Male C57/BL6 mice (8–10 weeks old, 22–25 g) were purchased from the Experimental Animals Center of Southern Medical University. tMCAO model was used as focal cerebral ischemia model (Liu et al., 2009). Briefly, mice were anesthetized with isoflurane (5% for induction and 1.5%–2.5% for maintenance). A 6-0 silicone rubber-coated nylon monofilament (YuShun Biological Technology Co. Ltd.) was inserted from a cut of the right external carotid artery, then monofilament was introduced into the internal carotid artery through the carotid bifurcation and advanced 9–11 mm to occlude the MCA for 90 min. Transcranial LDF (Moor VMS-LDF, Moor Instruments, United Kingdom) was used to measure the cerebral blood flow (CBF) in the MCA territory (1 mm posterior and 5 mm lateral to the bregma on the right parietal skull). A 75% or higher decrease in the regional CBF was considered as successful occlusion. Rectal temperature was maintained at 37 ± 0.5°C during entire surgical procedures by a heating pad. During the occlusion period, the anesthesia was disrupted and mice were placed back to their cages. Mice were re-anesthetized to remove the monofilament, ligate the cut and close the incision after 90 min occlusion. Mice in sham group adopted an exactly same procedure except the MCA occlusion.

DEX (Jiangsu Hengrui Medicine Co., Ltd., China; Gu et al., 2011), and rapamycin (RAPA, Sigma) were diluted or dissolved with 0.9% saline, then intraperitoneally injected into mice at dosages of DEX (25 μg/kg) and RAPA (80 μg/kg). 3-MA, Sigma-was dissolved in 0.9% saline and was injected intracerebroventricular into mice with 3 μl of a 20 mg/ml (Wang et al., 2014). At the onset of the reperfusion, the treatment groups, including tMCAO+DEX, tMCAO+3-MA, tMCAO+DEX+3-MA, tMCAO+RAPA and tMCAO+DEX+RAPA were administrated with DEX, 3-MA, DEX combined with 3-MA and RAPA and DEX combined with RAPA respectively. 2-Methoxyestradiol (2ME2, Selleck) was dissolved in 0.5% dimethylsulfoxide (DMSO) and intraperitoneally injected 30 min after tMCAO (Schaible et al., 2014). Sham and tMCAO groups without drug treatment were injected with the same volume of 0.9% saline or DMSO.

The neurological deficit scores of the mice were evaluated 1 day after tMCAO by researchers who were blinded to the experimental groups. The neurological grading scores range from 0 to 5 (0, no deficit; 1, forelimb flexion; 2, as for 1, plus decreased resistance to lateral push; 3, unidirectional circling; 4, longitudinal spinning or seizure activity; and 5, no movement; Jin et al., 2015).

Infarct size was measured 1 day after tMCAO. The mice were deeply anesthetized, decapitated and their brains (n = 8/group) were rapidly removed and sliced into six sections using a rodent brain matrix. The brain slices were stained with 2% 2,3,5-Triphenyltetrazolium chloride (TTC, Sigma) at 37°C for 15 min, then fixed in 4% paraformaldehyde for 1 day (Yang et al., 2013). The infarct size was analyzed using ImageJ (1.37v, Wayne Rasband, available through the National Institutes of Health) by researchers who were blinded to the experimental group. To exclude the effect of cerebral edema, the infarct size was normalized to the non-ischemic hemisphere and expressed as percentage of the contralateral hemisphere.

To examine the formation of autophagosomes and morphology changes in the neurons after tMCAO in cerebral cortex, transmission electron microscopy examination was performed as previously described (Liu et al., 2016). Briefly, mice were euthanized with an overdose isoflurane and transcardially perfused with icy PBS (pH 7.4) followed by 4% paraformaldehyde in PBS (pH 7.4) 1 day after tMCAO. Brain tissue (1 mm³) from the peri-ischemic area of cerebral cortex were fixed with ice-cold glutaraldehyde (2.5% in 0.1 M cacodylate buffer, pH 7.4) for 1 h, and post-fixed in OsO₄ and embedded in epoxy resin. Ultrathin sections (70–80 nm) were stained with uranyl acetate and lead citrate and examined in a CM-120 electron microscopy (Philips, Holland).

Twenty-four hours after ischemia onset, animals were sacrificed with an overdose of isoflurane and transcardially perfused with saline and 4% paraformaldehyde. After removal from the skull, brains were post-fixed in 4% paraformaldehyde for 48 h. The brains were embedded in paraffin, and coronal sections were prepared for immunostaining. The sections were washed in PBS several times and placed in 3% H₂O₂ for 15 min, and then incubated in 10% normal goat serum (Sigma) for 1 h to block nonspecific binding of IgG and incubated overnight at 4°C with NeuN (1:200, Millipore) and light chain 3 (LC3; 1:200, Santa Cruz) antibody. Then, the sections were washed in PBS and incubated with appropriate secondary antibody. Finally, the stained cells were observed with a fluorescent microscopy and captured using a CCD camera.

Primary cultured cortical neurons were prepared from the cerebral cortex of day 15 C57BL/6 mouse embryos. Mice were obtained from the Experimental Animals Center, Southern Medical University. Primary cultured cortical neurons were prepared as described previously (Pavlovski et al., 2012). Briefly, cortices were dissected, mechanically minced and minced cortical tissues were then dissociated by mild trypsinization (0.02% w/v) at 37°C for 10 min followed by trituration in a DNase solution (0.004% w/v) containing a soybean trypsin inhibitor (0.05% w/v) to prepare single cell suspension. The single cell suspension was re-suspended in a DMEM medium (5 mM KCl, 31 mM glucose, and 0.2 mM glutamine) supplemented with insulin, penicillin and 10% FBS (DMEM-FBS). Then cells were plated on poly-L-lysine-coated 6-well plates at 1.0 × 10⁶ cells/ml of neurobasal medium (Gibco) with 2% B-27 (Gibco) and 0.5 mM Glutamax (Gibco) for culturing. Half of the medium was changed after the first 3 days and three times a week thereafter. Neurons were cultured for 7–9 days before experiment, which may allow percentage of neurons to reach ~90% (Hu et al., 2013).

β-III Tubulin, is regarded as a neuron-specific marker, which was used to verify neuronal purity by flow cytometry assay (Supplementary Figure S1). 1 × 10⁶ cells/100 μL primary cultured neurons were harvested into tubes. Cells were fixed with flow cytometry fixation buffer and permeabilized with flow cytometry permeabilization buffer. After washing, the cells were incubated with beta-III Tubulin (1:200, Proteintech) for 45 min at room temperature in the dark and subsequent analyzed by flow cytometry (BD bioscience, San Hose, CA, USA).

To induce OGD, the primary neurons were washed three times and incubated with glucose-free Earle’s balanced salt solution (EBSS) and placed within a hypoxic chamber (Forma Scientific, USA) which was continuously maintained with 95% N₂ and 5% CO₂ at 37°C to obtain 1% O₂ (Cheon et al., 2016) for 6 h. EBSS group was exposed to EBSS in the standard incubator. OGD was terminated by replacing the medium to neurobasal medium with 2% B27 in the normoxic incubator. At the end of OGD, the treatment groups, including OGD+DEX, OGD+3-MA, OGD+DEX+3-MA, OGD+RAPA and OGD+DEX+RAPA, were administrated DEX, 3-MA, DEX combined with 3-MA and RAPA, and DEX combined with RAPA, respectively, to reach a final concentration of 1 μM DEX (Dahmani et al., 2010), 1 mM 3-MA (Shi et al., 2012) or 10 nM RAPA (Wang et al., 2012) for 4 h reoxygenation. In the groups of the Control group, Control+DEX, Control+3-MA and Control+RAPA were given saline or the same concentration of DEX, 3-MA and RAPA. In Group OGD+DEX+2ME2, a final concentration of 1 μM 2ME2 was applied given reoxygenation (Ryou et al., 2015). To evaluate the effect of DEX in the acute stage of ischemia, neurons incubated was end up at 4 h reoxygenation. All experiments were at least duplicated three times biologically.

To determine the time duration of OGD and the concentration used for each drug, cell viability assay was performed (Supplementary Figure S2). The viability of the primary neurons was measured by CCK-8 assay (Dojin Laboratories, Japan). Briefly, neurons were seeded in a 96-well plate and kept in a 5% CO₂ incubator at 37°C overnight. At the end of treatment in each group, CCK-8 assay was performed according to the manufacturer’s instruction. Absorbance was determined at 450 nm by using a microplate reader (BioTek, Winooski, VT, USA). Results are shown as percentages of the control group.

The apoptosis of the neurons were stained by Annexin V/PI (Beijing 4A Biotech Co. Ltd.) according to the manufacturer’s instruction. Briefly, before analysis by flow cytometry, 5 μl of Annexin-V-Fluorescein isothiocyanate (FITC) labeling reagent and 10 μl of propidium iodide (PI) were added to the medium and the neurons were incubated at room temperature for 10 min in the dark. 1 × 10⁵ cells for each sample were analyzed a flow cytometer (BD bioscience, San Hose, CA, USA).

The neurons grown on coverslips were fixed by 4% methanol for 15 min and 0.25% Triton for 10 min, incubated in 2% Bovine Serum Albumin(Sigma) for 1 h to block nonspecific binding of IgG. Then the cells were reacted with primary antibodies: rabbit anti-Beclin1 (1:200, Abcam), rabbit anti-BCL-2 (1:200, Santa Cruz), mouse anti-NeuN (1:200, Millipore) at 4°C overnight followed by the appropriate secondary antibodies with a mixture of goat anti-rabbit Alexa-Fluor 488 and goat anti-mouse Alexa-Fluor 594(1:200, Invitrogen). The cells were subsequently incubated with the nuclear staining dye 4,6-diamidino-2-phenylindole (DAPI, Vector Laboratories, Burlingame, CA, Burlingame, CA, USA). Finally, the stained cells were observed with a fluorescent microscopy (Nikon ECLIPSE 80i, Japan).

Western blot was performed as previously described (Qin et al., 2010). The brain tissues from cortex of peri-ischemic area and primary cultured cortical neurons were homogenized in lysis buffer. Ultrasonic crushed tissues for three times with low power. After reaction for 20 min on ices, the samples were centrifuged at 10,000× g at 4°C for 10 min and the supernatant was preserved at −80°C. Before immunoblotting, the protein concentrations were determined with a BCA detection kit (Pierce, 23225) and adjusted to the equal concentrations across different samples. Protein samples (30 μg) were separated by SDS-polyacrylamide gels, then transferred to PVDF membranes. After blocked with 5% fat free milk, the membranes were incubated with a p62 (1:750, Cell Signaling), LC3 (1:500, Santa Cruz), Beclin 1 (1:1000, Abcam), BCL-2 (1:1000, Santa Cruz) and HIF-1α (1:1000, Cell Signaling), the mammalian target of rapamycin (mTOR), p-mTOR (1:1000, Sigma) and β-actin (1:5000, Cell Signaling). Blots were visualized using an Image Lab 5.1 software (Bio-Rad, Hercules, CA, USA) and were analyzed using ImageJ.

Data are expressed as mean ± SEM. Differences were evaluated by one-way analysis of variance (ANOVA; three or more groups). When only two groups were compared, an unpaired t-test was used. P < 0.05 was considered statistical significance. Statistical analyses were performed using SPSS 20.0 Statistics (IBM SPSS Statistics for Version 20.0, IBM Corp, North Castle, NY, USA).

The experimental plan was showed in Figure 1A. We first examined whether DEX reduce infarction and improve neurological scores at 1 day after tMCAO. Furthermore, we utilized 3-MA and RAPA to investigate whether neuroprotection of DEX after tMCAO was related with regulation of autophagy. The data showed that treatment of DEX or autophagy flux inhibitor, 3-MA, significantly reduced the brain infarct volume. Moreover, post-treatment with the autophagy inducer, RAPA, was not protective and DEX could significantly reversed the brain infarction even when combined with RAPA (Figures 1B,C). Similarly, neurological deficit scores showed that DEX improved neurological outcomes after tMCAO, and this protection could be enhanced or attenuated by 3-MA or RAPA, respectively (Figure 1D). These data showed that DEX may regulate autophagy to produce neuroprotection in ischemic brain injury.

To clearly evaluate the effect of DEX, 3-MA and RAPA on autophagy, ultrastructural changes in mouse cortical neurons were examined with transmission electron microscopy at 1 day after tMCAO (Figure 2A). The number of intact mitochondria (arrow heads), autophagosomes (broad arrows) and lysosomes (narrow arrows) in each group of cortical neurons were counted and analyzed (Figures 2B–D). In neurons, DEX, 3-MA or their combination treatment after tMCAO respectively increased intact mitochondria and decreased autophagosomes dramatically compared with tMCAO group, but RAPA treatment after tMCAO decreased intact mitochondria and increased autophagosomes significantly and this effect could be partly reversed by DEX. To confirm whether DEX inhibited neuronal autophagy to produce neuroprotection after tMCAO, we elucidated the distribution pattern of LC3 positive cells in neurons by tissue immunostaining. Our results revealed that DEX, 3-MA and DEX+3-MA significantly decreased LC3 positive cells compared with tMCAO group. Meanwhile, RAPA and the combination of RAPA and DEX after tMCAO respectively had no effect on the distribution pattern of LC3 positive cells compared with tMCAO group (Figures 2E,F). Consequently, our results indicated that DEX inhibited neuronal autophagy to produce neuroprotection at 1 day after tMCAO.

Because autophagy and apoptosis could occur in the same injured neurons (Tian et al., 2010), we next investigated whether DEX affected cell apoptosis or viability using flow cytometry and CCK-8 assay. Flow cytometry detection revealed that DEX, 3-MA and DEX+3-MA significantly inhibited apoptosis of neurons following OGD compared with OGD group (Figures 3A,B). Then, CCK-8 assay results revealed that DEX, 3-MA and DEX+3-MA significantly increased viability. Meanwhile, RAPA and the combination of RAPA and DEX had no effect on viability of neurons following OGD (Figure 3C). Collectively, DEX increased the viability and inhibited the apoptosis of neurons following OGD, and this protection could be enhanced or attenuated by 3-MA or RAPA, respectively.

To examine the autophagy in primary neurons, we confirmed the expression of autophagy related proteins by cell immunofluorescence and immunoblotting. These data showed that DEX, 3-MA and DEX+3-MA treatment significantly increased Bcl-2 and decreased Beclin 1 expression respectively in cortical neurons following OGD compared with OGD group. RAPA had no effect on Bcl-2 and Beclin 1 expression level, but the Bcl-2 increased and Beclin 1 decreased, respectively when combined with DEX (Figure 4). Additionally, the expression of other autophagic proteins, p62 and LC3, were increased and decreased respectively by DEX in cortical neurons following OGD with or without the combination with 3-MA or RAPA (Figures 4A,B). These results indicate DEX attenuated the activation of autophagy in neurons following OGD. The autophagic inhibition effect of DEX was enhanced or attenuated by 3-MA or RAPA, respectively.

To examine which signaling pathways involved in the DEX-mediated autophagy, we measured expression of mTOR and HIF-1α following OGD. In the neurons, the expression of HIF-1α in OGD group and OGD+DEX group was elevated compared with control group, with an enhanced elevation in DEX+OGD group than in OGD group. DEX had no effect on p-mTOR/mTOR expression in neurons following OGD (Figure 5A). Furthermore, 2ME2, an inhibitor of HIF-1α, was applied to determine whether DEX inhibit autophagy was caused by up-regulation of HIF-1α. The immunoblotting results suggested that the DEX-induced inhibition of autophagy was reversed by 2ME2 (Figure 5B). To confirm that the neuroprotection effect of DEX could be neutralized by 2ME2, we tested the effects of 2ME2 following stroke and DEX treatment in the ischemic brain. Similar to the results in vitro, infarct volume showed an enlargement in DEX+2ME2 group than DEX group (Figures 6A,B). Moreover, our immunostaining results from brain section demonstrated that the number of LC3 positive cells decreased in the DEX after ischemia-reperfusion injury. When using 2ME2 to block the HIF-1α, the number of LC3 positive cells increased compared to DEX group (Figures 6C,D). The immunoblotting results from brain tissues also suggested that HIF-1α was significantly up-regulated after DEX treatment following tMCAO (Figures 6E,F). DEX-induced inhibition of autophagy were reversed by 2ME2 after tMCAO (Figures 6G,H). Collectively, our in vitro and in vivo results suggested up-regulation of HIF-1α may play a key role in DEX-mediated inhibition of autophagy.