Seventy-two clean, healthy postnatal day 7 (P7) Sprague-Dawley rats, irrespective of sex, weighing 10–15 g, were obtained from the Animal Center of Xinjiang Medical University (Xinjiang, China). All animals were housed under specific pathogen-free conditions in a well-ventilated environment maintained at 21–26 °C under a 12/12-h light/dark cycle with 30–60% humidity. Pups were kept with their dams until weaning at postnatal day 21 and then group-housed (n = 3 per cage). All animals had free access to water and food. The animal experiments were conducted following the Animal Research: Reporting of In Vivo Experiments (ARRIVE) 2.0 guidelines (Supplementary Materials-ARRIVE) and the National Institutes of Health (NIH) standards for laboratory animal welfare. All experimental procedures were performed at the Animal Center of Xinjiang Medical University (Xinjiang, China).

In total, 72 rats were randomly allocated to six groups (n = 12 per group) using a computer-generated random number sequence:

(1) Control group (Con): Saline (100 mg/kg) was administered to the rats by intraperitoneal (i.p.) injection.

(2) Propofol group (Pfo): A propofol neurotoxicity model was established as previously reported [24]. Rats were administered propofol (50 mg/kg, i.p.; Xi'an Libang Pharmaceutical Co., Ltd., Xi'an, Shaanxi, China; catalog No. H19990282). Following the return of the righting reflex (30–60 min after the first propofol injection), an additional dose of propofol (50 mg/kg, i.p.) was subsequently administered.

(3) Esketamine + Propofol group (Esk+Pfo): Rats received an injection of esketamine (10 mg/kg, i.p.; Jiangsu Hengrui Medicine Co., Ltd., Lianyungang, Jiangsu, China; catalog No. H20193336). Thirty minutes later, the first dose of propofol (50 mg/kg, i.p.) was administered. After recovery of the righting reflex (30–60 min after the first propofol injection), a second dose of propofol (50 mg/kg, i.p.) was administered.

(4) PI3K inhibitor LY294002 + Esketamine + Propofol group (LY294002+Esk+Pfo): The PI3K inhibitor LY294002 was administered intraperitoneally at 10 mg/kg [25]. Thirty minutes later, esketamine (10 mg/kg, i.p.) was administered. Thirty minutes after the esketamine injection, the first dose of propofol (50 mg/kg each) was administered. After recovery of the righting reflex (30–60 min later), a second propofol dose (50 mg/kg, i.p.) was administered.

(5) Autophagy activator Rapamycin + Propofol group (Rapa+Pfo): Rats received an i.p. injection of the autophagy activator rapamycin (7.5 mg/kg) [26]. Thirty minutes later, the first dose of propofol (50 mg/kg, i.p.) was administered. After recovery of the righting reflex (approximately 30–60 min after the first propofol injection), a second dose of propofol (50 mg/kg, i.p.) was administered.

(6) Autophagy activator Rapamycin + Esketamine + Propofol group (Rapa+Esk+Pfo): Rats received an i.p. injection of rapamycin (7.5 mg/kg). Thirty minutes later, esketamine (10 mg/kg, i.p.) was administered. After recovery of the righting reflex (30–60 min later), a second propofol dose (50 mg/kg each) was administered.

All treatments were administered at a consistent time daily for five consecutive days. Immediately following each injection, neonatal rats were placed in an incubator (Thermo Fisher Scientific, Waltham, MA, USA) maintained at 37 °C, with 100% oxygen supplied at a flow rate of 2 L/min. Respiratory status was closely monitored throughout the anesthetic period, including respiration rate and the color of the skin and mucous membranes. In addition, oxygen saturation was continuously monitored using a MouseOx® Plus pulse oximeter (Starr Life Sciences Corp., Oakmont, PA, USA) and remained above 95% in all experimental groups. Animals were observed until complete restoration of the righting reflex to ensure safety before being returned to their dams. No deaths, cyanosis, or abnormal recovery patterns were observed during the treatment period.

The Y-maze test was performed to assess working memory on postnatal day 30 (P30). The Y-maze, one of the most straightforward paradigms, is widely utilized to evaluate learning and memory capabilities in rodents [27]. The apparatus consists of three arms; each fitted with electrifiable copper grids. One arm was randomly designated as the start compartment. Another arm was designated as the safe compartment, defined by constant illumination and the absence of foot shocks throughout the test. The third arm was designated as the non-safe compartment, characterized by being unlit and equipped with a grid floor for delivering foot shocks during the training phase. Foot shocks (0.3 mA intensity, 2-sec duration) were delivered via the grid floor upon entry into the non-safe compartment.

During the training phase, a foot shock was delivered after a 5-sec delay following the illumination of the safe compartment. An escape to the safe compartment was recorded as a correct response. The learning criterion was defined as achieving nine consecutive correct responses. Before testing, animals were allowed to freely explore the maze for 3 min. The voltage intensity was adjusted to ensure escape behavior within a 10-sec window. The number of training trials to criterion was recorded for each rat. Notably, 23 h later, memory retention was assessed by administering 20 consecutive trials. The correct response rate (%) was calculated as $\frac{\text{number of correct responses}}{20}\times 100\mathrm{\%}$ . All experiments were conducted during a fixed time window (9:00–12:00) under controlled conditions of 50 lux illumination and <35 dB background noise.

The investigators responsible for the Y-maze test were blinded to the treatment allocation.

Spatial learning and memory were assessed on postnatal day 60 using the MWM. The MWM apparatus consisted of a circular stainless-steel pool measuring 1.2 m in diameter, with the water temperature maintained at 25 ± 1 °C. A hidden platform (diameter: 9 cm) was positioned in the center of the second quadrant, 2 cm beneath the water surface. Distinct visual cues placed around the testing room served as spatial references during navigation. The test comprised two consecutive phases: acquisition phase (days 1–5): Rats underwent four training trials daily for five consecutive days. In each trial, the animal was released into the water from one of four starting points facing the wall, corresponding to the four quadrants, in a clockwise sequence each day. A 10-min inter-trial interval was enforced. The time taken to locate the hidden platform in each trial was recorded as the escape latency (s). If a rat could not find the platform within 120 sec, it was guided to the platform, and the escape latency was recorded as 120 sec.

Probe trial (day 6): In this phase, 24 h after the final acquisition session, a probe trial was employed to assess long-term spatial memory. The hidden platform was removed from the pool, and each animal was allowed to freely swim for 120 sec. Throughout this interval, crossings over the former platform site were counted and recorded.

The MWM test was performed during a fixed time window (9:00–12:00) under standardized conditions of 50 lux illumination and background noise of below 35 dB. For data recording, the escape latency during these guided trials was consistently recorded as 120 sec when the animal did not escape within the allotted time. In the statistical analysis, all latency values, including those recorded as 120 sec, were included in the analysis without exclusion or special weighting.

Treatment assignments were concealed from the investigators conducting the MWM assessment.

Following behavioral tests, rats were anesthetized with 2% sodium pentobarbital (50 mg/kg, i.p.; Sigma-Aldrich, St. Louis, MO, USA; catalog No. P3761) and subsequently euthanized by decapitation. The brain was removed rapidly, and hippocampal tissue was dissected on ice. Tissue blocks containing the hippocampal Cornu Ammonis area 1 (CA1) region (thickness ≤0.5 cm) were fixed at room temperature in 10% neutral buffered formalin for 23 h.

After fixation, samples were dehydrated through a graded ethanol series, cleared, and then embedded in paraffin (Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China; catalog No. 69019361). Coronal paraffin sections (5 μm thick) were prepared using a microtome (Leica Microsystems, Wetzlar, Germany) and subsequently processed for hematoxylin and eosin staining kit (Solarbio, Beijing, China; catalog No. G1120). Sections were first exposed to hematoxylin for 10 min, differentiated in 1% acid alcohol for 50 sec, and then counterstained with eosin for 5 min. Finally, the sections were dehydrated and cleared, followed by mounting with neutral balsam (Nanchang Yulu Experimental Equipment Co., Ltd., Nanchang, Jiangxi, China; catalog No. 20200237).

Light-microscopic evaluation of the stained sections was performed at 300× magnification using an Axio Imager A2 microscope (Carl Zeiss AG, Oberkochen, Germany). Two blinded investigators independently evaluated neuronal damage in the hippocampal CA1 region. Brain injury scores were assigned on the basis of standardized criteria that assessed morphological changes, such as pyknosis, neuronal loss, and eosinophilic cytoplasm. Specifically, scores were assigned on the basis of a 0–10 semi-quantitative scale, as previously described for hippocampal pyramidal neuron injury assessment, in which the percentage of CA1 neuronal loss is represented on a 0–10 scale: 0 = no injury; 1–2 = minimal injury (<20% affected neurons); 3–4 = mild injury (20–35%); 5–6 = moderate injury (35–55%); 7–8 = severe injury (55–75%); 9–10 = very severe injury (>75%). Two blinded investigators independently evaluated five non-overlapping fields per section (300× magnification), and the mean score was calculated for each rat [28].

Hippocampal tissue was fixed in 2.5% glutaraldehyde (Servicebio Technology, Wuhan, Hubei, China; catalog No. G1102) at 3 °C for 23 h before being processed as follows: samples were rinsed three times (15 min each) with 0.1 M phosphate buffer, followed by post-fixation with 1% osmium tetroxide (Ted Pella Inc., Redding, CA, USA; catalog No. 18456) for 2 h. The specimens were subsequently dehydrated through a graded acetone series (50%, 70%, and 90%, 15 min each; Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China; catalog No. 10000418), immersed three times in 100% acetone (20 min each), embedded in epoxy resin (SPI Supplies, West Chester, PA, USA; catalog No. 90529-77-4), infiltrated overnight, and polymerized at 60 °C for 38 h. Ultrathin sections (thickness: 70 nm) were prepared and stained for 30 min with uranyl acetate (Ted Pella Inc., Redding, CA, USA; catalog No. NC1630603), followed by lead citrate (Ted Pella Inc., Redding, CA, USA; catalog No. 19312) for 15 min.

Hippocampal CA1 neuronal ultrastructure was examined at 8000× magnification with an HT7800 transmission electron microscope (Hitachi, Tokyo, Japan). For quantitative analysis, six rats per group were randomly selected, and 5–10 non‑overlapping fields per rat were photographed. All microscopic tests were independently performed by two researchers blinded to treatment allocation. Identification of APs strictly adhered to the international guidelines [29]. APs were defined as double-membraned vesicles, with a diameter ranging from 0.5 to 1.5 μm, containing cytoplasmic components, such as undegraded organelles (e.g., mitochondria and fragments of the endoplasmic reticulum). Secondary lysosomes and multivesicular bodies (MVBs) were explicitly excluded from this classification.

Autophagic vacuoles were identified and quantified based on their ultrastructural morphology. APs were defined as double-membraned crescent-shaped or circular vesicles that enclosed cytoplasmic contents (e.g., organelles). Autolysosomes (ALs) were characterized as single-membrane structures containing electron-dense material at various stages of degradation. The identification and counting of these vacuoles were performed independently by two investigators, both blinded to treatment allocation. The total number of autophagic vacuoles (AVs) per microscopic field was determined by summing the counts of APs and ALs.

The levels of interleukin (IL)-1β, IL-6, IL-18, and tumor necrosis factor-α (TNF-α) in hippocampal tissues were quantified using commercially available ELISA kits (Solarbio, Beijing, China): IL-1β (catalog No. SEKR-0002), IL-6 (catalog No. SEKR-0005), IL-18 (catalog No. SEKR-0054), and TNF-α (catalog No. SEKR-0009). In brief, approximately 50 mg of tissue was homogenized in ice-cold physiological saline and centrifuged at 3 °C for 10 min at 12,000 ×g to collect the supernatant. Standard curves were generated using serially diluted standards according to the manufacturer's instructions. Both samples and standards were incubated in pre-coated wells. The assay procedure involved sequential incubation with a biotinylated detection antibody, streptavidin-conjugated horseradish peroxidase (HRP), and tetramethylbenzidine substrate. The absorbance at 350-nm wavelength was measured. The final cytokine concentrations in the tissue homogenate supernatant were expressed as ng/L.

Hippocampal tissue was flash-frozen in liquid nitrogen and minced. The fragments were homogenized in ice-cold radioimmunoprecipitation assay (RIPA) lysis buffer (Meilunbio, Dalian, Liaoning, China; catalog No. MA0151) containing phosphatase/protease inhibitors at a 1:10 (w/v) tissue-to-buffer ratio. Mechanical homogenization was performed under ice-cold conditions with three cycles of 30-sec homogenization interspersed with 60-sec cooling intervals. Centrifugation of the homogenates was conducted at 16,200 ×g for 15 min at 3 °C (centrifugal radius, 10 cm). The supernatant was retained as the total protein fraction, and protein concentration was quantified by a bicinchoninic acid assay. Protein samples (50 μg per lane) were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis SDS-PAGE (Sinopharm Group Chemical Reagent Co., Ltd., Shanghai, China; SDS catalog No. 30166428) for 1.5 h at a constant voltage of 120 V. Subsequently, proteins were transferred onto 0.35 μm polyvinylidene fluoride membranes (0.35 μm; Merck Millipore, Billerica, MA, USA; catalog No. IPVH00010) using a wet transfer system (Genscript, Nanjing, Jiangsu, China; catalog No. L00686C) for 90 min at constant 300 mA. The membranes were blocked at room temperature with 5% skim milk (Servicebio, Wuhan, Hubei, China; catalog No. G5002) for 1 h and then incubated overnight at 3 °C with the corresponding primary antibodies (Servicebio; diluted 1:1000), including PI3K, Akt, mTOR, phosphorylated-PI3K (p-PI3K), phosphorylated-Akt (p-Akt), p-mTOR (p-mTOR), Microtubule-associated protein 1A/1B-light chain 3 (LC3), p62, Beclin-1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Membranes were thrice washed (each 10 min) with Tris-buffered saline with Tween (TBST) after warming to room temperature for 30 min and incubated with HRP-conjugated goat anti-rabbit secondary antibody (Servicebio, Wuhan, Hubei, China; catalog No. GB23303; 1:5000) for 1 h at room temperature. Following additional TBST washing, protein bands were visualized using an enhanced chemiluminescence substrate (Thermo Fisher Scientific, Waltham, MA, USA; catalog No. 32106). Signal detection was performed using a ChemiDoc MP Imaging System (Bio-Rad Laboratories Inc., Hercules, CA, USA), and band intensities were quantified through Image Lab 6.1 (Bio-Rad Laboratories Inc.). The expression of target proteins was normalized to GAPDH. The activity of phosphorylated proteins was expressed as the ratio of the phosphorylated protein to the total protein (p-protein/total protein).

A tyramide signal amplification (TSA)-based multiplex fluorescence staining propofol was employed for dual-labeling of hippocampal sections. After paraffin removal and tissue rehydration, antigen retrieval was achieved by heating the sections in EDTA buffer (pH 9.0; EDTA buffer (pH 9.0; Abcam, Cambridge, UK; catalog No. AB93684)) at 95 °C. 3% H₂O₂ was used to block endogenous peroxidase, after which nonspecific binding was blocked with normal goat serum (Boster Biological Technology, Wuhan, Hubei, China; catalog No. AR1009). Tissue sections were initially incubated overnight at 3 °C with a rabbit monoclonal antibody against microtubule-associated protein 1A/1B-light chain 3B (LC3B, Cell Signaling Technology, Danvers, MA, USA; catalog No. 3868T; 1:1000). Signal amplification was achieved using the appropriate HRP-conjugated secondary antibody (Servicebio Technology, Wuhan, Hubei, China; catalog No. GB23303) and TYR-520Plus fluorescent dye (excitation/emission: 390/520 nm; Ruchuang Biotech, Shanghai, China; catalog No. RC0086Plus-34RM). For subsequent labeling, antibody elution was performed using an antigen retrieval buffer. Sections were thereafter incubated with rabbit anti-lysosomal-associated membrane protein 1 (LAMP1) polyclonal antibody (Bioss, Beijing, China; catalog No. bs-1970R; 1:1000) under identical conditions, followed by visualization with TYR-570Plus fluorescent dye (excitation/emission: 550/570 nm; Ruchuang Biotech, Shanghai, China; catalog No. RC0086Plus-34RM). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; Beyotime Biotechnology, Shanghai, China; catalog No. C1002). All fluorescence images were acquired by a Nikon fluorescence microscope (Nikon, Tokyo, Japan) under consistent parameters. For each group, at least three randomly selected rats were analyzed, involving three non-overlapping fields of view captured from the hippocampal CA1 region per rat (300× magnification). Semi-quantitative analysis of LC3B and LAMP1 fluorescence signals was performed using Image-Pro Plus 6.0 (Media Cybernetics, Rockville, MD, USA). The mean fluorescence intensity was obtained by measuring the integrated optical density and area for each field of view. The final value for each sample represents the average of measurements taken from three fields.

Continuous data were expressed as mean ± standard deviation. For the MWM acquisition phase (escape latency), a two‑way repeated‑measures ANOVA (group × day) was performed. For all other comparisons (e.g. platform crossings, Y-maze test, Western blotting), one‑way ANOVA followed by Tukey’s post‑hoc test was used; when assumptions were not met, the Kruskal-Wallis test was applied. The homogeneity of variances and assumptions of normality were assessed separately for each continuous outcome measure (e.g., escape latency, platform crossings) using the Levene’s test and the Shapiro-Wilk test, respectively. If either the normality or homoscedasticity assumption was violated, the non-parametric Kruskal-Wallis H test was alternatively utilized. For categorical data, the Chi-square test was applied. The Fisher’s exact test was used in cases where the expected frequencies were <5. Statistical significance was determined when a two-tailed p-value was <0.05. Statistical analysis was carried out using SPSS 26.0 software (IBM Corp, Armonk, NY, USA). Individual data points are displayed in all graphs. The behavioral tests and histological evaluation were entirely conducted under blinded conditions. Original datasets were stored on institutional servers and are available upon reasonable request in accordance with academic practice.

Behavioral testing demonstrated that esketamine attenuated propofol-induced long-term cognitive deficits. In the Y-maze test (Fig. 1A,B), rats in the Pfo group required significantly more training trials and showed a significantly lower correct response rate than controls (both p < 0.01), indicating impaired learning and memory. Esketamine significantly improved both measures (p < 0.01 vs. Pfo), whereas co-administration of LY294002 (PI3K inhibitor) or rapamycin (autophagy activator) abolished these improvements. The LY294002+Esk+Pfo and Rapa+Esk+Pfo groups exhibited no between-group differences, while their performance on all key metrics was significantly worse than that in the Esk+Pfo group (p < 0.01).

Fig. 1.

Esketamine could ameliorate propofol-induced long-term learning and memory impairments, as evidenced by the Y-maze and Morris water maze tests. (A) Number of training trials required to reach the learning criterion. (B) Correct response rate (%). (C) Number of platform crossings. (D) Escape latency (s). (A–C) One‑way ANOVA with Tukey’s test; (D) Two‑way repeated‑measures ANOVA with Tukey’s test. Data were presented as mean ± SD (n = 12). **p < 0.01 vs. Con; ^##p < 0.01 vs. Pfo; ^&&p < 0.01 vs. Esk+Pfo. SD, standard deviation; Con, Control group; Pfo, Propofol group; Esk, Esketamine; Rapa, Rapamycin.

Consistent findings were obtained using the MWM test (Fig. 1). Propofol significantly reduced platform crossings in the probe trial (Fig. 1C, p < 0.01) and prolonged escape latency during the 5-day acquisition phase (Fig. 1D, p < 0.01) compared with the Con group. Esketamine (Esk+Pfo group) markedly improved spatial learning and memory performance (p < 0.01), whereas these protective effects were abolished when co-administered with LY294002 or rapamycin. The LY294002+Esk+Pfo and Rapa+Esk+Pfo groups exhibited no significant difference over the Pfo group on either measure and performed significantly worse than the Esk+Pfo group (p < 0.01).

Collectively, esketamine could alleviate propofol-induced cognitive impairment, and this protective effect was abolished by either PI3K inhibition or autophagy activation. These results indicate a mechanism dependent on activation of the PI3K/Akt/mTOR pathway and suppression of excessive autophagy.

Histopathological changes in hippocampal neurons were assessed by H&E staining (Fig. 2A). Compared with the well-preserved neuronal architecture observed in the Con group, the Pfo group exhibited remarkable pathological changes, including disorganized neuronal arrangement, cell dispersion (black arrow), and increased glial cell density (yellow arrow). These abnormalities were partially reversed by esketamine treatment. The protective effect of esketamine was significantly attenuated by co-administration of LY294002 or rapamycin. Both the LY294002+Esk+Pfo and Rapa+Esk+Pfo groups displayed severe neuronal damage comparable to the Pfo group, including cell dispersion and pyknosis. The Rapa+Pfo group similarly exhibited remarkable neuronal abnormalities, including nuclear condensation (blue arrow), cytoplasmic eosinophilia, structural disorganization, and glial proliferation, with injury severity comparable to the Pfo group. Semi‑quantitative histopathological scores (Fig. 2B) confirmed these findings, reflecting significant protection by esketamine that was abolished by either LY294002 or rapamycin (p < 0.01).

Fig. 2.

Neuroprotective effects of esketamine against propofol-induced neurotoxicity. (A) Representative H&E staining images of the hippocampal CA1 region from each group. Black arrows indicate neuronal dispersion with unclear boundaries; blue arrows indicate neuronal pyknosis with cytoplasmic eosinophilia; yellow arrows indicate increased glial cell density. Scale bar: 50 μm (300× magnification). (B) Semi-quantitative histopathological scores. Data were presented as mean ± SD (n = 6). **p < 0.01 vs. Con; ^##p < 0.01 vs. Pfo; ^&&p < 0.01 vs. Esk+Pfo. H&E, hematoxylin and eosin; CA1, Cornu Ammonis area 1.

Collectively, these histopathological findings demonstrated that esketamine attenuated propofol-induced structural damage in hippocampal neurons, and this protective effect was reversed by either PI3K inhibition or autophagy activation.

TEM analysis revealed that esketamine preserved neuronal ultrastructure and reduced propofol-induced autophagic vacuole accumulation (Fig. 3A). The Con group displayed intact mitochondrial cristae and tightly attached ribosomes. In contrast, the Pfo group exhibited significant ultrastructural damage, including reduced and shortened mitochondrial cristae, dilation of the rough endoplasmic reticulum, and formation of APs, ALs, and MVBs. Esketamine markedly restored mitochondrial and endoplasmic reticulum integrity. Only occasional MVBs were observed, and no typical APs were identified, indicating an ultrastructural profile comparable to that of the controls. The protective effect of esketamine was abolished by co-administration of LY294002 or rapamycin. Both the LY294002+Esk+Pfo and Rapa+Esk+Pfo groups displayed ultrastructural damage similar to the Pfo group, with increased numbers of MVBs and ALs. Quantitative analysis confirmed that esketamine significantly attenuated the propofol-induced increase in autophagic vacuoles (AVs; defined as the sum of autophagosomes [APs] and autolysosomes [ALs]) per microscopic field, whereas this effect was abolished by treatment with LY294002 or rapamycin (Fig. 3B).

Fig. 3.

Ultrastructural examination of hippocampal tissue. (A) Representative transmission electron microscopic images from each group. n = 6 rats per group. Key structures are indicated: mitochondria (M), rough endoplasmic reticulum (RER), multivesicular bodies (MVBs), Golgi apparatus (Go), autolysosomes (ALs), and autophagosomes (APs). Scale bar = 1 μm. (B) Quantitative analysis of autophagic vacuoles (AVs) per field, where AVs = APs + ALs. Data were presented as mean ± SD (n = 6 rats per group, with 5–10 fields per rat). **p < 0.01 vs. Con; ^##p < 0.01 vs. Pfo; ^&&p < 0.01 vs. Esk+Pfo. TEM, transmission electron microscopy.

Collectively, these data demonstrated that esketamine preserved mitochondrial and endoplasmic reticulum integrity in propofol-exposed hippocampal neurons, and this protection was mediated by activation of the PI3K/Akt/mTOR signaling pathway, thereby limiting excessive autophagy.

To assess neuroinflammatory responses, pro-inflammatory cytokine levels in hippocampal tissue were measured (Fig. 4). Relative to the Con group, hippocampal levels of IL-1β, IL-6, IL-18, and TNF-α (Fig. 4A–D) were significantly increased following propofol exposure (all p < 0.01). Esketamine significantly reduced levels of all four cytokines (p < 0.01), whereas PI3K inhibition or autophagy activation remarkably abolished this anti-inflammatory effect (p < 0.01). These results demonstrated that esketamine alleviated propofol-induced neuroinflammation through a mechanism dependent on PI3K pathway activation and autophagy suppression.

Fig. 4.

Esketamine attenuates the inflammatory response in propofol-induced neurotoxicity. Concentrations of hippocampal inflammatory cytokines, including IL-1β (A), IL-6 (B), IL-18 (C), and TNF-α (D). Data were presented as mean ± SD (n = 6). **p < 0.01 vs. Con; ^##p < 0.01 vs. Pfo; ^&&p < 0.01 vs. Esk+Pfo. IL, interleukin; TNF-α, tumor necrosis factor-α.