This was a case-control study. From March 2019 to March 2020, 86 patients who were hospitalized in the Neurology Department of Jiangbin Hospital and initially diagnosed with ACI-induced MCI were enrolled in our study. The diagnosis of ACI conforms to the diagnostic criteria of the Guidelines for The management of Ischemic Stroke and Transient Ischemic Attack 2008 [13]. All patients had evidence of ACI on nuclear magnetic resonance (MRI) or computed tomography (CT). Based on the results of MRI/CT, the territories of cerebral infarct were divided into three, the territory of the anterior cerebral artery (ACA), the territory of the middle cerebral artery (MCA) and the territory of posterior circulation (PC). In this study, all patients had a single territory of cerebral infarct. The diagnosis of MCI follows Petersen et al.’s criteria [14]. The exclusion criteria were as follows: patients with multiple territories of cerebral infarct; patients with diabetes mellitus, malignant tumor, metabolic disorders, recurrent stroke, or massive ischemic stroke with severe impairment of consciousness. This study adhered to the principles of the Declaration of Helsinki. Ethical approval was granted by the institutional review board of Jiangbin Hospital (No. LW-2022-09). All patients were randomly divided into the control (Ctrl) and NBP groups. Informed consent was obtained from patients or their guardians.

Patients in both groups received conventional symptomatic treatment, including blood pressure and blood glucose control, hypolipidemic, anticoagulants, and microcirculation improvement drugs, supplemented with cognitive function rehabilitation training. Patients in the Ctrl group received a placebo. Patients in the NBP group received NBP soft capsule (H20050299, CSPC-NBP Pharmaceutical Co., Ltd., Shijiazhuang, Hebei, China) (0.2 g orally thrice daily). Efficacy was evaluated after eight weeks of therapy.

The brains of the two groups were scanned by MRI (GE Signa HDXT 1.5T MR, General Electric Company, Boston, MA, USA)/CT (SIEMENS SOMATOM Definition FLash CT, SIEMENS, Amberg, Germany) before and after treatment, and the changes in lesions were compared. The montreal cognitive assessment scale (MOCA) was used to assess cognitive function in both the groups before and after treatment. Cognitive function was evaluated based on memory, attention, orientation, language impairment, naming ability, visual-spatial executive ability, abstract thinking, and delayed memory. A total score of 30, 26, or above was considered normal; 18–26, MCI; 10–17, moderate impairment; and 10, severe impairment. Activities of daily living (ADL) were assessed in both groups before and after treatment. The total scores were 100, 100 represent normal, $>$60 was mild disorder, 41–60 was moderate disorder, 20–40 was severe disorder, and 20 was disability.

Neurons (CP-R107, Procell Life Science & Technology Co., Ltd., Wuhan, Hubei, China) were cultured according to the protocol described by Wang et al. [15]. Leica microscopes (DMi1, Leica, Shanghai, China) were used to observe cell morphology. OGD model was performed following the method of Tasca et al. [16]. After 24 h of exposure to OGD, neurons were replaced with normal media. The cells were divided into OGD and OGD + NBP groups. The drugs and concentrations added to each group were as follows: no drugs were added to the OGD group, and NBP was added to the OGD + NBP group at concentrations of 0.1, 1, 10, 50, and 100 µM. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) kit (AR1156, Boster, Wuhan, Hubei, China) was used to calculate the cell survival rate after 24 h. The optimal concentration was selected for subsequent experiments.

The patch-clamp system was used to obtain the action potentials (APs). The primary method described by Li et al. [17]. The amplitude and frequency of neuronal APs were collected and analyzed.

Protein extraction, protein concentration detection, sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE), membrane transfer, antibody incubation, and visualization of immunoreactive bands were performed as described by Li et al. [18]. The primary antibodies used were anti-TLR4 antibody 1:1000 (ab8376, Abcam, Shanghai, China), anti-HMGB1 antibody 1:1000 (ab190377, Abcam, Shanghai, China), anti-interleukin (IL-6) antibody 1:1000 (ab259341, Abcam, Shanghai, China), anti-tumor necrosis factor-alpha (TNF-$\alpha$) antibody 1:1000 (ab205587, Abcam, Shanghai, China), and anti-GAPDH antibody 1:10000 (D110016, Sangon Biotech, Shanghai, China). Subsequently, the nitrocellulose (NC) membrane was incubated with horseradish peroxidase (HRP)-conjugated secondary antibody 1:8000 (Proteintech, SA00001-2, Wuhan, Hubei, China) for 1 h at room temperature. Protein levels were determined by normalizing to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) loading controls. Each experiment was conducted in triplicates. The original figures of Western Blot can be found in the Supplementary Materials-IB.

Protein extraction, protein concentration detection, construction of antigen-antibody complexes, and immunoblotting (IB) were performed as described by Li et al. [18]. The antigen-antibody complex was constructed using an anti-TLR4 receptor antibody 1:30 (Abcam, ab252430). The primary antibodies for IB were anti-TLR4 antibody 1:1000 (Abcam, ab8376) and anti-HMGB1 antibody 1:1000 (Abcam, ab18256). The Immunoprecipitation (IP) protein band for TLR4 was used to normalize the pull-down protein levels obtained from Co-IP. The co-precipitated protein levels between each group were then compared. The original figures of Western Blot can be found in the Supplementary Materials-IP.

The fixation, penetration, and blocking of the neurons were based on the methods described by Li et al. [17, 18]. A multiplex fluorescent immunohistochemistry kit (abs50012, Absin, Shanghai, China) was used to stain neurons [18]. The primary antibodies for Immunofluorescence (IF) were anti-TLR4 antibody 1:100 (Abcam, ab8376) and anti-HMGB1 antibody 1:100 (Abcam, ab18256). A Leica 8.0 laser confocal microscope (TCS SP8, Leica, Wetzlar, Germany) was used to acquire fluorescence signals of TLR4 (green) and HMGB1 (red). The ImageJ plug-in (version 1.54b; University of Wisconsin, Madison, WI, USA) was used to calculate the Pearson’s correlation coefficients (PCCs) between TLR4 and HMGB1.

Statistical analysis was conducted using Statistical Package for Social Sciences (SPSS, version 25.0, IBM, Chicago, IL, USA) software. The measurement data are expressed as mean $\pm$ standard deviation. The independent sample t-test, paired samples t-test, and One-Way analysis of variance (ANOVA) were used to assess the significance of the differences between the groups. The chi-square test assessed the significance of the difference in percentage. Statistical significance was set at p 0.05.

The baseline characteristics were exhibited in Table 1. The Ctrl group included 20 male and 19 female participants aged between 43 and 78 (66.00 $\pm$ 9.10) years. The disease course ranged from 1 to 10 days (5.97 $\pm$ 2.33 days). The number of the patients with history of hypertension was 31 (79.49%). The number of the patients with history of smoking was 13 (33.33%). The number of the patients with the territory of ACA was 12 (30.77%). The number of the patients with the territory of MCA was 19 (48.72%). The number of the patients with the territory of posterior circulation (PC) was 8 (20.51%).

The NBP group comprised 26 male and 17 female participants aged between 44 and 76 (62.44 $\pm$ 8.88) years. The disease course ranged from 1 to 10 days (6.09 $\pm$ 2.16 days). The number of the patients with history of hypertension was 33 (76.74%). The number of the patients with history of smoking was 16 (37.21%). The number of the patients with the territory of the ACA was 10 (23.26%). The number of the patients with the territory of the MCA was 28 (65.12%). The number of the patients with the territory of posterior circulation 5 (11.63%). There were no differences in the baseline characteristics of the Ctrl and NBP groups, such as gender, age, disease course, hypertension, smoking and the territories of cerebral infarct, between the two groups (p $>$ 0.05).

Before treatment, no significant differences in MOCA were observed between the control and NBP groups. Table 1 presents the pretreatment composition of the Ctrl group, including 0 normal patients, 0 patients with MCI, 13 with moderate cognitive impairment, and 30 with severe cognitive impairment. After treatment, there were 0 normal patients, 1 patient with MCI, 15 with moderate cognitive impairment, and 28 with severe cognitive impairment in the Ctrl group (Table 2). The pretreatment NBP group comprised 0 normal patients, 1 patient with MCI, 12 moderate cognitive impairment patients, and 24 severe cognitive impairment patients. After treatment, there were 0 normal patients, 2 patients with MCI, 20 with moderate cognitive impairment, and 17 with severe cognitive impairment in the NBP group. The Ctrl group achieved a MOCA score of 8.81 $\pm$ 2.44 after treatment compared to 8.51 $\pm$ 2.30 before treatment (p = 0.000, Table 3). Similarly, the NBP group achieved a MOCA score of 10.79 $\pm$ 3.79 after treatment compared to 8.44 $\pm$ 3.83 before treatment (p = 0.000, Table 3). The Ctrl group revealed no significant improvement in memory, attention, orientation, language impairment, naming ability, visuospatial executive ability, abstract thinking, or delayed memory (Table 3). Conversely, the patients in the NBP group demonstrated significant improvements in memory, attention, orientation, and delayed memory (Table 3). Before treatment, there were no significant differences in ADL between the Ctrl and NBP groups (Ctrl versus NBP, p $>$ 0.05). Before treatment, there were 0 normal patients, 12 patients with mild disorder, 17 with moderate disorder, 4 with severe disorder, and 10 with disability in the Ctrl group. After treatment, there were 0 normal patients, 12 patients with mild disorder, 17 with moderate disorder, 6 with severe disorder, and 8 with disability (Table 4). Before treatment, there were 3 normal patients, 4 with mild disorder, 19 with moderate disorder, 2 with severe disorder, and 11 with disability in the NBP group. After treatment, there were 3 normal patients, 4 with mild disorder, 20 with moderate disorder, 2 with severe disorder, and 10 with disability in the NBP group (Table 4). The Ctrl group achieved an ADL score of 49.19 $\pm$ 27.45 after treatment compared to 47.44 $\pm$ 28.12 before treatment (p = 0.004, Table 5). The NBP group improved the ADL score from 42.18 $\pm$ 31.03 pretreatment to 45.13 $\pm$ 28.20 post-treatment (p = 0.000, Table 5). Patients in the Ctrl group revealed no significant improvement in stool, urination, grooming, toilet use, eating, transfer, movement, dressing, ascending and descending stairs, and bathing after treatment (Table 5). However, patients in the NBP group demonstrated improvements in grooming and eating abilities after treatment (Table 5).

Before treatment, CT and MRI scans revealed distinct cerebral ischemic lesions in both groups without statistically significant differences in the ischemic area. The Ctrl group exhibited an infarct size of 276.81 $\pm$ 249.90 mm², whereas the NBP group exhibited an infarct size of 198.25 $\pm$ 103.32 mm² (versus Ctrl, p = 0.064; Fig. 1). After treatment, no significant alterations in infarct size were observed in the Ctrl group (versus before treatment, p = 0.084; Fig. 1) or the NBP group (versus before treatment, p = 0.073; Fig. 1).

Fig. 1.

Changes in cerebral infarct size: the infarct size of the Ctrl group was 276.81 $\mathrm{\pm }$ 249.90 mm², while the infarct size of the NBP group was 198.25 $\mathrm{\pm }$ 103.32 mm². After treatment, infarct size did not change significantly in the Ctrl and NBP groups. Red arrows: infarct size. DWI, diffusion weighted imaging.

In the OGD group, cellular edema was observed in the neurons, and their number was also reduced. The OGD + 1 µM NBP, OGD + 10 µM NBP, and OGD + 50 µM NBP groups exhibited decreased edema and increased number of neurons (Fig. 2A). To further clarify the protective effects of NBP on neurons, we performed an MTT assay. The changes in neuronal survival were most significant in the OGD + 1 µM NBP, OGD + 10 µM NBP, and OGD + 50 µM NBP groups (versus OGD, p $>$ 0.05; Fig. 2B). The survival rate of neurons was highest in the OGD + 10 µM NBP group. Therefore, we performed electrophysiological tests on this group of neurons using whole-cell patch-clamp recordings. The data revealed that the amplitudes of the APs were higher in the neurons of the OGD + 10 µM NBP group versus those of the OGD group (p = 0.001; Fig. 2C). However, there were no differences in the frequency of neuronal APs between OGD group and OGD + 10 µM NBP group (p $>$ 0.05; Fig. 2C).

Fig. 2.

Morphology, survival and action potentials of neurons. (A) Neuronal morphology: in the OGD group, cellular edema was observed in the neurons, and the number was also reduced. The OGD + 1 µM NBP, OGD + 10 µM NBP, and OGD + 50 µM NBP groups exhibited decreased edema and increased number of neurons. Red arrows: swollen and ruptured neurons. Scale Bar: 500 µm. (B) MTT assay: the changes in neuronal survival were most significant in the OGD + 1 µM NBP, OGD + 10 µM NBP, and OGD + 50 µM NBP groups (versus OGD, #, p 0.01). The survival rate of neurons was highest in the OGD + 10 µM NBP group. (C) The amplitudes of the APs were higher in the neurons of the OGD + 10 µM NBP group (versus OGD, #, p 0.01). There were no differences in the frequency of neuronal APs between OGD group and OGD + 10 µM NBP group. OGD, oxygen and glucose deprivation; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; APs, action potentials.

We extracted total protein from neurons in the OGD and OGD + 10 µM NBP groups for the IB assay. The results demonstrated that TLR4, HMGB1, and IL-6 expression in the OGD + 10 µM NBP group did not change significantly. However, TNF-$\alpha$ expression was significantly decreased (versus OGD, p = 0.00; Fig. 3A). We identified the relationship between TLR4 and HMGB1 using the Co-IP assay. These findings revealed a significant decrease in the interaction between TLR4 and HMGB1 (versus OGD, p = 0.000; Fig. 3A). We observed co-expression of TLR4 and HMGB1 in neurons using an IF assay. To further verify the relationship between TLR4 and HMGB1, PCC was used to evaluate the interactions between these proteins. In the OGD + 10 µM NBP group, the PCC between TLR4 and HMGB1 was decreased (versus OGD, p = 0.000; Fig. 3B).

Fig. 3.

The TLR4/HMGB1 pathway was inhibited by NBP. Neuronal inflammation and the TLR4/HMGB1 pathway between OGD group and OGD+10 µM NBP group (A) The IB results indicated that TLR4, HMGB1, and IL-6 expressions did not change significantly in the OGD + 10 µM NBP group, but TNF-$\alpha$ expression was significantly decreased (versus OGD, *, p 0.05). The results of the Co-IP assay revealed that the interaction between TLR4 and HMGB1 decreased (versus OGD, #, p 0.01). (B) The PCC between TLR4 and HMGB1 decreased (versus OGD, #, p 0.01). Scale Bar: 50 µm. TLR4, toll-like receptor 4; HMGB1, expression of high mobility group box 1; IL-6, interleukin-6; TNF-$\alpha$, tumor necrosis factor-alpha; PCC, Pearson’s correlation coefficient; Co-IP, Co-Immunoprecipitation; IB, Immunoblotting; IgG, immunoglobulin G; GADPH, Glyceric acid phosphate dehydrogenase; DAPI, 4${}^{\prime }$,6-diamidino-2-phenylindole.