Male mice 12 months of age were used in this study referred to previous published papers and the groups of mice were randomly assigned (Wen et al., 2020). C57BL/6J mice (male, 12 months old) were obtained from Medical Animal Experiment Center of Nanchang University. They were bred and maintained in the Animal Resource Center of the Faculty of Nanchang University. They had free access to food and water in a room with an ambient temperature of 22 ± 2°C and a 12:12–h light/dark cycle. They were fasted for 12 h before surgery. Neonatal mice were used within 3 days. All animal procedures were performed in strict accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, the United States) and the guideline of the Institutional Animal Care and Use Committee of Nanchang University.

Exploratory laparotomy was performed to establish a postoperative delirium and PND model under aseptic conditions with isoflurane anesthesia, as described previously (Peng et al., 2016; Wen et al., 2020). Briefly, the mice were anesthetized with 1.8% isoflurane and oxygen at 2 L/min in an induction chamber for 30 min and then fixed supine on the operating table (Zuo et al., 2020). After disinfection of the abdomen, a sterile hole towel was spread. An incision of about 1.5 cm was made in the middle of the abdomen. The spleen, stomach, and duodenal organs were treated with caution following the principles of sterility to avoid damaging the organs of mice. The surgery was performed for about 30 min. The surgical incision was sterilized with iodophor and sutured. The mice were placed in a temperature-stable environment. After the mice woke up, they were placed in their original cages. DEX was administered (10, 20, and 30 μg/kg, intraperitoneally) once a day for 3 days before surgery. An incision of about 1.5 cm was made in the middle of the abdomen and stitched as soon as possible in the sham control group.

LPS (L2630), 3-MA (M-9281), Dexmedetomidine (SML-0956), MG-132 (C-2211), and rapamycin (R8781) were obtained from Sigma–Aldrich. ATP (S1985), Bafilomycin A1 (S1413), MCC950 (S7809), and LY294002 (S1105) were purchased from Selleck. Green fluorescent protein (GFP)-red fluorescent protein (RFP) MAP1LC3B double fluorescent adenovirus was obtained from HANBIO (HB-AP2100001, Shanghai, China). Hoechst (33342) was obtained from ThermoFisher. The primary antibodies for immunoblotting detection were as follows: NLRP1 (Cell Signaling Technology, 4990S, 1:1,000); NALP2 (Abcam, ab36850.1:1,000); NLRP3 (AdipoGen, AG-20B-0014-C100,1:1,000); NLRC4 (Abcam, ab201792, 1:1,000); AIM-2 (Santa Cruz, sc-137967, 1:1,000); Caspase1 (AdipoGen, AG-20B-0042, 1:1,000), IL-1β (R&D,AF-401-NA, 1:1,000), PSD-95 (Cell Signaling Technology, 3450, 1:1,000), ionized calcium-binding adaptor molecule-1 (IBA-1) (Wako, 019-19741, 1:1,000), BCL2-associated X (BAX) (Cell Signaling Technology, 14796, 1:1,000), BCL-2 (Cell Signaling Technology, 3498, 1:1,000), apoptosis-associated speck-like protein (ASC) (Santa Cruz, sc-22514, 1:1000), MAP1LC3B (Cell Signaling Technology, 2775, 1:1,000), Sequestosome 1 (SQSTM1) (Cell Signaling Technology, 5114s, 1:1,000), Autophagy related 5 (ATG5) (Signalway Antibody, 44254-4, 1:1,000), ATG7 (Signalway Antibody, 38148, 1:1,000), Ubiquitin (Abcam, ab7780, 1:1,000), and GAPDH (Santa Cruz, sc-32233,1:3,000). The secondary antibodies were obtained from Invitrogen (A16124, 32230, and A11058,1:3,000). Specific primary antibodies for immunofluorescence staining were as follows: glial fibrillary acidic protein (GFAP) (Abcam, ab7260, 1:1,000) and MAP-2 (Abcam, ab5392,1:500). The information on NLRP3 (1:500), Ubiquitin (1:500), and IBA-1 (1:500) antibodies was consistent with previous information. Fluorescence-specific secondary antibodies were obtained from Invitrogen (A11008 and A21422,1:1,000). The antibodies used for immunohistochemical staining were the same as those for immunoblotting.

The cells and the hippocampal tissues were lyzed in the RIPA lysis buffer (Cell Signaling Technology) plus protease inhibitor (Halt Protease Inhibitor Single-Use Cocktail, Thermo Fisher Scientific). The protein content was detected using the BCA Protein Assay Kit (Beyotime). The proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to a PVDF (Millipore, IPVH00010) membrane activated with methanol, After blocking with 5% milk at room temperature for 1 h, PVDF membranes were incubated with various specific primary antibodies as described earlier in antibody dilution (NCM Universal Antibody Diluent, New Cell and Molecular Biotech Co., Ltd.) at 4°C overnight. The PVDF membrane was washed with TBST containing Tris-HCl, NaCl, and Tween 20, and incubated with HRP-conjugated secondary antibody. After incubation for 1 h at room temperature, the proteins were visualized and detected using BeyoECL Moon (Beyotime) and analyzed with an ImageQuant LAS 4000 imaging system (GE Healthcare, PA, United States).

The brain slices were washed three times with phosphate-buffered saline (PBS) and then incubated with 0.3% Triton X-100 in PBS supplemented with 5% bovine serum albumin (BSA) for 1 h at room temperature. Then, the brain slices were incubated with the specific primary antibody at 4°C overnight. After incubating with the fluorescent secondary antibody at room temperature the next day, the brain slices were placed on the slide and observed under a stereomicroscope (Olympus, Tokyo, Japan).

The hippocampal tissue slices were washed three times with PBS, followed by 3% H₂O₂, for 15 min to quench the endogenous peroxidase activity, and then incubated with 0.3% Triton X-100 in PBS supplemented with 5% BSA for 1 h at room temperature. The brain slices were incubated with the specific primary antibody at 4°C overnight. After incubating the slices with the secondary antibody the next day, they were incubated with diaminobenzidine for 5 min. PBS was used to stop the diaminobenzidine color reaction. The brain slices were dehydrated in alcohol and xylene and observed under a stereomicroscope (Olympus). The proportion of positive area was calculated using ImageJ software.

Experimental method referring to previous studies (Li et al., 2021), after the mice were decapitated under deep anesthesia, the brains were quickly removed and washed with distilled water. The hippocampal tissue was immersed in the A and B impregnation solution (#PK401, FD NeuroTechnologies, MD, United States) and stored at room temperature for 14 days in the dark. Next, the brain tissue was transferred to the C solution and stored at room temperature in the dark for 7 days. Finally, it was embedded with the OCT glue to make a brain slice of about 100–200 μm, which was pasted on a glass slide containing C solution, treated with gelatin, and stored in the dark. After dyeing with D and E solution, the slices were dehydrated in alcohol and xylene and observed under a stereomicroscope (Olympus). The pictures taken were analyzed with the NeuroJ plugin in ImageJ software.

New object recognition experiments are used to evaluate the ability of mice to recognize new objects in the environment. First, two identical objects were put in an autonomous mobile box (40 × 27 × 20 cm³), close to the sides of the box, and then the experimental animals were put into the box. The mice were allowed to explore freely for 5 min, and the time spent exploring each object was recorded (exploration was defined as sniffing or touching the object in a radius of 0–2 cm with its nose). After 5 min, one object was replaced with a novel one, the mice were allowed to explore for another 5 min, and the time spent by the animals exploring the novel and old objects was recorded.

The hippocampus of neonatal mice aged 1–3 days were removed and separated from the meninges and basal ganglia. Tissues were dissociated with 0.25% trypase (Amresco, OH, United States) at 37°C for 10 min and terminated using Dulbecco's Modified Eagle Medium/F-12 medium (Gibco, Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) (Gibco, Thermo Fisher Scientific) and 1% penicillin/streptomycin. The cells were plated on 25 cm² flasks (Thermo Fisher Scientific) at a density of 3 × 10⁶ cells/mL. After 1 day, the culture medium was replaced with the fresh medium every 3 days, for about 10–12 days. The cells were shaken vigorously, and the culture medium was aspirated and transferred to a centrifuge tube. The culture medium was centrifuged at room temperature at 1,500 rpm for 10 min. The resuspended microglia were plated at a density of 5 × 10⁵ cells/mL in a 12-well plate pre-plated with polylysine.

Primary microglial cells were seeded in a 96-well plate. After the microglia were iron-walled overnight, different concentrations of DEX were added. Three replicates of each group were used, and 10 μl of cell counting kit 8 reagent (Bimake) was added to each well. Next, the cells were incubated for another 2 h until the color turned orange. The 96-well plate was placed in a full-wavelength microplate reader (Thermo Fisher Scientific), and light absorption at 450 nm was detected. The degree of cell proliferation and toxicity were calculated.

Cell culture was performed referring to literature concerned (Wei et al., 2020).16 day pregnant mice were sacrificed and E16 embryos were dissected out in 1X PBS, and the meninges and basal ganglia were removed under a microscope. Tissues were digested with 0.025% trypsin for 20 min at 37°C and terminated with 10% FBS. The cell suspension was passed through a 40 μm cell sieve and centrifuged for 5 min at 1000 g. The cell pellet was resuspended in neurobasal medium (Gibco, Thermo Fisher Scientific) supplemented with 2% B-27 (Gibco, Thermo Fisher Scientific) and 0.5 mM L-glutamic acid. The cells were seeded on a plate pre-coated with poly-D-lysine hydrobromide (Sigma–Aldrich). The neuronal cell culture could be used for experiments until 7 days, and fresh media needed to be replaced every 3 days.

The cell slides in the 24-well plate were washed three times with 0.1 M PBS and fixed with 4% paraformaldehyde for 15 min. The paraformaldehyde was aspirated, washed with PBS, and blocked with PBS containing 5% BSA for 1 h. Then, the cell slide was incubated with the specific primary antibody at 4°C overnight, and the fluorescent secondary antibody was incubated for 1 h at room temperature the next day. After aspirating the secondary antibody and covering the cell slide with a glass slide, the cells were observed under the stereomicroscope (Olympus).

The apoptosis of cells was detected using an AV-PI kit, following the manufacturer’s protocols (Vazyme, Annexin V–FITC/PI Apoptosis Detection Kit). The cells were incubated with AV-PI for 10 min, washed with PBS, and then analyzed using a Guava easyCyte System 8 (Millipore 25801, CA, United States).

The collected 1 ml of the cell supernatant was centrifuged at 13,000 g for 5 min, and 600 µl of it was carefully transferred into another EP tube. An equal volume of pre-cooled methanol and one-fourth volume of chloroform were added to the supernatant, quickly vortexed, and shaken. Then, the supernatant was centrifuged at 13,000 rpm for 5 min. The supernatant was discarded carefully. Subsequently, 500 µl of pre-cooled methanol was added to the pellet. The liquid in the EP tube was quickly shaken and centrifuged at 13,000 rpm for 5 min, and the supernatant was discarded carefully. The centrifuge tubes with cell pellet were placed in an oven at 37°C for 5 min, and then 40 µl of 2.5 × loading buffer was added to dissolve the protein. After boiling at 95°C for 5 min, the precipitate was used for immunoblotting.

Mouse primary microglia were lyzed in the buffer (Cell Signaling Technology). The proteins were immunoprecipitated with specific primary antibodies (1 μg antibody per 100 μg of total protein), followed by incubation with G Agarose (Santa Cruz Biotechnology) at 4°C overnight. Proteins were eluted from the beads prepared for Western blot analysis.

The primary microglia were transfected with a tandem fluorescent mRFP- GFP-MAP1LC3B adenovirus (HanBio, HB-AP2100001) following the manufacturer’s protocol. The GFP green fluorescence and RFP red fluorescence of cells indicated the hindrance and smoothness of the autophagy process. Six fields of view in each group of cells were chosen for detection. After the cells were processed, they were fixed with 4% polymethanol, rinsed with PBS, stained with Hoechst, and photographed. The ratio of green fluorescent dots to red fluorescent dots was calculated to reflect the process of autophagy.

All data were expressed as mean ± SEM and analyzed in a blind way using the Student t test, one-way analysis of variance (ANOVA), or two-way ANOVA. In all studies, n represents the number of samples per group. A p value < 0.05 indicated a statistically significant difference.

Neuroinflammation is one of the main pathological factors for most neurodegenerative diseases and acute surgical anesthesia stress (Bright et al., 2019). Recent studies indicated that inflammasome-mediated neuroinflammation was vital in surgical anesthesia stress–induced PND (Wei et al., 2019). Furthermore, some studies detected mature IL-1β levels in the peripheral blood, indicating a significant activation of inflammation in many diseases (Amdur et al., 2019). This study aimed to clarify the link between neuroinflammation and cognitive impairment caused by surgery and anesthesia. A PND model of mice was established for abdominal exploration. The hippocampal brain areas related to learning and cognition were evaluateds in vivo. The immunoblotting results showed that the expression levels of NLRP3 and AIM2, especially NLRP3 expression levels, in the hippocampus of mice significantly increased in the surgery group compared with the sham group (p < 0.05) (Figures 1A,B). However, the expression levels of NLRP1, NLRP2, and NLRC4 were not significantly different between the two groups. Similarly, CASP1 and IL-1β, the components of NLRP3 inflammasome in the hippocampus, had high expression levels (Figures 1C,D). Furthermore, the immunofluorescence staining was used to detect the expression of GFAP and IBA-1, which are the markers of astrocyte and microglia, to investigate whether NLRP3 expression on the glia of the hippocampal dentate gyrus was activated by surgery. As shown in Figures 1E,F, NLRP3 on microglia was significantly activated in the surgery group, and the degree of co-expression of IBA-1 and NLRP3 was higher than that of GFAP and NLRP3. In summary, these results suggested that NLRP3 expression was related to a surgery-induced inflammatory response.

Several animal and clinical studies have shown significant anti-inflammatory effects of DEX. Given that neuroinflammation is an important factor leading to PND, a PND model and a pre-administration group of DEX were established to explore the effect of DEX on inflammatory response and cognitive function in mice. As shown in Figure 2A, DEX was administered (30 μg/kg, intraperitoneally) once a day for 3 days before the surgery. The immunoblotting results showed that the expression of CASP1/IL-1β in the hippocampal tissue significantly increased in the surgical group compared with the sham group. DEX treatment attenuated the protein expression in a dose-dependent manner (Figure 2B). Surgical stress reduced the expression of postsynaptic density-95 (PSD-95), and the pre-protection of Dex can increase the expression of PSD-95. (Figures 2C,D). The inflammatory reactions were usually accompanied by the activation of glial cells. The expression of GFAP and IBA-1 was detected to investigate whether DEX had an effect on glia activation induced by surgery and anesthesia. Significant glial cell activation was observed in the surgery group, but DEX (30 μg/kg) administration significantly attenuated the activation of both astrocytes and microglia (Figures 2E,H). Golgi-Cox staining was used to examine the density of dendritic spines in the hippocampal region, which was thought to be closely related to learning and memory. The spine density was lower in the surgery group than in the sham group. DEX (30 μg/kg) treatment improved the decrease in spine density induced by surgery. However, the total dendritic length did not differ between these groups (Figures 2I–K). Finally, the behavioral analysis of spatial learning memory was performed using Y maze and novel object recognition tests to assess whether the efficacy of DEX improved against cognitive impairment in mice. The behavioral results showed that the exploration time of the DEX pre-protected group (30 μg/kg) in the new arm was longer than sham group in the Y maze experiment. Similarly, the experimental data showed that DEX pre-protected group (30 μg/kg) of the time spent exploring new things was longer than sham group in the new object recognition (Fig. 2L–2N). Collectively, these results suggested that DEX mitigated surgery-induced inflammation and reversed cognitive dysfunction caused by anesthesia and surgery.

The results indicate that DEX could reduce surgically induced nerve damage caused by inflammation in vivo. Hence, the neuroprotective effect of DEX were further verified in vitro. First, the effect of DEX on microglia activation was examined in LPS- and ATP-primed mouse primary microglia after determining the optimal concentration of DEX through cell viability and cell proliferation experiments in vitro (Figure 3A). As shown in Figures 3B,C, the treatment of DEX blocked IBA-1 expression in a concentration-dependent manner. Therefore, a dose of 10 μM was chosen for subsequent experiments. Previous studies reported that activated microglia damaged neurons through multiple pathways: production of superoxide directly caused cell necrosis or apoptosis; release of matrix metalloproteinases caused hypoxic-ischemic neuronal damage, reducing nutrition; and production of protective brain-derived neurotrophic factors and insulin-like growth factors indirectly increased neuronal apoptosis. Therefore, the cells were incubated with microglial conditioned media (MCM) mixed with 2% B-27 neurobasal medium (1:1) for 24 h to investigate the neuroprotective effects of DEX. The proteins of primary neurons were detected by immunoblotting. The expression of apoptosis-related protein BAX (apoptosis regulator) was upregulated, and the expression of BCL-2 (apoptosis regulator) was downregulated after incubation with LPS and ATP, but this response was reversed in the DEX treatment group (Figures 3D–F). In addition, the Annexin V/PI staining indicated that the neurons incubated with LPS and ATP MCM promoted the occurrence of early apoptosis. However, DEX protected neurons by reducing the release of microglial inflammatory factors, thus minimizing the occurrence of early apoptosis (Figures 3G,H). Similarly, the results of microtubule-associated protein2 (MAP-2) immunofluorescence showed that the microglial inflammation induced by LPS and ATP shortened the axon length of neuronal cells, but the culture medium of microglial cells treated with DEX reduced neuronal damage and improved axon length (Figures 3I,J). Taken together, these results indicated that DEX reduced microglia activation and neuronal cell apoptosis in vitro.

NLRP3 expression in the hippocampus was measured to investigate the mechanism underlying the neuroprotective effect of DEX. As shown in Figures 4A,B, the pre-administration of DEX reduced NLRP3 expression induced by surgery in a dose-dependent manner in a mouse model. At the same time, primary microglia were cultured and stimulated with LPS (100 ng/ml, 5 h) to verify the experimental results in vitro, which was the first activation signal of inflammation, prompting the expression of NLPR3 protein. Similarly, DEX attenuated the expression level of NLRP3 in primary microglia (Figures 4C,D). Considering the involvement of NLRP3 in inflammation, MCC950 (an inhibitor of NLRP3) was used to suppress NLRP3 expression in primary microglia. Also, LPS and ATP were administered to induce the activation and assembly of NLRP3 inflammasomes. IL-1β expression significantly reduced after interfering with NLRP3 expression compared with that in the group without MCC950 treatment; DEX reversed this effect (Figures 4E–G). NLRP3 inflammasome is composed of NLRP3 protein, pro-CASP1, and ASC, which are important in developing inflammation. Therefore, the present study further examined the effect of DEX on the expression of IL-1β and ASC after LPS and ATP stimulation. The results showed that DEX reduced CASP1 activation, IL1β maturation, and ASC oligomerization induced by LPS and ATP dose-dependently (Figures 4H–K). These results confirmed that DEX reduced the maturation of CASP1, production of IL-1β, and oligomerization of ASC, and then inhibited the activation and assembly of NLRP3 inflammasome.

To limit the dissipation in experimental animals on the mechanism research, the primary microglial cells were stimulated with LPS to activate the original protein complexes of NLRP3 so as to further explore the intrinsic mechanism by which DEX inhibited NLRP3 inflammasome. The results showed that DEX reduced NLRP3 expression in a dose-dependent manner, but had no effect on the expression of the components of NLRP3 inflammasome assembly, suggesting that DEX might be responsible for the degradation pathway of NLRP3 (Figures 5A,B). It was speculated that DEX might be responsible for the degradation pathway of NLRP3 to verify the effectiveness of two major degradation pathways (ubiquitin-proteasome and autophagy-lysosome pathways) in eukaryotes. The ubiquitin-proteasome and autophagy-lysosome pathways are two major degradation pathways in eukaryotes. They were blocked by treating the primary microglia with 3-MA and MG132, respectively. The immunoblotting analysis confirmed that the degradation of NLRP3 was inhibited by 3-MA, but not by MG132 (Figures 5C,D). These data indicated that DEX inhibited NLRP3 expression via promoting the autophagy pathway to degrade the protein. Simultaneously, DEX reduced NLRP3-mediated IL-1β secretion in an autophagy-dependent manner (Figures 5C,D). During autophagy induction, microtubule-associated protein 1 light chain 3 beta (MAP1LC3B; MAP1LC3B-I), a common marker of autophagosome formation, was covalently conjugated to phosphatidylethanolamine and converted into a fatty form (MAP1LC3B--II) on the elongated phagocyte membrane. Therefore, the increased conversion of MAP1LC3B-I into MAP1LC3B-II might serve as a sign of autophagosome formation. Next, the expression of MAP1LC3B was detected, revealing that DEX promoted autophagy and increased the production of MAP1LC3B-II slightly (Figures 5E,F). Sometimes, the decrease in MAP1LC3B-II levels was due to the rapid depletion of autophagy flux. This possibility was based on the accumulation of MAP1LC3B-II in the absence and presence of autophagy flux blockers such as Bafilomycin A1 (BafA1). DEX increased the accumulation of MAP1LC3B-II significantly in the presence of BafA1, indicating that it promoted an increase in the autophagy flux (Figure 5E). Next, the changes in the expression of ATG genes were compared between the control (CON) and DEX groups following LPS treatment. As shown in Figures 5G,H, DEX upregulated the LPS-stimulated reduction of the expression of autophagy related 5 (ATG5) and autophagy related 7 (ATG7). It also reduced the LPS-induced blockade of SQSTM1 protein and increased the conversion of MAP1LC3B-I into MAP1LC3B-II (Figures 5G,H).

On autophagy induction, red and green fluorescence co-localization was observed in the cells. In the late stage of autophagy, lysosomes and autophagosomes formed autophagolysosomes, intracellular pH decreased, GFP fluorescence decreased in acidic environments, and RFP was relatively stable. Hence, the two-point ratio of GFP–RFP could be used to evaluate the progress of the autophagy flow. Tandem fluorescent-mRFP-GFP-MAP1LC3B was transfected into primary microglial cells. In the present study, the green fluorescence of DEX-treated cells weakened, and the spots of red fluorescence increased. This effect was similar to that of the autophagy inducer rapamycin. BafA1 further inhibited the DEX-mediated reduction of red fluorescence of GFP (Figures 5I,J). Therefore, DEX promoted not only the formation of early autophagosomes in microglial cells but also the formation of autophagolysosomes in late stages. Taken together, these results showed that DEX promoted the degradation of NLRP3 and the NLRP3-mediated IL-1β secretion through the autophagy pathway, thus improving the inflammatory response.

In the initial phase of autophagy, STSQTM1 (a receptor protein) is vital in the degradation of substrate proteins, which are linked to ubiquitin-modified substrates and then transported to autophagosomes for degradation. The results showed the involvement of SQSTM1 in the effect of DEX on NLRP3 degradation via the autophagy pathway. Similarly, attempts were made to investigate whether the ubiquitination degradation pathway was a key regulatory mechanism for the activation and assembly of NLRP3 inflammasomes. The experimental result in this part of the study confirmed that DEX-induced degradation of NLRP3 protein was regulated by SQSTM1-related autophagy signals. Considering the importance of ubiquitination for substrate protein degradation, the role of DEX in mediating the degradation of NLRP3 ubiquitination was analyzed. First, the immunofluorescence data revealed that LPS and ATP inhibited the fluorescence signal of ubiquitin protein in primary microglia; also, DEX enhanced the fluorescence signal of ubiquitin protein and inhibited the fluorescence signal of NLRP3 (Figure 6A). The ubiquitinated substrate was removed by autophagy via the proteasome pathway. Therefore, the present study further explored the mechanism of DEX-induced degradation of NLRP3. The DEX-induced degradation of NLRP3 was inhibited by autophagy inhibitor 3-MA and lysosomal inhibitor LY294002, resulting in the failure of degradation of ubiquitinated NLRP3 and the increase in its deposition in the cytoplasm (Figure 6B). In contrast, the ubiquitin-proteasome inhibitor MG132 had no significant effect on this process (Figure 6C). The results showed that DEX accelerated the degradation of NLRP3 protein by promoting the ubiquitin-mediated autophagy process.