Identification of the biological function of miR-9 in spinal cord ischemia-reperfusion injury in rats

Experimental animals

Sprague–Dawley rats (male, 200–250g) were obtained in the present study. All rats were maintained for at least 1 week before the surgical procedures, with freely available rodent chow and water at 22–24 °C and 50–60% relative humidity, under a 12h/12h light-dark cycle. The present study had approval from the Ethics Committee of China Medical University, Shenyang (CMU 2020391), and were carried out in conformity with the National Institutes of Health Guide for the Use and Care of Laboratory Animals (NIH Publications No.80-23, revised 1996).

Rat model

To create SCII rat models, a cross-clamped aortic arch was used as previously reported (Chen et al., 2020a; Chen et al., 2020b). Pentobarbital sodium (50 mg/kg) was intraperitoneal injected for anesthetizing rats. SCII was induced by the aorta occlusion for 14 min. The same surgical procedures without occlusion were performed on sham-operated rats.

Euthanization

Rats were euthanized by sevoflurane overdose and ensured that it was effective by pinching the tail with tweezers. Any movement in the rat showed that pain could still be felt so enough time was allowed for the anesthesia to fully work before sacrificing the mice. Rats that survived the study or were excluded were bred for other experiments.

Quantification of miRNA expression

Total RNA was extracted from segments L4–L6 of the spinal cord with using Trizol reagent (Takara, Otsu, Japan). RNA was reverse-transcribed into cDNA using the Prime Script® miRNA cDNA Synthesis Kit (Takara, Tokyo, Japan) (Li et al., 2016b; Wang et al., 2015). The levels of miR-9 were detected using SYBR Premix qRT-PCR on a PCR System (Corbett Research, Australia) (Chen et al., 2020a). The primers used for miR-9 in the present study were as follows: forward: 5′-CGCGCTCTTTGGTTATCTAGCTGTATG-3′ (Yao, Wang & Zhang, 2018). Relative miR-9 expression was normalized to U6 expression levels using the 2^−ΔΔCt method.

Intrathecal administration

Thoracic laminectomy was performed for intrathecal pretreatment (Li et al., 2015b). Pretreatment with a synthesized miR-9 mimic (5′-UCUUUGGUUAUCUAGCUGUAUGA-3′), inhibitor (5′-TCATACAGCTAGATAACCAAAGA-3′) and negative control (NC) was previously described (Li et al., 2015b). Rats were intrathecally injected with 10 μl of the oligonucleotides (500 pmol/10 μl) and EntransterTM-in vivo transfection regent (Engreen, Beijing, China) (Wang et al., 2020). Intrathecal injection was applied in vivo prior to ischemia induction once a day for three consecutive days according to the results of our preliminary experiment(Li et al., 2016b; Li et al., 2015b). In order to inhibit the expression of MAP2K3, SB203580 (5 μl, dissolved in 0.1 nmol/μl solution using 1% DMSO) once daily for 2 consecutive days (Yao, Wang & Zhang, 2018; Zhou et al., 2016) at the same time. Notch2 siRNA or NC RNA were provided by Jima Inc. (Shanghai, China). For suppressing the expression of Notch2, two days before ischemia, intrathecal infusion of 5 μg of siRNA(5′- CCTCCCATCGTGACTTTCCAGCTTA-3′)or control RNA at a concentration of 1 μg/μl once a day was carried out, as directed by the manufacturer (Chen et al., 2020b; Meng et al., 2015; Wang et al., 2020).

Experimental protocol

Neurological evaluation

Two observers evaluated the movement function of the rat lower limb based on the BBB scoring after SCII as described previously (Bao et al., 2018; Basso, Beattie & Bresnahan, 1995).

Bioinformatics analysis for potential target genes of miR-9

At first, targets of miR-9 were predicted by means of databases miRDB and Targetscan. Target genes were identified for miR-9 in both databases. Then the raw data of GEO Series (GSE)138966 [species: Rattus norvegicus; Platforms: GPL22396Illumina HiSeq 4000 (Rattus norvegicus)] was obtained from Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/). 3 sham-operated samples and 3 SCII samples at 48 h post-SCII were included in GSE138966(Ding et al., 2020). The differentially expressed genes (DEGs) between sham-operated samples and SCII samples were obtained. DEGs are those genes with an |log2FC| ≥ 1 and p < 0.05. For the purpose of analyzing the diversification of DEGs expression, the heatmaps of DEGs were drawn by heatmap function. We chose the intersection of the up-regulated DEGs and the above predicted genes as the potential miR-9 target genes and conducted bioinformatics analysis.

DAVID 6.8 (https://david.ncifcrf.gov/) is a database with annotation, visualization, and integrated discovery functions (Huang, Sherman & Lempicki, 2009). The annotation table is its main analysis tool, which contains functional annotation charts, functional annotation clustering, and functional annotation table subtools (Wang et al., 2021). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation of DEGs was carried out through the annotation tool. The biological functional coherence and biological attributes of the putative target genes were determined based on GO analysis. KEGG is a collection of databases dealing with genomes, diseases, biological pathways, drugs and chemical materials (Altintas et al., 2016). The biological processes and pathways of potential miR-9 target genes were analyzed using the DAVID online database. p < 0.05 was selected as the cut-off criterion with statistic difference. According to the relations of genes and statistically significant pathways as well as the relations among genes, pathways and miR-9,we also built a miRNA–pathway–gene network

Evans blue (EB) extravasation

EB cannot normally pass through BSCB and thus its presence in spinal cord tissue indicates BSCB disruption (Goldim, Della Giustina & Petronilho, 2019). At the time point of 48 h after SCII, the intravenous injection of EB (45 mg/kg) was performed. The rats were euthanized after 1 h circulation. Segments L4–L6 were homogenized with trichloroacetic acid, and the tissues were centrifuged (Li et al., 2015b). The absorption of the supernatant was measured at 632 nm with a microplate reader (BioTek, Winooski, USA) (Fang et al., 2015). EB staining was visualized using a fluorescent microscopy (Leica, German) with a green filter (Chen et al., 2020a).

Water content of spinal cord

Water content (%) of segments L4–L6 was obtained using the following formula: (wet weight − dry weight)/wet weight × 100% (Li et al., 2018b). Spinal cord tissues were rapidly removed to measure wet weight. Then at 105 °C for 48 h the tissues were dried and measure dry weight.

TUNEL (terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling) Assay

10-mm-thick sections were subjected to fluorometric TUNEL assay using the commercial kit (Beyotime, Haimen, China) to detect the apoptotic DNA strand breaks (Bao et al., 2018). The sections were fixed with 4% neutral-buffered formaldehyde for 20 min and incubated with proteinase K for 20 min, followed by the labeling reaction for 2 h. Then, the nuclei were stained with DAPI and images captured by fluorescence microscopy.

Western Blotting

The protein expression levels of MAP2K3, Notch2 in segments L4–L6 of spinal cord tissues were detected with Western Blotting. Spinal cord tissues were collected at 48 h after SCII. Total proteins were extracted by using protein lysis buffer. Rabbit monoclonal anti-MAP2K3 (1:5,000, abcam), rabbit monoclonal anti-Notch2 (1:500, Santcruz), rabbit monoclonal anti-cleaved caspase-3 (1:500, abcam) and HRP-conjugated secondary antibodies (Beyotime, Haimen, China) were used.

Enzyme-linked immunosorbent assays (ELISAs)

The rats were euthanized at 48 h after SCII. ELISA kits purchased from Signalway Antibody Company (College Park, maryland, USA) were used to test the levels of IL-6 and IL-1β according to the instructions of the manufacturer.

Statistical analysis

Data were expressed as mean ± standard deviation and analyzed by SPSS 15.0 (IBM, USA). All variables were calculated with one-way ANOVA followed by the Newman–Keuls post hoc test. p < 0.05 was defined significant.

Expression of miR-9

At first, we verified the expression of miR-9 at 24, 48 and 72 h after SCII by qRT-PCR. The results showed that miR-9 levels significantly decreased at 24 and 48 h (p < 0.01). At 48 h after SCII, miR-9 expression demonstrated the lowest level (p < 0.01), as presented in Fig. 1A. These results implicated that aberrant expression of miR-9 might be related with SCII.

miR-9 mimic improved neurological function following SCII

At 48 h after SCII, we assessed the levels of miR-9 by qRT-PCR after intrathecal injection of miR-9 mimic for 3 days. miR-9 mimic significantly increased the expression of miR-9 (p < 0.05) (Fig. 1B). In addition, SCII induced severe neurological deficits of lower extremities (p < 0.01) and miR-9 mimic improved neurological function after SCII (p < 0.01) according to BBB scores (Fig. 1C).

miR-9 mimic attenuated BSCB leakage following SCII

Under a fluorescent microscope, EB extravasation exhibits red color (Fang et al., 2015; Li et al., 2014b). EB extravasation at 48 h after SCII was more (p < 0.01 versus the sham rats). In contrast, EB extravasation was reduced with miR-9 mimic pretreatment (p < 0.01), as depicted in Fig. 2A. EB fluorescence density and content were calculated, as depicted in Figs. 2B and 2C. Spinal cord edema was evaluated based on water content. SCII induced spinal cord edema, whereas miR-9 mimic reduced spinal cord edema (p < 0.05) (Fig. 2D).

miR-9 mimic reduced apoptosis and neuroinflammation following SCII

We explored effects of miR-9 mimic on apoptosis at 48 h after SCII. Figures 3A–3B demonstrates the TUNEL results. Apoptosis rate increased following SCII (p < 0.01). Lower apoptosis rate was found in operated rats subjected to miR-9 mimic pretreatment compared to the SCII group (p < 0.01). The expression of cleaved-caspase3 protein increased at 48 h after reperfusion (p < 0.01), whereas miR-9 mimic significantly attenuated this downregulation (p < 0.01). as presented in Fig. 3C.

We assessed activation of IL-6 and IL-1β expression by ELISAs after miR-9 mimic pretreatment. The data meant that IL-6 and IL-1β were all upregulated at 48 h after SCII. Intrathecal pretreatment with miR-9 mimic evidently reduced this upregulation (p < 0.01). The findings are presented in Figs. 3D–3E.

Bioinformatics analysis of potential miR-9 targets

To study the participation of miR-9, target mRNAs of miR-9 were predicted by means of databases Targetscan and miRDB. 327 potential target genes were selected for miR-9 in 2 databases (Table S1). In addition, 4,829 differentially expressed genes (DEGs) (2,650 upregulated and 2,179 downregulated) were attained with a |log2FC| ≥ 1 and p < 0.05 between sham-operated samples and SCII samples in GSE138966 were obtained (Table S2). A clustered heatmap of top 50 upregulated and downregulated DEGs in GSE138966 is displayed in Fig. 4A. 87 intersection genes of the up-regulated DEGs and the above predicted genes were identified as potential miR-9 target genes (Table S3 and Fig. 4B). We further conducted GO and KEGG pathway analysis to detect the key target genes of miR-9.

GO analysis was performed to investigate the enrichment of miR-9 potential target genes on biological processes. Several biological processes such as “central nervous system development”, “regulation of growth” and “response to cytokine” (p < 0.05), were significantly enriched. The results are shown in Fig. 5A. Enrichment of miR-9 potential target genes was also conducted by KEGG pathway analysis. Several significantly representative enriched pathways for potential target genes of miR-9 were identified. The results were notably enriched in several pathways, including “Notch signaling pathway”, “MAPK signaling pathway”, “Focal adhesion” and “Prolactin signaling pathway” (p < 0.05), as shown in Fig. 5B and Table 1. The miRNA–pathway–gene network is demonstrated in Fig. 5C. It was noted that MAP2K3 and Notch2 were enriched in MAPK signaling pathway and Notch signaling pathway, respectively (Fig. S1).

miR-9 mimic or SB203580 improves neurological function and alleviates the upregulation of cytokines after SCII

Figure 6A demonstrated the neurological function which was evaluated according to BBB scores. SCII induced neurological deficits of lower limbs. Compared to the SCII group, miR-9 inhibitor decreased BBB scores (p < 0.01). SB203580 enhanced neurological function recovery (p < 0.01). The miR-9 inhibitor + SB203580 group showed higher BBB scores (p < 0.01 vs SCII group). The results implied SB203580 could eliminate neurological deficits caused by miR-9 inhibitor. The protein expression of MAP2K3 increased at 48 h after SCII (p < 0.01), whereas miR-9 mimic notably reduced this increase (p < 0.01) and miR-9 inhibitor exacerbated the increase. SB203580 recuded the protein expression of MAP2K3 (p < 0.01) and SB203580 also reversed the effect of miR-9 inhibitor, as presented in Fig. 6B. Next, we detected the expression levels of IL-6 and IL-1β in eight groups. As shown in Figs. 6C–6D, miR-9 inhibitor increased IL-6 and IL-1β (all p < 0.01 vs SCII group), while SB203580 reduced the release of IL-6 and IL-1β (all p < 0.01 vs SCII group). In addition, IL-6 and IL-1β expression were reduced in the miR-9 inhibitor+ SB203580 group compared with the miR-9 inhibitor group.

miR-9 mimic or Nocth2 siRNA improves neurological function and reduces the protein expression of cleaved-caspase3 after SCII

Figure 7A showed the neurological function which was evaluated according to BBB scores. SCII induced severe neurological deficits of lower limbs (p < 0.01); meanwhile, si-Notch2 relieved neurological damage induced by SCII (p < 0.01), while miR-9 inhibitor aggravated neurological deterioration (p < 0.05). And the miR-9 inhibitor + si-Notch2 showed higher BBB scores (p < 0.01 vs SCII group) indicating si-Notch2 could eliminate the damage caused by miR-9 inhibitor. The effect of miR-9 on Notch2 protein expression was similar to that on MAP2K3 protein expression, as present in Figs. 7B–7C. Then we assessed the expression of cleaved-caspase3 in SCII. As shown in Figs. 7B–7D, intrathecal injection of Notch2 siRNA reduced the increase of cleaved-caspase3 (vs SCII group; p < 0.01). In contrast, miR-9 inhibitor increased cleaved-caspase3 (vs SCII group; p < 0.01), while cleaved-caspase3 expression were reduced in the miR-9 inhibitor + Notch2 siRNA group, and were comparable in the SCII and miR-9 NC groups.