2.1 Mouse anesthesia and treatment

The animal protocol was approved by the Tianjin Medical University General Hospital Institutional Animal Care and Use Committee (Tianjin, China; clearance number, 2018-X6-12). The number of animals included in this study was reduced as much as possible. P6 and P60 female wild-type (WT) and ApoE-knockout (ApoE-KO) C57BL/6J mice were obtained from the Nanjing Biomedical Research Institute of Nanjing University (Nanjing, China), and female ApoE3-targeted replacement (ApoE3) and ApoE2-targeted replacement (ApoE2) C57BL/6J mice were obtained from the Shanghai Model Organisms Center (Shanghai, China). All animals were housed in cages in a controlled setting (five adult mice or one adult mouse with litter per cage) with standard food and appropriate water (temperature 21–23°C, humidity 50–60%, 12:12 h light–dark cycle).

The mice were divided into two groups: control and sevoflurane anesthesia. They were administered sevoflurane or a control therapy from postnatal day (P) 6 to P8 or P60 to P62, and hippocampal tissues were harvested at P8 or P62. Various groups of mice were employed in behavioral research. These mice were administered sevoflurane or a control condition from P6 to P8 or P60 to P62, and the Morris water maze (MWM) test was conducted from P30 to P36 or P84 to P90. The mice in the sevoflurane group were anesthetized with 3% sevoflurane plus 60% oxygen twice daily for 3 days in a special sealed resin box chamber (20 cm × 15 cm × 7 cm) with an inlet and outlet, whereas the control mice received only 60% oxygen (balanced with nitrogen) at an equal rate of flow in a chamber similar to the anesthesia mice (Lu et al., 2017; Tao et al., 2014). The concentrations of sevoflurane and oxygen were continuously monitored using a variable anesthetic gas monitor (Vamos; Drager Medical AG & Co. KgaA, Germany) during sevoflurane anesthesia or control administration. The temperature of the anesthetic chamber was maintained at 37 ± 0.5°C using a warming pad placed beneath the chamber.

2.2 Brain tissue harvest

The mice were given sufficient time to recuperate from anesthesia. At P8 or P62, each mouse was decapitated and hippocampal tissues were collected. ELISA and western blot analyses were performed on the acquired hippocampal tissues. After freezing, the harvested hippocampal tissues were homogenized with immunoprecipitation buffer (M-PER® Mammalian Protein Extraction Reagent, Cat#78501, Thermo Scientific, MA, USA), protease and phosphatase inhibitor cocktails (Cat#19541400, Roche, Switzerland), and phosphatase inhibitor cocktail (Cat#539131, Millipore, MA, USA). Lysates were collected and centrifuged at 12,000 rpm for 10 min.

2.3 Protein quantification

A bicinchoninic acid protein assay kit was used to assess the total protein content (BCA, Pierce, Iselin, NJ, USA).

2.4 Real-time polymerase chain reaction (RT-PCR)

Total RNA concentration was calculated using a spectrophotometer. In each group, the All-in-one First-Strand cDNA synthesis kit (Cat#AORT-0050, Gene Copoeia, USA) was used to convert 500 ng of mRNA into cDNA, which was subsequently amplified by PCR using Real Master Mix on the iQTM5 machine (SYBR Green). The 2-Ct technique was used to calculate relative mRNA levels. The following primer sequences were used for several genes: ApoE, reverse primer, 5′-CATGTCTTCCACTATTGGCTCG-3′ and forward primer, 5′-GACCCAGCAAATACGCCTG-3′; GAPDH, reverse primer, 5′- AGGTCGGTGTGAACGGATTTG-3′ and forward primer, 5′-TGTAGACCATGTAGTTGAGGTCA-3′.

2.5 Enzyme-linked immunosorbent assay

A mouse ApoE ELISA kit (Cat# E-EL-M0135, Elabscience, TX, USA) was used to measure total ApoE expression levels in the hippocampi of mice in each group. In a nutshell, each well was filled with 100 μl of the standard or sample and incubated for 90 min at 37°C. After that, 100 μl of particular antibody was added to each cell and incubated for another 60 min at 37°C. After that, each well was filled with 100 μl of enzyme conjugate and incubated for 30 min at 37°C. Ninety microliters of the substrate solution was added after washing five times. The reaction was stopped 15 min later with 50 μl of terminating solution.

2.6 Western blot analysis

Full-length ApoE, ApoE fragments, and Tau5, AT8, and PHF1 expressions in the hippocampus was detected using western blotting. Using a 4–20% double TIS polyacrylamide gel (Cat#M00655, Gen Script Biotech Corp., Nanjing, China) or a standard XT 4–12% double TIS gel (Bio-Rad, USA), the samples were separated from SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, USA). The membranes were then blocked with 5% skimmed milk powder and identified using the following primary antibodies: Tau5 (1:1000, Cat#ABN454, Millipore, USA), full-length ApoE and ApoE fragments (1:4000, Cat#178479, Calbiochem, Germany), PHF1 (1:1000, Cat#3Ab184951, Abcam, UK), and AT8 (1:2000, Cat#MN1020, Thermo Fisher Scientific, USA). After washing at 4°C overnight, the membranes were incubated at 37°C for 1 h with goat anti-mouse antibody (1:5000, Cat#31430, Invitrogen, USA). The nitrocellulose membranes were then scanned and photographed using Bio-Rad image analysis equipment after being washed in TBST and dripped with ECL chemiluminescent liquid (Cat#34577, Invitrogen, USA). The ratio of the integral optical density of the target band to that of the β-actin band represents the level of expression of the target protein. A 100% change refers to control or sevoflurane levels for comparison between the experimental settings.

2.7 Immunofluorescence

The animals were anesthetized with 3% sevoflurane for 5 min and perfused transcardially with PBS, followed by 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4 for immunofluorescence of brain slices. The brain tissues of the mice were removed and stored at 4°C in paraformaldehyde. For immunostaining, 10 mm frozen slices of mouse brain hemispheres were used. The sections were incubated overnight at 4°C with anti-ApoE (1:4000; Cat#178479, Calbiochem, Germany) and anti-AT8 antibodies (1:1000; Cat#MN1020, Thermo Fisher Scientific, USA), followed by immunostaining with Alexa Fluor® 488 donkey anti-goat IgG (1:500; Cat#A-11055, Thermo Fisher Scientific, USA) and Alexa Fluor^® 594 goat anti-mouse IgG (1:500; Cat#A-11032, Thermo Fisher Scientific, USA) for 1 h at room temperature in dark. Finally, the slices were coated with 4′6-diamidino-2-phenylindole, dihydrochloride (DAPI; Cat#104139, Abcam, UK) and incubated in a humidified dark room for 10 min before being observed in mounting fluid under a fluorescence microscope.

2.8 Morris water maze

MWM experiments were performed as described in our previous study (Yu et al., 2020). Briefly, P30 and P84 mice were examined in the MWM four times each day for 7 days (P30 to P36 and P84 to P90). Escape latency was measured daily. On the last day, the platform was disassembled and the platform crossing times were recorded. Each mouse was kept in a holding cage for 5 min after each experiment to dry before being returned to its home cage.

2.9 Statistical analysis

MWM escape latency was reported as mean ± standard deviation (SD) in separate groups, platform crossing times were presented as median with interquartile range, and other data as mean ± SD. The number of samples was 10 per group for behavioral studies, four per group for RT-PCR, four or six per group for western blot analyses, four per group for ELISA, and three per group for immunohistochemistry analyses. The interaction between time and group variables (based on escape delay) between the two groups in the MWM was investigated using two-way ANOVA with repeated evaluations. In addition, the Mann-Whitney test was used to compare platform crossing times among all groups in the MWM. There were no missing data for MWM variables (escape latency and platform crossing time) throughout the data analysis. Finally, to compare the two groups for other biochemical data, the unpaired t-test (if the values were in Gaussian distribution) or Mann–Whitney test (not-Gaussian distribution) was applied. Note that p < .05 was judged statistically significant, and the double-tail test was used for statistical significance. GraphPad Prism (version 5.0) and SPSS statistical software (version 21.0) were used for statistical analyses.

3.1 Sevoflurane anesthesia induced total ApoE, full-length ApoE, and ApoE fragments increase, Tau phosphorylation enhancement, and cognitive impairment in young (P6), but not adult (P60) mice

We initially evaluated the mRNA and protein levels of total ApoE in the hippocampus of P6 and P60 animals after sevoflurane and control treatments, as shown in Figure 1(a). We observed that sevoflurane anesthesia substantially enhanced the mRNA and protein levels of total ApoE in P6 mice (p < .05 vs. P6 + Control) but not in P60 animals (p > .05 vs. P60 + Control), as compared to the control group (Figure 1b and c). Furthermore, full-length ApoE and ApoE fragments levels were both substantially higher in the sevoflurane group compared to the control treatment only in P6 mice (p < .05 vs. P6 + Control), but not in P60 mice (p > .05 vs. P60 + Control) (Figure 1d–f).

Next, we examined the total Tau (Tau5) and phosphorylated Tau (AT8 and PHF1) levels in the hippocampus of P6 and P60 mice under sevoflurane anesthesia. The findings showed that P6 mice had greater levels of total Tau than P60 animals (p < .05 vs. P6 + Control). Furthermore, sevoflurane anesthesia enhanced hippocampus AT8 and PHF1 levels in P6 mice (p < .05 vs. P6 + Control), but not in P60 animals (p > .05 vs. P60 + Control) (Figure 1 g-1j).

Finally, we investigated whether variations in ApoE and Tau phosphorylation might cause cognitive impairment in P6 and P60 animals. Immunofluorescence labeling of the hippocampus revealed that the sevoflurane group had higher expressions of ApoE and AT8 proteins than the control group in P6 mice (p < .05 vs. P6 + Control), but not in P60 mice (p > .05 vs. P60 + Control) (Figure 1k,m ). In addition, as shown in Figure 1a, we used the MWM to examine cognitive skills in young (P30 to P36) and adult mice (P84 to P90) after sevoflurane anesthesia for 7 days. We discovered that, compared to control therapy, sevoflurane anesthesia did not cause neurocognitive impairment in P60 mice (p > .05 vs. P60 + Control), but caused neurocognitive impairment in P6 mice (p < .05 vs. P6 + Control) (Figure 1n and o). These data imply that sevoflurane anesthesia increased total ApoE, full-length ApoE, ApoE fragments, and phosphorylated Tau protein levels in the hippocampus, which might have led to neurocognitive impairment in P6 mice, but not in P60 mice.

3.2 ApoE deficiency could not mitigate Tau phosphorylation and cognitive impairment induced by sevoflurane anesthesia in P6 mice

For further study whether ApoE could be the main reason behind sevoflurane-induced brain damage and cognitive dysfunction in P6 mice, 6-day old WT and ApoE-KO mice were used in the research as shown in Figure 2(a). The results of RT-PCR and ELISA showed that there were significant changes in total ApoE mRNA and protein levels between sevoflurane and control treatment groups in WT mice, but not in ApoE-KO mice (p < .05 vs. WT + Control; p > .05 vs. ApoE-KO + Control) (Figure 2b and c). In WT mice, but not ApoE-KO, the western blot findings revealed clear differences in full-length and fragmented ApoE expressions between sevoflurane and control therapy (p < .05 vs. WT + Control; p > .05 vs. ApoE-KO + Control group) (Figure 2d and f). In addition, ApoE-deficient mice had higher AT8 and PHF1 levels than WT mice (p < .05 vs. WT + Control), and sevoflurane treatment could induce AT8 and PHF1, but not Tau5 expression (although not statistically significant) compared with the control treatment in ApoE-KO mice (Figure 2 g–m ). Furthermore, compared to control therapy, sevoflurane anesthesia might have exacerbated cognitive impairment in both WT and ApoE mice (p < .05 vs. WT + Control; p < .05 vs. ApoE-KO + Control) (Figure 2n and o). These data demonstrated that ApoE could not be the therapeutic target of sevoflurane-induced brain damage and cognitive impairment in young mice.

3.3 Sevoflurane-induced phosphorylated Tau enhancement and cognitive impairment in ApoE3 mice could be mitigated in ApoE2 mice

To investigate whether ApoE fragments were the primary cause of sevoflurane-induced brain injury and cognitive impairment in P6 mice, 6-day-old ApoE3- and ApoE2-target replacement animals were used in the investigation, as shown in Figure 3(a). Sevoflurane anesthesia increased total ApoE mRNA and protein levels in both ApoE3- and ApoE2-target replacement mice, according to RT-PCR and ELISA data (p < .05 vs ApoE2 + Control; p < .05 versus ApoE2 + Control) (Figure 3b and c). The western blot findings revealed significant changes in full-length ApoE expression between sevoflurane and control therapy in ApoE3, but not ApoE2, mice (p < .05 vs. WT + Control; p > .05 vs. ApoE-KO + Control) (Figure 3d and f). In fact, the quantity of ApoE fragments produced by ApoE2 target-replacement mice was low, regardless of the use of sevoflurane. Furthermore, in ApoE2 mice, sevoflurane anesthesia did not enhance AT8 and PHF1 levels compared to the control treatment (p > .05 vs. ApoE2 + Control); however, in ApoE3 mice, there were greater AT8 and PHF1 expressions in the sevoflurane group than in the control group (p < .05 vs. ApoE3 + Control) (Figure 3 g--m ). Furthermore, when compared to the control group, sevoflurane anesthesia might have exacerbated cognitive impairment only in ApoE3-target replacement mice but not in ApoE2-target replacement mice (P 0.05 vs. ApoE3 + Control; p > .05 vs. ApoE2 + Control) (Figure 3n and o). These findings suggest that the fragments, but not full-length ApoE, are the main reason for enhanced Tau phosphorylation and neurocognitive impairment following sevoflurane anesthesia in vivo.