Metformin (CAS: 1115-70-4), compound C (CAS: 866405-64-3), and LPS (L2880) (all from Sigma–Aldrich, St. Louis, MO, USA) were utilized in this experiment. Metformin was dissolved in normal saline and used in the animal experiments. For the cell experiments, metformin, compound C (an AMPK inhibitor) and LPS were dissolved in PBS and used to treat microglia. AMPK (polyclonal, 2532), p-AMPK (recombinant, 2535), SIRT1 (recombinant, 9475), LC3 (polyclonal, 2775), Beclin1 (polyclonal, 3738), and p62 (polyclonal, 5114) antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). A CD68 (monoclonal, ab955) antibody was obtained from Abcam (Cambridge, MA, USA). A glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (monoclonal, 60004-1-Ig) was purchased from Proteintech (Wuhan, China). Dulbecco’s modified eagle’s medium (DMEM) and fetal bovine serum (FBS) were obtained from Thermo Fisher Scientific, Inc (Waltham, MA, USA).

Male Sprague‒Dawley (SD) rats aged 18 months and weighing 500 g each were purchased from the Shanghai Laboratory Animal Center of the Chinese Academy of Science (Shanghai, China) and housed in accordance with industry standards at 22–24 °C on a 12-hour light–dark cycle. All procedures were approved by the Shanghai Huadong Hospital’s Animal Research Committee (approval number: 2022110029S) and carried out in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals. All rats were randomly divided into four groups (n = 12 per group): the control, metformin, operation, and operation plus metformin groups. The rats underwent splenectomy to establish a POCD animal model. Before surgery, the rats were fasted for 12 hours. The animals were pretreated with 200 mg/kg metformin by oral gavage once daily for two weeks.

After anesthetization via intraperitoneal injection of 2% pentobarbital sodium (Sigma-Aldrich, St. Louis, MO, USA) (40 mg/kg), the rats were placed on a sterile operating table. The skin on the abdomen was prepared, and after disinfection, we made a 1.5 cm incision along the midline of the abdomen. The spleen was located, and then the blood vessels of the root of the spleen were ligated. After successful splenectomy, we closed the abdominal incision with sterile sutures. Finally, the rats were placed on a heated blanket for recovery. The Morris water maze (MWM) test was conducted three days later.

To prevent infection after surgery, the training and test phases of the MWM test were performed from postoperative day 3 to postoperative day 8. The protocol was carried out as described previously [22]. The MWM consisted of a circular pool with a diameter of 180 cm and height of 50 cm. The pool was filled halfway with water, and the temperature of the water was maintained between 21 and 23 °C. The circular pool was divided into four quadrants (northeast, northwest, southeast, southwest). The platform, which had a diameter of 12 cm, was placed in the northwest quadrant slightly below the water surface. In our research, the process of training and testing in water maze was performed between 8 AM and 11 AM. There are opaque curtains that completely block the circular pool, so it can avoid visual interference. The animals’ movements were captured by a camera placed above the maze’s center, and Any-Maze software was used to analyze the data, which included the swimming speed, escape latency, and amount of time spent in the target quadrant. The test lasted for six days. The rats were placed in the water in a randomly selected quadrant near the pool wall on the first five days, i.e., during the training phase. If a rat could not locate the platform within 60 seconds, it was placed on the platform and kept there for 20 seconds to help it develop a memory of the platform location. On the final day, the platform was removed, the rats were placed in the water in the southeast quadrant, and the percentage of time spent in the target quadrant was recorded. After completing the water maze test, the rats were killed for sampling with inhaling excess carbon dioxide.

BV2 microglial cells were purchased from the Shanghai Laboratory Animal Center of the Chinese Academy of Science (Shanghai, China). The cells were identified in terms of morphology, phenotype, and function. The mycoplasma testing has been done for the cell lines used in the experiments. The cells were cultured in high-glucose DMEM containing 10% FBS and 1% penicillin‒streptomycin. The cells were grown in a 37 °C incubator containing 5% CO${}_2$. As described in a previous study, the microglia were stimulated with 500 ng/mL LPS [30]. The concentration of metformin used to treat the LPS-stimulated microglia was 2 mM, as described in previous research [31]. In the present study, the microglia were divided into five groups: the control, metformin, LPS, LPS plus metformin, and LPS plus metformin plus compound C groups.

Following the in vitro and in vivo interventions, protein was extracted from rat cortex samples and BV2 cells using 4 °C RIPA buffer (Beyotime Institute of Biotechnology, Shanghai, China). The proteins were then centrifuged at a speed of 12000 $\times$g for 20 min at 4 °C. A BCA protein assay kit (Beyotime, Shanghai, China) was used to determine the protein concentration. The proteins were separated by electrophoresis (8%~12% SDS–PAGE) and transferred to PVDF membranes. The membranes were blocked with 5% bovine serum albumin (BSA) for one hour and the incubated with LC3 (1:1000; item No. 2775), Beclin1 (1:1000; item No. 3738), p62 (1:1000; item No. 5114), AMPK (1:1000; item No. 2532), p-AMPK (1:1000; item No. 2535), SIRT1 (1:1000; item No. 9475), CD68 (1:1000; item No. ab955), and GAPDH (1:5000; item No. 60004-1-Ig) overnight at 4 °C. The membranes were washed with tris-buffered saline with 0.1% Tween (TBST) before being incubated with secondary antibodies for two hours. Finally, an ECL assay kit (Tanon, Shanghai, China) was used to visualize the protein bands. We used QuickBlock™ Primary and Secondary Antibody Dilution Buffer for Western Blot to dilute the primary and secondary antibody (Beyotime, Shanghai, China). The concentration of secondary antibodies was diluted 1:5000. ImageJ version 1.38 (NIH Image J system, Bethesda, MD, USA) was used for quantitative analysis.

In addition to Western blotting, rat cortex samples were used for immunofluorescence. They were removed immediately, stored in 4% paraformaldehyde, embedded in paraffin, and sliced into approximately 5 µm sections for staining. A primary antibody against LC3 (1:200) was used. The cortical sections were washed with 4 °C PBS and placed in a wet box. After that, they were incubated with the LC3 antibody. Following a second wash in PBS, the sections were incubated at room temperature with a FITC-conjugated anti-rabbit IgG secondary antibody. DAPI (4${}^{\prime }$,6-diamidino-2-phenylindole) was used to stain the nuclei. The secondary antibody was Alexa Fluor 555-labeled Donkey Anti-Rabbit IgG (H+L) (1:500, Beyotime, Shanghai, China).

To measure the ROS level, we used Reactive Oxygen Species Assay Kit (Beyotime, Shanghai, China). BV2 microglia were incubated with antibodies against oxidative stress indicators in 6-well plates for 30 min. The cells were then removed and washed 3 times with PBS for 3 minutes each. After that, we used a fluorescence microscope (TCSSP5, Leica Microsystems GmbH, Wetzlar, Germany) to determine the fluorescence intensity.

Statistical analysis was performed with GraphPad Prism 6.0 (GraphPad Software Inc., La Jolla, CA, USA). The continuous variables in our experimental data have been tested for normality. And all data in this study conform to normal distribution. The means and standard error of the mean (SEMs) are displayed for all data. Multiple group comparisons were analyzed with repeated measures analysis of variance (ANOVA), followed by a Bonferroni’s post hoc test. An unpaired, two-tailed Student’s t-test was applied to compare between two groups. The results were considered significant at p 0.05.

The experimental layout is depicted in Fig. 1A. The rats were pretreated with metformin for two weeks. Three days after the operation, cognitive tests, including the positioning navigation and spatial exploration experiments, were performed. According to a previous study [21, 32], surgery and inhalation anesthetics can both lead to cognitive dysfunction in rodent models, especially in aged animals, and the same result was found in our study. The operation clearly had a significant impact on the learning and memory abilities of old rats, as evidenced by the fact that the escape latency was significantly longer (Fig. 1B) and the percentage of time spent in the target quadrant was lower in the operation group than in the other groups (Fig. 1C). The rats pretreated with metformin did not display significant cognitive impairment, and they spent more time in the target quadrant (Fig. 1C) and had a shorter escape latency (Fig. 1B) than the rats in the operation group. Swimming speed was not significantly different among the four groups (Fig. 1D), indicating that the surgery had no impact on the motor ability of the aged rats. These results suggested that metformin may prevent age-related cognitive decline in rats after surgery.

Fig. 1.

Metformin alleviated surgery-induced learning and memory impairment. (A) Timeline of metformin intervention and the MWM test (n = 12 each group). Before splenectomy, the rats received metformin by oral gavage once daily for two weeks and then underwent behavioral tests. (B) The escape latency refers to the average time required to find the target platform in the training phase. (C) During the testing phase, the metformin-treated rats spent a greater percentage of time in the target quadrant. (D) The average swimming speed was used to assess differences in athletic ability in the MWM test. The data are shown as the mean $\pm$ SEM; *p 0.05 compared to the control group; p 0.05, the operation group compared to the operation+metformin group. Statistical differences were considered significant at *p 0.05, p 0.05, **p 0.01, or ***p 0.001. D, Day; MWM, Morris water maze; Con, Control; Met, metformin; OP, operation.

To elucidate the possible mechanism by which metformin can prevent POCD, we investigated whether microglial autophagy occurs in the cerebral cortices of aged rats. Autophagy was activated in the cortical homogenates of rats pretreated with metformin, as demonstrated by a significant increase in LC3 and Beclin1 levels (Fig. 2A,B). In contrast, p62 expression in the cortex was significantly higher in the operation group than in the other groups (Fig. 2A,B). One of the most commonly used autophagy markers is LC3, and the LC3II/I ratio was used to assess the level of autophagy in the present study. We used immunofluorescence staining to evaluate the immunoreactivity of LC3 (Fig. 2E). Consistently, the operation group showed a lower fluorescence intensity, while metformin pretreatment reversed the surgery-induced reduction in autophagy (Fig. 2E). Microglia are important immune cells in the central nervous system that can be activated by external stimulation, and one of the markers of microglial activation is CD68. Once activated, microglia release inflammatory factors such as IL-1$\beta$. The findings demonstrated that the operation group had significantly higher levels of microglial activation in the cortex and IL-1 expression than the other groups (Fig. 2C,D). However, these changes were mitigated by metformin pretreatment (Fig. 2C,D).

Fig. 2.

Metformin increased autophagy and inhibited microglial activation in the cerebral cortex. (A,B) Western blot analysis and quantification of Autophagy associated proteins (LC3, p62, Beclin1) expression in the cortex in the control, metformin, operation and operation+metformin groups. (C,D) Western blot analysis and quantification of IL-1$\beta$ and CD68 levels in the cortex in the control, metformin, operation and operation+metformin groups. (E) Immunofluorescence staining of LC3 in each group. The data are shown as the mean $\pm$ SEM. The scale bar represents 50 mm. Statistical differences were considered significant at *p 0.05, **p 0.01, ***p 0.001 or ****p 0.0001. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; DAPI, 4${}^{\prime }$,6-diamidino-2-phenylindole.

Microglia typically have a round or oval shape under physiological conditions. After pretreatment with 2 mM metformin, microglia had a nearly identical morphology as those in the control group, indicating that at this dose, metformin had little effect on cell viability. After LPS stimulation, microglial cells underwent obvious morphological changes, exhibiting an irregular shape and protrusions. However, metformin prevented these morphological changes in microglia to some extent (Fig. 3A).

Fig. 3.

Metformin changed the morphology and oxidative stress status of microglia. (A) Images of microglial morphology. (B) Immunofluorescence staining of ROS in each group. The scale bar represents 50 mm. Met, metformin; LPS, Lipopolysaccharide; ROS, reactive oxygen species.

Microglial state, in addition to cell morphology, was changed. LPS can induce oxidative stress in cells. The results are displayed in Fig. 3B. The fluorescence intensity of ROS was measured to determine ROS levels. In the LPS group, ROS levels were the highest. Although the ROS fluorescence intensity in the metformin pretreatment group was comparable to that in the control and metformin groups, the comparison between the metformin pretreatment group and the LPS group suggested that metformin had a protective effect, lowering the levels of ROS.

Microglia were stimulated with LPS in vitro to simulate neuroinflammation in rats with POCD, and changes in autophagy in the cells were assessed. The results were similar to those observed in vivo. Western blotting revealed that the levels of LC3 and Beclin1 were significantly increased in microglia pretreated with metformin, indicating that autophagy was triggered in these cells (Fig. 4A–E). However, LPS significantly inhibited autophagy in microglia (Fig. 4A–E).

Fig. 4.

Metformin increased autophagy and activated the AMPK/SIRT1 signaling pathway in vitro. (A–E) Western blot analysis and quantification of LC3, p62, Beclin1 and IL-1$\beta$ levels during LPS-induced neuroinflammation. (F–I) Western blot analysis and quantification of phosphorylated and total AMPK and SIRT1 levels during LPS-induced neuroinflammation. Statistical differences were considered significant at **p 0.01, ***p 0.001 or ****p 0.0001. AMPK, Adenosine monophosphate-activated protein kinase; SIRT1, Sirtuin 1.

According to a previous report, the regulation of autophagy is intimately tied to the AMPK/SIRT1 signaling pathway [26]. Thus, we further investigated whether this signaling pathway participates in LPS-induced neuroinflammation in microglia. In this part of the experiment, metformin pretreatment clearly enhanced the AMPK/SIRT1 interaction in microglia, as shown by an increase in AMPK and SIRT1 phosphorylation (Fig. 4F,H,I). However, compound C (a popular AMPK inhibitor) inhibited metformin-induced activation of the AMPK/SIRT1 signaling pathway, as evidenced by a reduction in AMPK and SIRT1 phosphorylation (Fig. 5C,D). Additionally, in microglia treated with LPS, compound C increased the expression of p62 while decreasing the expression of LC3 and Beclin1 (Fig. 5A,B). The ROS fluorescence intensity was not significantly different between the LPS plus metformin group and the compound C group (LPS+Met+CC). Nevertheless, the ROS fluorescence intensity in these groups was lower than that in the LPS group. According to the aforementioned findings, metformin activates the AMPK/SIRT1 pathway to reduce POCD-induced autophagy inhibition in cortical microglia.

Fig. 5.

During LPS-induced neuroinflammation, metformin promoted autophagy by activating the AMPK/SIRT1 signaling pathway and alleviated microglial oxidative stress in vitro. (A,B) Western blot analysis and quantification of LC3, p62 and Beclin1 levels during LPS-induced neuroinflammation in the LPS, LPS+operation, and LPS+operation+compound C groups. (C,D) Western blot analysis and quantification of phosphorylated and total AMPK and SIRT1 levels during LPS-induced neuroinflammation in the LPS, LPS+operation, and LPS+operation+compound C groups. (E) Immunofluorescence staining of ROS in each group. The scale bar represents 50 mm. Statistical differences were considered significant at *p 0.05, **p 0.01, or ***p 0.001. CC, Compound C.