STAT3-induced upregulation of lncRNA MEG3 regulates the growth of cardiac hypertrophy through miR-361-5p/HDAC9 axis

Cardiac hypertrophy is closely correlated with diverse cardiovascular diseases, augmenting the risk of heart failure and sudden death. Long non-coding RNAs (lncRNAs) have been studied in cardiac hypertrophy for their regulatory function. LncRNA MEG3 has been reported in human cancers. Whereas, it is unknown whether MEG3 regulates the growth of cardiac hypertrophy. Therefore, this study aims to investigate the specific role of MEG3 in the progression of cardiac hypertrophy. Here, we found that MEG3 contributed to the pathogenesis of cardiac hypertrophy. MEG3 expression was remarkably strengthened in the mice heart which undergone the transverse aortic constriction (TAC). Moreover, qRT-PCR analysis revealed that MEG3 was upregulated in the cardiomyocytes which were treated with Ang-II. Silenced MEG3 inhibited the increasing size of hypertrophic cardiomyocytes and reversed other hypertrophic responses. Mechanically, MEG3 could affect cardiac hypertrophy by regulating gene expression. Mechanically, we found that MEG3 could be upregulated by the transcription factor STAT3 and could regulate miR-361-5p and HDAC9 by acting as a ceRNA. Finally, rescue assays were made to do further confirmation. All our findings revealed that STAT3-inducetd upregulation of lncRNA MEG3 controls cardiac hypertrophy by regulating miR-362-5p/HDAC9 axis.


Materials and Methods
Animals and histological analysis. Vital River Laboratory Animal Company (Beijing, China) provided C57BL6 mice (SPF, male, 18-25 g, 8 weeks) for this experiment. All experiments were conducted in accordance with the principles approved by the institutional Animal Care and Use Committee of The Third Affiliated Hospital of Guangxi Medical University. The experimental procedures were conducted abiding by the Care and Use of Laboratory Animals published by The Third Affiliated Hospital of Guangxi Medical University. A total of 10 CH mouse models used in this study were induced by transverse aortic constriction (TAC). This step conformed to a previous study 8 . In brief, intraperitoneal ketamine (85 mg/kg) and xylazine (10 mg/kg) were used to anesthetize the mice. Then, the transverse thoracic aorta of the mice was dissected and the surgery silk thread and a 26-gage needle were used to suture the wound. The sham group was generated in the age-matched mice whose aorta was not sutured. All mice recovered on a ventilator at 37 °C after the operation. One week later, the hearts of mice were collected for further analysis.
Cardiomyocyte culture and treatment. Cardiomyocytes were isolated from newborn mice hearts as previously described 14,15 . After the quick washing, the isolated hearts were sliced into some small pieces. Then the tissues were transferred into HEPES-buffered saline solution complex (0.1% trypsin +0.14 mg/ml collagenase) (Roche, USA) at 37 °C. Dulbecco's modified Eagle's medium/F12 (Invitrogen, USA) was utilized for re-suspending the dissociated cells. After centrifugation, cardiomyocytes were pre-plated at 37 °C for 1 h for isolation. Next, cardiomyocytes were collected for the next experiments. The expression level of MEG3 was treated with different dose of angiotensin II (Ang-II, Sigma, USA). The highest expression of MEG3 was observed in 150 nM. Therefore, subsequent experiments were carried out in cardiomyocytes which were treated with 150 nM of Ang-II.
Immunofluorescence staining and cell surface area assay. At  RNA pull-down assay. Cardiomyocytes were transfected with biotinylated miRNA. Two days later, cells were collected for next steps. M-280 streptavidin magnetic beads (Invitrogen) was utilized for incubation of the cell lysates. We purified the bound RNAs with TRIzol reagent (Invitrogen) for qRT-PCR analysis.
Dual luciferase reporter assay. Cells were seeded in 24-well plates (3 × 10 4 cells/well). The next day, the cells were co-transfected with pmirGLO-MEG3-WT or pmirGLO-MEG3-MUT reporter plasmids and miR-361-5p mimics. The sequence of MEG3 containing the predicted binding site with miR-361-5p; then we cloned the putative and mutated sequences of the binding site into a pmirGLO Dual-luciferase miRNA target expression vector to shape the reporter vector pmirGLO-MEG3 -wild type (MEG3-WT) and pmirGLO-MEG3 -mutated type (MEG3-MUT). Lipofectamine 2000 was taken advantage of co-transfecting miR-142-3p mimics or miR-NC into the pmirGLO-MEG3-WT or pmirGLO-MEG3-MUT in cells. Two days after transfection, Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) enabled us to measure the relative luciferase activity.
Western blot analysis. Proteins were lysed from heart tissues and cells with RIPA buffer (Beyotime, China) on ice and they were centrifuged at 4 °C for 20 min. Protein concentration was determined with a BCA protein assay kit (Beyotime, China). Subsequently, SDS-PAGE was utilized to separate proteins (30 μg each lane). Then proteins were transferred onto PVDF membranes (Millipore, USA). Afterwards, 5% defatted milk was used to block the membranes in TBST buffer. Next, the membranes were exposed to primary antibodies at 4 °C overnight.
After washing, the membranes were incubated with secondary antibodies for 45 min at normal temperature. GAPDH was used as an internal control. Finally, an enhanced chemiluminescence detection system (Pierce, Rockford, USA) was used to detect proteins.
Statistical analysis. Data were displayed as the mean ± SD from at least three independent experiments.
All experiments were independently conducted at least three times. Statistical significance between two groups was determined with the Student's t test. Experimental data among multiple groups were compared and analyzed with One-way ANOVA (SPSS 20.0). Expression correlation was analyzed by Spearman's correlation analysis. Differences were considered to be statistically significant only when P value less than 0.05.

MEG3 plays a positive role in CH in vivo.
In has been reported that the abnormal expression of MEG3 is closely related to the occurrence and progression of human diseases 11,16 . This study aims to explore the special function of MEG3 in cardiomyocyte hypertrophy. According to previous report, we knew that angiotensin II (Ang-II) can induce CH 17 . Here, we treated the cardiomyocytes with Ang-II. qRT-PCR was used to explore whether the expression of MEG3 in cardiomyocytes can be affected by the dose change of Ang-II (0, 50, 100, 150, 200 nM). As shown in Fig. 1A, the highest level of MEG3 was tested when cardiomyocytes were treated with 150 nM Ang-II. Therefore, we chose this dose of Ang-II to do next experiments. To obtain adequate evidence, we detected MEG3 expression in different cardiac cells (without Ang-II). As a result, the expression of MEG3 was more abundant in fibroblast than that in myocyte (Fig. 1B). Next, MEG3 was silenced in cardiomyocytes which were treated with or without Anh-II by specific shRNAs (sh-MEG3) and control shRNA (sh-NC) (Fig. 1C). The transfection efficiency was obtained after 48 hours. Immunofluorescence staining with anti-α-actin was used to detect the influence of silenced MEG3 on the cell surface area of cardiomyocytes. It was uncovered that Ang-II-induced bigger cell surface area was reversed by sh-MEG3 (Fig. 1D). Atrial natriuretic factor (ANF), brain natriuretic peptide (BNP), and β-myosin heavy chain (β-MHC) are three hypertrophic markers. Upregulation of these three markers in cardiomyocytes are taken as common cardiac hypertrophy responses. The expression levels of β-MHC, ANF and BNP were decreased by sh-MEG3 (Fig. 1E). All the above findings suggested that MEG3 positively regulates cardiac hypertrophy.
MEG3 is upregulated by the transcription factor STAT3. Based on the data of Fig. 1, high expression MEG3 might be a promoter in cardiac hypertrophy. Here, we investigated the molecular mechanism underlying MEG3 upregulation. The search result of UCSC (http://genome.ucsc.edu) suggested that STAT3 was found to be the upstream transcription factor of MEG3. The binding motif of STAT3 was obtained from JASPAR (http:// jaspar.genereg.net/) and illustrated in Fig. 2A. Top five bindings sequences of STAT3 to three parts of MEG3 promoter were shown in Fig. 2B. Subsequently, ChIP analysis revealed the affinity of STAT3 to the part 2 (P2) of MEG3 promoter (Fig. 2C). Luciferase activity analysis was used to make further confirmation. The result indicated that the luciferase activity of vector containing mutated biding site 2 (−1068~−1078) was not increased by STAT3 (Fig. 2D), indicating the binding relation between STAT3 and MEG3 promoter in site 2. The level of STAT3 was then examined in ten pairs of smouse hearts treated with Sham or TAC. Not surprisingly, STAT3 was expressed higher in TAC group (Fig. 2E). The positive expression correlation between STAT3 and MEG3 in TAC group was analyzed by Spearman's correlation analysis (Fig. 2F). Furthermore, qRT-PCR and northern blot assays revealed that the expression of MEG3 in cardiomyocytes was positively regulated by STAT3 (Fig. 2G,H), indicating the positive regulatory effect of STAT3 on the expression of MEG3.
MEG3 serves as a miR-361-5p sponge. It has been reported that lncRNAs can act as a ceRNA in human diseases by competitively sponging miRNAs 18,19 . In this study, we hypothesized that MEG3 exert function of ceRNA in cardiac hypertrophy. At first, we applied subcellular fractionation assay to detect the specific localization of MEG3 in cardiomyocytes. It was found to be mainly located in the cytoplasm (Fig. 3A). Next, we found out four miRNAs which can bind with MEG3 from starbase (http://www.lncrnablog.com/starbase-v2-0-for-decoding-rna-interaction-networks/). Next, we examined the expression level of all these four miRNAs in response to the silencing of MEG3. As a result, only miR-361-5p was obviously upregulated (Fig. 3B). Therefore, miR-361-5p was chosen to do next experiment. Next, miR-361-5p was found to be upregulated in cardiomyocytes which were treated with Ang-II (Fig. 3C). RIP assay revealed that both MEG3 and miR-361-5p tended to enrich in Ago2-immunoprecipitation (Fig. 3D). What's more, RNA pull down assay was conducted and indicated that MEG3 could only be pulled down by the wild type bio-labeled miR-361-5p, but not the mutated oligos (Fig. 3E). The binding sites between MEG3 and miR-361-3p was shown in Fig. 3F. Dual luciferase reporter assay was applied to further demonstrate the combination between them. After miR-361-5p and miR-NC were separately transfected into cardiomyocytes, we found that only the luciferase activity of wild type MEG3 (MEG3-WT) was obviously changed by miR-361-5p mimics (Fig. 3G). Next, MEG3 was overexpressed by pcDNA-MEG3 but was silenced by sh-MEG3 (Fig. 3H). The transfection efficiency was harvested after 48 hours. The expression levels of miR-361-5p were negatively regulated by MEG3 (Fig. 3I). Accordingly, we confirmed that MEG3 acts as a ceRNA by competitively binding with miR-361-5p in cardiac hypertrophy.
MEG3 upregulates HDAC9 through competitively binding with miR-361-5p. Bioinformatics analysis was applied to find the mRNA which constitutes a ceRNA model with miR-361-5p and HDAC9 in CH. We found 69 potential target mRNAs of miR-361-5p by using two public bioinformatics tools (picTarSites and miRandaSites) (Fig. 4A). Subsequently, the expression of these 69 mRNAs was examined in response to miR-361-5p mimics and sh-MEG3. As a result, only ten of them was obviously downregulated (Fig. 4B). Therefore, these ten mRNAs were chosen for further study. Next, the fold change of all these ten mRNAs in the mouse hearts in TAC and Sham group (Fig. 4C). HDAC9 showed the most obvious fold change. In addition, we detected the fold change of HDACs family (HDAC1, HDAC2, HDAC3, HDAC5, HDAC6, HDAC7 and HDAC9). As a result, HDAC9 exhibited the most obvious fold change (Fig. 4D). Moreover, HDAC9 has been proved to be targeted by miRNA in human disease 20 . Here, we supposed that HDAC9 can participate in a ceRNA model. To further demonstrate our hypothesis, the localization of HDAC9 was examined in cardiomyocytes. It was uncovered to be located in both nucleus and cytoplasm (Fig. 4E). Next, the binding sites between MEG3 and miR-361-5p as well as between miR-361-5p and HDAC9 was predicted by using bioinformatics analysis (Fig. 4F). The result of luciferase reporter assay manifested that the luciferase activities of HDAC9-WT was largely decreased by miR-361-5p mimics or sh-MEG3 (Fig. 4G). Pull down assay was conducted and suggested that HDAC9 could only be pulled down by the wild type bio-labeled miR-361-5p, but not the mutated oligos (Fig. 4H). The decreased expression level of HDAC9 caused miR-361-5p mimics was rescued by pcDNA-MEG3 (Fig. 4I). The expression level of HDAC9 increased by miR-361-5p inhibitors was partially reversed by sh-MEG3 (Fig. 4I). All these experimental results indicated that MEG3 can positively regulate HDAC9 via sponging miR-361-5p.
The relationships among MEG3, miR-361-5p and HDAC9. To understand whether MEG3 is differently expressed in CH samples, we tested the expression level of MEG3 in mice heart which undergone TAC or sham operation. The specific experimental methods in this step was in accordance with the previous study 21 . To demonstrate the correlations among MEG3, miR-361-5p and HDAC9, the expression levels of these three genes were separately tested in Sham and TAC group. MEG3 and HDAC9 were highly expressed in TAC group, while miR-361-5p was downregulated in TAC group (Fig. 5A). Next, the expression correlations among them were analyzed with Spearman's correlation analysis. The negative relevance between MEG3 and miR-361-5p as well as between miR-361-5p and HDAC9 were analyzed (Fig. 5B). Additionally, the positive correlation between MEG3 and HDAC9 was analyzed (Fig. 5B).

MEG3 improves the hypertrophy of cardiomyocyte by altering signals of miR-361-5p/HDAC9
axis. Based on the previous results, we could found the relevance among MEG3, miR-361-5p and HDAC9. In order to confirm the exact functions of this pathway in the progression of cardiac hypertrophy, rescue assays were designed and conducted in hypertrophic cardiomyocytes induced by Ang-II at 48 hours' transfection. To demonstrate that sh-MEG3 is not cytotoxic, the transfection efficiencies of mock, sh-NC and sh-MEG3 were detected by qRT-PCR ( Supplementary Fig 1). We found sh-MEG3 largely shrank the size of CH cells in comparison with sh-NC, while this tendency could be reversed by miR-361-5p inhibitors and pcDNA-HDAC9 (Fig. 6A). Next, the expression of β-MHC, BNP and ANP was also observed in response to Ang-II, the results were in accord with the former study (Fig. 6B-D). The decreased expression of these three elements was observed when MEG3 expression was interfered, while this tendency was changed over in both groups when CH cell was transfected into miR-361-5p inhibitors and pcDNA-HDAC9. According to all these results, we concluded that MEG3 improved the hypertrophy of cardiomyocyte by altering signals of miR-361-5p/HDAC9 axis.

Discussion
As a set of noncoding RNAs, lncRNAs were longer than 200nt. They have become the focus of researchers in the past decades. They are the vital molecule in researches because of their crucial role in complex biological process in human diseases 22 . It has been widely acknowledged that lncRNAs could act as regulators in human diseases. The role of them could be divided into two types: oncogene and tumor suppressor. It has been demonstrated that lncRNAs could regulate their downstream miRNAs to promote or suppress the occurrence and development of various diseases 23 . Growing number of lncRNAs have been verified to exert essential functions in heart failure [24][25][26] . While the study of the interaction between lncRNAs and cardiac hypertrophy still limited. Some lncRNAs have been certified to be a ceRNA in cardiac hypertrophy, such as lncRNA MIAT 6 and lncRNA-GAS5 27 . Based on previous studies, we supposed MEG3 might be a ceRNA in CH.
MiRNAs are a group of RNA molecules which were coded by endogenous genes. They regulate genes at the post-transcriptional level. MiRNAs have been studied in human diseases, such as human cancers 28,29 and cardiac hypertrophy 14 . It has been certified that miRNAs could be shared by lncRNAs and mRNAs to regulate biological behaviors of human diseases. As a member of miRNAs, miR-361-5p has been reported to be a suppressor in progression of human diseases by targeting mRNAs [30][31][32][33] . In this study, we found that miR-361-5p is a target gene of MEG3 through using bioinformatics analysis, pull-down assay and dual luciferase reporter assay. The reverse correlation between MEG3 and miR-361-5p was analyzed and demonstrated in CH. Therefore, we initially judged that MEG3 could competitively bind with miR-361-5p in CH. It is widely reported that mRNAs can act as downstream genes of miRNAs in ceRNA pathway. In a ceRNA pathway, lncRNAs regulate biological behaviors by sponging with miRNAs to regulate mRNA. MRNAs have been reported to be crucial factors to regulate the biological processes of human diseases 34,35 . HDAC9 was found to be a motivator in the progression of human diseases 36,37 . It also could be targeted by miRNA to alter progression of human diseases 38,39 . In this study, we found that HDAC9 was highly expressed in TAC and Ang-II group. The positive correlation between MEG3 and HDAC9 was demonstrated to support our hypothesis. The binding sites among MEG3, miR-361-5p and HDAC9 were obtained by using bioinformatics analysis. The exact combination was confirmed by pull down assay and luciferase activity assay. Finally, rescue assays were conducted to certify the function of MEG3-miR-361-5p-HDAC9 axis in CH progression. All findings in this study verified that lncRNA MEG3 was activated by STAT3 and positively regulated CH via miR-361-5p/HDAC9 axis (Fig. 7). These findings may help to develop novel therapeutic target for CH.