LncRNA MEG3 targeting miR-424-5p via MAPK signaling pathway mediates neuronal apoptosis in ischemic stroke

Sema3A was overexpressed in IS

The heatmap was obtained by microarray analysis of GSE22255, the results were listed in Supplementary Table 1. Supplementary Figure 1A showed the top 10 differentially upregulated and downregulated mRNAs in IS. We screened out a variety of signaling pathways upregulated in control group and IS group (Supplementary Figure 1B). The significantly activated or suppressed pathways were represented by the dotplot and ridgeplot, and MAPK signaling pathway was one of the evidently suppressed pathways in IS (Supplementary Figure 1C, 1D). Sema3A along with DUSP1 and FOS were recognized as the top 3 aberrantly expressed molecules related to MAPK signal pathway (Supplementary Figure 1E). Later, we treated N2a cells with OGD/R to simulate IS in vitro. These 3 molecules were collected for qRT-PCR to cross-check the abnormal expression in control and OGD/R cell lines. DUSP1 and FOS displayed no significant difference in OGD/R group (Supplementary Figure 1F) while Sema3A expression rose over time of intervention (Figure 1A).

Figure 1. MRNA Sema3A was highly expressed in IS and correlated with cell viability and apoptosis. (A) The mRNA level of Sema3A in N2a cells climbed up with time after 6h, 12h, 24h of OGD/R treatment. *P<0.05 vs control. (B) Successful downregulation of Sema3A by si-Sema3A. *P<0.05 vs control. (C) Sema3A downregulation led to higher viability of OGD/R cells. *P<0.05 vs control, #P<0.05 vs siRNA-NC. (D, E) OGD/R cells have higher apoptosis rate compared with control groups while Sema3A downregulation is correlated with lower apoptosis level among OGD/R cells. *P<0.05 vs control, #P<0.05 vs siRNA-NC. (F) OGD/R cells tend to have higher level of c-caspase-3 as well as p-JNK and p-p38 in comparison with control group. Downregulation of Sema3A could counteract with this trend in OGD/R cells partly. *P<0.05 vs control, #P<0.05 vs siRNA-NC.

Sema3A mediated cell viability and apoptosis in OGD/R cells

To detect the function of Sema3A, we conducted a series of experiments by manually altering the expression of Sema3A. N2a cells used in the following experiments were under 2h OGD treatment and cultured under normal conditions for 24 h. It was demonstrated that Sema3A was successfully downregulated in the OGD/R cells by si-Sema3A (Figure 1B). In si-Sema3A groups, cell viability was notably higher than that in the control group (Figure 1C). TUNEL staining and flow cytometry assay both showed that downregulation of Sema3A led to less cell apoptosis of OGD cells (Figure 1D–1E). However, expression levels of phosphorylated JNK and p38 decreased after knockdown of Sema3A, indicating that Sema3A had an activating effect on the MAPK signaling pathway (Figure 1F).

MiR-424-5p was lower-expressed in IS and targeted Sema3A

We used TargetScan to predict Sema3A-related miRNAs and used DIANA TOOLS to verify miRNAs related to MAPK signaling pathway. As one of the common miRNAs related to Sema3A and MAPK signaling pathway, miR-424-5p was screened out by using Venny 2.1 (Figure 2A). The expression level of miR-424-5p in control and OGD/R group was detected by qRT-PCR. In contrast with control group, miR-424-5p expression demonstrated a decline along with the prolonged duration of treatment. (Figure 2B). TargetScan was used to predict the binding position between Sema3A and miR-424-5p (Figure 2C), and dual luciferase assay was applied to verify the interactive relationship (Figure 2D). We also tested the Sema3A expression after manually modifying the expression of miR-424-5p. MiR-424-5p upregulation by miR-424-5p mimics led to downregulation of Sema3A, while miR-424-5p downregulation by miR-424-5p inhibitor could promote Sema3A expression (Figure 2E, 2F). Hence, the expression of Sema3A could be regulated by miR-424-5p.

Figure 2. MiR-424-5p targeted with Sema3A and was lowly expressed in IS. (A) MiR-424-5p was among 11 miRNAs which both targeted Sema3A and related to MAPK signaling pathway. (B) The relative miR-424-5p expression declined with time of 6h, 12h, 24h of OGD/R treatment. *P<0.05 vs control. (C, D) The target relationship between Sema3A and miR-424-5p was predicted by Targetscan and verified by dual luciferase assay. *P<0.05 vs control. (E, F) Upregulation of miR-424-5p by miR-424-5p mimics led to lower Sema3A expression, while downregulation of miR-424-5p led to higher Sema3A expression. *P<0.05 vs control.

MEG3 was highly expressed in IS and interacted with miR-424-5p

We further analyzed the differentially expressed lncRNAs in GSE22255 (Supplementary Table 2), and disclosed that MEG3 was one of the most significantly overexpressed lncRNAs in IS group (Figure 3A). With the help of the MiRanda and LncRNA and Disease Database, we found that MEG3 is both related to miR-424-5p and IS (Figure 3B). In N2a cells, the expression of MEG3 increased along with OGD treatment time, and this trend could also be observed on the expression of Sema3A (Figure 3C). The binding position between MEG3 and miR-424-5p was predicted by MiRanda and the target relationship was verified by dual luciferase assay (Figure 3D). Forced downregulation of MEG3 by si-MEG3 led to a higher expression level of miR-424-5p and a lower expression level of Sema3A (Figure 3E–3G). All of these results revealed that MEG3 could regulate Sema3A expression by targeting miR-424-5p.

Figure 3. MEG3 was a highly expressed lncRNA in IS that bonded with miR-424-5p. (A) The heatmap from the microarray analysis of GSE22255 showed the significantly differentially expressed lncRNAs including MEG3 in the IS group. (B) MEG3 bound to miR-424-5p and was related to IS. (C) MEG3 expression was positively correlated with OGD/R intervention time. *P<0.05 vs control. (D) The target relationship between MEG3 and miR-424-5p was predicted and proved. *P<0.05 vs control (E) The effect of downregulation MEG3 by si-MEG3 was proved significant. *P<0.05 vs control. (F) MEG3 knockdown correlated with higher miR-424-5p expression. *P<0.05 vs control. (G) MEG3 knockdown correlated with lower Sema3A expression. *P<0.05 vs control.

MEG3 inhibited cell viability and promoted cell apoptosis by positively regulating Sema3A

To further verify the effect of potential MEG3/miR-424-5p/Sema3A/MAPK axis on the development of IS, a series of experiments detecting cell viability, apoptosis and MAPK signal pathway were conducted. In OGD/R-treated cells, cell viability decreased compared to the control group. si-MEG3 group where MEG3 expression was down-regulated showed recovered viability while a co-transfection of Sema3A could impair this restoration. On the contrary, miR-424-5p inhibitor led to a lower cell viability, which could be reversed by co-transfection of si-Sema3A (Figure 4A). According to TUNEL staining and flow cytometry results, apoptosis rate of OGD/R cells increased dramatically compared to the control group. In OGD/R treated groups, the downregulation of MEG3 restrained apoptosis, while miR-424-5p inhibitor promoted it. The effect of down-regulation of MEG3 and miR-424-5p could be reversed partially by co-transfection of Sema3A and si-Sema3A respectively (Figure 4B, 4C). To observe alteration in the downstream MAPK signal pathway, the level of cleaved caspase-3 and phosphorylated levels of JNK and p38 were detected through western blot. All of these markers displayed similar trends among different groups. OGD/R cells displayed higher cleaved caspase-3/caspase-3 ratio as well as p-JNK/JNK and p-p38/p38 ratio indicating enhanced apoptosis and activated MAPK signal pathway. This trend could be partially counteracted through down-regulation of MEG3 and reversed by co-transfection of Sema3A. Meanwhile, inhibition of miR-424-5p could promote apoptosis and MAPK signal pathway while co-transfection of si-Sema3A could help to counteract it (Figure 4D). These results further indicate that MEG3 activated the MAPK signaling pathway through regulating miR-424-5p/Sema3A axis.

Figure 4. The MEG3/miR-424-5p/Sema3A axis affected cell viability and apoptosis in OGD cells via the MAPK pathway. (A) OGD/R cells have lower cell viability compared with control group. Among OGD/R cells, MEG3 downregulation led to higher cell viability, which could be offset by a simultaneous overexpression of Sema3A; miR-424-5p downregulation by its inhibitor led to a lower cell viability compared among OGD/R cells, which could be neutralized by a simultaneous downregulation of Sema3A. *P<0.05 vs control, #P<0.05 vs OGD/R NC, $P<0.05 vs si-MEG3, &P<0.05 vs inhibitor (B, C) OGD/R cells tend to have fairly high apoptosis rate compared to control group. Among OGD/R cells, MEG3 downregulation led to lower cell apoptosis level, which could be neutralized by a simultaneous overexpression of Sema3A; miR-424-5p downregulation by its inhibitor mediated more cell apoptosis, which could be neutralized by a simultaneous downregulation of Sema3A. *P<0.05 vs control, #P<0.05 vs OGD/R NC, $P<0.05 vs si-MEG3, &P<0.05 vs inhibitor. (D) OGD/R cells tend to have higher cleaved caspase-3, p-JNK and p-p38 expression in comparison with control. Among OGD/R cells, the low expression of MEG3 led to a rise in expression of cleaved caspase-3, p-JNK and p-p38, miR-424-5p inhibitor led to drop of expression of cleaved caspase-3, p-JNK and p-p38. Sema3A overexpression could reverse the rise of si-MEG3, while si-Sema3A could reverse the drop of miR-424-5p inhibitor. *P<0.05 vs control, #P<0.05 vs OGD/R NC, $P<0.05 vs si-MEG3, &P<0.05 vs inhibitor.

MEG3 expression aggravated severity of IS by promoting Sema3A expression in vivo

In MCAO mice models, the MEG3 expression level was increased, compared to the sham group (Figure 5A). Before the ischemic reperfusion, si-MEG3 and si-MEG3+Sema3A transfection were conducted on mice in experimental groups. Results showed that 72h after the reperfusion, Sema3A co-transfection could almost reverse the downregulation of Sema3A by si-MEG3, but had no effect on the inhibition of MEG3 in the mice models (Figure 5B, 5C). The infarct area in MCAO group increased dramatically compared with Sham group. In si-MEG3 group, the infarct area of MCAO mice shrank and it could be reversed in the si- MEG3+Sema3A group (Figure 5D, 5E). At the same time, the neurological score rose significantly in comparison with sham group. si-MEG3 group could partly offset the increase in neurological score in MCAO mice, while Sema3A co-transfection counteracted this effect (Figure 5F). All the results above indicated MEG3 could aggravate IS progress by regulating Sema3A expression. According to the western blot results in Figure 5G, the expression of cell apoptosis marker cleaved caspase-3 was increased in MCAO mice in contrast with the control group, so did MAPK signal pathway markers-phosphorylated JNK and p38. This trend in MCAO mice could be offset by down-regulating MEG3, which could then be reversed by the co-transfection of Sema3A indicating that MEG3 promoted cell apoptosis and activated MAPK signaling pathway by modulating Sema3A.

Figure 5. MEG3 expression accelerated the process of IS by promoting Sema3A expression in vivo. (A) Expression level of MEG3 increased gradually at 4h, 8h, 24h point in the process of ischemia reperfusion (I/R stands for ischemic reperfusion). *P<0.05 vs sham. (B) Downregulation of MEG3 expression by si-MEG3 was proved significant. Simultaneous upregulation of Sema3A had little influence on the MEG3 expression. *P<0.05 vs control. (C) Sema3A expression was suppressed by si-MEG3, and the decrease was neutralized by manual Sema3A overexpression. *P<0.05 vs control, #P<0.05 vs si-MEG3. (D) Representative image of brain slices of three groups of rats. Normal tissues were a pink or red color, whereas the ischemic tissues were white. MCAO rats tend to have a bigger ischemic area in contrast with Sham group. Among MCAO rats, si-MEG3 group with MEG3 down-regulated have smaller ischemic area while it could be reversed by overexpression of Sema3A. (E) Quantified average percentage of the infarction area in the whole brain, the percentage was higher in MCAO rats compared with sham group. Among MCAO rats, the infarct volume decreased in si-MEG3 group, while simultaneous overexpression of Sema3A weakened the decrease. *P<0.05 vs sham, #P<0.05 vs MCAO, $P<0.05 vs MCAO+si-MEG3. (F) Neurological score rose in MCAO rats compared with sham rats. Among MCAO rats, si-MEG3 group had lower neurological score than MCAO control, while simultaneous overexpression of Sema3A neutralized the decrease of score. *P<0.05 vs sham, #P<0.05 vs MCAO, $P<0.05 vs MCAO+si-MEG3. (G) Expression of cleaved caspase-3, p-JNK and p-p38 increased significantly in MCAO group compared with control group. In si-MEG3 group, expression of cell apoptosis markers cleaved caspase-3 and phosphorylated levels of JNK and p38 decreased. Overexpression of Sema3A reversed the change. *P<0.05 vs sham, #P<0.05 vs MCAO, $P<0.05 vs MCAO+si-MEG3.

Microarray analysis

The data set of GSE22255 obtained from The Gene Expression Omnibus (GEO) was analyzed to reveal the differentially expressed genes and activated signaling pathway in IS tissues, and was first published on Dec 31, 2011 by Tiago Krug, et al [41]. This data set includes gene expression profiling of peripheral blood mononuclear cells (PBMCs) collected from 20 IS patients and 20 matched controls. The data were collected 6 months after the onset of stroke and were analyzed using Affymetrix microarrays. Important variables such as age and sex are matched between IS group and control group. Several researches regarding IS arose from this dataset since then.

Differential analysis of mRNAs was performed using "limma' package in R 3.4.1 (https://www.r-project.org/), and mRNAs with |log₂Fold change|>0.5 and P<0.05 were chosen for further analysis.

TargetScan database (www.targetscan.org) was employed to screen out the potential upstream regulator of target effector molecule. In addition, DIANA Tool (http://diana.imis.athena-innovation.gr/DianaTools/index.php) was applied to filter out microRNAs related to the MAPK signaling pathway. The intersection of these 2 sets were used to detect the microRNA for further research.

Once the microRNA was determined, we used MiRanda (http://www.microrna.org/) to predict potential lncRNAs interacting with microRNA mentioned above. Meanwhile, The LncRNA and Disease Database (http://www. cuilab.cn/lncrnadisease) was used to find out the IS associated lncRNAs. The lncRNAs of the intersection of these 2 sets were chosen for further study.

Gene set enrichment analysis (GSEA) analysis

Total normalized mRNAs expression data were uploaded into GSEA v3.0 software. KEGG pathway gene set was used to explore enriched pathways, and GO term gene set was used to explore enriched functions. We also employed R language “GSEABase” package to conduct data processing. The Benjamini-Hochberg Procedure was used to adjust P values. Statistic of default weighted enrichment was employed to manipulate data for 1000 times with normalized P<0.05 which was considered significantly enriched. Next, results at the top and on the bottom of GSEA reports were respectively chosen for curves using “ggplot2” package in R language.

Cell culture

We adopted the mouse brain neuroma cell line, N2a, to build oxygen-glucose deprivation (OGD) model. The cells were purchased from BeNa Culture Collection (Shanghai, China), and maintained in a medium of 90% DMEM (Invitrogen, Carlsbad, CA, USA) and 10% FBS (Invitrogen, Carlsbad, CA, USA), at 37°C in 5% CO₂.

Oxygen and glucose deprivation and reoxygenation (OGD/R) model

Cells were washed and then incubated in deoxygenated glucose-free DMEM. The cultures were then transferred to an anaerobic chamber filled with a gas mixture of 95% N₂/5% CO₂ at 37°C for 4h. At the end of OGD treatment, the medium was replaced with normal medium, and the cultures were returned to a normal atmosphere for 6, 12 and 24 hours respectively. Control cells were cultured under normoxic conditions without OGD treatment. With this condition, we harvested the 6, 12 and 24 hours of OGD/R cells for the detection of mRNAs, miRNAs and lncRNAs.

Cell transfection

The overexpression vectors of Sema3A constructed by GenePharma (Shanghai, China). MiR-424-5p mimics, inhibitor, MEG3 siRNAs and Sema3A siRNAs (listed in Supplementary Table 3) were designed and purchased from Sigma Aldrich (St. Louis, MO, USA). All RNAs were transfected into OGD/R model cells by the Lipofectamine 3000 Transfection Reagent (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer’s instructions.

Quantitative real-time reverse transcription-PCR (RT-PCR)

Total RNAs were extracted from cells and tissues by Trizol (Invitrogen, Carlsbad, CA, USA) according to the protocols. Then we carried out reverse transcription by the PrimeScript^® RT Kit (TAKALA, Dalian, China), and performed quantitative PCR in a 7500 Real-Time PCR System (Applied Biosystems, USA) using SYBR^® Premix Ex Taq™ II (TAKALA, Dalian, China). The expression of Sema3A, miR-424-5p and MEG3 was normalized to the level of GAPDH and U6. We calculated all results using the 2^−ΔΔCT method. The experiments were performed three times and the sequences were listed in Supplementary Table 4.

Cell counting Kit-8 (CCK-8) assay

48h after the transfection, cells were seeded in 96-well plates with a density of 2×10⁴ cells/well and incubated overnight with 5% CO₂ at 37°C. CCK-8 solution (Beyotime, Shanghai, China) was added (10 μL/well) to each well, and stored at 37°C for an additional 4h. Then, we detected the viability of cells per well through measuring the absorbance at 450nm using Elx800 Reader (Bio-Tek Instruments Inc., Winooski, VT, USA).

TUNEL staining

Cell apoptosis was determined by means of in situ cell Death Detection kit (Roche, Indianapolis, IN, USA). After the treatment, 0.2% Triton X-100 was added into cells for a 5min incubation. Then, the cells were incubated with the TUNEL reaction mixture containing the nucleotide mixture and terminal deoxynucleotidyl transferase (TdT). Subsequently, the cells were incubated in 0.3% H₂O₂ for 10min and stained with 0.5μg/mL DAPI in the dark for 5min at room temperature. TUNEL staining with a green fluorescence showed apoptotic cells, DAPI with a blue fluorescence showed the nuclei. The cells were observed and calculated under a fluorescence microscope (Olymbus, Tokyo, Japan). Five visual fields were chosen randomly for each specimen.

Flow cytometry

After transfection, OGD/R-treated N2a cells were incubated for 48h before cell apoptosis analysis. Incubated cells were stained with a FITC Annexin V Apoptosis Detection Kit (BD Biosciences, Bedford, MA, USA), and then put through FCM detection.

Western blot

We lysed the cells or tissues in RIPA protein extraction regent (Beyotime) with a protease inhibitor cocktail provided by Roche Diagnostics, Mannheim, Germany. Protein concentration was measured using BCA protein assay kit form Abcam (Cambridge, MA, USA). The equal amounts of proteins approximately 50μg were separated by 10% SDS-PAGE, and then were transferred to control membrane (Millipore, Billerica, MA, USA). Proteins were blocked with 5% milk at room temperature for 1h, and then incubated with primary antibodies, rabbit monoclonal JNK, p-JNK, p38, p-p38, caspase-3, cleaved caspase-3 and β-actin (anti-JNK, ab179461, 1:1000; anti-p-JNK, ab124956, 1:1000; anti-p38, ab170099, 1:1000; anti-p-p38, ab4822, 1:1000; anti-caspase-3, ab44976, 1:500; anti-cleaved caspase-3, ab2302, 1:1500; anti-β-actin, ab8227, 1:10000, Abcam, Cambridge, MA, USA) at 4°C overnight, and last incubated with Goat Anti-Rabbit IgG H&L (HRP) (ab6721, 1:10000, Abcam) at room temperature for 1h. We analyzed the protein bands by the Pierce ECL Substrate Western Blot Detection system (Bio-Rad Laboratories, Hercules, CA, USA).

Luciferase reporter assay

We predicted the sequences of wild-type (WT) Sema3A and MEG3 containing the targeting position of miR-424-5p. Then the synthesized Sema3A-WT, Sema3A-MUT, MEG3-WT and MEG3-MUT were cloned into the downstream gene via the dual luciferase reporter vectors (Promega, Madison, WI, USA). The configured luciferase vectors were named as Sema3A-WT, Sema3A-MUT, MEG3-WT and MEG3-MUT respectively. We co-transfected Sema3A-WT and Sema3A-MUT as well as MEG3-WT and MEG3-MUT with miR-control or miR-424-5p mimics using Lipofectamine2000 (Invitrogen, Carlsbad, CA, USA). We measured the luciferase activities by Dual Luciferase Assay (Promega, Madison, WI, USA) after transfection 48h.

Animals

Adult male C57BL/6J mice weighing 22 g -25 g were used for this study. Mice were allowed 1 week for adaption. The mice were arranged into four groups: the sham group, the MCAO group, the MCAO+si-MEG3 (intraventricular injection si-MEG3) group and the MCAO+si-MEG3+Sema3A group (intraventricular injection si-MEG3 and Sema3A). The methods for intraventricular injection were: i) mice were anesthetized with 10% chloral hydrate (3.2 ml/kg) by intraperitoneal injection and then placed on stereotactic frame; ii) si-MEG3 as well as si-MEG3+Sema3a were delivered to the right lateral ventricle through a Hamilton microsyringe (Hamilton Co., Reno, NV, USA); All experiments involving animals were in accordance with the institutional animal welfare guideline and approved by Institutional Animal Care and Use Committee of Qilu Hospital of Shandong University.

Mice model of middle cerebral artery occlusion (MCAO)

One day after the intraventricular injection of MEG3 siRNA and Sema3A, the mice were performed with the operation of MCAO. In brief, after anesthetized with ketamine (ip, 80-100mg/kg) as well as xylazine (ip, 10mg/kg), the right common carotid artery [42], external carotid artery (ECA) and internal carotid artery (ICA) were sequentially isolated. An incision was made in the distal region of the CCA, then a 6-0 nylon monofilament with a round tip was inserted into the ICA until the origin of the MCA was blocked. After 60 minutes of ischemia, the nylon monofilament was softly withdrawn to allow reperfusion. Laser Doppler flow was employed to monitor the blood flow and animals with a blood flow reduction less than 80% were excluded for further experiment. Other vital physical variables such as heart rate, body temperature and blood pressure were monitored through the operation. Sham underwent the same procedure except for the insertion of filament. At 72h after the reperfusion, the brain tissues were collected for TTC staining and expression level of MEG3, miR-424-5p, Sema3A and MAPK signaling pathway related proteins were detected through qRT-PCR and western blot. At 1, 3, 5, and 7 days following the reperfusion, the neurological deficits was assessed via a modified Bederson scoring system of five-point scale as described in a previous study [43].

TTC staining and measurement of infarct volume

Tri-phenyl-tetrazolium chloride (TTC) staining was used to measure cerebral infarction volume 72h after reperfusion. We incubated the 2-mm-thick dissected brain slices with a 2% TTC solution at 37°C for 30 min. Later, TTC-stained slices were photographed using a Nikon E950 digital camera attached to a dissecting microscope and determined the percentage of the infarct volume in the total brain volume in the digitized images with the Quantity One software package (Bio-Rad, Hercules, CA, USA). Normal tissues showed a pink or red color, and ischemic tissues were white.

Neurological score system

The modified Bederson scoring system [43] was used to quantify the neurological behaviors. We trained rats for 3 consecutive days before surgery, and used the score on the day before MCAO as the baseline. The test was performed at 1, 4 and 7 days following the reperfusion. A score of 0 meant no deficit; 1 for forelimb flexion; 2 for unidirectional circling after lifted by tail; 3 for spontaneous unidirectional circling; 4 for longitudinal rolling upon lifted by tail and 5 for spontaneous longitudinal rolling. Consequently, average score in each group reflected severity of neurological deficit - higher score meant severer deficit.

Statistical analysis

All the measurement data were expressed with mean ± standard deviation [44] and visualized by GraphPad Prism 6.0. We compared the difference between every two groups of data by t-test. One-way ANOVA was used to compare the differences among multiple groups. We repeated all experiments for at least three times, with P<0.05 indicating that the difference was significant statistically.

Ethical approval

All procedures performed in studies involving animals were in accordance with the ethical standards of Qilu Hospital of Shandong University. This article does not contain any studies with human participants performed by any of the authors.