A Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) dataset (GSE115697) that assessed circRNA expression patterns in NSCLC and was analyzed by using GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was obtained and used in the present study (20). Raw data were normalized by Quantile algorithm, limma package the R program (21). R software (version 4.0.3) was used for the data download and processing (22). The ‘GEOquery’ package was used to download the dataset expression matrix and platform file. The ‘limma’ package was used for differential analysis of microarray data. After confirming that the quality of the samples, the subsequent difference analysis was performed. Significant differentially expressed transcripts were screened using P<0.05 and fold-change ≥0.5 or ≤-1 The ‘ggplot2’ package was used to visualize the heatmap and volcano plot.

The mouse neuroblast (Neuro-2a cells; http://www.atcc.org/products/ccl-131) were obtained from the American Type Culture Collection. Mouse neuronal cells were cultured in DMEM (Invitrogen; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (Invitrogen; Thermo Fisher Scientific, Inc.) and antibiotics (100 U/ml penicillin, 100 µg/ml streptomycin, PAN Biotech), and maintained in a humidified incubator at 37°C and supplied with 5% CO₂ air.

For the establishment of the OGD/R cell model, the original culture medium was discarded, the neuronal cells were washed with PBS, placed in sugar-free DMEM medium without FBS, and incubated as above for 2 h. Next, the media was removed, and supplemented DMEM was added, and the cells were transferred to an incubator supplied with air consisting of 95% N₂, 5% CO₂, and 1% O₂ for 6 h (23). Subsequently, the media was replaced with supplemented DMEM and transferred back to the previous conditions for 12 h. In addition, cells cultured in normal DMEM at 37°C with 5% CO₂ in a humidified incubator were used as the control.

To induce circ_0000018 knockdown, OGD/R-treated cells and control cells were transfected with 50 nM circ_0000018 small interference RNA (si-circ_0000018; Guangzhou RiboBio Co., Ltd.) or non-targeting siRNA (si-NC, Guangzhou RiboBio Co., Ltd.). Similarly, 50 nM miR-871 mimic (agomiR-871), miR-871 inhibitor (antagomiR-871) and their blank controls (miR-NC and anti-miR-NC; all from Guangzhou RiboBio Co., Ltd.) were respectively transfected into neuronal cells. The overexpression vector (pcDNA-BCL2L11) was obtained by inserting the BCL2-like 11 overexpressed CDS sequence (oe-BCL2L11) into a pcDNA vector (Invitrogen; Thermo Fisher Scientific, Inc.), and the empty vector was used as the negative control. A total of 200 ng vector was transfected into neuronal cells using Lipofectamine^® 3000 (Invitrogen; Thermo Fisher Scientific, Inc.).

Male C57BL/6J mice (7–8 weeks, 18–25 g) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. The animals were randomly separated into groups and kept in a temperature-controlled room at 25±2°C with a 40–50% relative humidity and a 12 h light/dark cycle. All experimental procedures were approved by the Southeast University Animal Care and Use Committee (approval no. 20220110026).

The mice were randomly allocated into a sham group, tMCAO group, tMCAO+sh-NC group, or tMCAO+sh-circ_0000018 group (n=18 per group). After 7 days of acclimatization, a model of cerebral AIS was created in mice by tMCAO with modifications based on earlier reports (24,25). Briefly, 1% sodium pentobarbital (50 mg/kg) was injected intraperitoneally to anesthetize the mice. Following a midline skin incision, the right common carotid artery, external carotid artery, and internal carotid artery (ICA) were explanted, and the L1800 silicone wire (Jialing Biotechnology Co., Ltd.) was then inserted 10 mm into the ICA to obscure the origin of the MCA. Brain reperfusion was performed by withdrawing silicone filaments after 90 min. In the sham group, the mice underwent the same tMCAO procedure but without ICA occlusion. The sh-NC and sh-circ_0000018 (Shanghai GenePharma Co., Ltd.) were injected into the lateral ventricle 1 day prior to tMCAO. Throughout the procedure, the temperature of the rectum of the mice was kept at 37.0±0.5°C using a heat lamp (Jiaxing Nomoy Pet Products Co., Ltd.). Following the neurobehavioral scoring of all mice, the mice were euthanized by inhalation of carbon dioxide (35% volume displacement rate/min) inhalation following isoflurane anesthesia (induction percentage, 3%; maintenance percentage, 1%), and brain tissue was obtained for subsequent experiments.

The neurological deficits of mice were evaluated using the Longa biologic score 24 h after 1 h of brain ischemia and reperfusion (26). The scoring system used was: 0, both forelimbs are strong and symmetrically extended to the ground, both shoulders have the same resistance, and walking is normal; 1, internal rotation on the contralateral side of the surgery, forelimb tucked in, both shoulders resist in unison, and walking is normal; 2, internal rotation on the contralateral side of the surgery, forelimb inversion, decreased resistance on the contralateral side of the surgery when pushing both shoulders, and walking is normal; 3, when the surgical contralateral side is internally rotated, the forelimb is tucked in, and both shoulders are pushed; when pushing both shoulders, the resistance of the contralateral side of the surgery decreases, mice can walk around; and 4, no spontaneous movement on the opposite side of the surgery. The observers were blinded to the treatment.

After assessment of neurological damage, mice were euthanized by CO₂ inhalation following halothane anesthesia, and brains were rapidly collected, cut into 2-mm sections, and incubated with 2% TTC solution (MilliporeSigma) at 37°C for 10 min. Subsequently, 4% paraformaldehyde was used to fix the brain tissue for 1–2 days at room temperature, and then an image was taken. The infarct volume was calculated using ImageJ 1.8.0.345 (National Institutes of Health). Normal brain tissue appeared red when TTC reacted with dehydrogenase, while in the ischemic brain tissue, it appeared white due to reduced dehydrogenase activity. The infarct volume ratio as a percentage was calculated as follows: Infarct volume (%)=[(Sum of infarct area ×2 mm³)/(sum of total brain area ×2 mm³) ×100.

Total RNA was isolated from cultured cells, serum, and brain tissue using TRIzol^® reagent according to the manufacturer's protocol (Invitrogen; Thermo Fisher Scientific, Inc.). The PrimeScript RT kit was used to reverse transcribe total RNA to cDNA according to the manufacturer's protocol (Takara Bio, Inc.). The levels of circ_0000018, miR-871, and BCL2L11 were measured using qPCR with a SYBR green kit (Takara Bio, Inc.) in a Bio-Rad CFX96 system (Bio-Rad, Laboratories, Inc.). The procedure was as follows: 95°C for 3 min; followed by 39 cycles of 95°C for 15 sec, 60°C for 60 sec and 72°C-30 sec for mRNA, or 95°C for 15 sec and 60°C for 60 sec for miRNA; 95°C for 10 sec, followed by a melt curve analysis (60–95°C, 0.5°C increments for 20 sec) to confirm specificity of the PCR primers. The expression levels were normalized to the expressions of β-actin or U6. The 2^−∆∆Cq method was used to determine gene expression in the neuronal cells and brain tissues (27).

The sequences of the primers used were: circ_0000018 forward, 5′-CAAGATCACCTCCGATTGGT-3′ and reverse, 5′-TGTCTTCTGCTCCAGGATCTTT-3′; miR-871 forward, 5′-TGCGGTCTGACCGTGGTAAGACC-3′ and reverse, 5′-CCAGTGCAGGGTCCGAGGT-3′; BCL2L11 forward, 5′-CCCGGAGATACGGATTGCAC-3′ and reverse, 5′-GCCTCGCGGTAATCATTTGC-3; and U6 forward, 5′-CTCGCTTCGGCAGCAC-3′ and reverse, 5′-ACGCTTCACGAATTTGC-3′; cicr_0001646 forward, 5′-GGCAGATGGAAGCTTCTTGA-3′ and reverse, 5′-AGCAAGTTGACCCATTTTCC-3′.

Total protein was extracted from cells using RIPA lysis buffer (Cell Signaling Technology, Inc.) and quantified using a BCA Protein Quantification Kit (Abbkine Scientific Co., Ltd.). Equivalent quantities of protein (30 µg) was loaded per lane on a 10% SDS-gel, resolved using SDS-PAGE, and transferred onto PVDF membranes (MilliporeSigma), followed by blocking with 5% milk in TBST at room temperature for 2 h, and subsequent incubation with one of the primary antibodies at 4°C overnight. The corresponding secondary antibodies were used to incubate the membranes for 1 h at room temperature, followed by visualization using a chemiluminescence detection kit (Beyotime Institute of Biotechnology). The primary antibodies used in the present study were: Toll-like receptor 4 (TLR4; ab22048; 1:300), β-tubulin III (ab18207; 1:3,000), Bax (ab32503; 1;300), Bcl-2 (ab182858; 1:3,000), Caspase-3 (ab32351; 1:1,000), H2A.X Variant Histone (H2AX) (ab124781; 1:1,000), γH2AX (ab81299; 1:5,000) and β-actin (ab8226; 1:1,000) (all Abcam). The secondary antibody used in the present study were as follows: HRP-Conjugated AffiniPure Goat Anti-Rabbit/Mouse IgG H&L (cat. no. BA1056; 1:5,000) and HRP-Conjugated AffiniPure Donkey Anti-Rabbit IgG H&L (cat. no. BA1061; 1:2,000) (both Wuhan Boster Biological Technology, Co., Ltd.).

A total of 2 µg RNA was analyzed by incubation at 37°C for 30 min with or without 5 U/µg RNase R (Epicenter; Illumina, Inc), following purification with an RNeasy MinElute Cleaning Kit (Qiagen GmbH), and then assessed using RT-qPCR.

The nerve cells in logarithmic growth stage were used, and when the cell density reached 80–90%, the adherent cells were enzymatically dissociated into a single-cell suspension. The neuronal cells were exposed to 2 µg/ml actinomycin D (MilliporeSigma). The cells were then collected and total RNA was withdrawn. circRNA and mRNA stability was analyzed using RT-qPCR.

The cells in each group following treatment were seeded in 96 well plates with a cell density of 2×10⁴ cells/well. Following 0, 24, and 48 h of culture, 10 µl CCK-8 solution (Beyotime Institute of Biotechnology) was added to each well. Following culture at 37°C for 1.5 h, the absorbance of the cells at a wavelength of 450 nm was measured using an Elx800 reader (Omega Bio-Tek, Inc.) as a measure of the viability of the cells (28).

The OGD/R-treated neuronal cells were cultured for 48 h following transfection to assess apoptosis (18). After the culturing, cells were stained using an Annexin V FITC Apoptosis Detection Kit according to the manufacturer's protocol (BD Biosciences). The proportion of apoptotic cells in each group was analyzed using a flow cytometer (CytoFLEX; Beckman Coulter, Inc.). The flow data were analyzed by FlowJoTM v10.6.1 (BD Biosciences).

The online tool StarBase (https://starbase.sysu.edu.cn/) was used to predict the putative miR-871 binding sites in circ_0000018 and BCL2L11. A fragment of circ_0000018 untranslated region (3′-UTR) with wild-type (circ_0000018-WT) and mutant (circ_0000018-MUT) was introduced into the pmirGLO luciferase vector (E1330, Promega Corporation). Similarly, the BLC2L11-3′-UTR fragment with wild type (BLC2L11-WT) and mutant (BLC2L11-MUT) was introduced into the pmirGLO luciferase vector. Lipofectamine^® 3000 was used to co-transfect circ_0000018-WT/circ_0000018-MUT with BLC2L11-WT/BLC2L11-MUT and agomiR-871 or agomiR-NC into neuronal cells. Following transfection for 48 h, cells were collected, and the activity of luciferase was determined using a SpectraMax L fluorometer (Molecular Devices, LLC).

RNA pull-down assays were performed as previously reported (29). Briefly, neuronal cells (1×10⁷ cells) were collected and lysed. Glycosylated miR-871 probes were synthesized by Shanghai GenePharma Co., Ltd. and cultured with streptavidin agarose beads (Thermo Fisher Scientific, Inc.). The cell lysate of the miR-871 probe or oligo probe (control) was incubated overnight at 4°C. The bead-bound RNA complexes were purged with wash buffer, and the degree of enrichment of circ_0000018 extracted by the miR-871 probe was examined using RT-qPCR.

Data were analyzed using GraphPad Prism version 7.0 (GraphPad Software, Inc.). Comparisons between two groups were performed using a Student's t-test, and comparisons between multiple groups were performed using a one-way ANOVA with a post-hoc Tukey's test. Data are presented as the mean ± standard deviation of three repeats.

First, the GSE115697 dataset was downloaded from GEO and analyzed in R. The differentially expressed circRNAs in the tMCAO mice were obtained and expressed as a Volcano plot (Fig. 1A). Among these differentially expressed circRNAs, 10 circRNAs with the most notable differences in expression were selected and their levels in the OGD/R-treated cells were determined using RT-qPCR. The findings demonstrated that circ_0000018 and circ_0001646 levels were significantly increased in OGD/R treated cells compared with those in normal neuronal cells, while circ_0000018 levels were decreased. Of these two circRNAs, the difference in circ_0000018 levels expression was greater (Fig. 1B). Therefore, circ_0000018 was selected for subsequent experiments. In addition, following ribonuclease R (Fig. 1C) or actinomycin D (Fig. 1D) treatment, the linear RNA levels were significantly reduced, while no significant change was observed in circular RNA levels.

To explore the effect of circ_0000018 in vitro, neuronal cells were transfected with si-circ_0000018 1#, si-circ_0000018 2#, or si-NC. The RT-qPCR results confirmed that transfection with siRNA plasmids targeting circ_0000018 markedly reduced the levels of circ_0000018 in neuronal cells, and si-circ_0000018 2# was selected for the subsequent experiments given the better knockdown efficiency (Fig. 2A). Furthermore, to explore the effects of circ_0000018 knockdown on AIS in vitro, an ODG/R cell model was established using transfected neuronal cells. CCK-8 assays showed that the knockdown of circ_0000018 markedly reduced neuronal cell activity (Fig. 2B). In addition, flow cytometry analysis indicated that apoptosis of neuronal cells was induced by OGD/R treatment, whereas circ_0000018 knockdown attenuated the apoptosis of OGD/R-stimulated neuronal cells (Fig. 2C).

Using bioinformatics analysis, the top-7 miRNAs targeted by circ_0000018 were obtained. Next, based on PCR, only miR-871 expression was decreased in the OGD/R-treated neuronal cells (Fig. 3A). miR-871 was selected for further study. In Fig. 3B, the binding site between circ_0000018 and miR-871 is shown. Dual luciferase assays indicated that agomiR-871 markedly decreased the luciferase activity of the circ_0000018-wt reporter vector, with no influence on the MUT reporter vector (Fig. 3C). The RNA pull-down assays also showed that miR-871 and circ_0000018 could bind to each other (Fig. 3D). Furthermore, it was discovered that the levels of miR-871 were markedly increased in the OGD/R-treated neuronal cells transfected with si-circ_0000018 compared with cells transfected with si-NC (Fig. 3E). These findings revealed that circ_0000018 negatively regulated the expression of miR-871 by specifically binding to miR-871.

For the purpose of further investigating the roles of a circ_0000018/miR-871 axis on the injury of the OGD/R-treated cells, the OGD/R-treated neuronal cells were transfected with si-circ_0000018 and antogomiR-871. The transfection efficiency of antogomiR-871 was verified by RT-qPCR, and it was confirmed that antogomiR-871 transfection significantly decreased the miR-871 levels (Fig. 4A). The results of the CCK-8 assay indicated that si-circ_0000018 markedly reduced the survival of OGD/R-treated neuronal cells, while co-transfection with antagomiR-871 reduced this inhibitory effect (Fig. 4B). Similarly, the inhibitory effect on cell apoptosis induced by circ_0000018 downregulation was markedly alleviated by antagomiR-871 in OGD/R-treated neuronal cells (Fig. 4C). Together, these results showed that downregulation of miR-871 can partially reverse the effects of si-circ_0000018 on the apoptosis and proliferation of neuronal cells following treatment with OGD/R. Proof of transfection for agomiR-871 and agomiR-NC in neuronal cells was obtained by RT-qPCR (Fig. S1A).

There is a general consensus that miRNAs exert their biological effects by targeting mRNAs (30–32). Therefore, StarBase was used to find potential target genes for miR-871. Several binding sites were obtained between miR-871 and BCL2L11 (Fig. 5A). Next, a dual luciferase assay performed in neuronal cells confirmed this binding relationship. The findings demonstrated that the upregulation of miR-871 markedly decreased the luciferase activity of the BCL2L11-wt reporter vector, whereas the luciferase activity of BCL2L11-mut was not decreased (Fig. 5B). The RNA pull-down assay demonstrated that miR-871 and BCL2L11 could specifically bind to each other (Fig. 5C). In addition, RT-qPCR and western blotting showed that the introduction of agtagomiR-871 decreased the effects of si-circ_0000018 on the mRNA levels of BCL2L11 in neuronal cells (Fig. 5D and F), revealing that circ_0000018 could regulate the levels of BCL2L11 by sponging miR-871. Consistently, it was found that BCL2L11 expression was increased in OGD/R-treated neuronal cells compared with the corresponding control group (Fig. 5E and G), which demonstrated that BCL2L11 may be involved in OGD/R-treated cell trauma in vitro.

To verify the potential influence of the circ_0000018/miR-871/BCL2L11 axis on the neuronal injury induced by ODG/R, rescue experiments were performed. Transfection efficiencies were determined by RT-qPCR (Fig. 6A and B). In addition, CCK-8 assays revealed that the proliferation of OGD/R-treated neuronal cells transfected with agomiR-871 was significantly increased, and this was abrogated by BCL2L11 overexpression (Fig. 6C). The flow cytometry results suggested that the treatment with ODG/R promoted apoptosis, whereas transfection of agomiR-871 in ODG/R-treated cells decreased apoptosis; further overexpression of BCL2L11 significantly reversed the effects of agomiR-871 on stimulating cell apoptosis (Fig. 6D).

Finally, the contribution of circ_0000018 in vivo in the tMCAO mouse models was evaluated. RT-qPCR showed that circ_0000018 (Fig. 7A) and BCL2L11 (Fig. 7C) expression were significantly increased while miR-871 expression (Fig. 7B) was decreased in mouse brain tissues in the tMCAO group. Mice in the tMCAO group presented with increased neurological damage and brain infarction when compared with the sham group (Fig. 7D-F), and knockdown of circ_0000018 significantly ameliorated these changes. RT-qPCR results showed that the expression of circ_0000018 in tissues decreased after transfection with the plasmid with knockdown of circ_0000018 (Fig. S1B).