All experimental procedures were approved by the Animal Care and Use Committee of Fujian University of Traditional Chinese Medicine (FUTCM) and followed Chinese Specifications for the Production, Care and Use of the Laboratory Animals. All animal experiments were permitted, according to the License no. SYXK (Min) 2012-007. Throughout the experiments, the utmost effort was made to minimize animal suffering.

Adult male Sprague-Dawley rats (n=90; weight 230–250 g) were housed in an environmentally controlled room at FUTCM (22±2°C, with a 12 h light/dark cycle). The rats were provided ad libitum access to standard rat chow and water. Focal cerebral ischemia was induced by MCAO, as previously described (14), with minor modifications. Briefly, left MCAO was performed using an occluding suture (diameter, 0.26 mm) for 2 h, and following 2 h of MCAO-evoked ischemia, the suture was slowly drawn back to allow reperfusion. The ipsilateral cerebral blood flow was measured using laser Doppler flowmetry (MoorDRT4; Biopac Systems, Inc. Goleta, CA, USA). The MCAO model was considered successful only when the drop in cerebral blood flow was ≥80% of baseline during occlusion. This blood flow rate was maintained for at least 1 h, with the exception of the 0 h time-point.

The rats were randomly divided into the following five groups (n=18/group): i) Sham operation group (Sham); ii) MCAO model group (MCAO); iii) EA group (MCAO + EA); iv) EA combined with dimethyl sulfoxide (DMSO) group (EA + DMSO); v) EA combined with miR-9 inhibitors using a random number table method (EA + miR-9 inhibitors) (15).

The MCAO + EA, EA + DMSO and EA + miR-9 inhibitors groups were stimulated with an electric potential difference of 4 V, and dense disperse wave of 1- or 20-Hz (adjusted to the muscle twitch threshold) using an EA apparatus (G6805; SMIF, Shanghai, China) for 30 min once daily. The rats were immobilized using a surgical fixation device (Beijing Science and Technology Development Co., Ltd., Beijing, China) (12). EA was performed by stimulating the right paralyzed limb at Zusanli (ST36) and Quchi (LI11) acupoints with a depth of 2–3 mm (12,14). The treatment was performed 24 h following the surgery. All animals were sacrificed by cervical dislocation following anesthesia with 10% chloral hydrate (Fujian Academy of Integrative Medicine, Fuzhou, China) at 72 h following the surgery.

Intraventricular injections were performed in the EA + DMSO and EA + miR-9 inhibitors groups 30 min prior to MCAO. The animals were anesthetized with 10% chloral hydrate and were subsequently placed in a stereotaxic apparatus (68001; RWD Life Science Co., Ltd., Shenzhen China). A volume of 7 μl miR-9 inhibitors (10 mM/l in DMSO) or DMSO were injected into the left lateral ventricle in rats (16). Stereotactic coordinates were as follows: Anteroposterior, 0.8 mm; mediolateral, 1.5 mm; depth, 3.5 mm.

Neurological deficits in the MCAO rat model were scored using the Longa et al (17) method following resuscitation. The scores for the neurological behavioral were as follows: 0, Absence of neurological injury symptoms; 1, flexion of the right front paw; 2, body turning towards the right side while walking; 3, falling down on the right side; 4, complete loss of consciousness with the inability to walk. Rats scoring 0 and 4 were excluded from the present study.

Brain tissues were dissected using surgical apparatus (Fujian Academy of Integrative Medicine). The brain tissues were frozen at −20°C for 20 min and cut into five 2 mm sections in the coronal plane at the level of the optic chiasm. The sections were preserved in 2% TTC (Sigma-Aldrich, St. Louis, MO, USA) phosphoric acid buffer [0.2 mol/l (pH 7.4)], as previously described (12) and placed in an incubation box at 37°C in the dark. The sections were alternately overturned twice and were removed following two 15 min long alternations. The normal tissues were stained red using TTC; the infarct loci were not stained and were white in color. Images of the sections were captured using a digital camera (Canon SX20; Canon, Inc., Tokyo, Japan). The percentage of the infarct volume in the total brain volume was calculated using a Motic Med 6.0 image analysis system (Motic Group Co., Ltd., Xiamen, China).

The tissue sections were embedded with paraffin and subjected to conventional gradient dewaxing, followed by HE staining for 3 min at 25°C. Following rinsing with flowing water for 1–2 min, the tissue sections were placed in a 75% hydrochloric acid/alcohol mixture to highlight hematoxylin for 30–45 sec. The tissue sections were subsequently rinsed with flowing water for 10–20 min until the blue color returned. Successive treatment with 95% ethanol and acidified eosin-ethanol staining were performed for 1 min at 25°C. Finally, the tissue sections were dehydrated with gradient ethanol, rendered transparent with xylene and sealed with neutral balsam. The morphological changes were observed under a light microscope (DM500; Leica Microsystems, Wetzlar, Germany).

The paraffin-embedded brain-tissue sections were treated with microwave heat-induced epitope retrieval. Following three washes in phosphate-buffered saline (PBS; pH 7.4), the specimens were incubated for 1 h at 37°C in a 1:120 dilution of fluorescein isothiocyanate-labeled rabbit anti-rat NF-κB p65 antibody (cat. no. ab16502; Abcam, Cambridge, UK). Washing was repeated, as above. The nuclei of all cells were counterstained with 0.1 μg/ml 4′,6-diamidino-2-phenylindole (DAPI) and incubated for 5 min at 25°C. Following a further three washes in PBS, the tissues were mounted in Prolong Gold Antifade reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Images were captured using a fluorescent microscope (Olympus BX61; Olympus Corp., Tokyo, Japan) at a magnification of ×200.

The total RNA was extracted from the ischemic cortex using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA). A SYBR Green^® microRNA Reverse Transcription kit (Guangzhou RiboBio Co., Ltd., Guangzhou, China) was used to reverse transcribe the RNA into cDNA prior to qPCR. The PCR primer kit used for miR-9 was rno-miR-9 (miRQP0825; GeneCopoeia, Guangzhou, China) and the PCR primer kit used for U6 was rno-miR-U6 (RQP047936; GeneCopoeia). The GeneAmp PCR System 9600 (Thermo Fisher Scientific, Inc.) was used. The PCR cycling conditions were 95°C for 10 sec and 38 cycles at 56.5°C for 20 min. The relative expression of miR-9 was quantified using the 2^−ΔΔCq method. The reference gene was rno-miR-U6 (RQP047936; GeneCopoeia, Guangzhou, China).

Cell lysis buffer (1 ml; Invitrogen; Thermo Fisher Scientific, Inc.) and PMSF (10 μl) were added to 100 mg brain tissue at 25°C for 15 min for protein extraction. The protein concentration was determined using Coomassie brilliant blue staining (Invitrogen; Thermo Fisher Scientific, Inc.). The proteins were denatured by heating and 50 μg protein was used for 12% sodium dodecyl sulfate-polyacrylamide gel (Promega Corp., Madison, WI, USA) electrophoresis. The proteins were transferred onto nitrocellulose or polyvinylidene difluoride membranes (Promega Corp.) and were sealed with 5% non-fat milk at room temperature for 2 h. The membranes were incubated with NF-κB p65 (dilution, 1:800; cat. no. ab16502; Abcam), IκBα (dilution, 1:800; cat. no. ab7217; Abcam), TNF-α (1:800; cat. no. ab6671; Abcam), IL-1β (1:800; cat. no. ab200478; Abcam) and β-actin (1:1,000; cat. no. ab189073; Abcam) primary antibodies at 4°C overnight. The membranes were washed and horseradish peroxide-labeled secondary antibodies were added. The cells were incubated in an oscillation incubator at 37°C for 1 h. The color was developed by the addition of electrochemiluminescence solution (Invitrogen; Thermo Fisher Scientific, Inc.), and the images were visualized using a Bio-Image system (Bio-Rad Laboratories, Inc., Hercules, USA). Gray scale analysis was performed for the target bands using Image-Pro Plus software (version 7.0; UVP, LLC, Upland, CA, USA).

The psiCHECK2-NF-κB-luc and psiCHECK2-NF-κB-Mut-luc firefly luciferase reporter vectors (GeneCopoeia Inc.), which contain the intact or mutated putative miR-9 recognition sequence from the 3′-UTR of NF-κB, respectively, cloned downstream of the firefly luciferase gene were constructed. The following primers were designed (Shanghai Sangon Biological Engineering Co., Ltd., Shanghai, China) to PCR the 3′-UTR of human NF-κB from the total RNA extracted from primary cortical neuronal cells: NF-κB, forward: 5′-gcuuuaaaaaaaggagaaaa-3′ and reverse: 5′-acccatctcacccattcttg-3′; nf-κb mutant, forward: 5′-cuggcuuuaaaaaauccucuuaa-3′ and reverse: 5′-tgacacagcaactcctttgg-3′.

Each PCR product was ~1 kb and covers the putative miR-9 recognition sequence or mutated sequence from the 3′-UTR of NF-κB. Clones were selected following screening by restriction digestion with XhoI and NotI. Primary cortical neuronal cells were seeded into 96-well white assay plates (Corning Inc., Corning, NY, USA) at a density of 30,000 cells/well 1 day prior to transfection. A total of 10 ng of each reporter construct was co-transfected with the miR-9 mimics or a mimic negative control (GeneCopoeia Inc.) at a final concentration of 100 nM into primary cortical neuronal cells using Lipofectamine 2000 (Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. Following incubation for 48 h, firefly and Renilla luciferase activities were measured using the Dual-Glo luciferase assay system, according to the manufacturer's protocol (Promega Corp.). Renilla luciferase activity was normalized against firefly luciferase activity for each sample. All experiments were performed in triplicate and three wells were used for each condition in each experiment.

All quantitative data are expressed as the mean ± standard error and were analyzed using SPSS 17.0 statistical analysis software (IBM, Corp., Armonk, NY, USA). The data were subjected to a one-way analysis of variance between different groups, followed by the least significant difference t-test. P<0.05 was considered to indicate a statistically significant difference. All final results were analyzed in a blinded manner.

In the present study, the neuroprotective effect of EA treatment was examined using neurological deficit scores. Table I shows that all MCAO rats demonstrated obvious neurological signs, as compared with the rats in the Sham group (P<0.05), suggesting successful model construction. No significant differences were observed in the neurological signs among the MCAO, EA + DMSO and EA + miR-9 inhibitors groups at 24 h following MCAO and reperfusion. However, 72 h after MCAO and reperfusion, the neurological deficits of all rats in the EA + DMSO group exhibited a significant improvement compared with the MCAO and EA + inhibitors groups (P<0.05). These results suggested that EA treatment effectively alleviated the symptoms of neurological deficit of rats with cerebral ischemia. Furthermore, it was shown that miR-9 inhibitors suppressed the EA-triggered recovery of neurological function, while DMSO did not.

TTC-stained brain sections were evaluated 72 h after MCAO and reperfusion. The results showed that the sections from the Sham group were stained red, whereas the unstained white infarct area was visible on the left side of the brain in the other groups. Compared with the Sham group, the infarct volume in the MCAO+EA group was significantly decreased (P<0.05; Fig. 1). Furthermore, the infarct volume of the EA + DMSO group was lower compared with that of the MCAO and EA + miR-9 inhibitors groups (P<0.05; Fig. 1), suggesting that the miR-9 inhibitors suppressed the cerebral protective efficacy of EA treatment.

To further investigate the neuroprotective efficacy of EA treatment, its anti-inflammatory effect was examined using HE staining. As expected, no histopathological abnormalities and inflammatory cells were observed in the Sham group (Fig. 2A). By contrast, in the infarct core and bounding region of the cerebral cortical area of the MCAO group, the glial and neuron cells exhibited interstitial edema, were shrunken and showed condensed nuclei (Fig. 2B and C), which was ameliorated by EA treatment (Fig. 2D). Furthermore, compared with the MCAO group, considerably fewer inflammatory cells were infiltrated into the cerebral infarct areas in the EA + DMSO group (Fig. 2E), whereas inflammatory cells were infiltrated in the EA + miR-9 inhibitors group (Fig. 2F), suggesting that miR-9 inhibitors suppressed EA-alleviated cerebral inflammation in MCAO rats.

In order to investigate the effect of EA on the miR-9, the expression of miR-9 in the peri-infarct cerebral cortex was assessed using qPCR. Fig. 3 shows that the relative expression of the miR-9 in MCAO, MCAO + EA, EA + DMSO and EA + miR-9 inhibitors groups were significantly decreased compared with the Sham group 72 h after MCAO and reperfusion (P<0.05). In the MCAO + EA and EA + DMSO groups, however, a significant increase was observed compared with the MCAO group (P<0.05). In addition, the expression of miR-9 in the EA + miR-9 inhibitors group was found to be lower compared with that in the MCAO + EA and EA + DMSO groups (P<0.05).

To investigate the mechanism of the anti-inflammatory effect of EA, the localization and expression of NF-κB signaling pathway-associated factors in the peri-infarct cortex was evaluated. The NF-κB p65 subunit was visualized using immunofluorescent staining, and the cells were counterstained with DAPI. NF-κB nuclear translocation was recognized by the colocalization of the p65 subunit with DAPI. The cerebral ischemic injury resulted in the nuclear translocation of the NF-κB p65 subunit in the MCAO group, which was not observed in the sham operation group. However, EA inhibited the NF-κB nuclear translocation and reduced the number of NF-κB p65-positive cells, whereas of NF-κB p65-positive cells decreased in the EA + miR-9 inhibitors group, as compared with the EA and EA + DMSO groups (Fig. 4A). The levels of expression of NF-κB p65, TNF-α and IL-1β in the EA group were significantly decreased compared with the Sham group (P<0.05). By contrast, the IκBα expression levels were increased in the EA group compared with those in the MCAO group (P<0.05). In the EA + miR-9 inhibitors group, the expression levels of NF-κB p65 not only increased, but also exceeded those of the MCAO group, suggesting that the administration of miR-9 inhibitors significantly promoted the expression of NF-κB p65. Furthermore, the differences in IκBα expression among the EA, EA + DMSO and EA + miR-9 inhibitors groups were found to be statistically significant (P>0.05), suggesting that miR-9 inhibitors did not alter the expression of NF-κB upstream-related protein IκBα.

The target genes of miR-9 were predicted with an online algorithm using miRWalk database. Fig. 5A shows that the site where miR-9 binds to NF-κB was predicted and identified as ACCAAAG. The interaction of miR-9 and NF-κB was observed using a dual-luciferase reporter vector assay (Fig. 5B). The results showed that miR-9 bound to 3′UTR of the NK-κB gene (P<0.05), however, not to the mutant 3′UTR (P>0.05).