Animals

All animal experiments were approved by the Animal Committee of Shandong University Qilu Hospital (Qingdao, China) (KYDWLL-202107). 18-month-old C57BL/6 male mice were used in the study and housed in a 23–25 °C, 12-h light–dark cycle environment, and all mice were allowed to eat and drink freely (2–3 mice/cage). A total of 162 aged mice were utilized in the study. Mice were randomly assigned to Intervention groups. The number of mice that contributed data for analysis in each experiment was specified in the figure legends.

Anesthesia and surgery

A modified aseptic exploratory laparotomy was performed under isoflurane anesthesia in a chamber prefilled with 3.5% isoflurane and 100% oxygen while maintaining spontaneous ventilation. After the mice lost their upright reflex, they were placed on their right side and anesthetized using 1.5% isoflurane for 20 min. Surgery was performed under 1.5–2% isoflurane anesthesia, and an oxygen mask was worn during surgery. A 1-cm midline vertical incision was made in the abdomen to penetrate and explore the abdominal cavity. The operator manipulated the organs and organ musculature and removed approximately 2 cm of the intestine, which was rubbed vigorously with two sterile swabs for 30 s. The bowel was placed back into the abdominal cavity and the wound was closed with a 4–0 sterile surgical suture. Polysporin (Johnson & Johnson Inc.) was used to relieve and prevent postoperative pain. The duration of surgery was 15 min, the total duration of isoflurane anesthesia was 1.5 h. The sham-operated mice were anesthetized with isoflurane, and the abdominal area was shaved and cleaned as described above, without making any incision. OI (50 mg/kg, Cat# SML2338, Sigma-Aldrich) was administered intraperitoneally immediately 1 h after surgery. The dosage of OI used in the study were based on previous reports [17,18,19] and our pilot study. In addition, for the Nrf2 antagonist treatment, aged mice received intraperitoneal injections of ML385 (30 mg/kg, Cat# SML1833, Sigma-Aldrich), a specific Nrf2 antagonist, 1 h after surgery [20], and the same amount of dimethyl sulfoxide and normal saline were administered intraperitoneally to mice in other groups. 5-ethynyl-2′-deoxyuridine (EDU) (100 mg/kg, C0081L, Beyotime) was intraperitoneally administered for three consecutive days before the euthanasia of a separate cohort of aged mice. A schematic illustration of the experimental procedure is outlined in Fig. 1.

Fig. 1

Experimental protocols. Aged mice were administered intraperitoneally with 4-Octyl itaconate (OI) or/and ML385 (an antagonist of Nrf2) 1 h after exploratory laparotomy under isoflurane anesthesia. Intestinal contents, blood, and brain tissue were collected for 16S rRNA gene sequencing, metabolomics, ELISA, and immunofluorescence analysis 24 h after surgery. Another cohort of aged mice underwent the open-field test (OFT) and new object recognition (NOR) test on postoperative days 5 and 6. Subsequently, 5-ethynyl-2′-deoxyuridine (EDU) was intraperitoneally injected into aged mice for 3 days, and brain tissue was collected for immunofluorescence analysis 9 days after surgery

Cell culture and treatment

Fetal mice were removed at 17 days of gestation, and the heads were cut off and placed in ice-cold Hank’s balanced salt solution (HBSS). The brain tissues were separated with micro tweezers and placed into another Petri dish containing ice-cold HBSS. The meninges were peeled off with microforceps, and the hippocampal tissue was separated and transferred to a centrifuge tube containing serum-free Dulbecco’s modified Eagle medium (DMEM). The tubes were centrifuged at 1000 rpm for 5 min, the supernatant was removed by aspiration, and 1 mL of 0.05% tryptic digest was added to each centrifuge tube, which was lightly inverted several times and then digested for 20 min in a 37 °C incubator. The reaction was terminated by adding serum-containing medium to each tube, and the tubes were inverted lightly until the brain tissue was separated. After a 5-min centrifugation at 1000 rpm, the supernatants were discarded, and the pellet was resuspended with medium. Glass coverslips were coated with poly-D-lysine for 6 h and then washed with sterile phosphate-buffered saline (PBS) three times. The cells were applied to the cover slips, and the medium was half-exchanged the following day. The cells were cultured for 14 days for experiments.

Incubation with lipopolysaccharide (LPS) and isoflurane

Primary microglia and hippocampal neurons were placed in an incubator at 37 °C. Cells were treated with 1 μg/mL of LPS and exposed to 3% isoflurane in pure oxygen for 6 h. The isoflurane concentration was continuously monitored during this period using a Datex-Ohmeda ULT-SV analyzer. Primary microglia and neurons were treated with 250 μM OI or 5 μM ML385 1 h after POCD modeling, and the cells were collected after 24 h for experimental assays [21].

Enzyme-linked immunosorbent assay (ELISA)

The cytokines IL-6 and IL-1β secreted by microglial culture supernatant and from the hippocampus of aged mice were assessed 24 h post-operation using ELISA kits, according to the manufacturer’s instructions (Elabscience) [22, 23].

Ultra-high performance liquid chromatography-mass spectrometry/mass spectrometry analysis

Twenty-four hours after surgery, the intestinal contents of Sham, POCD, and POCD + OI groups were collected and weighed. The samples were placed in a 4 °C autosampler throughout the analysis. The samples were analyzed using a SHIMADZU-LC30 ultra-high performance liquid chromatography (UHPLC) with an ACQUITY UPLC®HSS T3 (2.1 × 100 mm, 1.8 µm) column (Waters, Milford, MA, USA). The injection volume was 4 μL, the column temperature was 40℃, and the flow rate was 0.3 mL/min. The chromatographic mobile phase A was a 0.1% formic acid solution, and mobile phase B was an acetonitrile solution. The chromatographic gradient elution program was as follows: 0–2 min, 0 B; 2–6 min, B varied linearly from 0 to 48%; 6–10 min, B varied linearly from 48 to 100%; 10–12 min, B was maintained at 100%; 12–12.1 min, B varied linearly from 100 to 0%; 12.1–15 min, B was maintained at 0%.

The positive (+) and negative (−) modes of each sample were detected by electrospray ionization. The samples were separated by UPLC, and mass spectrometry analysis was performed using a QE Plus mass spectrometer (Thermo Scientific). Ionization was performed using a heated electrospray ionization source. Ionization conditions were: spray voltage, 3.8 kv ( +) and 3.2 kv ( −); capillary temperature, 320 ± ; sheath gas, 30 ( ±); auxiliary gas, 5 ( ±); probe heater temperature, 350 ( ±); S-Lens RF level 50; mass spectrometry acquisition time, 15 min; parent ion scanning range: 75–1050 m/z; primary mass spectral resolution, 70,000@m/z 200; AGC target, 3e6; and primary maximum IT, 100 ms. Secondary mass spectrometry analysis was performed according to the following conditions: secondary mass spectra (MS2 scans) of the 10 highest intensity parent ions were triggered after each full scan, secondary mass spectral resolution: 17,500@m/z 200, AGC target: 1e5, 2-stage maximum IT: 50 ms, MS2 activation type: HCD, isolation window: 2 m/z, normalized collision energies (step): 20, 30, and 40.

Microbial taxonomy and metabolite biomarker correlation analysis

Microbiome bioinformatic analysis was performed using QIIME 2 2019.4, according to the official tutorial (https://docs.qiime2.org/2019.4/tutorials/). Briefly, the raw sequence data were demultiplexed using the Demux plugin, followed by primer cutting using the CutAdapt plugin. The sequences were then quality-filtered, denoised, merged, and the chimera was removed using the DADA2 plugin. Non-singleton amplicon sequence variants (asv) were aligned with MATT and used to construct phylogenies using fasttre2. α-Diversity metrics (Chao1, Observed species, Shannon, Simpson, Faith’s phylogenetic diversity, Pielou’s uniformity, and Good’s coverage) and β-diversity metrics (weighted UniFrac, unweighted UniFrac, Jaccard distance and Bry–Curtis dissimilarity) were estimated using diversityplugin. The Greengenes 13_8 99% operational taxonomic unit reference sequence was classified using the classiy-sklearn naïve Bayesian classifier in the feature classifier plugin.

Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay and immunofluorescence staining

Frozen sections were fixed with 4% paraformaldehyde (PFA) for 30–60 min, washed twice with PBS for 10 min each, incubated with 0.5% Triton X-100 at room temperature for 5 min, mixed with terminal deoxynucleotidyl transferase, fluorescent labeling solution, and TUNEL detection solution according to the manufacturer’s instructions (C1088, Beyotime), incubated at room temperature for 60 min, washed three times with PBS, and blocked for confocal microscopy. The slices were sealed and observed under a confocal microscope.

Immunofluorescence

Brain tissues were fixed with 4% PFA and sectioned in 40-μm cryosections, rinsed 3 times in PBS, blocked in bovine serum album (BSA) for 1 h, and incubated with primary antibodies against Nrf2 (rabbit, 1:500, ab31163, Abcam), Iba1 (rabbit, 1:500, ab178846, Abcam), GFAP (rabbit, 1:500, ab7260, Abcam), EDU (C0081L, Beyotime), doublecortin (DCX) (mouse, 1:100, Santa Cruz Biotechnology, sc-271390), extracellular signal-related kinases (ERK) (phosphoT202/T185, rabbit, 1:300, ab201015, Abcam), MAP2 (mouse, 1:1000, ab300645, Abcam) overnight at 4 °C. The tissues were washed 3 times in PBS, incubated with goat anti-mouse Alexa Fluor® 647 (1:500, ab150119, Abcam),, goat anti-rabbit Alexa Fluor® 488 (1:500, ab150081, Abcam), goat anti-mouse Alexa Fluor® 488 (1:500, ab150117, Abcam) or goat anti-rabbit Alexa Fluor® 647 (1:500, ab150083, Abcam) for 2 h, washed 3 times with PBS, sealed, and dried at 30–32 °C for 10–15 min for confocal imaging (Thermo Fisher, USA). Cellular immunofluorescence was performed as follows: immersed in 4% PFA for 15 min, blocked with BSA for 2 h, incubated with primary antibodies against Nrf2 (rabbit, 1:100, ab31163, Abcam), ERK (phosphoT202/T185, rabbit, 1:1000, ab201015, Abcam) and MAP2 (rabbit, 1:800, ab300645, Abcam) overnight, washed three times with PBS, and incubated with secondary antibody (AlexaFluor488- or AlexaFluor647-goat anti-rabbit IgG) for 2 h. The cells were then rinsed thrice with PBS. The cover glass was used for a Leica confocal microscope (Leica Camera, STELLARIS 5). The same settings were used for image processing [24].

Open-field test (OFT) and new object recognition (NOR) test

Behavioral testing was performed in an opaque square chamber on a white table above the ground (400 L × 400 mm W × 400 mm H) under dim light and quiet conditions. The NOR test was performed after the OFT. All mice underwent a 10-min acclimatization period in the apparatus on the first day, during which open-field data was recorded. The following day, two identical objects were placed in the apparatus to eliminate any potential presence of unique odors. The mice were positioned approximately 10 cm away from the edge of the apparatus, facing the center at an equal distance. A camera was set up to capture and record the exploration time of the two objects, defined as when their noses were within 2 cm of the objects or when they made direct contact with the objects. The number of times and length of time the mice explored the two objects were recorded for 5 min. To evaluate the effect on memory improvement, a 1-h interval passed after the completion of 5 min of recording. After the dwell time was completed, one of the objects was replaced by another new object in the apparatus, and the experiment was repeated for 5 min using the above experimental method. The exploration time of the mice was observed and recorded (“N” for the novel object and “O” for the old object). The NOR test was quantified as a discrimination index (N/N + O).

Statistical analysis

The sample sizes for each experiment were determined based on previous studies utilizing similar experimental paradigms. Large sample sizes were employed for behavioral experiments [25, 26], while small sample sizes were utilized for biochemical studies [9]. Statistical analyses were performed using GraphPad Prism software (version 8.0). All data were expressed as mean ± SD. A t-test was used to compare the statistical significance of two independent groups. One-way analysis of variance (ANOVA) followed by Tukey’s test was used to determine differences between groups. For the unpaired distribution and ANOVA, a rank-sum test was performed to check for significance. P < 0.05 was considered statistically significant [27].