Environmental enrichment mitigates sevoflurane-induced neurodevelopmental injury via cGAS–STING–dependent microglial modulation Authors: Feixiang Li, Bingqing Gong, Haiyan Wu, Yongyan Yang, Ying Luo, Yonghao Yu, Ding Li Published: 2026 ABSTRACT BackgroundNeonatal exposure to sevoflurane has been implicated in long-term neurodevelopmental abnormalities, yet the underlying mechanisms remain unresolved. This study sought to determine whether cGAS–STING–mediated microglial activation and aberrant synaptic pruning underlie sevoflurane-induced cognitive deficits and to assess how environmental conditions modulate these processes.MethodsNeonatal mice underwent sevoflurane exposure followed by rearing in enriched (EE) or impoverished (IE) environments. Cognitive function, synaptic structure, microglial activity, mitochondrial status, and cGAS–STING signaling were evaluated using behavioral tests, immunostaining, biochemical assays, and pharmacological inhibition.ResultsSevoflurane exposure induced cognitive impairment, microglial overactivation, mitochondrial dysfunction, and excessive synaptic pruning resulting from microglial overactivation. EE mitigated these abnormalities by preserving mitochondrial integrity and reducing mtDNA-driven cGAS–STING activation, thereby preventing the microglia-mediated imbalance in synaptic pruning and improving cognitive outcomes. In contrast, IE exacerbated mitochondrial injury, aggravated synaptic loss, and further worsened cognitive impairment.ConclusionSevoflurane disrupts neurodevelopment through a mitochondria–cGAS–microglia–synapse pathway. Environmental enrichment offers significant neuroprotection, highlighting both cGAS–STING signaling and early-life environmental modulation as promising targets for preventing anesthesia-related neurodevelopmental injury.Graphical abstract Introduction The neonatal period is a critical stage of brain development characterized by rapid neuronal growth, synaptogenesis, and synaptic refinement [1]. Proper regulation of synaptic pruning during this period is essential for the establishment of functional neural circuits and long-term cognitive competence. Disruption of these processes can result in persistent neurodevelopmental abnormalities [2, 3]. Increasing evidence has shown that repeated exposure to the commonly used anesthetic sevoflurane during early postnatal life induces neurodevelopmental toxicity, manifested as impaired synaptic structure and function, excessive synaptic pruning, and subsequent cognitive deficits in later stages of life [4, 5]. Microglia play a fundamental role in brain development by eliminating redundant synapses, thereby refining neural circuits during critical postnatal periods [6]. Exposure to sevoflurane during early postnatal development has been shown to alter synaptic structure and function in neonatal mice [7, 8]. These observations suggest that sevoflurane may interfere with the normal process of synaptic pruning, potentially by affecting microglial activity, leading to excessive synaptic elimination and subsequent cognitive dysfunction [9]. The cGAS–STING pathway is a key intracellular DNA–sensing system that coordinates innate immune responses to cellular stress and damage. Recent studies have identified cGAS–STING signaling as an important regulator of microglial activation [10–12]. Given that sevoflurane exposure can induce cellular stress and neuroinflammation, we hypothesize that it may activate the cGAS–STING pathway, leading to microglial overactivation and aberrant synaptic pruning. Consequently, excessive cGAS–STING–mediated microglial activity may contribute to cognitive impairment in the developing brains of neonatal mice. Environmental conditions play a critical role in shaping neural development and cognitive function during early life [13]. Environmental enrichment (EE), which provides enhanced sensory, cognitive, and physical stimulation, has been widely demonstrated to promote neurogenesis, synaptic plasticity, and mitochondrial function, thereby improving cognitive outcomes in various models of neurodevelopmental injury [14]. In contrast, an impoverished environment (IE), characterized by sensory deprivation and limited physical and social stimulation, has been associated with delayed neuronal maturation, reduced synaptic density, and impaired cognitive performance [15]. Recent studies suggest that EE can attenuate neuroinflammation and support synaptic recovery following exposure to anesthetics or neurotoxic agents, whereas IE may exacerbate neural damage and worsen functional deficits [16–18]. Our previous studies have demonstrated that EE can attenuate the neurotoxic effects of repeated sevoflurane exposure in neonatal mice. Based on these findings, we hypothesize that EE mitigates sevoflurane-induced cognitive deficits by suppressing cGAS–STING–mediated microglial activation and excessive synaptic pruning, whereas an impoverished environment may exacerbate these effects through enhanced mitochondrial dysfunction, cGAS–STING activation, and aberrant synaptic remodeling. Elucidating the mechanisms of sevoflurane-induced neurotoxicity and the protective role of environmental enrichment will advance our understanding of anesthetic-related developmental deficits and may inform strategies to prevent neurodevelopmental disorders in infants and young children. Materials and methods Animals Neonatal C57BL/6J mice were obtained from pregnant dams purchased from Huafukang Biotechnology Co., Ltd. (Beijing, China). Animals were maintained under controlled conditions (22 ± 2 °C, 50–60% humidity) on a 12 h light/dark cycle with ad libitum access to standard chow and water. Procedures were approved by the Institutional Animal Care and Use Committee and adhered to relevant ethical guidelines. Offspring derived from pregnant dams were used for all experiments. Because the sex of the offspring is randomly distributed at birth, both male and female pups were included in the study. Sevoflurane exposure, RU.521 treatment, and environmental intervention Sevoflurane exposure was performed as previously described [3]: mice received 3% sevoflurane in 60% oxygen for 2 h daily over 3 consecutive days, while controls were exposed to 60% oxygen alone. The cGAS inhibitor RU.521 (5 mg/kg) was administered intraperitoneally 2 h before each sevoflurane exposure. Based on our previous studies and existing literature, environmental interventions were implemented with slight modifications [3]. The EE included a variety of stimulating elements such as differently shaped toys, running wheels, ladders, tunnels, and platforms to provide enhanced sensory, cognitive, and social stimulation. Cages offered sufficient space to promote voluntary exploration and physical activity. The standard environment (SE) consisted of conventional cages with standard bedding and no additional objects. In contrast, the IE utilized smaller cages containing minimal bedding, without any toys or enrichment materials, thereby restricting physical activity and sensory stimulation. Starting at 8:00 a.m. daily, neonatal mice and their dams were placed in the assigned environmental conditions for 2 h per day during the 3 weeks following sevoflurane exposure. After each session, the pups and their mothers were returned to standard housing. On postnatal day 21 (P21), mice were weaned and housed according to sex. Littermates were group-housed under SE or EE conditions, whereas mice in the IE group were housed individually. All mice were subjected to behavioral testing or brain tissue collection 3 weeks after the completion of sevoflurane or oxygen exposure. Western blotting Hippocampal tissue was isolated and lysed in buffer containing protease and phosphatase inhibitors. Tissue was sonicated on ice, centrifuged at 12,000 rpm for 10 min at 4 °C, and the supernatant was collected. Protein concentrations were determined using a BCA assay, and equal amounts were loaded onto a 4–12% SDS-PAGE gel, separated, and transferred to PVDF membranes. Membranes were blocked with 5% skim milk for 2 h at room temperature, incubated overnight at 4 °C with primary antibodies (cGAS, 1:3000, Proteintech; STING, 1:3000, Proteintech; GAPDH/β-actin, 1:5000, Affinity), washed three times with TBST, and incubated with HRP-conjugated secondary antibodies for 2 h at room temperature. After additional TBST washes, bands were visualized using ECL and imaged. Semi-quantitative analysis of target protein levels relative to loading controls was performed using ImageJ. Immunofluorescence Mouse brains were fixed in 4% paraformaldehyde for 24 h, cryoprotected in 30% sucrose, and coronally sectioned at 20 μm. Free-floating sections were permeabilized with 0.3% Triton X-100 and blocked with 10% goat serum for 1 h at room temperature. Sections were then incubated overnight at 4 °C with primary antibodies against Iba1 (1:300), CD68 (1:200) and PSD95 (1:200, Abcam). Following PBS washes, sections were incubated with appropriate fluorescent secondary antibodies for 2 h at room temperature in the dark. Nuclei were counterstained with DAPI, sections were mounted, and images were acquired using a confocal laser scanning microscope. Images were acquired using an Olympus fluorescence microscope. All imaging parameters, including illumination intensity, exposure time, gain, and acquisition settings, were kept identical across experimental groups. Post-acquisition image processing, including adjustments to brightness, contrast, and saturation, was applied uniformly to all images. High-resolution images show representative synaptic puncta and microglial morphology across experimental groups. Transmission electron microscopy Ultrastructural analysis was performed using transmission electron microscopy (TEM). Brain tissue was fixed in 2.5% glutaraldehyde at 4 °C overnight, post-fixed in 1% osmium tetroxide for 1 h, and dehydrated through a graded ethanol series. Samples were infiltrated with epoxy resin, polymerized at 60 °C for 48 h, and sectioned at 70–90 nm. Ultrathin sections were double-stained with 2% uranyl acetate and lead citrate for 10 min each, then examined by TEM, and images were captured for analysis. MWM Spatial learning and memory were assessed using the classic Morris Water Maze (MWM). The 5-day training phase involved four daily trials in a water pool (22–24 °C) divided into quadrants, with a hidden platform submerged 1–2 cm below the surface. Mice were released from random start points and guided by visual cues; latency and swimming paths to reach the platform were recorded. On day 6, a probe trial without the platform measured memory retention by recording time spent in the target quadrant and the number of platform crossings. Y-maze The Y-maze test was performed as previously described [19]. The maze comprises three arms (40 cm × 10 cm × 15 cm) arranged at 120°. Mice were acclimated to the laboratory for 1 h. During habituation, the novel arm was blocked, and mice explored the start and familiar arms for 10 min. After 1 h, the novel arm was opened for a 5-min test, and behavioral metrics included time spent and distance traveled in the novel arm. Novel object recognition Recognition memory was assessed using the Novel object recognition test (NOR). All experiments were conducted in a quiet, dimly lit room. Mice were habituated to an open-field arena (40 × 40 × 40 cm) for 10 min without objects. During the training phase, two identical objects were placed in the arena, and mice explored freely for 10 min. After a 24 h retention interval, one familiar object was replaced with a novel object of similar size but different shape and texture. In the 5-min test phase, exploration time for each object—defined as nose-oriented contact or sniffing within 2 cm—was recorded using a video tracking system. Objects and the arena were cleaned with 75% ethanol between trials to eliminate olfactory cues. Assessment of mitochondrial function Hippocampal ROS and ATP levels were measured using commercial ELISA kits (Beyotime, Shanghai, China) following the manufacturers’ instructions. Mitochondria were isolated from hippocampal tissue, and mitochondrial membrane potential (MMP) was assessed using a dedicated assay kit (Solarbio, Beijing, China) according to the manufacturer’s protocols. Cytosolic mtDNA measurement Cytosolic genomic DNA was isolated from hippocampal tissue according to the manufacturer’s instructions. Cytosolic mitochondrial DNA (mtDNA) release was quantified by qPCR using mitochondria-specific primers targeting ND1 and CytB as previously described [12]. Statistical analysis Data were analyzed using GraphPad Prism 9.0 (GraphPad Software, USA). Normality was assessed with the Shapiro–Wilk test. Two-group comparisons used an independent t-test for normally distributed data or the Mann–Whitney U test for non-normal data. Multi-group comparisons with two independent factors were analyzed by two-way ANOVA followed by Bonferroni post hoc test when data were normal, or by the Kolmogorov–Smirnov test when non-normal. Data are presented as mean ± SD, with significance set at P < 0.05. Results Sevoflurane induces cGAS–STING activation, microglial response, and excessive synaptic pruning in neonatal mice To test our hypothesis, we first examined the effects of repeated sevoflurane exposure on the cGAS–STING pathway, with the experimental timeline shown in Fig. 1A. We observed that multiple sevoflurane exposures markedly increased the protein expression levels of cGAS (Fig. 1B, C; t(10) = − 3.237, P = 0.009) and STING (Fig. 1B, D; t(10) = − 2.935, P = 0.015) in the hippocampus of neonatal mice. Immunofluorescence analysis revealed that repeated sevoflurane exposure increased the number of microglia (Fig. 1E, F; t(10) = − 2.388, P = 0.038) and enhanced their phagocytic activity (Fig. 1E, G; t(10) = − 2.765, P = 0.02). We further observed that repeated sevoflurane exposure increased the points of contact between microglial markers and PSD95, indicating enhanced microglial engulfment of postsynaptic elements (Fig. 1H, I; t(10) = − 2.41, P = 0.037). Fig. 1Sevoflurane induces microglial activation, excessive synaptic pruning, and cognitive deficits in neonatal mice. A Experimental timeline depicting neonatal sevoflurane exposure and subsequent analyses. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). E–G Immunofluorescence images of Iba1⁺ microglia (red arrows) and the phagocytic marker CD68 in dorsal hippocampal CA1 (dCA1) (white arrows), with quantification of microglial number and Iba1–CD68 colocalization (n = 6/group). H, I Iba1 colabeled with the postsynaptic marker PSD95 in dCA1 (white arrows), with quantification of microglia–synapse interactions (n = 6/group). J–L Y-maze performance, including representative trajectories, time in the novel arm, and total locomotor distance (n = 10/group). M–O Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant Sevoflurane induces microglial activation, excessive synaptic pruning, and cognitive deficits in neonatal mice. A Experimental timeline depicting neonatal sevoflurane exposure and subsequent analyses. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). E–G Immunofluorescence images of Iba1⁺ microglia (red arrows) and the phagocytic marker CD68 in dorsal hippocampal CA1 (dCA1) (white arrows), with quantification of microglial number and Iba1–CD68 colocalization (n = 6/group). H, I Iba1 colabeled with the postsynaptic marker PSD95 in dCA1 (white arrows), with quantification of microglia–synapse interactions (n = 6/group). J–L Y-maze performance, including representative trajectories, time in the novel arm, and total locomotor distance (n = 10/group). M–O Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant Repeated sevoflurane anesthesia induces cognitive impairment in neonatal mice We further assessed the effects of repeated sevoflurane exposure on cognitive function in neonatal mice using behavioral tests. In the Y-maze test, repeated sevoflurane exposure significantly reduced the time spent exploring the novel arm (Fig. 1J, K; t(18) = 2.808, P = 0.012) without affecting the total distance traveled (Fig. 1J, L; t(18) = 0.413, P = 0.684). In the MWM, repeated sevoflurane exposure decreased the time spent in the target quadrant (Fig. 1M, N; t(18) = 3.785, P = 0.001) and the number of platform crossings (Fig. 1M, O; t(18) = 3.851, P = 0.001). These findings indicate that repeated sevoflurane exposure impairs cognitive function in neonatal mice without altering their locomotor activity. RU.521 attenuates microglial activation and excessive synaptic pruning induced by repeated sevoflurane exposure To investigate the role of the cGAS–STING pathway in microglial activation and synaptic pruning, we employed RU.521, a specific inhibitor of cGAS. The experimental design is illustrated in Fig. 2A. Western blot analysis showed that RU.521 did not affect the basal expression of cGAS (Fig. 2B, C; F(1,20) = 0.038, P = 0.848) or STING (Fig. 2B, D; F(1,20) = 0.002, P = 0.966) in the hippocampus of control neonatal mice but markedly reduced the elevated levels of cGAS (F(1,20) = 5.158, P = 0.034) and STING (F(1,20) = 6.924, P = 0.016) induced by repeated sevoflurane exposure. We further found that RU.521 did not affect the total number of microglia (Fig. 2E, H; F(1,20) = 0.002, P = 0.966) or the number of phagocytic microglia (Fig. 2G, H; F(1,20) = 0.049, P = 0.827) in the hippocampus of control mice. However, RU.521 significantly reduced both the total number of microglia (F(1,20) = 6.924, P = 0.016) and the number of phagocytic microglia (F(1,20) = 5.295, P = 0.032) in the hippocampus of mice subjected to repeated sevoflurane exposure. We further found that RU.521 did not affect the colocalization of microglial markers with PSD95 in the hippocampus of control mice (Fig. 2G, I; F(1,20) = 0.929, P = 0.347), but it significantly reduced such colocalization in mice exposed to repeated sevoflurane (F(1,20) = 6.278, P = 0.021), indicating that RU.521 markedly suppressed microglia-mediated engulfment of postsynaptic structures. Fig. 2RU.521 attenuates sevoflurane-induced microglial activation and synaptic pruning.A Experimental timeline for RU.521 administration. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). E–H Immunofluorescence images of Iba1⁺ microglia (red arrows) and CD68 in dCA1 (white arrows), with quantification of microglial number and Iba1–CD68 colocalization (n = 6/group). (G, I) Iba1 colabeled with PSD95 in dCA1 (white arrows), with quantification of microglia–synapse contacts (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant RU.521 attenuates sevoflurane-induced microglial activation and synaptic pruning.A Experimental timeline for RU.521 administration. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). E–H Immunofluorescence images of Iba1⁺ microglia (red arrows) and CD68 in dCA1 (white arrows), with quantification of microglial number and Iba1–CD68 colocalization (n = 6/group). (G, I) Iba1 colabeled with PSD95 in dCA1 (white arrows), with quantification of microglia–synapse contacts (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant RU.521 ameliorates sevoflurane-induced cognitive impairment and restores synaptic and dendritic spine density We further evaluated the effects of RU.521 on cognitive function using behavioral analyses. In the Y-maze test, RU.521 had no effect on the time spent exploring the novel arm in control mice (Fig. 3A, C; F(1,36) = 0.047, P = 0.83). However, it significantly increased the time spent exploring the novel arm in mice exposed to repeated sevoflurane (F(1,36) = 10.08, P = 0.003), without affecting the total distance traveled in the maze (Fig. 3A, D; F(1,36) = 0.028, P = 0.869). Similarly, in the MWM, RU.521 did not alter the number of platform crossings (Fig. 3B, E; F(1,36) = 0.195, P = 0.661) or the time spent in the target quadrant (Fig. 3B, F; F(1,36) = 0.35, P = 0.558) in control mice. In contrast, RU.521 significantly increased both the number of platform crossings (F(1,36) = 9.564, P = 0.004) and the time spent in the target quadrant (F(1,36) = 11.835, P = 0.001) in mice subjected to repeated sevoflurane exposure. Fig. 3Behavioral deficits and synaptic structural impairments following neonatal sevoflurane exposure.A, C, D Y-maze performance, including representative trajectories, time in the novel arm, and total locomotor distance (n = 10/group). B, E, F Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). (G, H) Electron microscopy images of synaptic ultrastructure in dCA1 (red arrows), with quantification of synapse number (n = 6/group). (I) Golgi staining in dCA1, showing representative dendritic spines and quantification of spine density (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant Behavioral deficits and synaptic structural impairments following neonatal sevoflurane exposure.A, C, D Y-maze performance, including representative trajectories, time in the novel arm, and total locomotor distance (n = 10/group). B, E, F Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). (G, H) Electron microscopy images of synaptic ultrastructure in dCA1 (red arrows), with quantification of synapse number (n = 6/group). (I) Golgi staining in dCA1, showing representative dendritic spines and quantification of spine density (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ns, not significant To further examine the effects of RU.521 on synaptic integrity, we analyzed the synaptic ultrastructure in the dCA1 region using transmission electron microscopy. RU.521 did not affect synapse number in control mice (Fig. 3G, H; F(1,44) = 0.001, P = 0.972). Repeated sevoflurane exposure significantly reduced synapse number (F(1,44) = 8.258, P = 0.006), and this reduction was reversed by RU.521 treatment (F(1,44) = 6.623, P = 0.013). Golgi staining further revealed that RU.521 had no effect on dendritic spine density in control mice (Fig. 3I; F(1,44) < 0.001, P = 1). However, repeated sevoflurane exposure markedly decreased spine density in the dCA1 region (F(1,44) = 8.897, P = 0.007), which was also reversed by RU.521 (F(1,44) = 7.353, P = 0.013). Collectively, these findings demonstrate that RU.521 effectively ameliorates sevoflurane-induced cognitive impairment and restores synaptic and dendritic integrity in the hippocampus of neonatal mice. Environmental enrichment attenuates sevoflurane-induced activation of the cGAS–STING pathway and microglia, and reduces excessive microglial synaptic pruning To investigate the effects of environmental conditions on the cGAS–STING pathway, microglial activation, and microglia-mediated synaptic pruning, mice were housed in either enriched or impoverished environments following repeated sevoflurane exposure. The experimental timeline is shown in Fig. 4A. Environmental enrichment attenuated the sevoflurane-induced upregulation of hippocampal cGAS (Fig. 4B, C; P = 0.036) and STING (Fig. 4B, D; P = 0.014) proteins, whereas an impoverished environment further elevated the expression of cGAS (P = 0.041) and STING (P = 0.041) in sevoflurane-exposed mice. Environmental enrichment also reduced both the total number of microglia (Fig. 4E, H; P = 0.08) and the number of phagocytic microglia (Fig. 4F, H; P = 0.039) in the hippocampus, while the impoverished environment increased both parameters (P = 0.026 and P = 0.041, respectively). Furthermore, environmental enrichment decreased the points of contact between microglial markers and PSD95 in the hippocampus of sevoflurane-exposed mice (Fig. 4G, I; P = 0.019), indicating suppressed microglial engulfment of postsynaptic structures. Fig. 4Effects of environmental change on microglial activation and synaptic pruning after repeated sevoflurane exposure.A Timeline of environmental change and repeated sevoflurane exposure. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). (E, F, H) Iba1⁺ microglia and CD68 immunofluorescence in dCA1, with quantification of microglial number (red arrows) and Iba1–CD68 colocalization (white arrows) (n = 6/group). (G, I) Iba1 colabeled with PSD95 in dCA1 (white arrows), with quantification of microglia–synapse interactions (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01 Effects of environmental change on microglial activation and synaptic pruning after repeated sevoflurane exposure.A Timeline of environmental change and repeated sevoflurane exposure. B–D Western blots of cGAS and STING in hippocampal tissue, with quantification (n = 6/group). (E, F, H) Iba1⁺ microglia and CD68 immunofluorescence in dCA1, with quantification of microglial number (red arrows) and Iba1–CD68 colocalization (white arrows) (n = 6/group). (G, I) Iba1 colabeled with PSD95 in dCA1 (white arrows), with quantification of microglia–synapse interactions (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01 Collectively, these findings demonstrate that environmental enrichment suppresses microglial activation and prevents excessive synaptic pruning induced by repeated sevoflurane exposure. Environmental enrichment ameliorates sevoflurane-induced cognitive impairment and restores synaptic and dendritic spine density To evaluate the effects of environmental conditions on cognitive function following repeated sevoflurane exposure, we conducted a series of behavioral assessments. In the NOR, repeated sevoflurane exposure significantly reduced the discrimination index in mice (Fig. 5A, C; F(1,36) = 46.639, P < 0.001). Environmental enrichment markedly increased the discrimination index in sevoflurane-exposed mice (P < 0.001), whereas the impoverished environment further decreased it (P = 0.008). Changes in environmental conditions did not affect the total distance traveled during the NOR (Fig. 5A, D; F(2,36) = 2.012, P = 0.149). Consistent results were observed in the MWM. Environmental enrichment increased both the number of platform crossings (Fig. 5B, E; P < 0.001) and the time spent in the target quadrant (Fig. 5B, F; P = 0.021) in sevoflurane-exposed mice, whereas the impoverished environment further reduced these measures (P = 0.031 and P = 0.025, respectively). Fig. 5Effects of environmental change on cognitive function and synaptic structure. A, C, D Novel object recognition test, showing representative trajectories, recognition index, and total locomotor distance (n = 10/group). B, E, F Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). G, H Electron microscopy images of synapses in dCA1 (red arrows), with quantification of synapse number (n = 6/group). I Golgi staining in dCA1, showing representative dendritic spines and quantification of spine density (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ***P < 0.001 Effects of environmental change on cognitive function and synaptic structure. A, C, D Novel object recognition test, showing representative trajectories, recognition index, and total locomotor distance (n = 10/group). B, E, F Morris water maze performance, including representative swim paths, time in the target quadrant, and platform crossings (n = 10/group). G, H Electron microscopy images of synapses in dCA1 (red arrows), with quantification of synapse number (n = 6/group). I Golgi staining in dCA1, showing representative dendritic spines and quantification of spine density (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ***P < 0.001 Transmission electron microscopy analysis revealed that environmental enrichment significantly increased the number of synapses in the dCA1 region of sevoflurane-exposed mice (Fig. 5G, H; P = 0.024), while the impoverished environment further decreased synapse numbers (P = 0.034). Similarly, Golgi staining showed that environmental enrichment markedly increased dendritic spine density in the dCA1 region (Fig. 5I; P = 0.027), whereas the impoverished environment further reduced it (P = 0.034). Together, these findings demonstrate that environmental enrichment mitigates sevoflurane-induced cognitive deficits and restores synaptic and dendritic integrity, while impoverished conditions exacerbate these impairments. Environmental enrichment improves mitochondrial function and morphology Because mitochondrial dysfunction can trigger the release of mitochondrial DNA (mtDNA), which in turn activates the cGAS–STING pathway, we next investigated whether environmental conditions influence mitochondrial dynamics and mtDNA release following sevoflurane exposure. Given the critical role of mitochondrial integrity in neuronal and synaptic function, we further evaluated the impact of environmental enrichment on mitochondrial activity by measuring multiple functional indicators. Repeated exposure to sevoflurane significantly reduced hippocampal ATP levels (Fig. 6A; F(1,20) = 12.165, P = 0.002) and MMP (Fig. 6B; F(1,20) = 8.143, P = 0.01), while markedly increasing ROS production (Fig. 6C; F(1,20) = 12.453, P = 0.002) in mice. Environmental enrichment significantly restored ATP (P = 0.001) and MMP (P = 0.044) levels and concomitantly attenuated ROS accumulation (P = 0.014) in the hippocampus of sevoflurane-exposed mice. In contrast, housing in an impoverished environment further exacerbated the detrimental effects induced by repeated sevoflurane exposure. Fig. 6Effects of environmental change on hippocampal mitochondrial function and morphology. A–C Quantification of hippocampal mitochondrial ATP, mitochondrial membrane potential (MMP), and reactive oxygen species (ROS) following environmental change and sevoflurane exposure (n = 6/group). D, E Electron microscopy of synaptic mitochondria in dCA1 (black arrows), with quantification of damaged mitochondrial membranes (n = 6/group). F, G Western blots of the mitochondrial fission marker p-DRP1Ser616 in hippocampal tissue (white arrows), with quantification (n = 6/group) H–J qPCR analysis of mtDNA copy number, including Nd1, Cytb, and D-loop regions in hippocampal tissue (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ***P < 0.001 Effects of environmental change on hippocampal mitochondrial function and morphology. A–C Quantification of hippocampal mitochondrial ATP, mitochondrial membrane potential (MMP), and reactive oxygen species (ROS) following environmental change and sevoflurane exposure (n = 6/group). D, E Electron microscopy of synaptic mitochondria in dCA1 (black arrows), with quantification of damaged mitochondrial membranes (n = 6/group). F, G Western blots of the mitochondrial fission marker p-DRP1Ser616 in hippocampal tissue (white arrows), with quantification (n = 6/group) H–J qPCR analysis of mtDNA copy number, including Nd1, Cytb, and D-loop regions in hippocampal tissue (n = 6/group). Data are presented as mean ± SD. *P < 0.05, **P < 0.01; ***P < 0.001 To further investigate the effects of environmental conditions on mitochondrial integrity, we performed transmission electron microscopy to evaluate the structural integrity of the mitochondrial membrane. Repeated sevoflurane exposure increased the number of damaged mitochondria in the dCA1 region (Fig. 6D, E; F(1,20) = 8.637, P = 0.008). Environmental enrichment significantly reduced mitochondrial membrane damage (P = 0.017), while an impoverished environment further exacerbated mitochondrial injury (P = 0.006). Because mtDNA leakage can activate the cGAS–STING pathway, we next examined the impact of environmental changes on mitochondrial fission and mtDNA release. Repeated sevoflurane exposure markedly upregulated the expression of phosphorylated DRP1 at Ser616 (p-DRP1Ser616) in the dCA1 region (Fig. 6F, G; F(1,20) = 28.772, P < 0.001), indicating enhanced mitochondrial fission. Environmental enrichment significantly reduced p-DRP1Ser616 expression (P = 0.016), whereas impoverished housing further increased it (P = 0.036). To determine whether mtDNA leaked into the cytoplasm, we isolated cytosolic DNA and measured the levels of mitochondrial genes Nd1, Cytb, and D-loop. Repeated sevoflurane exposure significantly elevated cytosolic Nd1 (Fig. 6H; F(1,20) = 52.657, P < 0.001), Cytb (Fig. 6I; F(1,20) = 198.196, P < 0.001), and D-loop (Fig. 6J; F(1,20) = 147.489, P < 0.001) levels. Environmental enrichment markedly reduced these increases (all P < 0.001), whereas an impoverished environment further enhanced cytosolic Nd1 (P = 0.002), Cytb (P < 0.001), and D-loop (P < 0.001) levels. Collectively, these findings indicate that environmental enrichment ameliorates sevoflurane-induced mitochondrial structural disruption, stabilizes mitochondrial function, and prevents mtDNA leakage. Discussion This study demonstrates that repeated neonatal sevoflurane exposure induces cognitive deficits in mice, linked to cGAS–STING–mediated microglial activation and excessive synaptic pruning. Aberrant microglial activation disrupted normal synaptic refinement, leading to abnormal synapse loss and impaired cognition. Notably, EE alleviated these effects, whereas an IE exacerbated them. Mechanistically, EE suppressed cGAS–STING–dependent microglial activation, restored mitochondrial homeostasis, and preserved synaptic integrity. Collectively, these findings indicate that sevoflurane-induced cGAS–STING activation drives microglia-mediated synaptic loss and cognitive deficits, while environmental modulation represents a promising non-pharmacological strategy to protect neurodevelopment. Microglia play a crucial role in sculpting neural circuits during early postnatal development through activity-dependent synaptic pruning [20, 21]. However, excessive or dysregulated microglial activation can disrupt this finely tuned process, resulting in aberrant synaptic elimination and persistent cognitive deficits [22, 23]. Our findings demonstrate that sevoflurane exposure during the critical window of brain maturation induces pronounced microglial overactivation and excessive synaptic pruning within the hippocampus. Notably, administration of the cGAS inhibitor RU.521 alleviated these effects, indicating that activation of the cGAS–STING pathway serves as a key driver of microglia-mediated synaptic loss and cognitive dysfunction in neonatal mice. The cGAS–STING pathway is a key cytosolic DNA sensor that regulates innate immune responses [24, 25]. In the CNS, mitochondrial injury and mtDNA release can activate cGAS, triggering STING-mediated signaling and proinflammatory gene expression [26–28]. Although extensively studied in infectious and neurodegenerative conditions, our data indicate that sevoflurane-induced mitochondrial dysfunction similarly activates cGAS–STING in developing microglia. This aberrant activation likely drives excessive synaptic pruning, linking mitochondrial stress to maladaptive neural circuit remodeling in the immature brain. Environmental factors exert profound influences on neurodevelopment, especially during periods of heightened plasticity [29, 30]. Consistent with previous studies, we found that EE enhanced cognitive performance and promoted synaptic plasticity, whereas IE exposure impeded these processes [31]. Importantly, EE markedly attenuated cGAS–STING activation and microglial overactivation in sevoflurane-exposed mice, suggesting that environmental stimulation confers neuroprotection by maintaining mitochondrial homeostasis and dampening innate immune signaling. Consistent with our previous work, sevoflurane disrupted mitochondrial membrane potential, elevated ROS production, and impaired ATP synthesis, culminating in neuronal and glial dysfunction [32]. In the present study, EE intervention mitigated these mitochondrial impairments and restored bioenergetic balance, as evidenced by reduced ROS accumulation and improved ATP levels. By stabilizing mitochondrial function, EE likely curtailed mtDNA leakage and prevented aberrant cGAS–STING activation, thereby limiting microglial hyperactivity and excessive synaptic pruning. Collectively, our study provides compelling evidence that neonatal sevoflurane exposure disrupts synaptic development through cGAS–STING–dependent microglial activation and excessive synaptic pruning, leading to enduring cognitive impairment. Environmental enrichment preserves mitochondrial integrity and suppresses cGAS–STING signaling to confer neuroprotection, whereas impoverished rearing conditions exacerbate mitochondrial dysfunction and neuroinflammation. These results identify the cGAS–STING pathway as a critical mediator of anesthesia-induced neurodevelopmental injury and underscore the profound impact of environmental context on shaping brain health during early life. Several limitations of this study should be acknowledged. Although our findings suggest a link between cGAS–STING pathway activation and excessive synaptic pruning, the specific downstream effectors regulating microglial phagocytic activity remain incompletely defined. Furthermore, while these mechanisms were demonstrated in neonatal mice, it remains uncertain whether similar pathways operate in human infants exposed to anesthetics. Importantly, the present study does not directly demonstrate cell type–specific activation of cGAS–STING within microglia. Nevertheless, our in vivo observations indicate a microglia-associated mechanism contributing to synaptic and behavioral alterations. In future studies, isolating microglial cells after in vivo exposure to assess pathway activation in a microglia-specific manner will be a key direction to further strengthen the mechanistic link. Nevertheless, our study provides a mechanistic framework in which mitochondrial dysfunction triggers cGAS–STING–dependent microglial activation, thereby driving excessive synaptic pruning and cognitive impairment. These results advance our understanding of anesthesia-induced neurotoxicity and highlight potential strategies to prevent neurodevelopmental deficits following early-life anesthetic exposure.