NOX2-mediated reactive oxygen species are double-edged swords in focal cerebral ischemia in mice

Apocynin (178385, Sigma, St Louis, MO, USA) was dissolved in dimethyl sulfoxide and then diluted with sterile saline to the desired concentrations. Different concentrations of apocynin were intraperitoneally (i.p.) administered to the animals at a dose of 2.5 mg/kg 8 h after reperfusion as described by Qin et al. [22]. 3-MA (M9281, Sigma-Aldrich, St Louis, MO, USA) was dissolved in physiological saline and administered (15 mg/kg, i.p.) 1 and 3 h after occlusion [23]. Rapamycin (553210, Sigma-Aldrich, St Louis, MO, USA) was stored at room temperature and dissolved in DMSO with gentle heating to yield a clear, colorless solution and then injected (10 mg/kg, i.p.) immediately after MCAO [24].

The MCAO model was established as previously described [25]. In brief, C57BL/6J wild-type mice were anesthetized with 5% isoflurane in O₂ by a facemask, followed by ligation of the left middle cerebral artery with a 6-0 monofilament (Doccol Corp., Redlands, CA, USA). After 1 h of occlusion, the monofilament was removed to initiate reperfusion. A homeothermic heating pad was used to monitor and stabilize the body temperature at 37 ± 0.5 °C. The same procedure, but without monofilament ligation, was performed on sham-operated mice.

The mice were deeply anesthetized, euthanized with an overdose of isoflurane and decapitated 3, 7 and 14 days after MCAO. The brains were collected after transcranial perfusion with saline followed by 4% paraformaldehyde. After postfixation with 4% paraformaldehyde for 72 h, brain tissues were cut into 50-µm coronal sections, dipped in 0.1% cresyl violet solution for 30 min, and then rinsed in distilled water. The stained sections were fixed by serial dehydration in ethanol (70%, 90%, 100%) and xylene. The infarct volume was measured and analyzed by a blinded observer using ImageJ v1.37 (NIH, Bethesda, MA, USA), as described previously [26, 27], and the data were normalized and are presented as a percentage of the nonischemic hemisphere to correct for edema.

Neurological deficit scores were evaluated 3, 7 and 14 days after MCAO as described previously [28]. The score ranged from 0 (without observable neurological deficit) to 4 (no spontaneous motor activity and loss of consciousness).

Accelerating rotarod (SD Instruments, San Diego, CA) instruments were used to test the motor coordination of the mice. Each mouse was placed on a 2.75 cm diameter rotating rod every other day before MCAO onset for a total of 9 training days, and the rotation speed of the rod increased from 5 to 10 rpm every 5 min. The time between the beginning of the mouse staying on the rod and falling from the rod was determined up to a maximum duration of 300 s. After the training period, MCAO surgery was conducted, and rotarod testing was performed by a blinded observer on Days 0, 3, 7, 10, and 14 poststroke. The scores were calculated by averaging the three repetitive time records of each mouse each day.

The separation of ischemic core and penumbra was performed as previously descriptions with minor revision based on the average infarct volume measured by TTC or cresyl violet staining in our pilot experiments and the heterogeneity of penumbra as a dynamic and changeable process over time [29‐33]. Mice were deeply anesthetized and euthanized with an overdose of isoflurane, the brain tissue was quickly collected and put on ice and the olfactory bulb, cerebellum and low brain stem were removed. Firstly, the brain tissue was cut on a coronal plane 3 mm backward from the top of the frontal lobe into three slices with thickness of 3 mm (section 1), 4 mm (section 2) and 3 mm (section 3), respectively. Then, the section 2 was taken out and the midline was identified between the ipsilateral (left) and contralateral (right) hemispheres. Next, a sagittal cut 1.5 mm far from the midline was conducted from top to bottom on the ipsilateral hemisphere, and a transverse diagonal cut was also made at around “1 o'clock” position. Thus, the tissue outside “1 o'clock” was the infarct core area, and the cortical tissue between sagittal cut and “1 o'clock” was the ischemic penumbra (Additional file 1: Fig. S1).

To assess ROS production, the brain was carefully and quickly isolated, cut into 4.0 μm sections and placed on chilled microscope slides. The samples were incubated in physiological saline containing 10 μmol dihydroethidium (DHE; Sigma–Aldrich) for 30 min at 37 °C in the dark. The brain sections were washed twice with PBS and placed under an automatic fluorescence microscope (BX63, Olympus Optical Ltd, Tokyo, Japan).

Immunofluorescence analysis was performed as previously described [26]. Ischemic and sham-operated mice were euthanized and perfused with cold PBS, followed by fixation with 4% paraformaldehyde for 2 days. The ischemic brains were cut into 50-μm sections, and the free-floating slices were blocked with 0.1 M PBS containing 5% fetal bovine serum and 0.3% Triton X for 1 h at room temperature. After being washed, the slices were incubated at 4 °C overnight with the following primary antibodies: anti-NLRP3 (1:200; ab4207, Cell Signaling Technology, Boston, USA), anti-LC3B (1:200; ab104224, Abcam, Cambridge, England), and anti-CD31 (1:100, GB11224, Servicebio, Wuhan, China). The slices were then rinsed and incubated with an Alexa 594-conjugated antibody (1:200; ANT030, Millipore, Billerica, MA) or an Alexa 488-conjugated antibody (1:200; ANT024, Millipore, Billerica, MA) for 2 h at room temperature. After being thoroughly rinsed, the nuclei were stained with DAPI (94010, Vector Laboratories, Burlingame, CA, USA). All slices were photographed using a confocal fluorescence microscope (BX63, Olympus Optical Ltd, Tokyo, Japan). The number of immunoreactive cells in predefined areas was quantified using ImageJ software (Media Cybernetics Inc., Rockville, MD, USA). Six different fields for each mouse and six mice for each group were counted. All counts were conducted by blinded observers.

Western blotting was carried out as previously described [25]. Cortical sections 1.0 to 2.0 mm from ipsilateral brain tissue were harvested and homogenized in cold RIPA buffer (C1053, Applygen, Beijing, China) containing a protease inhibitor cocktail (G2006, Servicebio, Wuhan, China). The homogenates were centrifuged at 4 °C at 10,000×g for 30 min, and then the supernatants were harvested. Protein levels were determined with a BCA kit (G2026, Servicebio, Wuhan, China). Protein samples (20 μl/lane) were separated by electrophoresis on 4–15% sodium dodecyl sulfate–polyacrylamide gels and then transferred onto PVDF membranes (Millipore, Billerica, MA, USA). The membranes were then placed into 5% nonfat milk in PBS/0.1% Tween and blocked for 1 h, followed by incubation overnight with mouse anti-p62 (1:300, #16177, Cell Signaling Technology, Boston, USA), anti-NLPR3 (1:1000; ab4207, Cell Signaling Technology, Boston, USA), anti-ASC (1:1000; 67824, Cell Signaling Technology, Boston, USA), anti-CL-caspase-1 (1:500; 89332, Cell Signaling Technology, Boston, USA), anti-HMGB1 (1:1000; ab18256, Abcam, Cambridge, England), anti-IL-1β (1:500; ab8320, Abcam, Cambridge, England), anti-NOX2 (1:1000; ab18256, Abcam, Cambridge, England), anti-Beclin-1 (1:1000; ab18256, Abcam, Cambridge, England), anti-PI3K (1:1000; ab8378, Abcam, Cambridge, England), anti-p-PI3K (1:1000; ab8378, Abcam, Cambridge, England), anti-Akt (1:1000; ab8378, Abcam, Cambridge, England), anti-p-Akt (1:1000; ab8378, Abcam, Cambridge, England), anti- NF-κB p65 (1:1000; ab8378, Abcam, Cambridge, England), anti-phosphorylated NF-κB p65 (1:1000; 3033S, Cell Signaling Technology, Boston, USA) and anti-VEGF (1:1000; 2042744, Millipore, Billerica, MA) at 4 °C. After being washed with PBS/0.1% Tween, the membrane was incubated with IRDye-labeled secondary antibodies (1:10,000; c60405-05, Li-Cor Bioscience, USA) at room temperature for 1–2 h. Images were acquired with an Odyssey western blot analysis system (LI-COR, Lincoln, NE, USA). The relative band intensity was calculated using Quantity One v4.6.2 software (Bio–Rad Laboratories, Hercules, USA) and then normalized to the GAPDH loading control. These experiments were performed three times.

We first examined the effects of the NOX2 inhibitor apocynin on brain infarction and functional recovery. Three different doses of apocynin (1.5, 2.0 and 2.5 mg/kg) were studied. The results showed that 2.0 mg/kg apocynin reduced the infarct volume to approximately half that of the vehicle group; thus, this dose was selected for subsequent experiments (Additional file 2: Fig. S2). Then, we examined and compared the infarct volume by TTC staining at 3 days and Cresyl Violet staining at 7 and 14 days in WT mice and APO-treated mice subjected to MCAO. The results showed that the infarct volumes were significantly reduced in APO-treated mice compared with those in WT mice 3, 7 and 14 days after reperfusion (Fig. 1A, B). Although the neurological scores of APO-treated mice were lower within 3 days of stroke than those of WT mice, they worsened on Days 7 and 14 after stroke, when the neurological scores of WT animals had recovered to near normal levels (Fig. 1C). We then conducted a rotarod test to evaluate the effect of APO treatment on motor coordination. Similar to the neurological scores, both behavioral tests indicated that APO treatment improved behavioral performance on Day 3 but hindered functional recovery on Days 7 and 14 (Fig. 1D, E). Therefore, WT mice exhibited a significantly faster recovery than APO-treated mice.

As NOX2 is one of the main sources of ROS in the CNS, we evaluated NOX2 expression and ROS levels by using western blotting and DHE staining, respectively, to test whether the levels of ROS were related to NOX2. The results showed that the production of NOX2 and ROS were markedly increased at 3 days after stroke and gradually decreased between Day 3 and Day 14 (Fig. 2A and B). Furthermore, apocynin injection significantly suppressed the production of ROS and inhibited the expression of NOX2 (Fig. 2).

In the core of the ischemic tissue, which is clinically defined as the area with regional cerebral blood flow < 20%, and outside the ischemic core, brain tissue is still partially perfused, although at a reduced rate. This region, which is referred to as the ischemic penumbra, is often defined by a reduced perfusion rate that is, however, greater than that observed in the ischemic core. In the treatment of ischemic stroke, it is urgent to rescue neurons in the penumbra region. To investigate the biphasic mechanism of ROS, we performed RNA sequencing on the penumbral zone of the ischemic cortex of MCAO mice on Day 7. Bioinformatics analysis showed that 1616 genes were upregulated and 106 genes were downregulated in ischemic tissue compared to those in the sham group. Furthermore, apocynin increased the expression of 284 genes and reduced the expression of 2019 genes compared to MCAO (Fig. 3A). GO analysis of the top 20 genes under ischemic conditions showed that these genes were closely related to several biological processes, such as the regulation of leukocyte migration, T cell activation, and the regulation of the innate immune response (Additional file 3: Fig. S3). Furthermore, the genes upregulated by NOX2 inhibition were associated with the following biological processes: neurotransmitter transport, neutrophil-mediated immunity and regulation of neuronal projection development (Additional file 4: Fig. S4). In addition, we obtained two datasets related to differentially expressed genes between the vehicle group and sham group, as well as the apocynin treatment group and vehicle group, and then compared these datasets to further analyze a new dataset of differentially expressed genes. Next, the enrichment of biological functions was assessed in this new dataset by GO analysis, and these genes were involved in membrane-bound organelles, intracellular organelles and cytoplasmic parts (Fig. 3B). Moreover, using this method, we also found that apocynin inhibited angiogenesis (data not shown). Angiogenesis and inflammatory reactions in the penumbra of the ischemic cortex are essential steps for improving recovery or exacerbating brain injury following ischemic stroke, respectively. We performed confocal analysis of CD31 (an angiogenesis biomarker) in samples and revealed that angiogenesis in the penumbra gradually increased from Day 3 to Day 14 following ischemic stroke (Fig. 4A, C). After apocynin treatment, CD31⁺ cells were further significantly increased on Day 3 after stroke compared with those in the vehicle group. However, apocynin markedly decreased CD31⁺ cells on day 7 and 14 following stroke compared with the vehicle (Fig. 4A, C).

The NLRP3 inflammasome reflects inflammatory status. The immunofluorescent staining results showed that NLRP3⁺ positive cells were increased at 3 days but gradually decreased from 7 to 14 days following stroke in animals without apocynin (Fig. 4B, D). However, apocynin treatment significantly reduced NLRP3 expression on Day 3 after stroke but gradually increased NLRP3 levels from 7 to 14 days (Fig. 4B, D).

Autophagy, which is the intracellular mechanism by which cells digest and recycle unfolded proteins and dysfunctional organelles, is emerging as a major target of ROS and NOX enzymes [34, 35]. To explore the underlying mechanisms of the biphasic effects of ROS in the development of ischemic stroke, we focused on autophagy. The western blot results showed that the autophagy-related protein Beclin-1 was similar to NOX2 and ROS and significantly increased on Day 3 after stroke but gradually decreased between Day 3 and Day 14, and this effect was significantly inhibited by the NOX2 inhibitor apocynin (Fig. 5). Moreover, p62/SQSTM1 is an autophagosome adapter protein that is degraded by autophagy, whereas p62 accumulates when autophagy is inhibited. We also measured the expression of p62 by Western blotting, and the results showed that the levels of p62 were opposite to those of NOX2 and ROS, decreasing on Day 3 after stroke but gradually increasing between Day 3 and Day 14 (Fig. 5). We then measured the expression of another autophagy-related protein, LC3B, by immunofluorescent staining and confocal microscopy. The results suggested that LC3B was dramatically increased in the ischemic brain during the early stage of stroke but slightly increased during the late stage, which was also similar to the trend of NOX2/ROS and was inhibited by apocynin administration. These findings suggest that autophagy after stroke is likely dependent on NOX2/ROS activity.

As an intracellular bulk degradation system, autophagy contributes to pathologies associated with neuroinflammation regulation in mouse models of ischemic stroke [36, 37]. We then hypothesized that NOX2-derived ROS resolve inflammation during the delayed phase of stroke by activating autophagy. We used the autophagy inducer rapamycin and the autophagy inhibitor 3-methyladenine (3-MA) to induce and inhibit autophagy, respectively. The confocal immunostaining results showed that rapamycin induced autophagy but decreased the number of NLPR3⁺ cells; however, 3-MA inhibited autophagy activation and increased NLRP3 expression on Days 7 and 14 after MCAO (Fig. 6A). To study whether NOX2 inhibition regulates NLRP3 expression via autophagy inhibition, apocynin was injected in combination with either 3-MA or rapamycin. Confocal microscopic analysis showed that apocynin increased NLRP3⁺ cell numbers in the groups treated with apocynin plus 3-MA and rapamycin at 7 and 14 days after stroke compared with rapamycin or 3-MA alone (Fig. 6A). In addition, the western blot results indicated that apocynin promoted the expression of NLRP3 inflammasome-associated proteins, including NLRP3, ASC, CL-caspase-1 and IL-1β, in animals treated with the autophagy inhibitor 3-MA (Fig. 7). In contrast, apocynin did not significantly change the expression of NLRP3 inflammasome-associated proteins (Fig. 7) in mice treated with the autophagy agonist rapamycin. These results strongly indicate that apocynin exacerbates the inflammatory response by inhibiting autophagy.

Autophagy plays an important role in angiogenesis. We thus investigated whether the NOX2/ROS pathway enhances angiogenesis by modulating autophagy during brain recovery after stroke. The confocal microscopy results suggested that autophagy inhibition by 3-MA reduced staining of the angiogenesis marker CD31. However, rapamycin treatment significantly increased CD31 staining on day 7 and 14 following stroke (Fig. 8), and this effect was reduced by apocynin. We further found that apocynin inhibited VEGF expression in animals treated with rapamycin (Fig. 9). Consistently, apocynin further inhibited the expression of CD31 and VEGF 7 and 14 days following stroke in mice that received 3-MA (Fig. 9). Taken together, these results suggest that apocynin inhibits angiogenesis by inhibiting autophagy.

RNA sequencing and KEGG pathway analysis indicated that genes related to the NF-kappa B signaling pathway, NOD-like receptor signaling pathway and Toll-like receptor signaling pathway were enriched at 7 days after MCAO (Additional file 5: Fig. S5 and Additional file 6: Fig. S6). After apocynin treatment, the differentially upregulated genes were enriched in the PI3K–Akt signaling pathway, NF-kappa B signaling pathway and MAPK signaling pathway, as shown by KEGG pathway analysis. Next, the two groups of differentially expressed genes were further compared, and the results showed that apocynin treatment resulted in the upregulated genes being enriched in several key signaling pathways, including the PI3K–Akt signaling pathway, cytokine–cytokine receptor interactions and the N-Fkappa B signaling pathway (Fig. 9A). These findings indicated that the PI3K–Akt signaling pathway may participate in NOX2-mediated tissue recovery during the long-term stage of brain ischemia. The PI3K/Akt/NF-kB pathway is well known to be involved in autophagy-dependent responses to angiogenesis and inflammation inhibition. In addition, to further confirm whether NOX2 modulates these processes via the PI3K–Akt signaling pathway during the long-term recovery of ischemic stroke, we assessed the levels of pathway-related proteins by western blot analysis. We found that rapamycin promoted the protein expression of phosphorylated PI3K, Akt, NF-kB and VEGF, downregulated the autophagy-associated protein Beclin-1 and upregulated p62, while 3-MA inhibited expression of these factors (except p62) compared with the vehicle (Fig. 9B). However, apocynin reduced the levels of these protein that were increased by rapamycin and further attenuated the levels of proteins that were already reduced by 3-MA (Fig. 9). Figures 7 and 8 show that the activation of the NLRP3 inflammasome and its associated proteins, including NLRP3, ASC, CL-caspase-1, HMGB1 and IL-1β, was inhibited by apocynin treatment alone or combined with 3-MA or rapamycin. Therefore, combining Fig. 9 with Figs. 7 and 8, our results suggest that the PI3K/Akt/NF-kB signaling pathway is involved in angiogenesis and inflammation inhibition induced by NOX–ROS-dependent autophagy during the recovery stage of ischemic stroke.