irradiated with X-rays by WB Imaging Instrument (5260, Tanon, Shanghai, China). Image J software (1.8.0, NIH, USA) was used to analyse protein expression levels, with GAPDH used as the reference.

A2780 and SKOV3 cell lines were grown in 6-well plates. Wild-type (WT) and mutant fragments (MUT) from the 3${}^{\prime }$-untranslated region of PART1 (PART1- WT: GUAAUCCCAGCACUUUGGGAGG; PART1-MUT: GUAAUCCCAGCACUCGCA CGAG) and MYB (MYB-WT: GAAACUUUUCAUGAAUGGGAGA; MYB-MUT: GAAACUUUUCAUGAACACACCA) that were related to the binding site for miR-150-5p were synthesized and added to pMIR-REPORT vectors. Reporter gene assay was conducted as per our previous study [26]. Normalization of luciferase activity was carried out in relation with Renilla luciferase activity.

Female BALB/c nude mice aged 4–5 weeks were obtained from Animal Core Facility, Nanjing Medical University. Each mouse (5 mice per group) received a subcutaneous injection in the right armpit with a 200 µL suspension containing 5 $\times$ 10${}^5$ A2780 ovarian cancer cells. Tumor growth rate was evaluated at regular intervals starting from the 6th day. Tumor size was quantified with vernier calipers, where L was the longest diameter and W the longest transverse diameter perpendicular to the longest diameter when the tumor is viewed as an ellipse. Tumor volume was subsequently estimated with the formula: V = $\pi$/6$\times$L$\times$W$\times$W. Animal experimentation conformed with the guidelines for “Animal Research: Reporting of In Vivo Experiments (ARRIVE)”. The animal ethics approval number was 1811051, and the approval date was April 2018.

Experiments were conducted in triplicate. SPSS 15.0 (IBM, Chicago, USA) was used for statistical analyses, with values expressed as the mean $\pm$ standard deviation (SD). One-way-ANOVA analysis of variance was used to evaluate the differences among at least three groups. Least Significance Difference (LSD) was used with ANOVA. Student’s t test was used to determine the differences between groups. Pearson correlation analysis was used to analyse correlations between data. The normal distribution was tested using the Shapiro-Wilk (S-W test) by SPSS. p-values 0.05 were considered to show statistical significance.

LncRNA PART1 expression in clinical specimens was quantified by qRT-PCR. Fig. 1A shows the expression was significantly higher in EOC relative to normal ovarian surface epithelial (OSE) tissue. LncRNA PART1 expression was also higher in advanced stage EOC relative to lower stage tumors (Fig. 1B). Finally, the expression level in three EOC cell lines (OVCAR3, SKOV3, A2780) was greater relative to the OSE cell line (IOSE80), particularly in SKOV3 and A2780 cells (Fig. 1C).

Fig. 1.

Elevated expression of lncRNA PART1 in EOC detected with qRT-PCR. (A) PART1 expression in EOC and normal ovarian tissue. (B) PART1 expression in EOCs at different FIGO stages. (C) PART1 expression in EOC (OVCAR3, SKOV3, A2780) and ovarian surface epithelial (OSE) (IOSE80) cells. *p 0.05, **p 0.01.

PART1 expression in SKOV3 and A2780 cells was reduced following transfection with a knockdown vector (sh-PART1). A control vector (sh-NC) was also used, and transfection efficiency confirmed using qRT-PCR. Among them, sh-1, sh-2, sh-3 and sh-4 were compared with sh-NC, respectively, and sh-2 had the best transfection efficiency (Fig. 2A,B). Results from colony forming assays and CCK-8 assays demonstrated that lncRNA PART1 knockdown inhibited EOC cell proliferation (Fig. 2C–E). Furthermore, transwell assay revealed that lncRNA PART1 knockdown also inhibited the migration and invasion ability of cells (Fig. 2F–I). Hence, the above results imply that PART1 enhances the proliferation, migration and invasion of EOC cells in vitro.

Fig. 2.

LncRNA PART1 knockdown reduces the proliferation, migration and invasion of A2780 and SKOV3 cells. (A,B) PART1 expression in cells transfected with sh-PART1 (including sh-PART1-1, sh-PART1-2, h-PART1-3 and sh-PART1-4) as measured by qRT-PCR. (C) Proliferation of cells following transfection with sh-PART1 or sh-NC (control), as determined using colony-forming assay. (D,E) Proliferation of cells following transfection with sh-PART1 or sh-NC, as determined by CCK8 assay. (F,G) Migration of cells following transfection with sh-PART1 or sh-NC, as determined by transwell assay. (H,I) Invasion of cells following transfection with sh-PART1 or sh-NC, shown using transwell assay. *p 0.05, **p 0.01.

The findings from starBase revealed the presence of binding sites between lncRNA PART1 and miR-150-5p (Fig. 3A). Luciferase reporter gene assay also demonstrated that miR-150-5p decreased luciferase activity in lncRNA PART1-WT cells, but not lncRNA PART1-MUT cells (Fig. 3B). miR-150-5p expression was significantly reduced in EOC tissue relative to normal tissue (Fig. 3C). Pearson’s analysis demonstrated an inverse correlation between PART1 and miR-150-5p levels in EOC tissue (Fig. 3D). Similar results to EOC tissues were obtained for miR-150-5p expression in OSE and EOC cells. The expression levels in the three EOC cell lines (OVCAR3, SKOV3, A2780) were lower than those in the OSE80 cell line, especially in SKOV3 and A2780 cells (Fig. 3E). miR-150-5p expression was increased in SKOV3 and A2780 cells in which lncRNA PART1 was down-regulated (Fig. 3F). Together, these results indicate lncRNA PART1 is able to sponge miR-150-5p in EOC cells.

Fig. 3.

LncRNA PART1 can sponge miR-150-5p. (A) Putative binding sites between miR-150-5p and PART1 identified with online databases. (B) miR-150-5p reduces luciferase activity of PART1-WT, but not PART1-MUT, as shown by luciferase reporter gene assay. (C) miR-150-5p expression in EOC and normal ovarian tissue, as quantified by qRT-PCR. (D) Correlation analysis of relative expression levels for PART1 and miR-150-5p in clinical specimens. (E) miR-150-5p expression levels in EOC cells (OVCAR3, SKOV3, A2780) and OSE cells (IOSE80), as measured with qRT-PCR. (F) miR-150-5p expression in A2780 and SKOV3 cells after transfection with sh-PART1 or with sh-NC (negative control), as measured with qRT-PCR. *p 0.05, **p 0.01.

Functional assays were performed to determine whether lncRNA PART1 acts in A2780 and SKOV3 cells by binding to and therefore effectively removing miR-150-5p. Transfection efficiencies for miR-150-5p inhibitor, sh-PART1 and corresponding controls were confirmed using qRT-PCR (Fig. 4A,B). CCK-8 assays revealed that lncRNA PART1 down-regulation decreased EOC cell proliferation. This was partially reduced by miR-150-5p inhibitor (Fig. 4C,D). Similarly, results from transwell assays revealed miR-150-5p inhibitor can attenuate inhibition of migration and invasion caused by lncRNA PART1 downregulation (Fig. 4E–H). These findings suggest that lncRNA PART1 can enhance EOC cell proliferation, migration and invasion through binding and thus inactivation of miR-150-5p.

Fig. 4.

LncRNA PART1 enhances the progression of EOC by binding miR-150-5p. (A,B) miR-150-5p expression in A2780 and SKOV3 cells after transfection with miR-150-5p inhibitor, sh-PART1 or the corresponding controls, as determined with qRT-PCR. (C,D) A2780 and SKOV3 cell proliferation after transfection with miR-150-5p inhibitor, sh-PART1 or the corresponding controls, as measured with the CCK8 assay. (E,F) Migration of A2780 and SKOV3 cells following transfection with miR-150-5p inhibitor, sh-PART1 or the corresponding controls, as measured with the transwell assay. (G,H) Invasion of A2780 and SKOV3 cells following transfection with miR-150-5p inhibitor, sh-PART1 or the corresponding controls, as measured with the transwell assay. *p 0.05, **p 0.01.

We investigated miR-150-5p target genes to test the ceRNA hypothesis. 22 genes were screened out by StarBase v3.0. Then three of these genes, MTCH2, MYB and NDC1, were found to be elevated in ovarian cancer tissues relative to normal ovarian tissues through the GEPIA database (Supplementary Fig. 1). MYB is one of the more classical, malignant progression-promoting oncogenes in a variety of tumors, such as: breast, liver, colon and lung cancers [27, 28, 29, 30]. Next, MYB was selected as the object of study (Fig. 5A). The luciferase reporter gene assay revealed that miR-150-5p decreased luciferase activity for MYB-WT, but there was no significant effect for MYB-MUT (Fig. 5B). The expression of MYB was elevated in EOC compared to OSE tissue (Fig. 5C). Moreover, MYB expression was inversely associated miR-150-5p expression in EOC specimens, as shown by Pearson’s correlation analysis (Fig. 5D). Three EOC cell lines (OVCAR3, SKOV3, A2780) were compared with OSE cell line (OSE80), respectively. qRT-PCR also demonstrated that MYB expression in EOC cells was higher than in OSE cells (Fig. 5E). Together, the data indicate that MYB is a target for miR-150-5p. Functional assays showed the impacts of si-MYB on the proliferation, migration and invasion of EOC cells were partly attenuated by miR-150-5p inhibitor (Fig. 6A–H). Overall, these findings demonstrate that miR-150-5p can reduce proliferation, migration and invasion of EOC cells through targeting MYB.

Fig. 5.

MYB is a target for miR-150-5p. (A) Putative binding sites for miR-150-5p on MYB, identified with online databases. (B) miR-150-5p reduces luciferase activity of MYB-WT in EOC cell lines, but not for MYB-MUT, as shown by luciferase reporter assay. (C) MYB expression levels in EOC and normal ovarian tissue, as measured with qRT-PCR. (D) Correlation between MYB and miR-150-5p expression in clinical specimens, as determined using Pearson’s analysis. (E) MYB expression in EOC (OVCAR3, SKOV3, A2780) and OSE (IOSE80) cells, as measured with qRT-PCR. **p 0.01.

Fig. 6.

MYB enhances the proliferation, migration and invasion of A2780 and SKOV3 cells in vitro. (A,B) MYB expression in A2780 and SKOV3 cells following transfection with si-MYB, or si-MYB plus miR-150-5p inhibitor, as measured with qRT-PCR and Western blotting. (C,D) Proliferation of A2780 and SKOV3 cells following transfection with si-MYB, or si-MYB plus miR-150-5p inhibitor, as measured with the CCK8 assay. (E,F) Migration of A2780 and SKOV3 cells following transfection with si-MYB, or si-MYB plus miR-150-5p inhibitor, as determined using the transwell assay. (G,H) Invasion by A2780 and SKOV3 cells following transfection with si-MYB, or si-MYB plus miR-150-5p inhibitor, as measured using transwell assays. **p 0.01.

To investigate whether lncRNA PART1 is involved in EOC growth in vivo, female nude mice received subcutaneous injections of A2780 cells following transfection with sh-PART1 or sh-NC (negative control). Tumor size was estimated every four days and the tumor growth rate determined after 30 days. sh-PART1 was found to suppress cancer growth in the nude mice compared to controls (Fig. 7A–C). In addition, mean tumor volume and weight were smaller with sh-PART1(Fig. 7D,E). Overall, these finding indicate that lncRNA PART1 can enhance the growth of EOC in vivo.

Fig. 7.

MYB promotes in vivo growth of EOC. (A,B) EOC growth in mice injected with A2780 cells that had been transfected with sh-PART1 or with sh-NC. (C–E) PART1 downregulation resulted in a lower tumor growth rate (volume and weight) than observed in controls. **p 0.01.