miR-204-5p is sponged by TUG1 to aggravate neuron damage induced by focal cerebral ischemia and reperfusion injury through upregulating COX2

Studies have reported that miR-204-5p is involved in multiple biological processes. However, little is known about the expression and mechanism of miR-204-5p in cerebral ischemia and reperfusion injury. This study found that miR-204-5p expression was significantly downregulated in the blood of patients with ischemic stroke, MCAO/R rat brains, and OGD/R neurons. Overexpression of miR-204-5p markedly reduced infarct volume and neurological impairment and alleviated the inflammatory response in vivo. miR-204-5p promoted neuronal viability and reduced apoptotic cells in vitro. Mechanically, miR-204-5p was negatively regulated by the expression lncRNA TUG1 upstream and down-regulated COX2 expression downstream. Therefore, the TUG1/miR-204-5p/COX2 axis was involved in ischemia and reperfusion-induced neuronal damage. This finding may provide a novel strategy for the treatment of cerebral ischemia and reperfusion injury.

Little is known about the expression and mechanism of miR-204-5p in CIRI and how TUG1 acts with miR-204-5p at the molecular level. We hypothesize that the TUG1/miR-204-5p/COX2 axis is involved in CIRI. Our present study aimed to investigate the possible mechanisms of TUG1/miR-204-5p/COX2 axis in CIRI. The results may provide insights on CIRI and offer a potential approach to improve IS treatment.

RESULTS
TUG1, miR-204-5p and COX2 mRNA were involved in the IS development Table 1 shows the characteristics of the subjects. Compared with the control group, TUG1 and COX2 mRNA increased in IS patients, accompanied by a decrease in miR-204-5p ( Fig. 1A-C). TUG1 and COX2 mRNA were positively correlated with NIHSS scores (Fig. 1D, F), while miR-204-5p was negatively correlated with NIHSS scores (Fig. 1E). The spatiotemporal expressions of TUG1, miR-204-5p, and COX2 mRNA were also evaluated in brain tissues of MCAO/R rats at different reperfusion time (Fig. S1D-F). These results indicated that the TUG1/miR-204-5p/COX2 axis could be involved in the CIRI occurrence and development.
TUG1 knockdown ameliorated brain injury, restrained inflammatory response and apoptosis in MCAO/R rats Rats were intracerebroventricular injected with LV-TUG1-RNAi two weeks before MCAO/R operation ( Fig. 2A, B). Infarct volume, Zea-Longa scores and brain injury were significantly reduced in LV-TUG1-RNAi injected rats (Figs. 2C-E and S2). The inflammatory  response was inhibited after treatment with LV-TUG1-RNAi, with a  decrease in COX2, IL-1β, TNF-α, and PGE2, and an increase in IL-10  (Figs. 2F, H, S3 and S4A-D). Apoptosis was restricted in LV-TUG1-RNAi group, with increased Bcl-2 and decreased Bax expression (Fig. 2F). These results suggested that TUG1 could participate in the CIRI pathogenesis by regulating the inflammatory response and apoptosis.
TUG1 knockdown protected neurons from OGD/R injury Primary cortical neurons were isolated from the cortex of SD rats and were exposed to OGD/R to simulate ischemia and reperfusion injury in vitro (Fig. 4A). Compared with the normal group, TUG1 was significantly upregulated in the OGD/R group (Fig. 4B). TUG1 overexpression and suppression in neurons were achieved by transfection with LV-TUG1 or LV-TUG1-RNAi, respectively (Fig. 4C). The impairment of cell viability in neurons treated with OGD/R was alleviated after TUG1 knockdown (Fig. 4D). Apoptosis and inflammation were inhibited in OGD/R neurons transfected with LV-TUG1-RNAi (Figs. 4E, G and S6). These results demonstrated that TUG1 knockdown protected neurons from OGD/R injury.

DISCUSSION
The injury of neurons contributes to the secondary brain ischemic injury and represents the leading cause of cerebral injury aggravation [27]. Therefore, we focused on the mechanism of neuron injury involved in CIRI. We found that miR-204-5p improved ischemia and reperfusion-induced neuron injury by downregulating COX2 expression.
Our previous study has shown that COX2 is upregulated in CIRI, and miR-211-5p could alleviate CIRI in rats by downregulating the expression of COX2 [27]. miR-204-5p and miR-211-5p have very similar nucleotide sequences with only one different nucleotide in the entire sequence, a plausible explanation for sharing common targets [28]. In the previous study, we detected that miR-204-5p was decreased in the MCAO/R model, while the exact mechanism of neuron injury has not been elaborated. This study confirmed that miR-204-5p was downregulated in the blood of IS patients and the brains of MCAO/R rats.
The modulating network in organisms is complex and other genes can regulate miR-204-5p. TUG1 was first identified to play an essential role in the formation of photoreceptors [29,30]. TUG1 sponges microRNA-9 to promote neuronal apoptosis by upregulating Bcl2l11 under ischemia, indicating the therapeutic potential of TUG1 in IS [25]. TUG1 knockdown significantly reduced the infarct volume and Zeal-Long scores, promoted cell viability, and decreased apoptosis in OGD/R neurons. miR-204-5p antagomir or inhibitors could abolish the improvement of CIRI  6). B Infarct region was visualized by TTC staining. C Quantitative analysis of brain infarct volume after MCAO/R in rats (n = 9). D Zea-Longa scores (n = 13). E Relative protein levels of COX2, Bcl-2 and Bax in rats (n = 6). F Relative expression of COX2 mRNA in rats detected by qRT-PCR (n = 6). Data are presented as the mean ± SD. **p < 0.01, ***p < 0.001.
after TUG1 knockdown. Our study revealed that TUG1 acted as a regulator of miR-204-5p in CIRI.
When the brain suffers from ischemia and reperfusion, COX and other enzymes that promote ROS production destroyed the dynamic balance of ROS and caused neuron injury [31]. The suppression of TUG1 or overexpression of miR-204-5p enhanced the SOD content and weakened the expression of COX2, effectively reducing damage from oxidative stress (Fig. S4E). MDA is cytotoxic, causing the cross-linking and polymerization of macromolecules such as proteins and nucleic acids. Our results showed that MDA increased significantly in MCAO/R rats (Fig. S4F). MDA is the final product of the lipid oxidation reaction and is regulated by many different enzymes that promote or inhibit ROS production. The toxic effect of COX2 appears to be mediated by PGE2 rather than ROS, although COX2 can generate both [32]. Therefore, the simple change in COX2 and SOD would not significantly affect the MDA content.
Studies have shown that IL-1β, IL-10, and TNF-α in brain tissues are produced mainly by inflammatory cells [33]. We did not detect IL-1β, IL-10 and TNF-α in primary cortical neurons. PGE2 is the primary PGE product of COX2 in the brain, specifically derived from the intermediate product PGH2 catalyzed by COX2. Under pathological conditions, PGH2 is further catalyzed into PGE2 by microsomal PGE synthase 1 (mPGES-1) [34]. PGE2 is an important mediator to mediate harmful effects in neurological diseases [35,36]. Due to the rapid induction of COX2 and mPGES-1, the level of PGE2 increases correspondingly after cerebral ischemia. At 24 h after MCAO, the PGE2 level of the ipsilateral cortex in mice was three times higher than that of the contralateral cortex [37]. Our results showed that TUG1 knockdown or miR-204-5p overexpression significantly reduced PGE2 content.
In conclusion, our study suggested that decreased miR-204-5p promoted the inflammatory response and apoptosis and aggravated the neuronal damage induced by CIRI. The TUG1/miR-204- 5p/COX2 axis was involved in CIRI. This finding may provide a novel strategy for treating CIRI.

MATERIALS AND METHODS Sample collection
The sample size was calculated based on our pre-experiments. A 1.4 (SD 2) relative difference was assumed in the relative expression of miR-204-5p between healthy volunteers and IS patients. At least 44 subjects in each group were needed (90% power, 5% significance) with a two-tailed test. Forty-six patients who suffered an initial IS were recruited from the Department of Neurology of Dianjiang People's Hospital of Chongqing (DPHC) from December 2019 to May 2020. Patients with a history of hemorrhagic stroke, peripheral artery occlusive disease, transient ischemic attack, impaired consciousness or epilepsy, severe lung, liver or kidney dysfunction, severe malnutrition, thyroid disease, primary or metastatic tumors, and those who were pregnant or breastfeeding were excluded. Blood samples were collected within two hours after patient admission and stored at −80°C until RNA extraction. Forty-six healthy volunteers from DPHC were enrolled. The experiment was approved by the DPHC Ethics Committee. All subjects or their legal representatives signed a written informed consent. were kept at 23°C and 70% humidity in a 12 h light/dark cycle with free access to food and water, and randomized by lottery without blinding. Considering statistical requirements and the minimization of experimental animals, at least fifteen rats were included in each group. All animal experiments were performed according to the center guidelines and in accordance with the Ethics Committee of Chongqing Medical University.

Animals and the middle cerebral artery occlusion/reperfusion (MCAO/R) model
Rats were first anesthetized with an intraperitoneal injection of 3% pentobarbital sodium (1 ml/kg). Then the common carotid artery (CCA), internal carotid artery (ICA), and external carotid artery (ECA) were carefully isolated. The CCA was blocked by a bulldog clamp. The ICA was intercepted by a suture line. A silicone rubber coated nylon monofilament was inserted into the ECA until it reached the origin of the middle cerebral artery. After 90 min of occlusion, the monofilament was removed for reperfusion. The rats were deeply anesthetized and brains were immediately removed for further analysis after reperfusion. The sham group underwent the same procedure without occlusion. Body temperature was maintained at 37 ± 0.5°C during operation.

Lateral intracerebroventricular injection
Rats were anesthetized and the right lateral cerebral ventricle was exposed to a stereotaxic apparatus. The coordinates were 0.8 mm anteroposterior, 1.5 mm lateral, and 3.5 mm dorsoventral to the bregma [38]. LV-TUG1-RNAi (1 × 10 8 TU/mL, GeneChem, China) or negative control were slowly injected into the lateral ventricle in 10 min (2.5 µL). The rats were then fed for 14 days and subjected to MCAO/R. miR-204-5p agomir, antagomir, and their negative controls were synthesized by GenePharma (Shanghai, China) and diluted in artificial cerebrospinal fluid (Dingguo, China). One day before the MCAO/R operation, 5 µL (100 µM) miR-204-5p agomir, antagomir, or heir negative controls were injected into the right lateral ventricle, respectively. The rats were sacrificed the next day to obtain brain tissue.

Neurological deficit assessment
Neurological deficit assessment was performed after reperfusion for 24 h using the Zea-Longa method: 0 points, no behavior disorder; 1 point, flexion of the left forelimb; 2 points, body turns to the left in a circle; 3 points, dump to the left; and 4 points, inability to walk autonomously with disturbance of consciousness.

Infarct volume measurement
The freezing brains were cut into slices and stained with 2% 2,3,5-triphenyl tetrazolium chloride (TTC, Sigma, St. Louis, USA) dissolved in phosphatebuffered solution (PBS) for 20 min at 37 o C. The stained slices were fixed with 4% paraformaldehyde dissolved in PBS (0.1 M, pH 7.2) and photographed with a camera. Infarct volume (%) was measured using Image J software (Bethesda, USA) using the formula: volume of the infarct area/volume of total brain × 100%.

Hematoxylin-eosin staining
The rats were anesthetized and transcardially perfused with 100 mL PBS and fixed with 50 mL of 4% paraformaldehyde. The whole brain was removed and stored in 4% paraformaldehyde, cut into 5 μm thick coronal sections, and then stained with hematoxylin and eosin solutions. The images were captured by a microscope (Olympus, Tokyo, Japan).

Superoxide dismutase (SOD) and malondialdehyde (MDA) assay
Brain tissues were thoroughly ground in cold saline and the suspension was centrifuged at 3500 rpm for 10 min. The SOD activity and MDA content were detected by commercial kits (Nanjing Jiancheng, China).

Cell cultures
Primary cortical cultures were prepared from E17 embryos from SD rats [39]. The cerebral cortex of the embryos was isolated and chopped immediately and then incubated with 0.25% trypsin-EDTA (Gibco, USA) at 37°C for 15 min. Digestion was stopped with Dulbecco's modified Eagle Medium/Ham's F-12 containing 10% fetal bovine serum (Biological Industries, Israel). The suspension was filtered with a 200 mesh screen and centrifuged for 5 min at 1000 r/min. Cells were suspended in neurobasal medium (Gibco, USA) supplemented with 2% B27 (Gibco, USA), 0.5 mM L-glutamine (Gibco, USA) and 1% penicillin-streptomycin (10,000 U/ml, Gibco, USA), and seeded at a density of 2 × 10 5 cells/cm 2 in plates precoated with poly-L-lysine (0.1 mg/ml, Sigma-Aldrich, USA). Cells were cultured at 37°C in humidified air (5% CO 2 /95%) and the medium was changed half every three days.

Oxygen and glucose deprivation and re-oxygenation
Cell medium was replaced with glucose-free DMEM (Gibco). Cells were incubated in an anaerobic incubator for 1 h, and then placed in a complete neurobasal medium in normoxia for another 24 h. Fig. 6 COX2 was modulated by both TUG1 and miR-204-5p. A Relative expression of COX2 after transfected with COX2 siRNAs in neurons (n = 5). B MTT assay was performed to assess cell viability after co-transfected with miR-204-5p inhibitor and COX2 si-3 (n = 6). C Relative protein levels of COX2, Bcl-2 and Bax in neurons after co-transfected with miR-204-5p inhibitor and COX2 si-3 (n = 5). Data are presented as the mean ± SD. ***p < 0.001.

Cell viability assay
The survival of neurons with OGD/R treatment was evaluated using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. MTT solution (5 mg/mL, Sigma-Aldrich) was added to each well and incubated at 37 o C for 4 h. The supernatant was discarded and the formazan was dissolved in 150 μL dimethyl sulfoxide. The optical density of each well was determined using a microplate reader (Thermo Scientific, USA) at 490 nm.

Quantitative real-time PCR
Total RNA was extracted using a trizol reagent (Vazyme, China). RNAs were reverse transcribed to cDNA and miRNA using an all-in-one cDNA synthesis supermix (Bimake, USA) and a miRNA first-strand cDNA synthesis reagent kit (Sangon, China). Expressions of mRNA and miRNA were measured using the 2 × SYBR Green qPCR Master Mix (Bimake). The primer sequences are shown in Table S1.
Neurons were fixed with 4% paraformaldehyde, permeabilized with 0.3% Triton X-100 in PBS and blocked with 5% goat serum for 2 h. The remaining steps were consistent with the method mentioned above.

Apoptosis analysis
Cell apoptosis was evaluated using the Fluorescein (FITC) tunel cell apoptosis detection kit (Servicebio, China). Cells were gently washed with PBS and fixed with 4% paraformaldehyde. The cells were then permeabilized with 0.1% Triton X-100 for 20 min. A 50 μL equilibration buffer was added and incubated at room temperature for 30 min. The balanced calibration buffer was removed and cells were incubated in 56 μL TdT incubation buffer at 37 o C for 2 h. The nuclei were then stained with DAPI for 8 min at room temperature. The cells were imaged by an Olympus fluorescence microscope. Fig. 7 COX2 was a target of miR-204-5p. A Presentation of the putative binding site of miR-204-5p and COX2, and the designed mutant sequence. B The relative luciferase activity of HEK 293 cells co-transfected COX2-WT or COX2 MT with miR-204-5p mimic or NC (n = 5). C The illustration of how TUG1/miR-204-5p/COX2 axis was involved in ischemia and reperfusion induced neuron injury. Data are presented as the mean ± SD. ***p < 0.001.

Dual-luciferase reporter gene assays
The supposed binding site of miR-204-5p and COX2 3′-UTR was cloned by PCR and inserted into a pmirGLO vector to create COX2 WT. The corresponding mutants COX2 MT were constructed according to the supposed binding site. HEK 293 cells were transfected with pmirGLO vectors and miR-204-5p mimic by lipofectamine 2000 (Thermofisher, USA). After 48 h, dual luciferase detection was performed using the Dual-Luciferase Reporter Gene Assay Kit (Promega, USA). Firefly luciferase activity was normalized to Renilla luciferase activity.

Statistical analysis
All data were expressed as mean ± standard deviation (SD). Data were from at least three independent experiments and analyzed using GraphPad Prism 5. A single comparison between two groups was analyzed by an independent t-test, and multiple group comparisons were analyzed with a one-way analysis of variance (ANOVA). The association of the two variables was evaluated using a Pearson correlation analysis with twotailed test, with p < 0.05 considered significant.

DATA AVAILABILITY
The datasets are available from the corresponding author on reasonable request. Supplementary information is available at Cell Death Discovery's website.