Five-month-old male senescence-accelerated mouse strain P8 (SAMP8) mice weighing 45-55 g were purchased from Beijing Huafukang Biotechnology Co. (Chica) and transported to the animal housing facility at Beijing Stomatological Hospital. Mice were kept in standard cages under standard conditions (room temperature of 25°C, 40% humidity) with a 12 h light/dark cycle for 2 months. Mice had free access to food pellets and water. All animal study protocols followed the National Institutes of Health guidelines for the care and use of laboratory animals and were approved by the ethics committee of Capital Medical University (IACUC: AEEI-2022-026).

After 1 week of adaptation to the new environment, SAMP8 mice (n = 12) were randomly divided into two groups of six animals (S group [sham operation group] and TL group [bilateral maxillary tooth loss]). The mice were anesthetized via intraperitoneal injection of 10% chloral hydrate (400 mg/kg). All bilateral molars in the maxillary region were removed in the TL group. The S group was subjected to artificially caused gingival damage to the bilateral maxillary molars but no tooth extraction. Finally, each SAMP8 mouse was injected with buprenorphine (50 mg/g) intramuscularly to alleviate suffering during surgery. Behavioral tests, including the Y-maze and novel object recognition, were conducted 8 weeks after the animal model was established. Following the completion of behavioral experiments, the mice were immediately euthanized, and hippocampal tissue was collected.

The Y-maze consists of three arms of 30, 8, and 15 cm in length, width, and height, respectively, with an angle of 120° between the two arms. Mice were placed at the end of one of the three arms and the order in which they entered each arm within 5 min was recorded. Entry of the mice into each arm is standardized as complete entry of all limbs. Mice sequentially entering a different arm than the previous two were considered a correct alternation, and the percentage of spontaneous alternations was used to assess the memory capacity of the mice. Spontaneous alternations = (number of correct alternations) / (total number of arm entries − 2) × 100%. For spontaneous alternation experiments, behavioral testing with a cycle of 1 day can begin after 30 min of environmental adaptation.

The NOR experiment was divided into training and test phases, and a 3-day environmental acclimatization period was required before the training phase. In the training phase, two equivalent A objects were placed on opposite sides of the test box, with the objects about 10 cm away from the wall of the box; the mice were then placed in the test box and removed after 5 min of free exploration. During the test phase, an A object (familiar object) was randomly replaced by a B object (novel object) in the test box, and the position remained unchanged; the mouse was then placed in the test box and removed after 5 min of free exploration. During the test phase, the mice’s exploration times for familiar and novel objects were recorded separately, and the recognition index (RI) was calculated to assess the memory ability. The RI was calculated as follows: RI = (new object exploration time − old object exploration time)/(new object exploration time + old object exploration time) × 100%. For the NOR experiment, an adaptation period (3 days), training period (1 day), and testing period (1 day) are required.

Following behavioral experiments, SAMP8 mice were anesthetized by intravenous administration of pentobarbital sodium (15–40 mg/kg) and euthanized. Right and left brain tissues were subsequently extracted. Some were frozen at −80°C for subsequent molecular biology experiments, some were fixed in 2.5% glutaraldehyde stationary solution for transmission electron microscopy, and some were fixed in 4% paraformaldehyde for 24 h and then subjected to gradient alcohol dehydration. The brain tissue was then embedded into paraffin for subsequent histopathological testing.

The murine microglial BV-2 cell line (cat NO. CL-0493) was obtained from Wuhan Pricella Biotechnology Co., Ltd. (Wuhan, China). BV-2 cells were cultured in DMEM medium (11965092, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, 1693361, Gibco, USA), 100 U/mL penicillin, and 100  U/mL streptomycin. All cells were incubated in a humidified atmosphere of 5% CO2 at 37°C.

siRNA, against mouse STING and negative control, were designed and synthesized by Sangon Biotech (Shanghai, China). siRNA transfection was conducted according to the manufacturer’s instructions. Briefly, the transfection reagent Lipofectamine RNAiMAX (13778100, Invitrogen) and siRNA were mixed in Opti-MEM. The cells were then changed to normal medium and continued culturing for 24–48 h before following the experiments.

According to a standard nonbiotinylated protocol, the paraffin-embedded tissue sections were deparaffinized and rehydrated. The sections were then incubated overnight at 4°C with the primary antibodies, including rabbit anti-TFAM antibody (1:1000, A3173 ABclonal, China). Then, they were incubated for 20 min at 37°C with anti-rabbit IgG before being treated with diaminobenzidine (DAB) for approximately 1 min at room temperature. After the sections were sealed with neutral balsam (Biosharp, China), the sections were ready for light microscopy (Olympus/BX61, Japan). For the histomorphometric analysis, we selected three regions of interest (ROIs) in the hippocampus and measured the area of the ROI, the area of positive cells in the ROI, and the IOD (the area of positive cells in the ROI/ the area of the ROI).

Following the standard instructions, we performed immunofluorescence staining of paraffin-embedded tissue sections using the Treble-Fluorescence immunohistochemical mouse/rabbit kit (RS0035, Immunoway, USA). The sections were incubated overnight at 4°C with the primary antibody, including rabbit IBA-1 antibody (1:1000, ab178846, Abcam, USA), rabbit GFAP antibody (1:5000, ab7260, Abcam, USA), rabbit NeuN antibody (1:100, ab177487, Abcam, USA), rabbit STING/TMEM173 antibody (1:100, A3575 ABclonal, China) and mouse GSDMD antibody (1:200, 66387-1-lg, Proteintech, China).

BV2 cells were seeded on glass coverslips in 6-well plates. After various treatments, cells were fixed with 4% paraformaldehyde for 20 min at room temperature and then washed three times with PBS. After permeabilization with 0.1% Triton X-100/PBS for 15 min, cells were rinsed with PBS, blocked with PBS containing 3% BSA for 1 h at room temperature, and then incubated with mouse GSDMD (1:200, 66387-1-lg, Proteintech, China) primary antibodies at 4°C overnight. The subsequent steps are the same as for paraffin sections.

The paraffin-embedded tissue sections were deparaffinized and rehydrated. Consequently, the sections were stained with 1% Cresyl Violet staining solution (G1430, Nissl Stain Kit, Solarbio, China) in a 56°C thermostat for 1 h. The sections were dehydrated by fractional xylene and ethanol, and finally sealed with neutral gum, dried, and observed under an optical microscope for photography. Brain tissue was observed under a microscope after drying and analyzed via imaging.

Viable and non-viable cells were detected using Calcein/PI Cell Viability/ Cytotoxicity Assay Kit (C2015M, Shanghai Beyotime Biotechnology Co., Ltd.). The cells in each of the experimental groups were observed under a fluorescence microscope (Olympus/BX61, Japan).

We took 1 mm x 1 mm x 3 mm hippocampal CA1 mid-section tissue and fixed it with 2.5% glutaraldehyde fixative before making electron microscope sections. The sections were embedded in epoxy resin and double-stained with uranyl acetate and lead citrate on an ultrathin microtome. The microstructure of the hippocampal tissue was observed and photographed under a JEM1200EX transmission electron microscope.

Mouse hippocampal tissues and cells were lysed in Pierce RIPA Buffer was added to prevent protein degradation. The protein concentration of the lysate was measured using the Pierce BCA Protein Assay (23227, Thermo Fisher Scientific, USA) to normalize the 2′3′-cGAMP concentrations measured by ELISA according to the manufacturer’s instructions (EIAGAMP, Thermo Fisher Scientific, USA).

SAMP8 mice were sacrificed, and brain tissue was isolated from the hippocampus on ice. Then, the hippocampi were frozen in liquid nitrogen and stored at −80°C. As for bv2 cells, after blowing the adherent cells, the culture medium was aspirated and centrifuged to collect the cells for subsequent RNA extraction. According to the manufacturer’s instructions, total RNA was extracted with an Ultrapure RNA Kit (CWBIO, China). First-strand cDNA synthesis was performed using an All-in-one 1st Strand cDNA Synthesis SuperMix Kit (Novoprotein, China). The results were analyzed following the 2^−ΔΔCt method using β-actin as the control. MagicSYBR Mixture (CWBIO, China) with a Bio-Rad CFX Connect real-time system (CFX Connect Optics Module, Bio-Rad) was used. All RT–PCR experiments were repeated three times.

The primer sequences used were as follows:

Mouse GSDMD: 5′-CCATCGGCCTTTGAGAAAGTG-3′ and 5′-ACACATGAATAACGGGGTTTCC-3′; mouse cGAS: 5′-AGGAAGCCCTGCTGTAACACTTCT-3′ and 5′-AGCCAGCCTTGAATAGGTAGTCCT-3′; mouse STING: 5′-GGCGTCTGTATCCTGGAGTA-3′ and 5′-TAGACAATGAGGCGGCAGTTAT-3′. Mouse beta-actin (β-Actin): 5′-AGATTACTGCTCTGGCTCCTAGC-3′ and 5′-ACTCATCGTACTCCTGCTTGCT-3′.

As with RT-PCR, we performed Western blot analysis using SAMP8 mouse hippocampal tissue and BV2 cells, respectively. Proteins were extracted with an Applygen total protein extraction kit. The protein concentration was normalized using Coomassie brilliant blue G-250 staining. Equal amounts of proteins were separated by SDS–PAGE on a 10% polyacrylamide gel, and the proteins were transferred to PVDF membranes. After blocking with 0.1% TBST containing 5% nonfat milk at room temperature for 2 h. Primary antibodies against GSDMD (1:5000, 66387-1-lg, Proteintech, China), GSDMD-N (1:1000, YT7991 Immunoway, USA), cGAS (1:1000, A8335 ABclonal, China), STING/TMEM173 (1:2000, A3575 ABclonal, China), TBK1/NAK (1:1000, 3504 Cell Signaling Technology, USA), p-TBK1/NAK (1:1000, 5483 Cell Signaling Technology, USA), NLRP3 (1:1000, 15101 Cell Signaling Technology, USA), IL-18 (1:1000, 57058 Cell Signaling Technology, USA) and GAPDH (1:1000, 2118 Cell Signaling Technology, USA) were added for overnight incubation. The membrane was rinsed with 0.1% TBST three times for 10 min each and incubated with a secondary antibody (1:3000, 7074S, Cell Signaling Technology, USA) at room temperature for 1 h. Color development was performed using an ECL kit. Images were acquired using the Bio-Rad Imager and analyzed with ImageJ to measure the integrated optical density (IOD) values of specific bands.

Statistical analysis was performed using SPSS Statistics V17.0 software (SPSS Inc.). Normality was tested using the Shapiro–Wilk test. Statistical significance was determined using statistical tests detailed in both the figure legends and the Results section. These tests included the following: unpaired two-tailed t-test (when normally distributed by the Shapiro–Wilk test and the SDs are not different), unpaired two-tailed t tests with Welch’s correction (when normally distributed but SDs were unequal), and Mann–Whitney U test (when not normally distributed). The statistical significance level was 0.05. All graphs were made using GraphPad Prism 8.

After bilateral upper molars extraction, the body weight of SAMP8 mice in the TL and S groups decreased after surgical trauma but returned to preoperative levels after 2 weeks (Figures 1A,B). This shows that the mice can recover from the trauma of the surgery and return to their preoperative body weight. During the 2 months following tooth extraction, SAMP8 mice showed a steady increase in body weight (Figure 1B). This shows that the mice can adapt to tooth loss and continue to gain weight, which suggests that the mice can maintain a healthy weight despite tooth loss.

To explore the effects of loss of occlusal support on the behavioral performance of SAMP8 mice, we tested learning and memory abilities by Y-maze and NOR. The Y-maze was used to evaluate spatial working memory capacity in rodents, While NOR was used to assess their short-term memory capacity. The percentage of spontaneous alternation was significantly lower in the TL group of mice than in the S group (Figure 1C; p = 0.0022). At the same time, the recognition index of the TL group mice in the new object recognition experiment also significantly decreased (Figure 1D; p = 0.0019), suggesting cognitive dysfunction in SAMP8 mice after tooth extraction. In addition, neurons in the CA1 area of the hippocampus in the TL group were loosely arranged and there were fewer Nissl vesicles in the cytoplasm than in the control group (Figure 2E). Transmission electron microscopy results showed altered cell morphology after loss of occlusal support (Figure 2A).

Tooth loss induced eminently neuronal damage, as evidenced by Nissl staining (Figure 2E), TEM (Figure 2A), and behavioral assays (Figures 1C,D), demonstrating simultaneous increases in the levels of GSDMD when compared with the S groups. We examined the cellular pyroptosis target protein GSDMD by immunofluorescence staining, qRT-PCR, and western blot. The results showed that the GSDMD mRNA expression was significantly up-regulated in the TL group compared with the S group, accompanied by changes in protein levels (Figures 2B–D). Next, we performed colocalization staining for GSDMD/microglial marker IBA-1 (Figure 3A), neuronal marker NeuN (Figure 3B), and astrocytic marker GFAP (Figure 3C) in the hippocampus tissues of the TL group and S group. The results showed that GSDMD expression was elevated and active in all cell types in the TL group. This suggests that tooth loss leads to increased expression of the pyroptosis markers, which further exacerbates the inflammatory response in the brain. In addition, excessive pyroptosis may lead to a further decline in cognitive performance.

cGAS and STING expression in hippocampal microglia was assessed by immunofluorescence staining. As shown in Figures 4A,B. Immunofluorescent double staining showed that the cGAS and STING are more frequently expressed in the activated microglia/macrophages (IBA-1+) after loss of occlusal support. This also indicates that the number of cGAS- or STING-positive cells in hippocampal microglia was significantly higher in the TL group than in the S group. Then, the qRT-PCR and western blot were used to detect the mRNA and protein expressions, and the results were consistent with those of immunofluorescence staining (Figures 4C–E). As shown in Figure 5F, this figure illustrates the classic activation process of the cGAS-STING signaling pathway. cGAS, activated by dsDNA, catalyzes the production of cyclic GMP-AMP (cGAMP). As a second messenger, cGAMP binds to the stimulator of interferon genes (STING), thereby activating STING. Activated STING recruits and activates TANK-binding kinase 1 (TBK1), which plays a key role in immune responses. In this study, the decrease of TFAM expression in the hippocampus following tooth loss indicates the potential presence of mitochondrial impairment and excessive accumulation of mtDNA (Figures 5A,B). Additionally, studies have found that, compared to the S group, the levels of TBK1, p-TBK1, and cGAMP in the hippocampus were significantly higher in the TL group. This suggests that the cGAS-STING pathway is more active after tooth loss (Figures 5C–E). Excessive activation of the cGAS-STING axis is highly likely to originate from the abnormal accumulation of mtDNA following loss of occlusal support.

To further observe the cGAS-STING signaling pathway and pyroptosis, the BV2 cells model was used in vitro (Figure 6A). As shown in Figure 6B, the qRT-PCR results showed that the use of STING siRNA efficiently reduced the expression of STING. Additionally, we also demonstrated that the protein levels of STING, p-TBK1, NLRP3, GSDMD, GSDMD-N, and IL-18 were markedly reduced in comparison with the si-NC groups (Figures 6C,D, 7A,B). Concurrently, GSDMD fluorescence staining results in BV2 cells showed the same trend (Figures 7C,D). What’s more, the delivery of STING siRNA also remarkably restrained cell apoptosis to some extent and decreased the counts of dying BV2 cells in the si-STING groups, as demonstrated by Calcein/PI and immunofluorescence staining (Figure 6E). The above results suggest that tooth loss can induce hippocampal microglia to exhibit pyroptotic features, and this process may be involved in pathological damage, which is possibly mediated by the increased expression of GSDMD-N and the increased number of inflammatory factor-positive cells. In addition, the cGAS-STING signaling pathway plays a crucial role in hippocampal microglia pyroptosis, which may contribute to the cognitive deficits observed in the TL group.