The experiments adhered to the ethical standards of the Institutional Animal Care Committee and were approved by the Animals Ethics Committee of Jilin University. Male Wistar rats (250–280 g) and newborn rats were obtained from the Experimental Animal Center of Jilin University. Animals were maintained in a specific pathogen-free animal breeding room at 24°C under a 12 h day/night cycle with free access to water and food. All possible measures were taken to avoid animals suffering at each stage of the experiment.

Astrocytes were obtained from the cerebral cortex of newborn rats as previously described (Schildge et al., 2013). Newborn Wistar rats were decapitated, and the cerebral cortices were isolated in cold Dulbecco’s Modified Eagle Media: Nutrient Mixture F-12 (DMEM/F12) medium. Then, the meninges were carefully removed, and the tissues were treated with 0.125% trypsin solution for 15 min at 37°C. DMEM/F12 containing 10% fetal bovine serum (FBS) was added, and the mixture was centrifuged at 1300 rpm for 5 min. The sediment was resuspended with DMEM/F12 containing 10% FBS. At a concentration of 10⁵/ml, cells were planted onto 75 cm² flasks in 15 ml DMEM/F12 containing 10% FBS and 1% penicillin/streptomycin and placed in an incubator (Thermo Scientific, Waltham, MA, United States) at 37°C with 95% air and 5% CO₂. After 24 h, the medium was changed in the flasks, and then half of the medium was changed every 3 days. After approximately 12 days, the astrocytic cultures reached confluency. Oligodendrocytes and microglia were deprived from astrocytic cultures by shaking on an orbital shaker for 6 h at 37°C (Schildge et al., 2013). The astrocytic cultures were treated with 0.25% trypsin solution for 3 min at 37°C. Then, the cells were harvested, and they were adjusted to a density of 2 × 10⁵ cells/ml and planted on flasks. The third generation of primary cultured astrocytes were used in our study. The purity of astrocytes was higher than 95%, as confirmed by immunofluorescence staining with a specific marker, the glial fibrillary acidic protein (GFAP) (ab7260, Abcam, United States). A representative result is shown in Supplementary Figure S1A.

As described previously (Ferrer-Acosta et al., 2017), oxygen-glucose deprivation/reoxygenation (OGD/R) is a classic in vitro model of I/R injury. Briefly, astrocytes were washed three times with glucose-free DMEM and cultured in the same medium in a hypoxia chamber with a mixture of 95% N₂ and 5% CO₂ for 12 h. Then, the astrocytes were cultured in normal DMEM medium and re-oxygenated under normoxic conditions (95% air, 5% CO₂) for 6 h.

The astrocytic cultures were divided into five groups: (1) a control group, stimulated with DMSO; (2) an OGD/R group, stimulated with DMSO during OGD/R injury; (3) an OGD/R + Vinp group, stimulated with Vinp (30 μM) (Gedeon Richter Pharmaceutical Co., Ltd., Budapest, Hungary) during OGD/R injury; and (4) an OGD/R + Vinp + LY group, stimulated with LY294002 (20 μM) (ab120243, Abcam, Cambridge, MA, United States) and Vinp during OGD/R injury; (5) OGD/R + Vinp + BKM group, stimulated with BKM120 (2 μM) (S2247, Selleck, Houston, TX, United States). LY and Vinp were dissolved in DMSO at a final concentration of 100 mM (Hong et al., 2013; Takac et al., 2013; Nivison-Smith et al., 2015), and BKM was dissolved in DMSO at a final concentration of 10 mM. As described above, all groups were stimulated with the same volume of DMSO, and for the control group 0.33% DMSO proved to have no obvious toxicity on astrocytes (Supplementary Figure S1B).

Commercial cell counting Kit-8 (CCK-8) (Do-jindo, Kumamoto, Japan) was used to detect cell viability (Ishiyama et al., 1997). Primary astrocytes cultured to the third generation were seeded in 96-well plates at a density of 10⁴/well. The 96-well plates were placed in a cell culture incubator for 24 h before being subjected to OGD/R. Thereafter, 10 μL CCK-8 reagent was added to each well. The 96-well plates were then placed in the cell culture incubator for 2 h, and the absorbance at 450 nm was measured by a microplate reader (Multiskan, Thermo Scientific, Waltham, MA, United States).

Cytotoxicity was determined by measuring the lactate dehydrogenase (LDH) of the cell culture supernatant using the Cytotoxicity Detection Kit (C0016, Beyotime, Shanghai, China) according to the manufacturer’s instructions (Lobner, 2000). Briefly, the sample maximum enzyme activity control wells were set according to the instructions. Astrocytic supernatants from each group were centrifuged. In each well of the 96-well plate was added 120 μL supernatant and 60 μL reagent. Then, the 96-well plate was incubated at room temperature for 30 min in the dark, and the absorbance at 490 nm was measured by a microplate reader. Experiments were repeated five times, and each experiment contained five duplicate wells for each astrocyte group.

The ROS Assay kit (S0033, Beyotime) was used to detect ROS in astrocytes (Eruslanov and Kusmartsev, 2010). Briefly, astrocytes were seeded at a density of 104 cells/well in 96-well plates. After exposure to OGD/R injury, 10 μM of DCFH-DA in serum-free DMEM medium was added to each well. After incubation for 30 min in the cell culture incubator, each well was washed three times with serum-free DMEM and examined by a microplate reader using excitation/emission wavelengths of 488/525 nm.

Astrocytic Nitric oxide (NO) release was detected using the NO Assay Kit (S0021, Beyotime) (Weissman and Gross, 2001). Astrocytes were seeded in 96-well plates. A total of 50 μL/well of Griess Reagent I and 50 μL/well of Griess Reagent II were added into each well immediately after the astrocytes were exposed to OGD/R injury. The standard curve was constructed according to the instructions. The absorbance at 540 nm was measured by a microplate reader.

Astrocytes were fixed with 4% paraformaldehyde at room temperature for 30 min and washed three times with PBS. After permeabilization with 0.2% Triton X-100 for 10 min and blocking with 10% goat serum in PBS for 1 h, the cells were incubated with rabbit anti-IL-1β (ab9722, Abcam; 1: 100), anti-TNF-α (ab66579, Abcam; 1: 100), anti-Bcl-2 (ab194583, Abcam; 1: 50), and anti-caspase-3 antibodies (ab13847, Abcam; 1:50) overnight at 4°C, followed by incubation with goat anti-rabbit IgG Fc (Alexa Fluor 647, ab150091, Abcam; 1:200) for 2 h at 25°C. The cells were then incubated with DAPI for 5 min and examined under a fluorescence microscope (OLYMPUS BX51, Tokyo, Japan).

Apoptosis was assessed by flow cytometry using an FITC Annexin V Apoptosis Detection Kit I (556547, Becton Dickinson, Franklin Lakes, NJ, United States) according to the manufacturer’s instructions (Frey, 1997). Briefly, cells were rinsed with ice-cold PBS and then resuspended in 100 μL binding buffer (10⁵ cells). A total of 5 μL Annexin V and 5 μL PI were added to each sample, and they were incubated for 15 min at 25°C in the dark. Then, 400 μL binding buffer was added to each tube and cells were immediately analyzed using a FACSC-LSR (Becton Dickinson) and evaluated with the Flow Jo 7.6 software.

The MCAO model, a classic in vivo model of I/R injury, was prepared as previously described (Longa et al., 1989). Briefly, Wistar rats were anesthetized with chloral hydrate (350 mg/kg, i.p.). Then, a midline incision in the neck was made to expose the left external and internal arteries (ECA and ICA). The ECA was cut between two ligations, and a 0.26 mm silicone-tipped filament (2636, Xinlong lnc., Beijing, China) was inserted into the ICA via the ECA at approximately 20 mm until a resistance was felt, which ensured the occlusion of the MCA. Then, the suture was tightened around the ECA stump and the incision was closed. After surgery, rats remained in the cage for 2 h. Then, the animals were anesthetized again, and the filament was removed. During surgery, the rats’ body temperature was maintained at a normal level by heating pads. After awakening, the rats were maintained in cages with free access to food and water for 12 h.

A total of 64 male Wistar rats were randomly divided into four groups: (1) a sham group: the rats were injected with 0.9% normal saline and were not subjected to MCAO; (2) an I/R group: the rats were injected with 0.9% normal saline and subjected to MCAO; (3) a Vinp + I/R group: the rats were injected with Vinp (10 mg/kg) and subjected to MCAO; and (4) a Vinp + I/R + LY group: the rats were initially injected with LY294002 (0.3 mg/kg) and then with Vinp 15 min later and subjected to MCAO. All injections were administered intraperitoneally 30 min prior to MCAO.

Neurological evaluation was performed after 2 h of ischemia and 12 h of perfusion by a researcher blinded to the experimental groups. Evaluation was performed using a modified form (Longa et al., 1989) as follows: (0) no deficits; (1) difficulty to fully extend the contralateral forelimb; (2) inability to extend the contralateral forelimb; (3) mild circling to the contralateral side; (4) severe circling; and (5) falling to the contralateral side. Finally, the rats were anesthetized and decapitated for the brain water content assay, TTC staining, western blot, immunofluorescence, SOD activity, and MDA content analyses.

2,3,5-triphenyltetrazolium chloride (TTC, Sigma, St. Louis, MI, United States) staining was used to visualize the ischemic infarction (Bederson et al., 1986). After decapitation, the brains were sliced into 2 mm sections, and each slice was incubated in a 2% solution of TTC at room temperature for 20 min and fixed in 4% paraformaldehyde. The brain sections were photographed using a high-resolution digital camera (Olympus). The infarct size was measured using the Image J software (NIH Image, National Institutes of Health, Bethesda, MD, United States). The percentage of the infarction size was calculated as described previously (Jackman et al., 2011).

The classic wet-dry method was used to measure brain water content (Agrawal et al., 1968). Immediately after the rats were sacrificed, the brains were taken and weighed to obtain the wet weight. The samples were then dried in an oven at 100°C for 48 h. They were then weighed again to obtain the dry weight. Water content = (wet weight−dry weight)/wet weight × 100%.

Commercially available detection kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) were used to detect SOD activity and MDA content according to the manufacturer’s instructions as previously described (Hou et al., 2016). Briefly, SOD activity was assessed using the xanthine oxidase method, and MDA content was measured with the thiobarbituric acid method. The samples were analyzed with a spectrophotometer (BioRad, San Diego, CA, United States).

After decapitation, the brains were harvested immediately, immersed into pre-chilled isopentane (Beijing Chemical Factory, Beijing, China), and placed inside a −80°C refrigerator for 10 min for snap-freezing. Then, the brains were embedded in optimum cutting temperature compound (Sakura Finetek Inc., Torance, CA, United States) and stored in the −80°C refrigerator. Subsequently, 10-μM section of the brain were obtained using a cryomicrotome (Leica, Nussloch, Germany). The sections were fixed with 4% paraformaldehyde at room temperature for 15 min and washed three times with PBS. After permeabilization with 1% Triton X-100 for 10 min and subsequent blocking with 10% goat serum in PBS for 1 h, the sections were incubated with mouse anti-GFAP (ab10062, Abcam, 1:500) and rabbit anti-TNF-α (ab66579, Abcam; 1: 200) overnight at 4°C, followed by incubation with Alexa Fluor 647-conjugated goat anti-mouse IgG (ab150115, Abcam; 1:200) and Alexa Fluor488-conjugated goat anti-rabbit IgG (ab150077, Abcam; 1:200) for 1 h at 25°C. Experimental negative control was a section without any primary antibody treatment. The slices were then incubated with DAPI for 5 min and examined under a confocal microscope (Leica TCS SP5, Nussloch, Germany).

The western blot analysis was conducted as previously described (Hou et al., 2016). The cortex in the same set of rats or the cultured astrocytes was crumbled and homogenized with ice-cold lysis buffer (RIPA: NaVO3: PMSF: NaF = 92:5:2:1). Proteins were extracted from the cerebral cortex tissue, and the protein concentrations were assayed. Each sample (50 μg) was loaded on a 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis apparatus and electrophoresis was carried out until the bromophenol blue dye reached the bottom of the gel. Then, the proteins were electro-transferred to polyvinylidene fluoride membranes, and the membranes were placed in 5% skim milk powder dissolved in TBS with 0.1% Tween-20 for 1 h. The membranes were incubated with anti-Cx43 (ab11370, Abcam; 1:1000), anti-p-Cx43 (PA5-37584, Thermo Fisher Scientific; 1:1000), anti-AKT (4691, Cell Signaling Technology, Danvers, MA, United States; 1:1000), anti-p-AKT (13038, Cell Signaling Technology; 1:1000), anti-IL-1β (ab9722, Abcam; 1: 1000), anti-TNF-α (ab66579, Abcam; 1: 1000), anti-IL-10 (ab9969, Abcam; 1:2000), anti-Bcl-2 (ab194583, Abcam; 1:500), anti-caspase-3 (ab13847, Abcam; 1:500), and anti-β-actin (ab13847, Abcam; 1:2000) antibodies diluted in 5% skim milk powder dissolved in TBST overnight at 4°C. The membranes were then washed with PBST and incubated with a horseradish peroxidase-conjugated secondary antibody for 1 h. The protein bands were quantified with the Quantitation One software (Bio-Rad Laboratories, Hercules, CA, United States).

All data are presented as the mean ± standard error of the mean (SEM) from at least three independent experiments using Graphpad Prism 6 (Inc., San Diego, CA, United States). Analysis was carried out by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc tests. ^∗P < 0.05, ^∗∗P < 0.01 or ^∗∗∗P < 0.001 denoted the significance thresholds.

To investigate the neuroprotective effects of Vinp in primary cultured astrocytes in vitro, we assessed cell viability in each group via the CCK-8 assay (Figure 1A). The results showed that approximately half of the astrocytes survived in OGD/R group compared with the control group (56.86 ± 2.62%, P < 0.001). Compared with OGD/R group, viability was significantly improved in astrocytes treated with OGD/R + Vinp (78.94 ± 2.78%, P < 0.001). However, this elevation was reversed in the OGD/R + Vinp + LY group compared with the OGD/R + Vinp group (61.77 ± 2.09%, P < 0.001). We also measured astrocytic injury by testing the amount of LDH released into the supernatant. This test showed that OGD/R injury significantly increased the release of LDH compared with that in the control group (35.77 ± 2.60% vs. 8.62 ± 0.75%, P < 0.001). Treatment with OGD/R + Vinp remarkably decreased the release of LDH compared with that in the OGD/R group (20.14 ± 1.99%, P < 0.001). In comparison with OGD/R + Vinp group, coadministration of LY294002, a PI3K inhibitor, resulted in an apparent increased release of LDH (38.63 ± 1.81%, P < 0.001). These findings indicated that Vinp could promote cell survival and reduce cell damage in astrocytes subjected to OGD/R, and the inhibition of the PI3K/AKT pathway could abolish the protection of Vinp.

Increased ROS production is considered an initial step in OGD/R injury (Dirnagl et al., 1999). To examine the effect of Vinp on OGD/R injury-induced oxidative stress in astrocytes, we detected intracellular ROS and NO released into the extracellular supernatant (Figures 1C,D). The results showed that intracellular ROS and extracellular NO was significantly increased after OGD/R injury compared with the control group (P < 0.001), but this elevation was reversed in the OGD/R + Vinp group compared with the OGD/R group (P < 0.01). Compared with the OGD/R + Vinp group, intracellular ROS and extracellular NO in the OGD/R + Vinp + LY group was markedly increased (P < 0.01). The data revealed that Vinp decreased oxidative stress in astrocytes induced by OGD/R injury, which was attenuated by a PI3K/AKT pathway inhibitor (LY294002).

As previously mentioned, the large amounts of ROS generated during OGD/R can induce inflammation by activating astrocytes (Chan, 2001; Duris et al., 2018). To investigate the anti-inflammatory effects of Vinp, we used immunofluorescence staining to observe the expression of TNF-α and IL-1β (Figures 1E,F). OGD/R injury resulted in significant increase in astrocytic TNF-α and IL-1β expression compared with control group, and this increase was blocked by treatment with Vinp during OGD/R injury. Moreover, LY reversed the effect of Vinp on astrocytes subjected to OGD/R, significantly increasing TNF-α and IL-1β expression. The results indicated that the PI3K/AKT pathway is involved in the anti-inflammatory effects of Vinp against OGD/R injury.

As a cascade event, increased ROS may induce excessive inflammatory responses which could activate pro-apoptotic pathways (Duris et al., 2018), and therefore we evaluated the anti-apoptotic effects of Vinp. Annexin V FITC/PI staining and flow cytometry were used to detect the astrocytic apoptotic rate (Figure 2A), and immunofluorescence staining was processed to observe the expression of caspase-3 and Bcl-2 (Figures 2B,C). We found that OGD/R injury resulted in a significant increase in astrocytic caspase-3 expression, a significant reduction in Bcl-2 expression, and an increase in apoptotic rate compared to the control group (P < 0.001). Compared with the OGD/R group, caspase-3 expression and astrocytic apoptotic rate effectively decreased while Bcl-2 expression increased in the OGD/R + Vinp group (P < 0.001). Furthermore, results showed LY reversed the effect of Vinp on OGD/R-treated astrocytes, significantly increasing astrocytic caspase-3 expression, reducing Bcl-2 expression, and increasing apoptotic rate (P < 0.001). These results indicated that the PI3K/AKT pathway is involved in the anti-apoptotic effects of Vinp against OGD/R injury.

Next, we prepared the classic MCAO models to evaluate the early effects of Vinp against cerebral I/R injury in vivo. As previously described, neurological deficit scores, brain water content, and infarction size were measured to evaluate the effects of Vinp against cerebral I/R injury after the rats were subjected to ischemia for 2 h and reperfusion for 12 h. In the I/R group, there were obvious symptoms of neurological deficits including deviation to the right, circling, and inability to fully extend the right upper limb. Treatment with Vinp did not significantly reduce the neurological deficit scores (Figure 3A, P = 0.18). The cerebral infarct size after MCAO is shown in Figures 3B,C. In the I/R group, significant cerebral infarction was observed compared with the sham group (32.92 ± 1.63%, P < 0.001), and this phenotypic alteration was mostly abrogated in the I/R + Vinp group compared with the I/R group (8.06 ± 1.10%, P < 0.001). Compared to the I/R + Vinp group, infarct size in the I/R + Vinp + LY group was significantly enlarged (28.04 ± 1.05%, P < 0.001). The results of the water content of brain tissues were consistent with the trend observed for infarction size (Figure 3D). The results revealed that Vinp decreased the infarction size and brain edema, while inhibition of PI3K/AKT reversed the protection of Vinp.

We further examined oxidative stress in ischemic cerebral cortices as the in vitro study, which is considered the initial step of cerebral I/R injury. SOD activity is an important antioxidant enzyme, and MDA content reflects oxidative damage (Chan, 2001), therefore, we examined SOD activity and MDA content (Figures 3E,F). Compared to the sham group, SOD activity significantly decreased, while MDA content increased in the I/R group (P < 0.001). Treatment with Vinp effectively increased SOD activity and decreased MDA content compared with the I/R group (P < 0.001), whereas LY reversed the effects of Vinp by decreasing SOD activity and increasing MDA content compared with the I/R + Vinp group (P < 0.01). The above findings suggested that Vinp attenuated oxidative stress induced by cerebral I/R injury, which is related to the PI3K/AKT pathway.

To validate if the response cascade was caused by cerebral I/R injury, we examined the inflammation and apoptosis in vivo. First, we performed immunofluorescent analysis to observe the reactive astrocytes and inflammatory cytokine by double immunostaining the brain cryosections with anti-GFAP and anti-TNF-α antibodies (Figure 4). Cerebral I/R injury resulted in a significant increase in the expression of GFAP and TNF-α, and the co-localization of GFAP and TNF-α was compared with that in the sham group. However, this increase was blocked by Vinp treatment. Furthermore, LY reversed the effect of Vinp in reactive astrocytes, significantly increasing the expression of GFAP and TNF-α, and consequently, the co-localization of GFAP and TNF-α. The results showed that Vinp treatment significantly decreased cerebral I/R injury-induced inflammation by reducing astrocyte activation. Then, western blot analysis was used to detect the expression of inflammation-associated proteins, IL-1β, TNF-α, and IL-10, and apoptosis-related proteins, caspase-3 and Bcl-2 (Figures 5A,B). IL-1β, TNF-α, and caspase-3 expression significantly increased while Bcl-2 expression significantly decreased in the I/R group compared with the sham group (P < 0.001). IL-1β, TNF-α, and caspase-3 expression decreased while IL-10 and Bcl-2 expression increased in the I/R + Vinp group compared to the I/R group (P < 0.001). Conversely, LY blocked the above effects of Vinp, significantly increasing IL-1β, TNF-α, and caspase-3 expression while decreasing IL-10 and Bcl-2 expression (P < 0.01). Overall, these results indicated that the PI3K/AKT pathway is involved in the anti-inflammatory and anti-apoptotic effects of Vinp against cerebral I/R injury.

In order to explore whether the above protective effects of Vinp are exerted by targeting Cx43 via the PI3K/AKT pathway, we examined the expression of Cx43, p-Cx43, AKT, and p-AKT. Compared to the sham group, I/R injury significantly repressed the expression of Cx43 and p-Cx43 and the p-Cx43/Cx43 ratio (Figure 5C, P < 0.01). Interestingly, Vinp could increase Cx43 and p-Cx43 expression and the p-Cx43/Cx43 ratio compared with the I/R group, indicating the activation of the Cx43 (P < 0.001). However, significant reduction of p-Cx43 expression and the p-Cx43/Cx43 ratio was observed in the I/R + Vinp + LY group compared with I/R + Vinp group (P < 0.01). The expression of p-AKT and the ratio of p-AKT/AKT were similar to those of p-Cx43 and p-Cx43/Cx43, while the AKT levels did not significantly differ among the groups (Figure 5D).

To further explore whether the abovementioned protective mechanism of Vinp are exerted by targeting the astrocytes, we examined the expression of Cx43, p-Cx43, AKT, and p-AKT in vitro astrocyte cultures treated with BKM120 (a specific class I PI3K inhibitor) (Figure 6). The results of Cx43, p-Cx43, AKT, and p-AKT expression, and the ratio of p-Cx43/Cx43 and p-AKT/AKT in each group of in vitro cultured astrocytes were consistent with those observed in vivo. Moreover, there was no significant difference between the OGD/R + Vinp + LY and OGD/R + Vinp + BKM groups for the abovementioned proteins and their phosphorylation. These results provide more evidence that Vinp protects against cerebral I/R injury by targeting astrocytic Cx43 via the PI3K/AKT pathway.