All male Sprague Dawley rats (weight 260–280 g) were purchased from Zhejiang Laboratory Animals Center (Hangzhou, China). Male C57BL/6 (25–30 g) mice were purchased from Comparative Medicine Centre (Yangzhou University, China). All experiments carried out in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. Animals used were approved by the Committee of Experimental Animals in Jiangsu Province and the Ethics Committee of China Pharmaceutical University.

Total of three separate animal experiments were carried out as shown in Supplement Figures 1A, B and C.

CBL was provided by Guangdong Long Fu Pharmaceutical Co., Ltd. (Zhongshan, China). Rabbit anti-phospho-CREB (1:1,000), anti-CREB (1:1,000), anti-PGC-1α (1:1,000), anti-β-actin (1:10,000) antibodies were purchased from the ABclonal company. Rabbit anti-phospho-ERK1/2 (1:1,000), anti-ERK1/2 (1:1,000), anti-phospho-JNK (1:1,000), anti-JNK (1:1,000), anti-phospho-p38 MAPK (1:1,000), anti-p38 MAPK (1:1,000) antibodies were purchased from Cell Signaling Technology (Beverly, MA). Compound 666-15, the inhibitor of CREB was purchased from MedChemExpress (MCE, Shanghai, China). RIPA lysis buffer and LDH kit were purchased from Beyotime Biotechnology (Nanjing, China). Trizol reagent and the cDNA synthesis kit were purchased from Vazyme (Nanjing, China). SYBR Green was purchased from Invitrogen (Camarillo, CA). LPS was purchased from Sigma-Aldrich (St. Louis, USA). Cell culture medium and supplements were purchased from Invitrogen (Carlsbad, CA, USA). TTC (2, 3, 5-triphenyltetrazolium chloride) was bought from Sigma-Aldrich.

The healthy male SD rats were randomly divided into a series of groups for transient middle cerebral artery occlusion (tMCAO) (n = 12–15 for each group of successfully treated rats). Firstly, the rats were treated with anesthesia in an isoflurane chamber with 3.5% isoflurane, and then 2% isoflurane was maintained through a mask in the operation. During the surgery, the animals were placed on a heating device to ensure normal body temperature (37°C). After accurate separation of the right common carotid artery (CCA), internal carotid artery (ICA), and external carotid artery (ECA), a monofilament nylon suture (about 0.24 mm in diameter) with a rounded tip was inserted through the ECA stump into the ICA and gently advanced to the MCA. Then, in order to monitor blood block, the Laser Speckle Imaging system (moorFLPI-2^™) was used to detect whether the baseline of brain blood flow was >75% reduced (Figure S2). Two hours after cerebral ischemia, filament was removed to restore blood flow (reperfusion). All rats had free access to food and water. At 3 h and 24 h after the ischemia, the separated groups of rats were intravenously injected with indicated dose of CBL (dissolved in saline) or saline alone. In total, 120 rats went through tMCAO operation, among which 6 rats were excluded for hemorrhagic transformation, 10 for unsuccessful occlusion and 9 rats died during the operation.

Rat neurological performance was measured by the following behavioral tests with minor modifications, in a blinded manner. After the tMCAO experiment, the neurological impairment of rats were assessed by Longa’s test: 0, normal function; 1, flexion of the torso and contralateral forelimb after lifting the animal by the tail; 2, circling to the contralateral side but normal posture at rest; 3, reclining to the contralateral side at rest; 4, absence of spontaneous motor activity. Sections were then stained with 2% TTC and placed in the dark area at 37°C. After the brain was stained, the infarct area turned white, while the normal brain tissue remained red. The ratio percentages of the infarct areas to the total brain areas were accessed by morphometric analysis Image-pro plus (Media cybernetics, MD, USA).

To assess the motor behavior recovery after tMCAO with or without treatment of CBL, rats were subjected to a rotating rod test. In this experiment, the rats were placed in a rotating rod from 4× rpm to 40× rpm during the training period, and finally each rat was kept for 300 s without dropping the rod. The latency to fall off the rotating rod was recorded three times daily till 14 days after brain ischemia. And the final data were analyzed as the average value.

The Corner test was carried out to detect the asymmetrical sensorimotor dysfunction of rats after tMCAO. The experimental device consists of two cardboards forming an angle of 30°. When the rats enter it, the bilateral tentacles will be stimulated, and the rats will stand and face the open end of the angle. Normal rats have the same chances of going left side or right one. However, the rats after tMCAO operation usually turn to the side of the brain injury. The numbers of left and right over 10 trials were recorded.

Modified neurological severity score is used mainly to determine the improvement of neurological deficit, thus we used this method to assess the neurological impairment of rats after tMCAO with or without the administration of CBL. The modified neurological severity score (mNSS) contains exercise test, tailing test, the placement test, proprioceptive test (deep sensation, pushing the paw against the table edge to stimulate limb muscles), balance beam test, the experiment of reflex loss and abnormal movement. The Maximum point is 18 and the overall points were assessed by the behavior of rats after tMCAO (normal score, 0; maximal deficit score, 18).

As previously reported, LPS can induce the neuroinflammation in mice (Leite et al., 2016). Firstly, mice were randomly separated into five groups (n = 6–9 of successfully treated rats per group). Mice were intraperitoneally injected with 20 mg/kg, 60 mg/kg, or 100 mg/kg CBL dissolved in the saline or saline alone daily in the 3 consecutive days. On the third day, at 3 h after the administration of CBL or saline in mice, LPS (0.33 mg/kg) dissolved in the saline was intraperitoneally injected according to the body weight of the mice. After 3 h, all the mice were subjected to Open field test to analyze the sickness behavior induced by the LPS. After the behavior test, the blood of the mouse was immediately taken for measurement of pro-inflammatory factors, and the whole cerebral cortex of the mouse was extracted for assessment of inflammatory mediators and related protein expressions.

To determine the locomotor activity, the mice after the intraperitoneal administration of 0.33 mg/kg LPS were put into the square open field box (50 × 50 cm). And the total distance traveled, the total time spent and other useful information in the inner zones were recorded by the overhead camera and the final results were analyzed by TopScan software (Anymaze^™, Stoelting Co). After experiment, the inner and bottom surfaces of the square box must be cleaned to avoid interference information. And the inner zones need to be consistently bright to avoiding interfering with the locomotor activity of mouse. Each mouse was shortly given an overall score for total locomotor activity, including total distance removed, central distance removed, central time spent, time moving towards central area, line crossings, time in the central area, and number exit the central area.

The TNF-α ELISA kit was bought from the Becton Dickinson Company (USA). Detection of TNF-α protein was carried out according to the manufacturer’s protocol for the quantification.

Fifteen newborn mice were disinfected with 75% ethanol, then the brains were put into ice-cold D-HBSS solution, and the vascular membranes were completely removed from it. After 20 min of trypsin digestion, DF₁₂ with 10% FBS was added to stop the digestion. Then, the mixture was centrifuged and the precipitation was added to 75 cm² flask coated with poly-L-lysine (PLL, Sigma-Aldrich), and incubated at 37°C and 5% CO₂. After 14 days of culture, primary mouse microglia were obtained and validated by immunostaining.

When finishing the cell experiments, 100 μl of Radio Immunoprecipitation Assay (RIPA) lysis buffer was added after washing cells with phosphate buffered solution (PBS). Cell protein samples were then collected as described previously (Gao et al., 2015) and separated by sodium salt-Polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to Poly Vinylidene Fluorid (PVDF) membranes. Hereafter, the membranes were blocked with 5% bovine serum albumin for 2 h and then incubated with rabbit antibody of anti-phospho-CREB, rabbit anti-PGC-1α, and rabbit anti-β-actin for 12 h. After washing with Tris buffer saline plus Tween (TBST), the membranes were incubated with the anti-rabbit IgG (1:10,000; Sunshine Bio, China) at room temperature for 1 h. The final bands were measured through chemiluminescence with Bio-Rad ChemiDoc XRS (Bio-Rad, Hercules, California, USA) and protein expression levels were quantified using densitometric analysis and normalized to the levels of β-actin protein.

In order to detect the level of expression of the target genes in cells and brain tissues, TRIZOL reagent was added to extract total RNA. Afterwards, according to standard protocol, isolated RNA was reverse-transcribed into cDNA using cDNA synthesis kit. Quantitative PCR (qPCR) was performed at 95°C for 10 min and 40 cycles of 95°C for 15 s, 60°C for 60 s with synthetic primers and SYBR Green. The final results were all normalized as fold change of the target gene/18s rRNA. The primers for qPCR are listed in Table 1.

Primary microglia were cultured at a density of 70–80% in 96-well plates and then treated with different concentrations of CBL for 24 h. Cell supernatants were then collected and assessed by using the Lactate dehydrogenase (LDH) kit and the absorbance was read at 570 nm according to the manufacturer’s instructions.

The statistical analysis of data was processed using GraphPad Prism Software 7 (La Jolla, CA) and all results are presented as mean ± SEM. Statistical analysis about the frequencies (Longa test and corner test) was put into effect using non-parametric Mann-Whitney test, and the distinction of multiple groups were executed through the One-way ANOVA consistent with Bonferroni’s test or the Two-way ANOVA followed by Bonferroni’s test. The differentiation was recognized significant if the P < 0.05 (Table S1).

As previously reported, CBL has been used in treating acute ischemic stroke (Zhang et al., 2010; Heiss et al., 2012). Thereby, in order to make sure whether the treatment with CBL can display neuroprotective effects in rats after tMCAO operation, the percentage of the infarct size and neurobehavioral score were analyzed by TTC staining and Longa test, respectively. The stroke group displayed a significant infarct size and neurological deficit. In comparison to stroke group, the infarct area and the neurological deficit of ischemic rats treated with CBL were significantly reduced in a dose-dependent manner (Figures 1A, B) and in a time-dependent manner (Figures 1C, D). It’s well established that neuroinflammation exerted a crucial role in pathologic process of acute ischemic stroke. Thus, we performed the experiment and found that the mRNA expression of the pro-inflammatory factors (TNF-α, IL-1β, and iNOS) in ipsilateral cerebral cortex was significantly decreased and the mRNA expression of the anti-inflammatory factors (CD206, YM1/2, and Arginase 1) was increased after CBL treatment at 60 mg/kg than the stroke group (Figures 1E, F).

We found that even at 14 day after stroke, CBL also significantly reduced the infarct volume of rats subjected to ischemia (Figure 2A). To further test the long-term functional recovery after CBL administration in tMCAO rat model, rats were fulfilled with series of experiments of mNSS, rotarod test, corner test, survival proportions, and body weight from the first day to the fourteenth days after tMCAO. Administration of CBL at 60 mg/kg post-ischemia significantly increased the survival proportions in rats after ischemia, but had no effect on weight loss (Figures 2B, C). The stroke group displayed obvious neurological deficits according to the experiments of mNSS, rotarod test, and corner test, which were ameliorated by the administration of CBL (Figures 2D–F).

As previously demonstrated, in tMCAO model, neuroinflammation exerts considerable influence (Leite et al., 2016). Therefore, we selected the model of LPS-induced neuroinflammation to verify again whether CBL possesses the anti-inflammatory effects (Zhao et al., 2017). In this model, all mice survived and displayed no abnormal behavior. In the Open field test, the group injected with LPS showed clear sickness behavior, which was ameliorated by the administration of CBL in a dose-dependent manner (Figure 3A).More specifically, treatment with CBL significantly increased the LPS-induced decrease in total distance, central distance, central time, time moving towards central area, line crossings, time in the central area, and number exit the central area (Figure 3B).

Furthermore, LPS administration promoted the protein expression of TNF-α in the serum of mice, which was reduced by CBL treatment in a dose-dependent manner (Figure 3C). Subsequently, we extracted the cerebral cortex of the mouse and found the gene expression of pro-inflammatory mediators has been largely elevated in the group of LPS injection (TNF-α, COX2, and iNOS). Compared with it, CBL administration evidently ameliorated this phenomenon (Figure 3C). Correspondingly, CBL significantly increased anti-inflammatory mediator gene expression (CD206, Arginase 1, and IL-10) (Figure 3D). These data demonstrated CBL could promote microglia activation towards the anti-inflammatory phenotype in the LPS-induced neuroinflammatory model.

To ascertain the anti-inflammatory effects of CBLin in vitro model, we stimulated primary mouse microglia with 100 ng/ml LPS. When the concentration arrange of CBL was from 0.5 to 5 μg/ml, it displayed no toxicity in primary microglia (Figure 4A). In order to further determine the optimal anti-inflammatory dose of CBL, we collected the cell culture medium of primary microglia after LPS stimulation and found that 5 μg/ml is the best concentration to inhibit inflammation (Figure 4B). Meanwhile, 5 μg/ml of CBL significantly decreased LPS-induced elevation of pro-inflammatory factor gene expression (TNF-α, IL-1β, iNOS, and COX-2) (Figure 4C), and increased anti-inflammatory mediator gene expression (Figure 4D). The above results implied that CBL promoted the switch of microglia towards an anti-inflammatory phenotype in primary microglia cultures.

LPS activated the phosphorylation of c-Jun N-terminal kinase (JNK), extracellular signal-regulated kinase 1/2 (ERK1/2) and p38 mitogen-activated protein kinase (p38 MAPK) in primary microglia in a time-dependent manner, and 1 h was the highest point after stimulation with LPS. In primary mouse microglia, CBL significantly inhibited LPS-induced phosphorylation of p38 MAPK and JNK, but displayed weakly inhibitory effect on ERK1/2 phosphorylation (Figure 5A).

As previously reported, MAP kinases can activate phosphorylation of CREB (p-CREB), while CREB is a transcription factor that regulates the expression of PGC-1α (Raefsky and Mattson, 2017). Therefore, we examined the effect of CBL on the protein expression of p-CREB and PGC-1α. The administration of CBL enhanced the expression of PGC-1α and p-CREB in primary microglia cells in the time-dependent manner (Figure 5B) and dose-dependent manner (Figure 5C). The maximum effective concentration of CBL was 5 μg/mL and the expression levels of related protein reached maximum at 9 h after exposure to CBL.

The compound 666-15 is a potent and selective CREB inhibitor which inhibits transcription activity of CREB. When primary microglia were treated with 666-15 for 12 h, the level of p-CREB protein was markedly decreased (Figure 6A). When stimulated with LPS, 666-15 also inhibited the expression of p-CREB and PGC-1α after the administration of CBL in LPS-stimulated primary mouse microglia (Figure 6B). The inhibition of CREB also reversed the effect of CBL on reducing pro-inflammatory factor gene expression (TNF-α, COX2) (Figure 6C) and increasing anti-inflammatory mediator gene expression (Arginase 1 and IL-10) (Figure 6C). These data indicate that the CREB-mediated PGC-1α activation may be involved in the anti-inflammatory effects of CBL in primary microglia.

In addition, we performed animal experiment to determine whether CREB inhibitor could prevent the anti-ischemic effect of CBL. As shown in Figure 6D, we found that the CREB inhibitor 666-15 compound could significantly reverse the beneficial effect of CBL in rats subjected to tMCAO, which further confirms that the CREB pathway is involved in the anti-ischemic effect of CBL.

In LPS-induced neuroinflammation model, the group administrated with CBL significantly enhanced the protein expression of PGC-1α and CREB phosphorylation (p-CREB) compared with LPS group (Figure 7A). Accordingly, in the ipsilateral brain cortex of rats after tMCAO, treatment with CBL markedly increased the protein expression of PGC-1α and CREB phosphorylation in comparison with the stroke group (Figure 7B).

In conclusion, the above results indicate that the administration of CBL exerted anti-inflammatory effects and ameliorated cerebral ischemic stroke injury through activation of the CREB/PGC-1α signaling pathway.