Animals

C57BL/6 female mice, aged 6 to 8 weeks, were sourced from the Experimental Animal Center of Guizhou Medical University. Mice were housed in a specific-pathogen-free environment at standard temperature (18–22 °C), moderate humidity (50–55%), and under a 12 h light/dark cycle. All mice were acclimated in the animal room for 1 week before being used in experiments. All protocols for animal experiments were approved by the Experimental Animal Ethics Committee of Guizhou Medical University (No. 2201201).

Murine CRS model and related experiments

The research was divided into three separate experiments.

In the first experiment, mice were assigned to four groups, each consisting of 10 mice: Control (Ctrl), CRS-3 m, CRS-4 m, and CRS-5 m. The experimental schedule is shown in Fig. 1A. A mouse model of CRS was created following an established method (Kim et al. 2011). In short, experimental group mice were sensitized by intraperitoneal (i.p.) injections of ovalbumin (25 mg, OVA, Sigma-Aldrich) in 200 µl of PBS with aluminum hydroxide (2 mg) on days 0 and 5. Subsequently, they underwent intranasal (i.n.) challenges with 20 µl of a 3% OVA solution for 13 weeks continuously. During the final 8 weeks, mice in the CRS-4 m and CRS-5 m groups (12 and 16 weeks respectively) were challenged with staphylococcal enterotoxin B (10 ng SEB, Toxin Technology, USA) and OVA via intranasal administration. The Ctrl group received PBS throughout the experiment for comparative purposes.

Fig. 1

Varying inflammation in established CRS mouse models. (A) Experimental procedure timeline. (B) HE and PAS staining of nasal mucosa from Ctrl and CRS groups (×400). (C) Eosinophil count in nasal mucosa. (D) Goblet cell count in nasal mucosa. (E) Behavioral experiment schematic. (F-H) Staining, eosinophil, and goblet cell counts for additional mouse groups (×400). Red arrows indicate eosinophils; blue arrows indicate goblet cells. Data are presented as mean ± SD. **P < 0.01, ***P < 0.001

For the second experiment, the mice were assigned to the subsequent groups (n = 10 per group): Ctrl, CRS-3 m, CRS-3 m + siRNA-HMGB1, CRS-3 m + siRNA-NC. The experiment schedule is shown in Fig. 4A. The CRS-3 m + siRNA-HMGB1 and CRS-3 m + siRNA-NC groups followed a treatment process similar to the CRS-3 m group, with a key difference occurring on the 103rd day of the study. On this day, mice in both groups underwent stereotactic brain injections: the CRS-3 m + siRNA-HMGB1 group received siRNA-HMGB1, while the CRS-3 m + siRNA-NC group received siRNA-NC (negative control siRNA).

In the third experiment, the mice were allocated to the subsequent groups (n = 10 per group): Ctrl, CRS, CRS + Met, CRS + Saline. The experiment schedule is shown in Fig. 7A. After establishing the CRS model, the CRS + Met group received daily intraperitoneal injections of a metformin solution (200 mg/kg) for one week, while the CRS + Saline group received equivalent injections of saline solution.

After the completion of modeling and related interventions, mice underwent behavioral testing and PET experiments, followed by euthanasia and tissue collection.

Stereotactic injection

After the establishment of the CRS mouse model, we used a mouse brain atlas to determine the location of the hippocampal region based on coordinates: 2.3 to 2.4 millimeters posterior from the suture, 1.5 to 1.6 millimeters lateral, and 2.0 millimeters in depth. Isoflurane was used to anesthetize the mice, after which the scalp was incised and a hole was drilled. Subsequently, we used a microsyringe to inject 2 microliters (concentration of 2.5 mM/L) of siRNA-HMGB1 or siRNA-NC (Genechem, China) into each side to achieve bilateral injections. After the injection, the incision is sutured. These are the primer sequences for siRNA-HMGB1, 5’ - UCUUGACCACAGAUCUUAATT − 3’ (forward) and 5’ - AGCCTTGTGTGTTTCTGCG − 3’ (reverse); for siRNA-NC, 5’ - CCAAGAACTTCCAGAACATAT − 3’ (forward) and 5’ - ATATGTTCTGGAAGTTCTTGG − 3’ (reverse).

Behavioral experimentElevated plus maze

Comprising two enclosed arms and two open arms, each with dimensions of 30 cm long, 6 cm wide, and 15 cm tall, the apparatus is utilized to gauge the subjects’ anxiety levels. Animals are introduced to the maze and permitted to explore for a duration of 5 min, during which their duration of stay in both the enclosed and open arms is meticulously recorded.

Open field test

A simple apparatus, measuring 44 cm × 44 cm × 30 cm, was employed to evaluate the movement and anxiety-related behavior of the subjects. Following a brief 30-second acclimatization period, the total distance covered and the duration spent in the central area (14.7 cm × 14.7 cm) were measured over a 5-minute observation period.

Tail suspension test

Mice were subjected to a tail suspension test, where they were secured by tape 1 cm from the tail’s tip, suspended 15 cm above the ground. To prevent climbing, small plastic tubes were secured around their tails. The duration of the test was set at 6 min, with immobility being quantified and analyzed during the last 4 min of the session.

Forced swim test

Individual mice were placed into separate clear glass tubes, each 6 cm in diameter and 30 cm tall, containing water to a height of 18 cm, with the temperature maintained at approximately 25 °C. The rodents were observed within these tubes for a duration of 6 min, with the period of immobility being quantified in the final 4 min of the observation.

Sucrose preference test

Prior to the experiment, the mice were accustomed to two matching water bottles for a day. Subsequently, each mouse underwent a 24-hour period of water and food deprivation, after which they were given access to a 2% sucrose solution in water. After 24 h, the bottles were weighed, and sucrose preference was determined by calculating the ratio of sucrose water intake to the total fluid consumption (sucrose water intake + pure water intake) and then multiplying by 100%.

Mouse PET experiment

Prior to PET imaging, the mice were fasted for 12 h to ensure the accuracy of the results. Subsequently, they received an injection of 200µCi 18F-FDG into the tail vein to track glucose metabolism. After injection, mice were kept under 1.5% isoflurane anesthesia for 1 h to maintain stability. Positioning followed for proper scanning posture. A static scan was performed using the InliView-3000B scanner for 10 min to obtain clear PET images, followed by a CT scan for anatomical details. CT settings were 50 kV, 0.5 mA, and 80ms exposure time. PET images were reconstructed in ListMode with a 140 × 140 × 140 matrix, 40 iterations, and 0.8 mm slice thickness. CT images were reconstructed using the FDK algorithm with a 512 × 512 × 512 matrix, 0.1367 pixel size, and 0.18 mm slice thickness. PMOD software was used to process images, quantify brain region values, and calculate standardized uptake values (SUVs) from the PET images.

Histopathological analysis

Tissues from mice, comprising brains and noses, were immersed in 4% paraformaldehyde for a period of 48 h. Brains were dehydrated through a graded alcohol series (75%, 80%, 95%, 100%), cleared in dimethylbenzene, paraffin-embedded, and sectioned. Sections were then dewaxed, deparaffinized, and stained with Nissl and hematoxylin-eosin (HE). The stained sections were dehydrated, cleared, and examined using an optical microscope. Nose tissues, after fixation, were decalcified in ethylenediaminetetraacetic acid for one week, paraffin-embedded, and sectioned at 4 μm. These sections were stained with HE and periodic acid-Schiff (PAS). Eosinophil and goblet cell counts were assessed in three randomly selected high-power fields (400×) under a microscope.

Immunofluorescence assay

Hippocampus tissues were lysed in RIPA buffer with phenylmethylsulfonyl fluoride and protease inhibitors. Post centrifugation at 12,000 rpm at 4 °C for 15 min, supernatants were harvested. Hippocampus tissues were preserved in 4% paraformaldehyde for 48 h, then embedded in paraffin and cut into 5 μm sections. After blocking with 10% goat serum for 30 min, the sections were incubated with primary antibodies (IBA-1, HMGB1, Notch1, Hes-1; all 1:500 dilution from Abcam) overnight at 4 °C. The next day, they were incubated with anti-mouse secondary antibodies (ThermoFisher, 1:800) for 1 h at room temperature, rinsed with PBS, and stained with DAPI. Confocal microscopy was used to capture fluorescent images, and three randomly selected high-power fields (400×) per section were evaluated by two independent, blinded observers. Mean fluorescence intensity of target proteins was quantified according to the guidelines using ImageJ software, version 1.52a (NIH, USA).

Cell culture and intervention

Magnetic bead sorting preserves the in vivo status of microglia and reflects their relationship with neuroinflammation. As documented in references (Nikodemova and Watters 2012; Liu et al. 2021), the isolation of microglia from the hippocampus was achieved after the establishment of the model. The hippocampal tissue was enzymatically digested, filtered through 70-µm cell strainers to obtain a single-cell suspension, and microglia were sorted using CD11b magnetic beads (Miltenyi Biotec, Germany). After resuspension, loading, and washing, the positive cells were collected from the sorting column. Then the cells were resuspended in medium containing 10% fetal bovine serum (Gibco, USA), plated on polylysine-coated culture plates, and cultured in a humidified 5%CO2 incubator at 37 °C.

Firstly, the cells were divided into four groups: Ctrl, CRS, CRS + siRNA-HMGB1, and CRS + siRNA-NC. Microglia were extracted from the hippocampus of control mice for the Ctrl group and CRS mice for the other groups. The Ctrl and CRS groups were untreated, while the CRS + siRNA-HMGB1 and CRS + siRNA-NC groups were transfected with their respective siRNAs under standard culture conditions.

In the second experiment, the groups were assigned to receive CRS, CRS + low-dose metformin (CRS + Met-L), CRS + medium-dose metformin (CRS + Met-M), or CRS + high-dose metformin (CRS + Met-H). Metformin concentrations were 0.02, 0.2, and 2 mmol/L for low, medium, and high doses, respectively, with consistent culture conditions. After the experiment, supernatants and adherent cells were collected from each group for analysis.

In the third experiment, the cell overexpression studies included CRS + Met, CRS + Met + OE-HMGB1 (HMGB1 overexpression), and CRS + Met + OE-NC (empty vector overexpression). Microglia from CRS mice were cultured with 0.2 mmol/L metformin. The CRS + Met group received only metformin, while the CRS + Met + OE-HMGB1 and CRS + Met + OE-NC groups were transfected with the HMGB1 plasmid (pcDNA-HMGB1) and empty vector (pcDNA-NC), respectively, under some culture conditions.

siRNA and plasmid transfection

Microglia were planted into 6-well plates at a concentration of 5 × 105 per well, ensuring even distribution. Gene silencing procedures were initiated once the cell confluence reached approximately 70–90%, as observed under a microscope. siRNAs (against HMGB1,5’-UCUUGACCACAGAUCUUAATT-3’, Genechem, China) or plasmids were mixed with Lipofectamine 2000 (Invitrogen, USA) in serum-free medium (Gibco, USA). The medium was replaced for the cells in the plate during the resting period. During the resting phase, the culture medium for the plated cells was refreshed. Post a 15-minute interval, 200 µL of the solution was dispensed into each well, and the 6-well plates were marked, gently agitated, and returned to the incubator. Following a 24-hour period, the cell density and condition were assessed. Subsequently, 1 mL of complete growth medium was added to each well for cell nourishment and incubation.

Western blot

Hippocampus tissues and microglial cells were processed in RIPA buffer, supplemented with protease and phosphatase inhibitors, to isolate total proteins. The protein concentration was determined using the BCA protein ssay kit (Solarbio, China). An equal amount of protein lysates (40 µg) was loaded into 10% SDS polyacrylamide, transferred onto a PVDF membrane, and then blocked in 5% non-fat milk for 1 h. The membranes were incubated overnight at 4 ℃ with the following primary antibodies: HMGB1 (1:1000, Abcam), Notch1 (1:1000, Abcam), Hes-1 (1:1000, Abcam) and GAPDH (1:1000, Abcam). The membrane was incubated with secondary goat anti-rabbit antibodies (1:3000, Abcam) for 1 h at room temperature after washing with TBS and 0.05% Tween-20. Protein levels were normalized to GAPDH and quantified using ImageJ software, version 1.52a (NIH, USA).

Quantitative real-time PCR analysis

Total cellular RNA from the samples was isolated using RNA isolation reagent (Servicebio, China), following the manufacturer’s instructions. The concentration and purity of the RNA were assessed using a spectrophotometer. A quantity of 1 µg of RNA was then reverse-transcribed into cDNA using the PrimeScript™ RT reagent Kit with gDNA Eraser (Takara, China), which served as a template for PCR. Quantitative real-time PCR was performed using the TB Green® Premix Ex Taq™ II (Takara, China) with specific primers according to the manufacturer’s protocol. The primer sequences are as follows:

HMGB1: forward, 5’- CACCGTGGGACTATTAGGAT − 3’; reverse, 5’- GCTCACACTTTTGGGGATAC − 3’; Notch1: forward, 5’- TGCCAGTATGATGTGGATGAG − 3’; reverse, 5’- GGTCCCTGTGTAACCTTCTGT − 3’; Hes-1: forward, 5’- AGCCCACCTCTCTCTTCTGAC − 3’; reverse, 5’- AGGCGCAATCCAATATGAAC − 3’; GAPDH: forward, 5’- GCCTCGTCTCATAGACAAGATG − 3’; reverse, 5’- CAGTAGACTCCACGACATAC − 3’. GAPDH served as an endogenous control. Relative levels of mRNA were presented as the relative fold change and calculated using the 2−△△CT method.

Enzyme linked immunosorbent assay (ELISA)

After drug treatments, hippocampus tissues and culture media were collected from the mice. Tissues from both hippocampi were homogenized, and 200 µL of the lysate was added. The solution was chilled on ice for 30 min prior to centrifugation at 12,000 rpm for a duration of 5 min. The supernatants from both the lysate and the culture media were retained for analysis. ELISA kits specific for mice (Servicebio, China) were used to determine the levels of TNF-α, interleukin (IL)-1β, and IL-6, following the manufacturer’s instructions.

Statistical analysis

All experimental data are expressed as mean ± standard deviation. To determine statistical differences between pairs of groups, Student’s t-test was applied. For comparisons involving more than two groups, one-way analysis of variance was conducted, with subsequent analysis by Bonferroni’s post hoc test. Statistical significance was set at a p-value of less than 0.05. The software GraphPad Prism 8.0 (GraphPad Software, USA) was utilized for all statistical analyses.