Mouse microglial BV-2 (cat. no. SNL-155) and hippocampal neuronal HT22 (cat. no. SNL-202) cells were obtained from Wuhan Sunncell Biotechnology. Cells were cultured in BV-2 cell-specific medium (cat. no. SNLM-155) and HT22 cell-specific medium (SNLM-202; both Wuhan Sunncell Biotechnology, respectively. The cultures were maintained at 37°C with 5% CO₂, and the medium was refreshed every 2 days.

Following normal culture for 24 h at 37°C, BV-2 and HT22 cells were placed in a glucose- and serum-free basal medium (cat. no. PM150270, Wuhan Pricella Biotechnology) and placed in a 95% N₂, 5% CO₂ hypoxic incubator at 37°C for 2 h. Cells were transferred to normal medium and maintained at 37°C with 95% oxygen and 5% CO₂ for 24 h to construct an OGD/R cell model (29). In addition, for the PPF group, PPF (2, 4, 8, 16, 32, 64 and 128 μM, cat. no. HY-B0649, MedChemExpress) was added to the culture medium during reoxygenation (30). In subsequent experiments, cells were treated with 8, 16 and 32 μM PPF during reoxygenation to evaluate its neuroprotective effects. According to the method described by Beaulieu et al (31), OGD-treated HT22 (in the lower chamber) and BV2 cells (in the upper chamber) were co-cultured at 37°C in a Transwell chamber (0.4 μm, Corning, Inc.) for 24 h at a 1:3 inoculation ratio of HT22:BV2 cells (Fig. 1A).

MFG-E8 small interfering (si)RNA (forward, 5'-CCAAUGUCUGGUGACUUUTT-3' and reverse, 5'-AAAGUCACCAGACAUUUGGTT-3'), MFG-E8 overexpression (OE) plasmid and negative control plasmids (si-NC: Forward, 5'-GAUGGAGAAGCUCGCUGATTT-3' and reverse, 5'-AAAUCAGCGAGCUUCUCCATT'; OE-NC) were synthesized by Sangon Biotech Co., Ltd. Plasmids (2 μg) were transfected into BV-2 cells for 48 h at 37°C using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. At 48 h post-transfection, BV-2 cells were fully lysed with RIPA lysis buffer (cat. no. 20-188, Sigma-Aldrich; Merck KGaA) and MFG-E8 protein expression was analyzed by western blotting to evaluate the transfection efficiency. For the OGD/R + OE-MFG-E8 + LPS (NF-κB agonist) group, cells were transfected with OE-MFG-E8, treated with OGD and exposed to LPS (1 μg/ml; cat. no. HY-D1056, MedChemExpress) during reoxygenation (32). For the OGD/R + si-MFG-E8 + PPF group, BV-2 cells were first transfected with si-MFG-E8, then subjected to OGD and treated with PPF during reoxygenation.

BV-2 and HT22 cells were inoculated into 96-well plates (1×10⁴ cells/well, 100 μl). When the cells were completely adherent (12-16 h), PPF (2, 4, 8, 16, 32, 64 and 128 μM) or LPS (0.125, 0.250, 0.500, 1.000 and 2.000 μg/ml) were added for another 24 h at 37°C. The wells received 10% CCK-8 reagent (cat. no. HY-K0301, MedChemExpress), which was then thoroughly mixed and incubated in the dark for 2 h. Finally, the optical density at 450 nm (OD₄₅₀) was assessed by SpectraMax Mini microplate reader (Molecular Devices, LLC) to calculate cell viability.

BV-2 cells were resuspended with PBS to prepare a single cell suspension with a concentration of 1.0×10⁷ cells/ml. Anti-CD86 (1:100, cat. no. MA5-52361), anti-CD11b (1:100, cat. no. PA5-79533) and anti-CD206 antibody (1:400, cat. no. 17-2061-82; all Invitrogen; Thermo Fisher Scientific, Inc.) were added for 30 min under dark conditions at 4°C. Samples were detected with BD FACSCalibur flow cytometer and the FlowJo software (v10.8l; both BD Biosciences) for analysis.

The viability of HT22 cells was assessed by Calcein AM/PI detection kit (cat. no. C2015M, Beyotime Biotechnology). HT22 cells were washed with PBS, mixed with Calcein AM/PI detection working solution (500 μl) and incubated at 37°C for 30 min in a dark environment. Following incubation, staining was observed by fluorescence microscope (cat. no. BZ-X810, Keyence Corporation). Live (green) and dead (red) cell counts were analyzed using ImageJ 1.54h software (National Institutes of Health).

HT22 cells were plated into 12-well plates (1.0×10⁵ cells/well, built-in sterile coverslip). Cells were exposed to 4% paraformaldehyde (cat. no. 441244, Sigma-Aldrich; Merck KGaA) to fix the cells for 30 min at 25°C. 0.3% Triton X-100 solution (cat. no. T824275, Shanghai Macklin Biochemical Co., Ltd.) was added at 25°C for 10 min to increase permeability. TUNEL detection solution (cat. no. C1086, Beyotime Biotechnology) was added for 60 min at 37°C in a dark environment, then rinsed three times with PBS. The cells were incubated with DAPI (1 μg/ml, cat. no. D669617, Shanghai Macklin Biochemical Co., Ltd.) at 25°C for 5 min in the dark to stain cell nuclei, followed by three additional rinses with PBS. Once sealed with Anti-fade Mounting Medium (cat. no. HY-K1047, MedChemExpress), samples were observed using a fluorescence microscope, and 5 non-overlapping random fields of view were captured per sample to assess the apoptosis rate using the ImageJ 1.54 h software.

BV2 cells were mixed with glutaraldehyde solution (2.5%, cat. no. G6257, Sigma-Aldrich; Merck KGaA) and fixed in the dark at 4°C for 4 h. After being washed twice with sterile PBS to remove glutaraldehyde, the samples were dehydrated with ethanol solution (30, 50, 70, 80, 90%) for 10 min each, then dehydrated with anhydrous ethanol. Finally, isoamyl acetate (cat. no. 79857, Sigma-Aldrich; Merck KGaA) was used to replace anhydrous ethanol twice, for 10 min each. After centrifugation at 1,000 × g, 4°C for 5 min, the liquid was mixed with the cell pellet, and the mixture was dropped onto a silicon wafer, followed by drying at 37°C. The dried sample was sputter-coated with gold for 150 sec and morphological observations and image capture were performed by scanning electron microscopy (SEM; cat. no. S-4800, Hitachi Ltd.).

BV2 cells were seeded into a 12-well plate at a density of 1.0×10⁵ cells/well (built-in sterile coverslip). Cells were exposed to 4% paraformaldehyde for 30 min at 25°C. Hoechst 33342 (cat. no. C1025) and Yo-Pro-1 staining solution (500 μl; cat. no. C1356S; both Beyotime Institute of Biotechnology) were added at 37°C for 30 min. Cells were rinsed with PBS and photographed under a fluorescence microscope.

Caspase-1 enzyme activity was detected using the Caspase-1 Activity Assay kit (cat. no. E-CK-A381, Wuhan Elabscience Biotechnology) according to the manufacturer's instructions. BV2 cells (~1×10⁶) were collected and washed with PBS. A total of 100 μl Cell Lysis Buffer was added in an ice bath for 30 min. The lysed samples were centrifuged at 8,000 × g for 10 min at 4°C and the protein concentration in the supernatant was assayed by the BCA Protein Quantification kit. A total of 50 μl lysate was added to 45 μl Reaction Buffer and 5 μl Ac-YVAD-pNA were added for 2 h at 37°C. OD₄₀₅ was measured using a microplate reader and enzyme activity was calculated.

Wild-type (WT) and mutant-type (MUT) dual luciferase reporter plasmids for the NF-κB 3' untranslated region were synthesized by Sangon Biotech Co., Ltd. NF-κB-WT and si-MFG-E8, NF-κB-WT and si-NC, NF-κB-MUT, si-MFG-E8, and NF-κB-MUT and si-MFG-E8, respectively, were co-transfected into BV2 cells using Lipofectamine 3000 as aforementioned. Cells were collected after 48 h and assayed for luciferase activity using the Dual-Lucy Assay kit (cat. no. RG027, Beyotime Biotechnology) as previously described (33). Firefly luciferase activity was normalized to Renilla luciferase activity.

A total of 63 male C57BL/6 mice (age, 3 months; weight, 21-24 g) was obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. and raised at 22°C, humidity of 45-50%, 12/12-h light/dark cycle and food and water ad libitum. The mice were randomly divided into Sham (n=9) and experimental group (n=54) according to the random number table method. According to the method described by Liu et al (34), the tMCAO model was established. The mice received anesthesia with 2% isoflurane, followed by 1.5% isoflurane to maintain anesthesia and the surgical area was sterilized with 75% ethanol and 10% iodophor. In the experimental group, a small incision was made 4 mm from the bifurcation of right common carotid artery (CCA), with a suture placed around the internal carotid artery (ICA). The 0.16-mm diameter suture (cat. no. 1618-50, Beijing Cinontech Co., Ltd.) was gently pushed with ophthalmic tweezers, reaching an insertion depth of ~20 mm. The suture at the distal end of ICA and CCA was fastened to fix the suture. The incision was sutured and disinfected and a 6 mm suture was retained in vitro for reperfusion. After the blood supply was blocked for 45 min, the suture was removed. Blood flow in the mouse brain was monitored using Laser Speckle Doppler Flowmetry (PeriCam PSI Z; Perimed AB) (35). The body temperature of the mice was maintained at 37.0±0.5°C using a homoeothermic blanket throughout the surgical process. Sham group mice were sutured and disinfected, but tMCAO was not performed. To reduce post-operative pain in mice, buprenorphine (0.1 mg/kg; cat. no. PHR8512; Sigma-Aldrich; Merck KGaA) was administered subcutaneously 15 min before and 12 h and once 24 h after surgery.

At 45 min after tMCAO, mice were injected intraperitoneally with normal saline or 50, 100 or 200 mg/kg PPF (n=9 mice/group) (36). In addition, for the tMCAO + 200 mg/kg PPF + si-MFG-E8 group or the tMCAO + 200 mg/kg PPF + si-NC group, mice were microinjected with si-MFG-E8 or si-NC (166.67 μM) intracerebrally 2 days prior to surgery and were injected intraperitoneally 45 min postoperatively with 200 mg/kg PPF (37). Mice were sacrificed 1 day after PPF administration. A double-blind method was used during the animal experiments; none of the researchers responsible for feeding the mice, behavioral scoring or histological analysis were aware of the grouping information.

At 24 h after tMCAO, Longa scoring method was used to determine whether the modeling was successful (38). The Longa score was determined as follows: 0, normal, no neurological deficit; 1, inability to fully extend the contralateral forepaw, mild neurological deficit; 2, turning to the paralyzed side when walking, moderate neurological deficit; 3, body tilting toward the paralyzed side when walking, severe neurological deficit; 4, inability to walk spontaneously, loss of consciousness. A score of 4 represents mice with irreversible severe cerebral ischemic damage or loss of consciousness, which prevented subsequent behavioral scoring (corner test). Therefore, a score of 1-3 points was judged as successful modeling.

At 24 h after tMCAO, the sensorimotor and postural asymmetry neurological functions of mice were detected by corner test (39). The mice were placed at the corner entrance between two plates connected at a 30° angle. Once the mouse reaches the corner, it will retreat and turn right or left. Each mouse was tested 20 times. If the mice did not retreat when turning around, the experiment was abandoned. The number of right and left turns was recorded.

Following behavioral testing, mice were sacrificed by intraperitoneal injection of sodium pentobarbital (100 mg/kg) for 1 min, with cardiac and respiratory arrest used to determine death. Mouse brain tissue was frozen at −80°C for 20 min and sliced into 2-mm sections. Subsequently, the tissue sections were exposed to TTC solution (2%, cat. no. C0652, Beyotime Institute of Biotechnology) preheated to 37°C and incubated in a dark environment for 15 min. Images were captured and the cerebral infarction volume was evaluated through ImageJ 1.54h software.

Brain tissue was washed with pre-cooled PBS, the surface liquid was dried by filter paper and immediately weighed. The weighed tissue was placed in a drying tube and placed in an oven at 100°C for 24 h to dry (until the weight was constant) and weighed after cooling to room temperature. Brain water content (%) was calculated as follows: (Wet weight-dry weight)/wet weight ×100%.

Brain tissue was exposed to 4% paraformaldehyde for 48 h at 4°C, paraffin-embedded and serially sectioned (4 μm). The sections were dried and deparaffinized using xylene, then rehydrated using gradient ethanol. Following 10 min staining with hematoxylin, the sections were exposed to differentiation solution (cat. no. C0161s, Beyotime Institute of Biotechnology) for 30 sec. Sections were dyed with 1% eosin (cat. no. C0109, Beyotime Institute of Biotechnology) for 60 sec, dehydrated with gradient ethanol, transparentized with xylene and covered with neutral balsam. The ischemic penumbra was evaluated using a light microscope (Leica GmbH).

Brain tissue was exposed to 4% paraformaldehyde for 48 h at 4°C, paraffin-embedded and serially sectioned (4 μm), and then placed in warm water to expand. The slices were placed in a drying oven at 65°C for 90 min, dewaxed with xylene for 20 min and rehydrated in gradient ethanol. The liquid was removed, and sections were exposed to Nissl staining solution (cat. no. C0117, Beyotime Institute of Biotechnology) for 10 min at 25°C. After rinsing with tap water, tissue was immersed in 95% ethanol for dehydration for 4 min, then xylene. Following sealing with neutral balsam, samples were visualized with an inverted light microscope.

As described by Liu et al (34), the FJC degenerated neuron staining kit (cat. no. TR-100-FJT, Wuhan Anjiekai Biological Medicine Technology Co., Ltd.) was used to evaluate the neuronal degeneration in mice. The frozen sections of mouse brain tissue were treated with NaOH/ethanol solution (1:10) for dehydration for 4 min at 25°C, then treated with potassium permanganate (1:10 dilution) for 10 min at 25°C to enhance reagent permeability. Sections were then stained with FJC solution (1:10) in darkness for 10 min at 25°C. Sections were sealed with Anti-fade mounting medium, observed by fluorescence microscope and the proportion of FJC-positive cells (degenerated neurons) was counted using ImageJ 1.54h software (National Institutes of Health).

LDH is present in the cytoplasm and cannot penetrate the intact cell membrane. Therefore, the detection of LDH release can evaluate cell membrane damage (40). The release of LDH from HT22 cells was assessed using Cytotoxicity LDH assay kit (cat. no. HY-K1090, MedChemExpress). HT22 cells were centrifuged at 1,000 × g for 5 min at 4°C, and 50 μl supernatant was mixed thoroughly with 50 μl LDH working solution. Following incubation in dark for 30 min at 25°C, 50 μl Stop Solution was applied and OD₄₉₀ value was detected immediately by microplate reader. In addition, the fresh brain tissue of mice was cut into pieces (~1 mm³), mixed with PBS, homogenized, centrifuged at 8,000 × g for 10 min at 4°C, and supernatant was mixed with LDH working solution. The LDH release in brain tissue was determined as aforementioned.

Brain tissue was exposed to 4% paraformaldehyde for 48 h at 4°C, paraffin-embedded and serially sectioned (4 μm). The paraffin sections of mouse brain tissue were dewaxed with xylene, hydrated in a descending anhydrous ethanol series. For antigen retrieval, sections were immersed in 0.01 M citrate buffer (pH 6.0) and heated in a microwave oven at 95°C for 15 min, then naturally cooled to 25°C and rinsed three times with PBS. Then, 0.3% Triton X-100 was dropped on the surface for 10 min and 5% bovine serum albumin (BSA; cat. no. ST023; Beyotime Biotechnology) was added for 1 h at 25°C. Sections were exposed overnight to primary antibodies against Ionized calcium binding adaptor molecule 1 (Iba1, cat. no. PA5-27436, 1:500) and CD86 (cat. no. PA5-114995, 1:100) at 4°C. The sections were rinsed in PBS and incubated with FITC- (cat. no. F-2765, 1:100) or Cy3-conjugated goat anti-rabbit IgG (cat. no. A10520, 1:100) in darkness for 1.5 h at 25°C. All antibodies were obtained from Invitrogen (Thermo Fisher Scientific, Inc.). Following sealing with Anti-fade Mounting Medium, sections were observed by a fluorescence microscope and the fluorescence intensity was quantified using ImageJ 1.54h software.

BV2 cells were seeded in a 12-well plate (1.0×10⁵ cells/well, coverslips were placed in the plate) for 24 h to make cell climbing slides. The slides were exposed to paraformaldehyde for 20 min, rinsed with PBS and then incubated with 0.3% Triton X-100 at 25°C for 10 min, blocked with BSA and exposed to primary antibodies against NLRP3 (cat. no. MA5-32255, 1:1,000) and apoptosis-associated speck-like protein containing a CARD (ASC, cat. no. PA5-50915, 1:200, Invitrogen; Thermo Fisher Scientific, Inc.) overnight at 4°C. Subsequent immunofluorescence analysis was performed as aforementioned.

Mouse TNF-α (cat. no. PT512), IL-10 (cat. no. PI522), IL-6 (cat. no. PI326), transforming growth factor-β (TGF-β; cat. no. PT878), IL-1β (cat. no. PI301), IL-18 (cat. no. PI553) and IFN-γ (cat. no.) PI508; all Beyotime Biotechnology) ELISA kits were used to evaluate inflammatory factors levels in mouse brain tissue and BV2 cells, according to the manufacturer's instructions. Mouse Claudin-5 (Cla; cat. no. CSB-EL005507MO; Cusabio Technology LLC), Occludin (Occ; cat. no. ml063481) and zona occludens 1 (ZO-1, cat. no. ml037693; both Shanghai Enzyme-linked Biotechnology Co., Ltd.) ELISA kits were used to evaluate the tight junction proteins in mouse brain tissue, according to the manufacturer's instructions. The target areas of brain tissue (ischemic penumbra and core area and non-ischemic area) were separated by a blade. The brain tissue homogenate or cell supernatant and corresponding antibody were added to the ELISA plate at 37°C for 90 min. HRP-labeled streptavidin was introduced and incubated in the dark at 37°C for 20 min. Chromogenic agent A and B were incubated for 15 min in a dark environment at 37°C. Finally, 50 μl termination solution was introduced and mixed well to determine OD₄₅₀ value.

The fresh brain tissue of the mice was cut into pieces and then RIPA lysis buffer (cat. no. 20-188, Sigma-Aldrich; Merck KGaA) was added. For BV2 cells, pre-cooled PBS was used for gentle washing twice and then RIPA lysis buffer was added for lysis. After lysis, the protein concentration was detected by the BCA protein quantification kit. A total of 30 μg protein/lane was loaded and separated by 10% SDS-PAGE, transferred onto PVDF membranes and blocked with 5% BSA at room temperature for 2 h. Following rinsing with TBST containing 0.1% Tween-20, the membrane underwent overnight incubation with primary antibodies against synapsin-1 (SYP-1, cat. no. 51-5200), inducible nitric oxide synthase (iNOS; cat. no. PA5-17106l; both 1:1,000), arginase 1 (Arg 1, cat. no. PA5-29645, 1:5,000), postsynaptic density protein-95 (PSD-95; cat. no. 51-6900), brain-derived neurotrophic factor (BDNF; cat. no. PA5-85730; both 1:500), growth-associated protein 43 (GAP-43; cat. no. PA5-34943, 1:5,000), GSDMD (cat. no. PA5-116815; all Invitrogen; Thermo Fisher Scientific, Inc.), cleaved-caspase-1 (cat. no. HY-P80622; both 1:500, MedChemExpress), pro-caspase-1 (cat. no. PA5-87536, 1:2,000, Invitrogen; Thermo Fisher Scientific, Inc.), ASC (cat. no. ab283684), GSDMD-N (cat. no. ab215203; both 1:1,000; both Abcam), NLRP3 (cat. no. MA5-32255, 1:500), IL-1β (cat. no. P420B, 1:1,000), IL-18 (cat. no. PA5-79479, 1:2,000), phosphorylated (p-)NF-κB-p65 (cat. no. 44-711G, 1:1,000), MFG-E8 (cat. no. PA5-109955; all Invitrogen; Thermo Fisher Scientific, Inc.) and NF-κB-p65 (cat. no. ab76311; both 1:2,000, Abcam) at 4°C. The membrane underwent incubation with goat anti-rabbit antibody (cat. no. 31460, 1:10,000, Invitrogen; Thermo Fisher Scientific, Inc.) for 2 h at 25°C. Subsequently, the ECL working solution (cat. no. HY-K1005, MedChemExpress) was prepared and uniformly dropped onto the membrane, which was scanned using iBright CL1500 gel imaging system (Invitrogen; Thermo Fisher Scientific, Inc.). After scanning, the gray value was determined with ImageJ 1.54h software and normalized to GAPDH (cat. no. PA1-987, 1:1,000, Invitrogen; Thermo Fisher Scientific, Inc.).

All data are presented as the mean ± standard deviation of ≥3 independent experimental repeats. SPSS 26.0 software (IBM Corp.) was employed for statistical analysis. P<0.05 was considered to indicate a statistically significant difference. Prism software (GraphPad 9.0; Dotmatics) was used for plotting. The results were analyzed via unpaired t-test or one-way ANOVA with post hoc Tukey's test.

The present study reoxygenated BV-2/HT22 cells in the presence of PPF following OGD. BV-2 and HT22 cell viability markedly declined following OGD/R, as indicated by CCK-8 assay (Fig. 1B). PPF (2-64 μM) increased cell viability, but viability decreased at 128 μM PPF (Fig. 1B). In subsequent experiments, cells were treated with 8, 16 and 32 μM PPF to evaluate its neuroprotective effects. Flow cytometry revealed that following OGD/R, CD86 (a marker of M1-type microglia) positivity increased in BV-2 cells, accompanied by a significant rise in CD206-positive cells (a marker of M2-type microglia) Following PPF treatment, the number of CD86-positive cells decreased significantly, while CD206-positive cells increased significantly, suggesting that PPF regulated the phenotypic conversion of BV-2 cells (Fig. 1C-E). Western blotting verified that PPF decreased iNOS (an M1-type marker) protein levels in BV-2 cells while upregulating Arg 1 (an M2-type marker), consistent with flow cytometry findings (Fig. 1F and G). OGD/R-treated HT22 cells were co-cultured with BV-2 cells to simulate neuroglial interactions. Calcein AM/PI dual staining demonstrated that HT22 cell viability decreased notably following OGD/R treatment. When HT22 cells were co-cultured with BV-2 cells, their survival rate decreased further. PPF treatment effectively reversed this trend, increasing the proportion of viable HT22 cells (Fig. 1H-J). TUNEL staining revealed that OGD/R treatment significantly increased apoptosis in HT22 cells, with further elevation following co-culture with BV-2 cells (Fig. 1K and L). PPF treatment significantly decreased apoptosis in HT22 cells (Fig. 1K and L). Following OGD/R, LDH release in HT22 cells significantly increased. Co-culture with BV-2 cells further elevated LDH release, indicating exacerbated cell damage. PPF treatment markedly decreased LDH release, suggesting its protective effect on HT22 cells (Fig. 1M). In addition, PPF increased the levels of SYP-1, PSD-95, BDNF and GAP-43 in HT22 cells, suggesting synaptic function maintenance and neural repair in HT22 cells (Fig. 1N and O). The results indicate that PPF suppressed OGD/R-induced excessive activation in BV-2 cells, mitigated damage to HT22 cells and upregulated the expression of neurofunction-related proteins.

SEM revealed that following OGD/R, BV2 cells exhibited typical pyroptotic features, such as marked cellular swelling, irregular morphology, membrane pores and cytoplasmic contents leakage. Following PPF, the aforementioned morphological abnormalities in BV2 cells were attenuated (Fig. 2A). The levels of pyroptosis were assessed using Yo-Pro-1 and Hoechst 33342 staining. Following OGD/R, Yo-Pro-1-positive staining in BV2 cells was markedly increased, while PPF treatment decreased the number of Yo-Pro-1-positive cells (Fig. 2B and C). Immunofluorescence analysis revealed increased fluorescence intensity of NLRP3 and ASC following OGD/R, whereas PPF decreased the fluorescence intensity of both proteins in BV2 cells (Fig. 2D-F). Caspase-1 activity was significantly elevated in BV2 cells following OGD/R, whereas PPF decreased its activity (Fig. 2G). Moreover, following OGD/R, the protein levels of cleaved-caspase-1/pro-caspase-1, GSDMD-N/GSDMD, NLRP3, IL-1β, ASC, and IL-18 were markedly elevated in BV2 cells. By contrast, PPF lowered protein levels, suggesting that PPF inhibited NLRP3 inflammasome activation (Fig. 2H-K). ELISA showed that the levels of IL-1β and IL-18 in BV2 cells were significantly elevated following OGD/R, and PPF treatment decreased the levels of both (Fig. 2L). These results indicated that PPF suppresses NLRP3 inflammasome activation following OGD/R, thereby inhibiting pyroptosis in BV2 cells.

Previous studies have revealed that MFG-E8 is key in microglial inflammation (18,41). To investigate the impact of MFG-E8 on NF-κB signaling and NLRP3 activation, OE-MFG-E8 was transfected into BV2 cells, followed by OGD/R. Following OGD/R, MFG-E8 expression decreased in BV2 cells, while p-NF-κB/NF-κB levels increased, suggesting that OGD/R may downregulate MFG-E8 and activate NF-κB signaling (Fig. 3A-C). Transfection of BV2 cells with OE-MFG-E8 significantly elevated MFG-E8 protein levels, confirming successful establishment of the model (Fig. 3D and E). Following transfection with OE-MFG-E8, p-NF-κB/NF-κB, NLRP3 and ASC expression notably declined in BV2 cells (Fig. 3F-H). The present study examined the effects of different concentrations of NF-κB activator LPS on the protein levels of the NF-κB pathway in microglia and found that 0.25-2.00 μg/ml LPS effectively reduced p-NF-κB/NF-κB ratio (Fig. S1A and B). Additionally, LPS at 0.125-1.000 μg/ml had no adverse effect on BV-2 cell viability (Fig. S1C), therefore 1 μg/ml LPS was selected for subsequent experiments. The p-NF-κB/NF-κB, NLRP3 and ASC levels were significantly elevated in BV-2 cells following LPS treatment (Fig. 3I-K). Overexpression of MFG-E8 significantly decreased p-NF-κB/NF-κB levels in OGD/R-treated BV2 cells, whereas LPS increased p-NF-κB/NF-κB levels (Fig. 3L and M). Following OGD/R, immunofluorescence indicated increased fluorescence intensity of ASC and NLRP3 in BV2 cells. OE-MFG-E8 decreased the fluorescence intensity of both proteins, whereas LPS attenuated the effect of OE-MFG-E8 (Fig. 3N-P). In addition, following MFG-E8 overexpression, cleaved-caspase-1/pro-caspase-1, NLRP3, ASC, IL-1β, GSDMD-N/GSDMD and IL-18 protein levels were reduced notably in OGD/R-treated BV2 cells. However, LPS inhibited the impact of OE-MFG-E8 overexpression (Fig. 3Q-T). The results indicated that OE-MFG-E8 suppresses NF-κB signaling and NLRP3 activation in OGD/R-induced BV2 cells.

The present study investigated whether PPF inhibits pyroptosis by regulating MFG-E8 in BV-2 cells. si-MFG-E8 was transfected into BV-2 cells, which were subjected to OGD/R and exposed to PPF for 24 h. PPF treatment significantly increased MFG-E8 expression in BV2 cells, while simultaneously decreasing p-NF-κB/NF-κB levels, suggesting that PPF may inhibit NF-κB activation by upregulating MFG-E8 (Fig. 4A-C). Following transfection with si-MFG-E8, MFG-E8 protein expression markedly declined in BV2 cells, indicating successful establishment of MFG-E8 silencing (Fig. 4D and E). Silencing MFG-E8 significantly increased the luciferase activity of NF-κB-WT but had no significant effect on NF-κB-MUT, confirming that NF-κB was a downstream target of MFG-E8 (Fig. 4F). si-MFG-E8 attenuated the impact of PPF, causing a notable rise in p-NF-κB/NF-κB levels. This indicated that the inhibitory impact of PPF on NF-κB activation was dependent on MFG-E8 (Fig. 4G and H). PPF treatment significantly reduced the number of Yo-Pro-1-positive BV2 cells following OGD/R, whereas MFG-E8 silencing led to an increase in Yo-Pro-1 positivity, indicating that si-MFG-E8 attenuated the inhibitory effect of PPF on pyroptosis (Fig. 4I and J). ELISA revealed that levels of proinflammatory factors TNF-α, IL-1β and IL-6 were markedly raised in BV2 cells following OGD/R, while IL-10 levels decreased. PPF treatment reversed this effect, whereas silencing MFG-E8 attenuated the anti-inflammatory effect of PPF (Fig. 4K-N). Additionally, PPF markedly decreased cleaved-caspase-1/pro-caspase-1, NLRP3, ASC, IL-1β, GSDMD-N/GSDMD and IL-18 levels in OGD/R-treated BV2 cells, whereas si-MFG-E8 markedly increased these protein levels (Fig. 4O-R). These results suggested that PPF suppresses pyroptosis by upregulating MFG-E8 in BV2 cells, thereby inhibiting NF-κB/NLRP3 pathway activation.

The present study established a tMCAO mouse model to investigate whether PPF effectively improves neuronal injury in vivo (Fig. 5A). The Longa scoring system results demonstrated that neurological deficit scores were significantly lower in the PPF-treated group compared with the tMCAO group, indicating that PPF improved neurological function in tMCAO mice (Fig. 5B). The corner test demonstrated that PPF effectively mitigated sensorimotor deficits in tMCAO mice, reduced turning bias and enhanced limb coordination and spatial perception (Fig. 5C). Moreover, PPF significantly reduced brain tissue water content in tMCAO mice, indicating that PPF alleviated ischemia-induced cerebral edema (Fig. 5D). TTC staining revealed well-defined white infarct lesions in the cerebral cortex of tMCAO mice. Following PPF treatment, the white infarct areas significantly decreased (Fig. 5E and F). HE staining revealed notable brain tissue damage in the penumbral region of tMCAO mice, characterized by disorganized cell arrangement and pronounced interstitial edema. PPF mitigated the extent of brain tissue damage (Fig. 5G). Nissl staining demonstrated a significant reduction in Nissl-positive cells within ischemic penumbra of tMCAO mice, indicating severe neuronal damage. PPF markedly increased the number of Nissl-positive cells (Fig. 5H and I). Additionally, the number of FJC-positive degenerating neurons in the ischemic penumbra of tMCAO mice was markedly elevated. Following PPF treatment, the number of FJC-positive cells markedly declined, indicating that PPF suppressed neuronal degeneration following ischemia (Fig. 5J and K). The results indicate that PPF improves neurological deficits in tMCAO mice and effectively mitigated pathological damage in brain tissue.

Immunofluorescence colocalization revealed a notably raised proportion of CD86⁺ Iba1⁺ cells in the ischemic penumbra of tMCAO mouse brains, suggesting a shift of microglia toward a proinflammatory phenotype (M1 type). Following PPF treatment, the proportion of CD86⁺ Iba1⁺ cells significantly decreased, indicating PPF suppressed proinflammatory activation of microglia in tMCAO mice after ischemia (Fig. 6A and B). In tMCAO mouse brain tissue, levels of proinflammatory factors (TNF-α, IFN-γ) were markedly raised in the ischemic core (Fig. 6C), penumbra (Fig. 6D) and non-ischemic regions (Fig. 6E), while anti-inflammatory factor (IL-10, TGF-β) levels notably decreased. PPF treatment reversed this phenomenon. LDH levels were markedly elevated in the ischemic penumbra of tMCAO mice, while PPF effectively decreased these levels, suggesting that PPF mitigated the extent of cellular damage (Fig. 6F). tMCAO mice exhibited significantly decreased levels of Cla, Occ, and ZO-1, whereas PPF markedly increased these tight junction protein levels, confirming its protective effect on the blood-brain barrier (BBB; Fig. 6G). Moreover, tMCAO mice exhibited significantly decreased levels of SYP-1 and PSD-95 in brain tissue and elevated levels of BDNF and GAP-43; PPF treatment reversed these effects (Fig. 6H and I). These results indicated that PPF alleviates neuronal injury by suppressing microglial activation in tMCAO mice, preserving BBB integrity and modulating the expression of proteins associated with neuronal function.

The present study investigated whether PPF exerts neuroprotective effects by regulating MFG-E8 expression, thereby identifying the key target of PPF intervention. MFG-E8 protein expression was notably decreased in tMCAO mice, whereas p-NF-κB/NF-κB levels were significantly elevated. Furthermore, NLRP3 inflammasome and pyroptosis-associated protein levels were markedly elevated. Following PPF treatment, MFG-E8 levels markedly elevated in the ischemic penumbra, while p-NF-κB/NF-κB and pyroptosis-associated protein levels significantly decreased. This indicated that PPF may suppress NF-κB/NLRP3 pathway activation and hinder pyroptosis by upregulating MFG-E8 (Fig. 7A-F). At 3 days after intracerebroventricular injection of si-MFG-E8, the expression of MFG-E8 protein in mouse brain tissue was significantly decreased (Fig. 7G and H). Additionally, MFG-E8 silencing markedly attenuated the protective impact of PPF on the ischemic penumbra in tMCAO mice. Compared with tMCAO + 200 mg/kg PPF +si NC group, p-NF-κB/NF-κB levels were elevated and cleaved-caspase-1/pro-caspase-1, NLRP3, ASC, IL-1β, GSDMD-N/GSDMD and IL-18 levels were markedly increased in tMCAO + 200 mg/kg PPF +si MFG-E8 group, confirming MFG-E8 as a key molecular target for PPF-mediated neuroprotective effects (Fig. 7I-N). These findings suggest that PPF suppressed abnormal activation of the NF-κB/NLRP3 pathway by upregulating MFG-E8, thereby decreasing pyroptosis. This may represent a key molecular mechanism underlying its neuroprotective effects.