Oxygen glucose deprivation/re-oxygenation-induced neuronal cell death is associated with Lnc-D63785 m6A methylation and miR-422a accumulation

Oxygen glucose deprivation/re-oxygenation (OGD/R) induces neuronal injury via mechanisms that are believed to mimic the pathways associated with brain ischemia. In SH-SY5Y cells and primary murine neurons, we report that OGD/R induces the accumulation of the microRNA miR-422a, leading to downregulation of miR-422a targets myocyte enhancer factor-2D (MEF2D) and mitogen-activated protein kinase kinase 6 (MAPKK6). Ectopic miR-422a inhibition attenuated OGD/R-induced cell death and apoptosis, whereas overexpression of miR-422a induced significant neuronal cell apoptosis. In addition, OGD/R decreased the expression of the long non-coding RNA D63785 (Lnc-D63785) to regulate miR-422a accumulation. Lnc-D63785 directly associated with miR-422a and overexpression of Lnc-D63785 reversed OGD/R-induced miR-422a accumulation and neuronal cell death. OGD/R downregulated Lnc-D63785 expression through increased methyltransferase-like protein 3 (METTL3)-dependent Lnc-D63785 m6A methylation. Conversely METTL3 shRNA reversed OGD/R-induced Lnc-D63785 m6A methylation to decrease miR-422a accumulation. Together, Lnc-D63785 m6A methylation by OGD/R causes miR-422a accumulation and neuronal cell apoptosis.


Introduction
Ischemic stroke is a leading cause of human morbidity and mortality around the world 1,2 . As the prevalence of stroke rises more effective treatment strategies are urgently required [2][3][4] . The main pathogenesis of stroke, ischemia-reperfusion, induces significant oxidative injury to surrounding neurons 5,6 , which can be mimicked in vitro by an oxygen and glucose deprivation (OGD)/ re-oxygenation (OGD/R) procedure applied to cultured neurons [7][8][9][10] .
MicroRNAs (miRs) are a class of non-coding 21-25 nt mRNA-interfering molecules 11,12 , that regulate the expression of target genes by binding to their 3′-UTR (3′untranslated region) 11,12 . miR dysregulation is detected in cerebral pathogenesis [13][14][15] , and provide biomarkers in the diagnosis and prognosis of cerebral diseases [13][14][15] . Circulating stroke-associated miR profiles reflect temporal progression and specific etiologies of ischemic stroke 14,15 . The brain-enriched microRNA-422a (miR-422a) is upregulated in acute ischemic stroke, independent of age, severity, or confounding metabolic complications 14 . In the acute phase of stroke plasma miR-422a is significantly increased, and then downregulated in the sub-acute phase 16 . Similarly, following ischemia-reperfusion, miR-422a is downregulated in PC12 cells 17 . Here we explore whether OGD/R stimulation can affect miR-422a expression, and examine its potential functions in mediating OGD/R-induced neuronal cell death.
N6-methyladenosine (m6A) modification is the most abundant internal methylation of RNA transcripts, required for RNA processing, stabilization, and various biological functions 18 . The writer complex, that includes the methyltransferase enzymes methyltransferase-like protein 3 (METTL3), METTL14, and WTAP, catalyzes m6A modification 19 . Conversely, m6A modification is removed by the demethylases (erasers) FTO and ALKBH5 19 . Recent in vivo and in vitro studies suggest that m6A modification is involved in the mechanism of brain ischemia-reperfusion injury 19 , and indicate that inhibition of m6A methylation can protect neurons against ischemia-reperfusion injury 19 . m6A modifications regulate the function and stabilization of LncRNAs by providing a binding site for the m6A reader proteins or by changing the structure of the local RNA [20][21][22] . We here identify that LncRNA D63785 (Lnc-D63785) m6A methylation and downregulation is the primary cause of miR-422a accumulation in OGD/R-treated neuronal cells.

Cell culture
The neuronal cells derived from SH-SY5Y neuroblastoma cells, obtained from the Cell Bank of Shanghai Institute of Biological Science (Shanghai, China), were cultured in DMEM plus 10% fetal bovine serum. For neuronal differentiation, SH-SY5Y cells were cultured in the BDNF plus glutamine medium (serum free) as previously-described 23 . After differentiation, over 95% of cells were neuronal cells. The SH-SY5Y cells were subjected to mycoplasma and microbial contamination examination every 3-4 months. Authentication by STR profiling, population doubling time, and morphology were routinely confirmed as well to verify the genotype. The primary murine neurons were provided by Dr. Di [23][24][25] , and cultured by the previously-described protocols 23 . At day-10 (DIV), over 95% of cells were cortical neurons. The study was approved by the Ethics Committee and IACUC committee of Soochow University.

OGD/re-oxygenation
We utilized a previously-described OGD/R procedure 7,24 . Briefly, neuronal cells were initially placed in an airtight chamber, equilibrated for 15 min with a continuous flux of gas (95% N 2 /5% CO 2 ). The chamber was sealed and placed in an incubator for additional 4 h of OGD. Neuronal cells were then re-oxygenated (OGD/R) for applied time periods. "Mock" cells were placed in norm-oxygenated DMEM containing glucose.

Lactate dehydrogenase (LDH) assay
Following the applied treatments, cell death was examined by measuring LDH released to the medium, using a simple two-step LDH enzymatic reaction kit (Takara, Tokyo, Japan). Medium LDH contents were always normalized to total LDH contents 23 .

TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) assay
Neuronal cells were seeded into the 96-well tissue-culture plates (at 3 × 10 4 cells/cm 2 ). Following OGD/R treatment, TUNEL In Situ Cell Death Detection Kit (Roche Diagnostics Co) was utilized to quantitatively examine cell apoptosis intensity. Cells were stained with TUNEL (Invitrogen, 5 μM). Cell nuclei were co-stained with DAPI and visualized through a fluorescent microscope (Leica, Shanghai, China). For each treatment, at least 500 cells in five random views (1 × 200 magnification) were counted to calculate TUNEL ratio (% vs. DAPI).

Mitochondrial depolarization assay
In stressed/dying cells with mitochondrial depolarization, the fluorescence dye JC-1 will form green monomers by aggregating in the mitochondria 26 . Following the applied treatments, neuronal cells were stained with JC-1 (5 μg/mL, Sigma), washed and tested immediately using a fluorescence spectrofluorometer (F-7000, Hitachi, Japan) machine at test-wavelength of 545 nm (green). The JC-1 fluorescence images, integrating both the green (at 545 nm) and red (at 625 nm) wavelength were presented.

Annexin V-FACS assay of apoptosis
Following the applied treatments, neuronal cells were stained with Annexin V and Propidium Iodide (PI), washed and tested immediately using a FACS machine (BD, Shanghai, China). Annexin V ratio was recorded.

RNA-pull down assay
The pull-down assay was performed as described previously 27 . Briefly, the miR-422a (see sequence in ref. 27 ) and the mutant miR-422a (see sequence in ref. 27 ) singlestranded RNAs were labeled with biotin (at 5′), and then were individually transfected to cultured SH-SY5Y cells. The achieved lysates (using the described lysis buffer 27 ) were pre-cleared and thereafter incubated with the beads coated with RNase-free BSA (Sigma) and yeast tRNA (Sigma) 27 . The beads were incubated and washed 27 . The bound RNAs were then analyzed by qPCR.

Quantitative real-time PCR (qPCR)
Total RNA was extracted by Trizol reagents (Invitrogen, Shanghai, China) and quantified. For each condition, 500 ng of RNA was utilized for reverse transcription by SYBR Green SuperMix 28 . qPCR was performed by the 7900HT Fast Real-Time PCR system (Applied Biosystems). qPCR quantification was through 2 -ΔCt method using the following formula: 2 -(Ct of target gene-Ct of reference gene) . The data presented were normalized to GAPDH. Expression of LncRNA D63785(Lnc-D63785) and miR-422a was normalized to U6 RNA. The primers of this study were listed in Table 1.

Western blotting
Neuronal cells were seeded into the six-well tissueculture plates at 70-80% confluence. The protein lysates (40 μg per treatment) were separated by 10-12.5% SDS-PAGE gels, then transferred to PVDF membranes (Millipore, Shanghai, China). Each PVDF membrane was blocked in PBST with 10% nonfat milk, thereafter incubated with the designated primary and secondary antibodies. ECL reagents (Roche, Shanghai, China) were added to detect signals under X-ray films, and Image J software (NIH) utilized to quantify the total gray of each protein band. For Western blotting, the same set of lysate samples were run in sister gels when necessary to test different proteins. The exact amount of protein lysates, 40 μg lysates per lane, were loaded in each lane.

miR-422a inhibition
The pre-miR-422a antisense ("antagomiR-422a", purchased from Applied Biosystems) was annealed and subcloned into the GV248 lentiviral vector (Shanghai Genechem Co.). It was thereafter transfected with lentivirus package plasmids mix to HEK293T cells. The antagomiR-422a lentivirus (lv-antagomiR-422a) was added to cultured SH-SY5Y cells. After 24 h, the transfection medium was removed and replaced with fresh medium. Puromycin was then added to select stable cells for 4-5 passages. miR-422a expression was tested by qPCR in the stable cells.
Transfection of oligonucleotides SH-SY5Y cells or the primary murine neurons were seeded into six-well plates at 60% confluence. Exact 100 pmol of scrambled negative control siRNA ("si-C"), Lnc-D63785 siRNA (two different sequences, "seq-1/-2", see ref. 27 ), miR-422a mimic, the mimic control sequence, miR-422a inhibitor sequence or the inhibitor control sequence were transfected to SH-SY5Y cells or the murine neurons with Lipofectamine 2000, respectively. After 24 h, transfection was repeated one more round. Expression of target genes was verified by qPCR.
N6-methyladenosine (m6A)-RNA immunoprecipitation and qPCR (MeRIP-qPCR) The m6A antibody and the rabbit IgG 29 were respectively conjugated to 20 μl Beads protein A/G mixed magnetic beads 29 . A 100 μg aliquot of fragmented total RNA was incubated with the antibody in immunoprecipitation buffer 29 plus with 40U RNase inhibitor. RNA was eluted from the beads as described 29 . Following phenol extraction and ethanol precipitation, the m6A-enriched RNA was reversely transcribed and tested by qPCR assays, with Lnc-D63785 level normalized to Input controls.
shRNA The human METTL3 short hairpin RNA [two shRNAs, with two nonoverlapping sequences ("s1/s2"), as reported early 30,31 ] was inserted into the GV369 construct (Genechem Co., Shanghai, China). The construct, along with the lentivirus package constructs, was transfected to HEK-293T cells for 24 h. The generated shRNA lentivirus was filtered, enriched, and added to cultured SH-SY5Y cells. Puromycin was then added to select the stable cells.
Silencing of METTL3 was verified by Western blotting. Control cells were treated with the lentiviral scramble control shRNA ("shC", Genechem, Shanghai, China).

Statistical analysis
The investigators were blinded to the group allocation during all experiments. Experiments in this study were repeated at least three times. Data were expressed as mean ± standard deviation (SD). Statistics were analyzed by one-way analysis of variance (ANOVA) followed by a Scheffe's f-test by the SPSS 21.0 software (SPSS Inc., Chicago, IL). To test significance between two treatment groups, a two-tailed unpaired T test (Excel 2007) was utilized. p < 0.05 was considered significant. All the protocols of this study were approved by Ethics Committee of Soochow University.
Previous studies show that miR-422a acts as a tumor suppressor in cancer cells by targeting mitogen-activated protein kinase kinase 6 (MAPKK6) 33 and myocyte enhancer factor-2D (MEF2D) 27 . Examining the mRNA and protein levels of these verified miR-422a targets in SH-SY5Ycells, we found that MAPKK6 and MEF2D mRNA (Fig. 1b) and protein (Fig. 1c) were significantly downregulated after OGD/R (Fig. 1b). Studies have shown that MEF2D is a key transcription factor required for neuronal survival 34 , essential for the expression of Bcl-w 35 , an anti-apoptotic Bcl-2 family protein in neurons 35 . We found that Bcl-w protein expression was also downregulated in OGD/R-stimulated SH-SY5Y cells (Fig. 1c).

Discussion
There is an urgent need to identify novel biomarkers for early diagnosis and prognostic evaluation of ischemic stroke. Recent studies have explored the potential of abnormally-expressed circulating miRNAs as a diagnostic or prognostic biomarker for this disease 14,15 . It has been reported that circulating miR-422a, a brain-enriched miR, is significantly upregulated in patients with acute ischemic stroke 14 . Our results show that OGD/R induces significant miR-422a accumulation, mediating neuronal cell death and apoptosis. In SH-SY5Y cells and primary murine cortical neurons, miR-422a inhibition, by antagnomiR-422a, significantly attenuated OGD/R-induced viability reduction, cell death and apoptosis. Conversely, ectopic miR-422a overexpression mimicked OGD/R-induced actions, inducing neuronal cytotoxicity. Therefore, miR-422a accumulation may play a key role in the mechanism of OGD/R-induced neuronal cell injury.
LncRNAs are a class of non-coding RNAs with the length over 200 nt, generally considered to be mRNA-like transcripts 41,42 . LncRNA can function as ceRNAs to sponge target miRNAs, thereby antagonizing miRNAinduced activity 41,42 . The study by Zhou et al., has shown that Lnc-D63785 is a key ceRNA of miR-422a 27 . The results of the current study suggest that the OGD/Rinduced reduction of Lnc-D63785 is the primary cause of miR-422a accumulation and neuronal cytotoxicity. Lnc-D63785 directly associates with miR-422a in the neuronal cells. Restoring Lnc-D63785 expression, by a lentiviral construct, not only abolished OGD/R-induced miR-422a accumulation, but also attenuated neuronal cell death. Conversely, overexpression of amiR-422a-binding mutant Lnc-D63785 was unable to affect OGD/R-induced neuronal cell death/apoptosis. Importantly, siRNA-mediated knockdown of Lnc-D63785 induced miR-422a accumulation and neuronal cell injury, mimicking the neuronal cytotoxicity induced by OGD/R. Overall these results support that decreased Lnc-D63785 expression is responsible for OGD/R-induced miR-422a accumulation and subsequent neuronal cell death.

Conclusions
We conclude that OGD/R leads to Lnc-D63785 m6A modification and decreased expression, consequently resulting in miR-422a accumulation, downregulation of its targets (MEF2D-MAPKK6), and neuronal cell death/ apoptosis. role in study design, data collection and analysis, decision to publish, or preparation of the paper.
Author details 1