Methylene blue reduces postoperative delirium in mice through neuroinflammation suppression and blood-brain barrier repair Authors: Yixu Deng, Jie Hua, Hongcen Pan, Jun Zhang Published: 2025 ABSTRACT Postoperative delirium (POD) is a serious complication following anesthesia and surgery, with currently unclear pathogenesis and no effective treatments. In recent years, emerging evidence implicates that neuroinflammatory and blood-brain barrier (BBB) dysfunction play an important role in the pathogenesis of POD. Methylene blue (MB), an FDA-approved drug, has been reported to have anti-inflammatory and neuroprotective properties. We applied a POD model with a simple laparotomy to assess the POD-like behavior in adult mice. Behavioral tests were performed at 24 h preoperatively, 6 h and 24 h postoperatively. We evaluated neuroinflammation and the reactive states of microglia in the hippocampus and prefrontal cortex. We also evaluated the permeability of the BBB by detecting tight junction proteins, transcytosis, and the leakage of IgG and dextran. Our results found that POD-like behavior in adult mice mainly occurred at 6 h after surgery and gradually subsided at 24 h postoperatively. We found that MB (5 mg/kg) attenuated neuroinflammation and microglia activation in both the hippocampus and prefrontal cortex after anesthesia/surgery. By downregulating the protein levels of matrix metalloproteinase-inducing factor EMMPRIN and matrix metalloproteinase 9 (MMP9), up-regulating the expression of BBB tight junction proteins, and decreasing transcytosis, MB could improve the function of BBB and ameliorate the delirious-like behavior in mice after anesthesia/surgery. Our findings confirm the anti-inflammatory and BBB protective effects of MB in an inflammatory environment in vivo and provide better insights into the pathophysiology and potential treatment of POD.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-26108-8. Introduction Postoperative delirium (POD) is an acute cerebral dysfunction syndrome potentially associated with surgery, anesthesia, pain, comorbidities, or medication, manifesting with fluctuating and concurrent disturbance in cognition, attention, sleep-wake rhythm and consciousness level1. POD is associated with prolonged hospitalizations, increased healthcare costs, functional impairments, and even higher morbidity and mortality2,3. It mainly occur within the first postoperative week across all age groups, elderly population demonstrate heightened susceptibility4,5. Annually expanding surgical populations have driven greater surgical demand and increased POD incidence. Unfortunately, no effective treatments for POD have been identified because of the undefined underlying pathophysiology. Anesthesia and surgery have been reported to induce POD-like behavior in rodents6–8, in which neuroinflammation and blood-brain barrier (BBB) disruption play pivotal roles9–11.The influx of inflammatory cells and mediators into the central nervous system via the disrupted BBB lead to delirium-like behavior in mice11–15. The BBB consists of brain vascular endothelial cell (BVEC), basement membrane, pericytes, and astrocytic end-foot16. These components form tight junctions (Tjs) and adherent junctions (Ajs) that regulate molecular transit into brain tissue, establishing a critical interface between peripheral circulation and the central nervous system17. Yang et al. demonstrated laparotomy under 1.4% isoflurane anesthesia (anesthesia/surgery) decreased Tjs protein expression (claudin-5, occludin and ZO-1), without affecting Ajs proteins (VE-cadherin, E-cadherin, and p120-catenin)9. Hu et al. further showed that anesthesia/surgery-induced peripheral inflammation activated matrix metalloproteinase 9 (MMP9), which damaged BBB integrity through pericyte damage, Tjs degradation and enhanced transcytosis11. Collectively, these findings indicate that POD pathogenesis involves BBB dysfunction driven by anesthesia/surgery-induced inflammatory cascades. Methylene blue (MB), an FDA-approved medicine, has long been safely used clinically for many ailments. MB is a safe drug at low doses (< 10 mg/kg) and stable at room temperature, rapidly crosses the BBB, and is inexpensive. It also exerts anti-inflammatory, antioxidant, restoring mitochondrial function and neuroprotective effects, positioning MB as a promising candidate for treating central nervous system disorders18–21. Preclinical evidence indicated that low-dose MB (2 mg/kg) attenuated neuroinflammation and acute depressive-like behavior following diffuse brain injury in mice22. A single dose of MB at 5 mg/kg could protect sevoflurane-induced cognitive dysfunction in aged mice23. Another research found that 10 mg/kg MB administration reduced serum IL-6 levels and attenuated microglial activation in the brain of lipopolysaccharide (LPS)-administered mice23. And we also found that 5 mg/kg MB inhibited PHD3-dependent neuroinflammation by inhibiting the Siah2/Morg1/PHD3 pathway and could improve neurocognitive impairments in LPS-induced neurobehavioral deficits24. These all indicated that MB may reduce POD incidence by targeting neuroinflammation and cognitive impairment. The aim of our work was to investigate the protective effect of MB in a mouse model of POD. We found that MB alleviated neuroinflammation, attenuated microglial activation, restored impaired BBB disruption and improved POD-like behavior induced by anesthesia/surgery. Results MB attenuates postoperative delirium-like behavior induced by anesthesia/surgery in adult mice To evaluate whether anesthesia/surgery (A/S) induces delirium-like behavior in mice, we performed a battery of behavior tests with the food burial test, open-field test, and Y-maze test at 24 h before the procedure, 6 h and 24 h postoperatively. Composite Z-scores for each mouse in the four groups were calculated at 6 h and 24 h after surgery (Figs. 1A, B). We found that composite Z-scores in A/S group was significantly higher than that of control group and MB group, while administration of MB decreased the Z-scores at 6 h after surgery. No statistical difference of composite Z-scores was found in the four groups at 24 h after the procedure. It was suggested that postoperative delirium-like behavior in adult mice mainly occurred at 6 h after surgery and gradually subsided at 24 h postoperatively. Buried food test was performed to explore whether anesthesia/surgery affected the ability of mice to search for and eat food, depending on organized thinking, complete attention, and normal level of consciousness6. These behavioral disorders were associated with delirium. The latency to find the hidden food was markedly longer in the A/S group compared to the control group at 6 h after surgery, while administration of MB improved the impaired ability induced by A/S (p < 0.01, Fig. 1C). No significant changes were observed between the control group and the MB group. The open-field was used to analyze the locomotor ability and anxiety-like behavior in mice25. During the 5 min test session, there were no significant differences in the total distance and the freezing time among the four groups at the two time points after the procedure, which indicated that A/S did not affect the motor function of adult mice (Figs. 1D, G). However, the time spent in the center was significantly lower in the A/S group than those in the control group and MB group at 6 h after surgery (p < 0.001, Fig. 1E), and administration of MB ameliorated this phenomenon at 6 h after surgery (p < 0.05, Fig. 1E). No significant changes in the latency to center among the four groups at 6 h after surgery, while the latency to center area was significantly longer in the A/S group than those in the control group at 24 h postoperatively (p < 0.05, Fig. 1F). When compared with the control group, these parameters in the MB group were not affected (Figs. 1D-G). These findings suggested that A/S affected the anxiety behavior (time spent in the center) of the mice. Y-maze test was conducted to evaluate spatial learning and memory ability in mice as previously validated26. A/S significantly reduced the number of entries to the novel arm at 6 h after surgery (p < 0.01, Fig. 1H) as compared with the MB group. Additionally, A/S also decreased the time of duration in the novel arm at 6 h after surgery when compared with the control group (p < 0.01, Fig. 1I). However, MB administration did not change the performance of adult mice in the Y-maze test at two time points after surgery (Figs. 1H,I). In summary, no significant differences were observed between the MB group and the control group, but treatment with MB alleviated the impairment of POD behavior in adult mice caused by A/S in a fluctuating way. Fig. 1MB alleviates postoperative delirium-like behavior induced by anesthesia/surgery. (A, B) Changes in composite Z-scores after MB or NS treatment via intraperitoneal injection at 6 h and 24 h postoperatively. (C) Buried food tests, (D-G) open field tests, and (H, I) Y maze tests were monitored at 6 h and 24 h after anesthesia/surgery. Data are expressed as mean ± SEM. N = 12 per group. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates postoperative delirium-like behavior induced by anesthesia/surgery. (A, B) Changes in composite Z-scores after MB or NS treatment via intraperitoneal injection at 6 h and 24 h postoperatively. (C) Buried food tests, (D-G) open field tests, and (H, I) Y maze tests were monitored at 6 h and 24 h after anesthesia/surgery. Data are expressed as mean ± SEM. N = 12 per group. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB attenuates neuroinflammation and microglia activation in the hippocampus and prefrontal cortex To access the effects of MB on systemic inflammation, the inflammatory cytokines (TNF-α, IL-6 and IL-1β), complement system markers (C1q, C3 and C5a) and inflammasome markers (NLRP3 and ASC) were measured to assess the level of neuroinflammation in hippocampus and prefrontal cortex of mice at 6 h after surgery. As shown in Figs. 2, the mRNA levels of inflammatory cytokines (TNF-α, IL-6, and IL-1β), complement system markers (C1q and C3) and inflammasome markers (NLRP3) in the hippocampus and prefrontal cortex were increased in the A/S group compared with the control group, while MB mitigated the mRNA level of these inflammatory cytokines after A/S. Additionally, there was no significant difference of mRAN level of inflammatory cytokines between the MB group and the control group. The mRNA levels of marker C5a had no differences among the four groups in the hippocampus and prefrontal cortex (Figs. 2F, N). And the mRNA levels of marker ASC also had no significant differences among the four groups in the prefrontal cortex (Figs. 2P). Microglia activation was an essential early-phase change of neuroinflammation, and the change in immunoreactivity of Iba-1 could be used to express the reactive states of microglia. We found that immunoblotting for Iba1 revealed a visible decrease in Iba1 expression in hippocampus and prefrontal cortex in A/S + MB group, compared with the A/S group (p < 0.05, Figs. 3 A, B; Figs. 4 A, B). Correspondingly, MB attenuated microglial activation by transforming the states of Iba1. When compared with the control group, microglia in the A/S group presented amoeba-like morphology and the Iba1 immunoreactivity area in the A/S group was also increased in hippocampus and prefrontal cortex, while administration of MB restored the ramified shape of the microglia and decreased the cellular area (p < 0.05, Fig. 3C; Figs. 4C, D). There were no significant changes in Iba1 in the MB group. To further confirm neuroinflammation following A/S, we quantified reactive oxygen species (ROS) in the prefrontal cortex using Dihydroethidium (DHE) staining. Compared with the control group and MB group, the fluorescence intensity of ROS was significantly increased in the A/S group, while MB administration mitigated this anesthesia/surgery-induced ROS elevation (Figs. 4E, F). Fig. 2MB alleviates neuroinflammation in the hippocampus and prefrontal cortex in surgical mice. (A-H) Relative mRNA expression levels of TNF-α, IL-6, IL-1β, C1q, C3, C5a, NLRP3 and ASC in the hippocampus at 6 h after A/S (normalized to the control group). (I-P) Relative mRNA expression levels of TNF-α, IL-6, IL-1β, C1q, C3, C5a, NLRP3 and ASC in the prefrontal cortex at 6 h after A/S (normalized to the control group). Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates neuroinflammation in the hippocampus and prefrontal cortex in surgical mice. (A-H) Relative mRNA expression levels of TNF-α, IL-6, IL-1β, C1q, C3, C5a, NLRP3 and ASC in the hippocampus at 6 h after A/S (normalized to the control group). (I-P) Relative mRNA expression levels of TNF-α, IL-6, IL-1β, C1q, C3, C5a, NLRP3 and ASC in the prefrontal cortex at 6 h after A/S (normalized to the control group). Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. Fig. 3MB alleviates microglia activation in the hippocampus in surgical mice. (A, B) Relative protein level of Iba1 at 6 h after A/S (normalized to the control group) (the original blots/gels are presented in Supplementary Figure S1). (C) Activation of microglia in the hippocampus 6 h postoperatively. Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates microglia activation in the hippocampus in surgical mice. (A, B) Relative protein level of Iba1 at 6 h after A/S (normalized to the control group) (the original blots/gels are presented in Supplementary Figure S1). (C) Activation of microglia in the hippocampus 6 h postoperatively. Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. Fig. 4MB alleviates microglia activation in the prefrontal cortex in surgical mice. (A, B) Relative protein level of Iba1 at 6 h after A/S (normalized to the control group) (the original blots/gels are presented in Supplementary Figure S2). (C, D) Activation of microglia in the prefrontal cortex 6 h postoperatively. Four random sights of each section were selected for immunofluorescence intensity analysis. (E, F) The level of reactive oxygen species (ROS) in the prefrontal cortex was assessed by Dihydroethidium (DHE) staining. Four random sights of each section were selected for immunofluorescence intensity analysis. Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates microglia activation in the prefrontal cortex in surgical mice. (A, B) Relative protein level of Iba1 at 6 h after A/S (normalized to the control group) (the original blots/gels are presented in Supplementary Figure S2). (C, D) Activation of microglia in the prefrontal cortex 6 h postoperatively. Four random sights of each section were selected for immunofluorescence intensity analysis. (E, F) The level of reactive oxygen species (ROS) in the prefrontal cortex was assessed by Dihydroethidium (DHE) staining. Four random sights of each section were selected for immunofluorescence intensity analysis. Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates blood-brain barrier disruption induced by anesthesia/surgery Previous research reported that anesthesia and surgery caused increase of the BBB permeability and impairment of cognitive function in mice9. IgG, along with other plasma components, could escape from blood vessels and enter the brain under the breakdown of BBB. Therefore, IgG immunofluorescence staining, a visual indicator of vascular leakage and BBB breakdown, was conducted to access the presence of interstitial IgG. In our study, we found that compared with the control group, the fluorescence intensity of IgG in the A/S group significantly increased in both the hippocampus and prefrontal cortex of mice at 6 h after surgery (p < 0.001, Figs. 5 A, B; Figs. 6 A-C), which was ameliorated by MB treatment (p < 0.01, Figs. 5 A, B; Figs. 6 A-C). Additionally, we found that A/S increased the brain 10-kDa dextran levels of mice compared with those in the control and MB group in both the hippocampus and prefrontal cortex, while treatment with MB attenuated the A/S-induced increase in the extravascular dextran level in mice (Figs. 5C; Figs. 6D). To find inflammatory damage in the BBB components, we examined the effects of MB on the changes of the expression of tight junction associated components after anesthesia/surgery, including claudin-5, occludin and zonula occluden 1 (ZO-1). Using RT-qPCR, we observed that there was a visible decrease in the mRAN expression of claudin-5, occludin and ZO-1 in hippocampus and prefrontal cortex after anesthesia/surgery (p < 0.05, Figs. 5D-F and 6E-G), while treatment with MB improved the expression of these mRNA (p < 0.05, Figs. 5D-F and 6E-G). Similarly, for quantitative Western blot, we also found that anesthesia/surgery distinctly decreased the protein of claudin-5, occludin and ZO-1 in the brain of mice (p < 0.05, Figs. 5G–J; Figs. 6H–K), while administration of MB obviously attenuated the reduction in these tight junction proteins (p < 0.05, Fig. 5G-J; Figs. 6H–K). Additionally, MB also mitigated anesthesia/surgery induced dysregulation of the transcellular transport pathways of cavelolin-1 and Mfsd2a (p < 0.05, Figs. 5G, K-L; Figs. 6H, L-M). In other words, anesthesia/surgery decreased Mfsd2a expression, which triggered an upregulation of caveolin-1 expression, however, these changes were alleviated by MB. Moreover, administration of MB alone did not have any effects on the BBB. Fig. 5MB alleviates blood-brain barrier disruption in the hippocampus in surgical mice. (A, B) IgG leakage staining (red fluorescence) in the hippocampus. (C) Spectrophotometric quantification of brain dextran (10 kDa) levels in the hippocampus. (D-F) Relative mRNA expression levels of claudin5, occludin and ZO-1 (normalized to the control group) in the hippocampus. (G-L) The protein levels of the tight junction proteins (claudin5, occludin and ZO-1) and transcytosis relative proteins (caveolin-1 and Mfsd2a) in the hippocampus (the original blots/gels are presented in Supplementary Figure S3). Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates blood-brain barrier disruption in the hippocampus in surgical mice. (A, B) IgG leakage staining (red fluorescence) in the hippocampus. (C) Spectrophotometric quantification of brain dextran (10 kDa) levels in the hippocampus. (D-F) Relative mRNA expression levels of claudin5, occludin and ZO-1 (normalized to the control group) in the hippocampus. (G-L) The protein levels of the tight junction proteins (claudin5, occludin and ZO-1) and transcytosis relative proteins (caveolin-1 and Mfsd2a) in the hippocampus (the original blots/gels are presented in Supplementary Figure S3). Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. Fig. 6MB alleviates blood-brain barrier disruption in the prefrontal cortex in surgical mice. (A-C) IgG leakage staining (red fluorescence) in the prefrontal cortex. Four random sights of each section were selected for immunofluorescence intensity analysis. (D) Spectrophotometric quantification of brain dextran (10 kDa) levels in the prefrontal cortex. (E-G) Relative mRNA expression levels of claudin5, occludin and ZO-1 (normalized to the control group) in the prefrontal cortex. (H-M) The protein levels of the tight junction proteins (claudin5, occludin and ZO-1) and transcytosis relative proteins (caveolin-1 and Mfsd2a) in the prefrontal cortex (the original blots/gels are presented in Supplementary Figure S4). Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB alleviates blood-brain barrier disruption in the prefrontal cortex in surgical mice. (A-C) IgG leakage staining (red fluorescence) in the prefrontal cortex. Four random sights of each section were selected for immunofluorescence intensity analysis. (D) Spectrophotometric quantification of brain dextran (10 kDa) levels in the prefrontal cortex. (E-G) Relative mRNA expression levels of claudin5, occludin and ZO-1 (normalized to the control group) in the prefrontal cortex. (H-M) The protein levels of the tight junction proteins (claudin5, occludin and ZO-1) and transcytosis relative proteins (caveolin-1 and Mfsd2a) in the prefrontal cortex (the original blots/gels are presented in Supplementary Figure S4). Data are expressed as mean ± SEM. N = 3–4 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB inhibites the expression of EMMPRIN and MMP9 Considering the degradation of tight junction proteins depending on matrix metal proteins, the EMMPRIN/MMP9 pathway might mediate BBB damage. We observed that anesthesia/surgery induced a significant increase of EMMPRIN/MMP in both the hippocampus and prefrontal cortex (p < 0.001, Fig. 7), while MB treatment reversed this change (p < 0.01, Fig. 7). Therefore, the protective effect of MB on tight junction proteins and the BBB may be mediated through the inhibition of EMMPRIN/MMP9-dependent tight junction proteins degradation. Fig. 7MB inhibites the expression of EMMPRIN and MMP9. (A-C) The protein levels of EMMPRIN and MMP9 in the hippocampus. (D-F) The protein levels of EMMPRIN and MMP9 in the prefrontal cortex (the original blots/gels are presented in Supplementary Figure S5). Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. MB inhibites the expression of EMMPRIN and MMP9. (A-C) The protein levels of EMMPRIN and MMP9 in the hippocampus. (D-F) The protein levels of EMMPRIN and MMP9 in the prefrontal cortex (the original blots/gels are presented in Supplementary Figure S5). Data are expressed as mean ± SEM. N = 3 per cohort. A/S, anesthesia/surgery; MB, methylene blue. *P < 0.05; **P < 0.01; ***P < 0.001. Discussion In this study, we found that MB treatment ameliorated POD-like behavior in adult mice caused by A/S in a fluctuating way, via its anti-inflammatory and BBB-protecting effects. Our study showed that MB alleviated neuroinflammation in the hippocampus and prefrontal cortex according to the reduced expression of inflammatory cytokines and attenuated microglial activation in these brain regions. Moreover, MB treatment restored blood-brain barrier structure and permeability induced by A/S through inhibiting Tjs damage and transcytosis. In summary, the migration of peripheral inflammatory mediators after A/S into the brain induced neuroinflammation and POD-like behaviors, while MB could reduce the entry of peripheral inflammatory cytokines into the brain via repairing the impaired BBB, thus decreasing the incidence of POD. To our knowledge, this is the first study to report the neuroprotective effect of MB in a mouse model of POD. POD is an acute fluctuating brain disturbance after A/S. Our study found that POD mainly occurred at 6 h postoperatively and resolved by 24 h after surgery in adult mice, consistent with prior research6,8. In contrast, aged mice exhibited prolonged POD susceptibility extending to 24 h post-surgery7,27,28, indicating heightened vulnerability to postoperative brain dysfunction in geriatric populations. Additionally, previous studies also demonstrated that A/S lead to an age-dependent delirium-like behavior in rodents, and the reason was that A/S could induce age-dependent increasing in brain inflammation level6,29. Adult mice maintain minimal baseline neuroinflammation and demonstrate rapid postoperative recovery, explaining the transient delirious-like manifestations after A/S. Supporting this temporal pattern, a systematic review reported that cognitive deficits in adult mice aged around 2–3 months after postoperative 24 h were only about 50%, suggesting cognitive impairment in adult mice predominantly occurred during early recovery30. Another research reported that multiple A/S could not impair cognitive function in adult mice31. One reason for this negative result was that the time of behavioral test was too late (48 h after surgery). Another reason was that the increase of neuroinflammatory cytokines occurred among 6–24 h after surgery, maximally at postoperative 6 h, so central inflammation had subsided after postoperative 48 h. Therefore, cognitive evaluation at 6 h postoperatively is essential for detecting POD in adult mice. Accumulating evidence implicates neuroinflammation as a core pathogenic mechanism in POD. As anesthesia/surgery triggered neuroinflammation and cognitive impairment in adult mice32–34, this age group was selected for our experimental model. Surgical trauma-induced aseptic inflammation stimulates the innate immune system, triggering the inflammatory process which ultimately results in POD35,36. Postoperative upregulation of circulating pro-inflammatory cytokines (e.g., IL-1β, IL-6 and TNF-α) crosses the compromised BBB, overactivating microglia and triggering neuroinflammatory cascades in hippocampal and cortical regions37. Concurrently, microglia activation establishes a self-perpetuating cycle by generating additional pro-inflammatory factors38. Our results found that MB could decrease the mRNA level of neuroinflammatory factor and microglia activation after A/S. These findings are consistent with the potent anti-inflammatory effect of MB in many other disease models associated with inflammation, such as subarachnoid hemorrhage39, traumatic brain injury40 and lipopolysaccharide-induced neuroinflammation24. Collectively, these evidences indicate that MB’s potent anti-inflammatory activity modulates neuroinflammatory responses in the central nervous system. A fully functional BBB is essential for maintaining central nervous system homeostasis. The key processes that ensure the integrity of BBB are shutting down pathways for paracellular and intercellular leakage, which need maintenance of Tjs (ZO-1, occludin, and claudin-5) and suppression of transcytosis41,42. Mfsd2a inhibits transcytosis by translocating phospholipids to downregulate caveolin-1 formation. Consistent with existing literature15,17,43, our study confirmed that A/S downregulated Tj proteins and increased BBB permeability, correlating with delirium-like manifestations. We assessed BBB integrity by examining the expression of Tj proteins (ZO-1, occludin, and claudin-5), transcytosis regulators (Mfsd2a and caveolin-1) and extravasation of IgG and dextran. Our results showed that A/S-induced neuroinflammation compromised BBB function and triggered POD-like behavior, while MB could alleviate neuroinflammation, upregulate Tj proteins, inhibit transcytosis and ameliorate POD-like behavior. This work established that targeting A/S-induced BBB dysfunction represented a viable therapeutic strategy for POD prevention. The observed gender disparity in POD prevalence remains clinically contentious44,45. In neuroscience and related disciplines, the wider use of male rodents (especially rats and mice) has indeed been a long-standing practice. Female estrous cycles introduce significant neurobiological variability, as sex hormones modulate brain structure and function, including neuronal excitability, synaptic plasticity, neurotransmitter systems, and behavior, such as anxiety, depression, cognition, and social interaction46. If female animals at different stages of their estrous cycle are mixed in an experiment, these hormonal fluctuations introduce additional, significant within-group variability, increasing the “noise” in the experimental results. While estrogen demonstrated neuroprotective properties in cognitive studies47,48, this had paradoxically reinforced male-centric research paradigms. Notably, Professor Zongcong Xie’s research specifically incorporated female murine models, given their demonstrated enhanced vulnerability to sevoflurane-induced cognitive deficits following abdominal surgery in Alzheimer’s transgenic females49. Historically and currently, male rodents are more commonly used in neuroscience experiments primarily to avoid the significant variability introduced by the female estrous cycle, thereby simplifying experimental design, increasing statistical power, and continuing research tradition. However, this practice overlooks crucial biological sex differences, resulting in incomplete findings with limited translational relevance. Further studies should systematically integrate female animals through policy guidance and paradigm shifts, aiming to achieve more comprehensive and reliable biomedical knowledge. However, several limitations of our research should be acknowledged. First, the experimental model only utilized young male mice. We recognized that POD exhibited no gender predilection and demonstrated higher incidence in geriatric populations. Although the data supports our hypothesis, further studies are required to be conducted in both sexes and aged animals. Second, the precise mechanisms underlying MB’s promote of Tj development and regulation of transcytosis remain incompletely characterized. Key unresolved questions include: (1) How MB modulates Tj proteins localization and disruption in endothelial cells. (2) Its effects on EMMPRIN/MMP-9 localization in endothelial cells. (3) The potential role of MMP-9 inhibitors (e.g., SB-3CT, TIMP-1) in BBB function modulation. These processes merit rigorous investigation in subsequent research. Third, our assessment focused on microglial contributions to neuroinflammation without evaluating astrocytes, another critical central nervous system resident immune cell. Given documented microglia-astrocyte interactions in neurodegenerative pathologies50, the role of astrocytic responses in POD pathogenesis requires dedicated exploration. Finally, the dosage of MB was extrapolated from our prior sepsis model investigation and existing literature23,24. Optimal dosing for cognitive protection across different animal models necessitates further systematic determination. Conclusion This study demonstrated that MB could alleviate neuroinflammation and preserve the integrity of BBB, thereby ameliorating POD-like behavior in adult mice. These findings imply that MB maybe a promising drug candidate for the prevention of POD. Materials and methods All experiments were performed in accordance with all pertinent ethical regulations, including the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guideline. Experimental protocols were approved by the Animal Care and Use Committee of Shanghai Medical College, Fudan University (Approval No. 202404FD0003). Efforts were made to minimize the number of animals used and their suffering. Antibodies and reagents Primary antibodies for western blotting: anti-ZO-1 (1:5000), anti-occludin (1:5000), anti-claudin5 (1:5000), anti-Iba-1 (1:1000) and anti-β-actin (1:20000) were obtained from Proteintech. Anti-MMP9 (1:500) and anti-caveolin-1 (1:500) were obtained from Santa Cruz Biotechnology. Anti-Mfsd2a (1:1000) was obtained from Abcam; Anti-EMMPRIN (1:1000) was obtained from Affinity. Antibodies for immunofluorescence staining: Anti-Iba-1 (1:300) was obtained from Proteintech. Cy-3-conjugated goat anti-mouse IgG (1:200) was obtained from Affinity. Alexa Fluor 488 goat anti-rabbit IgG (1:500) was obtained from proteintech. Additional reagents: MB was obtained from Jiangsu Jichuan Pharmaceutical Co. Ltd. (Taixin, China). The protein extraction kit, bicinchoninic acid (BCA) assay reagents, Dihydroethidium (DHE), FITC-labeled dextran (ST2935), and protease and phosphatase inhibitors were purchased form Beyotime Institute of Biotechnology (Shanghai, China). DAPI Fluoromount-G was purchased from SouthernBiotech. Animals Male C57BL/6J mice aged 16 weeks old, were purchased from Suzhou Sibeifu Biotechnology Co., Ltd., Suzhou, China. All animals were housed in groups under standard laboratory conditions (22 ± 2 °C, 12-h light/12-h dark cycle) with ad libitum access to food and water. Mice were allowed one week to acclimatize to the laboratory environment prior to experiment. Experimental protocol Mice were randomly assigned to four groups: control group, anesthesia/surgery (A/S) group, A/S + MB group, and MB group. MB (Jiangsu Jichuan Pharmaceutical Co. Ltd., Taixin, China) was dissolved in normal saline. Mice in the MB group and A/S + MB group received intraperitoneal injections of MB solution (5 mg/kg) at 1 h before surgery and 4 h after surgery. Meanwhile, mice in the control group and A/S group received an equal volume of normal saline administered in the same manner and at the same time points. The MB dosage was based on previous studies using sevoflurane-induced cognitive dysfunction and lipopolysaccharide-induced neurocognitive impairment models, with minor modification23,24. Postoperative delirium mouse model Mice in the A/S group and A/S + MB group underwent a simple laparotomy under isoflurane anesthesia, following the methods described in previous studies with a modification by using 60% oxygen15,28. Arterial blood gas monitoring was omitted, as neither the surgical procedure51 nor anesthesia with 1.4% isoflurane52 significantly disturbed the blood pressure or blood gas values in mice. Specifically, anesthesia was induced and maintained using 1.4% isoflurane in 60% oxygen within a transparent acrylic chamber. Fifteen minutes after induction, mice were moved out of the chamber and placed on a heating pad for temperature maintenance during surgery. Isoflurane anesthesia was maintained via a cone device, and a 16-gauge needle was inserted into the cone near the nose of the mouse to monitor the concentration of isoflurane. Following limb immobilization, the abdominal skin was shaved and disinfected. Pre-incisional local anesthesia was administered using 1% lidocaine. A longitudinal midline incision was made from the xiphoid to the 0.5-cm proximal pubic symphysis on the skin, abdominal muscles, and peritoneum. Then, we explored abdominal organs for one min and sutured the incision layer by layer with 4 − 0 Vicryl sutures. At the end of the procedure, compound lidocaine cream (2.5% lidocaine and 2.5% prilocaine) was applied to the wound and then repeated every 8 h until the mice were euthanized, to relieve the pain associated with the surgery. The surgical procedure lasted approximately 10 min for each mouse. and the mouse was subsequently returned to the anesthesia chamber to continue receive the rest of the anesthesia, consisting of 1.4% isoflurane in 60% oxygen for up to 2 h. The rectal temperature of the mice was maintained at 37 ± 0.5 ◦C during the anesthesia/surgery. After recovering from the anesthesia, the mouse was returned to home cage with ad libitum access to food and water. The mice in the control group and MB group received 60% oxygen exposure for 2 h. Behavioral tests In accordance with previous study described, delirium-like behaviors were assessed using a battery of natural and learned behavior tests, including buried food test, open-field test, and Y-maze test, at 24 h before (baseline) and then at 6, 24 h after surgery, simulating certain aspects similar to the clinical diagnoses of patients with POD7,28,53. Mice were placed into the test environment for 1 h prior to tests and returned to their respective cages upon completion. All apparatuses were cleaned with 75% alcohol between subjects to eliminate residual odors (Fig. 8). At study completion, mice were placed in a sealed chamber, and CO2 was introduced at a flow rate of 20–30% chamber volume per minute until respiratory arrest was confirmed. Death was ensured by cervical dislocation as a secondary method. Fig. 8Experimental timeline. The mice received behavioral tests at 24 h before the anesthesia/surgery (baseline), and then at 6 h and 24 h after surgery. The mice in MB group and A/S + MB group were injected with MB solution at a dose of 5 mg/kg through intraperitoneal at 1 h before the surgery and 4 h after the surgery. Meanwhile, an equal volume of normal saline was given to the mice in the control group and A/S group at the same way. Mice were sacrificed for brain tissue collection at 6 h after the Anesthesia/Surgery. MB, methylene blue; NS, normal saline. Experimental timeline. The mice received behavioral tests at 24 h before the anesthesia/surgery (baseline), and then at 6 h and 24 h after surgery. The mice in MB group and A/S + MB group were injected with MB solution at a dose of 5 mg/kg through intraperitoneal at 1 h before the surgery and 4 h after the surgery. Meanwhile, an equal volume of normal saline was given to the mice in the control group and A/S group at the same way. Mice were sacrificed for brain tissue collection at 6 h after the Anesthesia/Surgery. MB, methylene blue; NS, normal saline. Buried food test The buried food test was performed as described in previous studies6,29 with modifications. Specifically, two days prior to testing, each mouse received two pieces of sweetened cereal. On the test day, mice were habituated for 1 h by placing their home cages in the testing room. The test cage was prepared with 3 centimeters of clean padding, in which one sweetened cereal pellet was buried 0.5 centimeter beneath the bedding surface at a randomized, non-visible location. The mouse was placed in the center of the test cage and measured the latency of the mouse to find and eat the food. Latency was defined as the time beginning when the mouse was placed in the test cage and ending when the mouse uncovered the food and grasped it in its forepaws and/or teeth, which was used as a measure of organized thinking and attention. When mice found the food pellet within 5 min, they were allowed to consume the pellet and then were returned to their home cage. If the mouse could not find the food within 5 min, the testing session ended, and the latency was recorded as 300 s for that mouse. To prevent the transmission of olfactory cues, we emptied the padding from the test cage and cleaned the cage with a 75% ethanol solution after each test. Open field test The open field test was carried out as described in previous studies with modifications29,54. The mouse was gently placed in the center of an open field chamber (40 × 40 × 40 centimeters) under dim light conditions and allowed to explore freely for 5 min. The movement parameters were automatically recorded and analyzed through a video camera connected to Any-Maze animal tracking system software (Noldus Ethovision). The total distance moved (centimeters), the time (seconds) spent in the center of the open field, freezing time (seconds), and the latency (the time in seconds for the mice to reach the center at the first attempt) to the center of the open field were recorded and analyzed. The floor of the open field was cleaned with 75% ethanol solution between each test. Y maze test The Y maze test assessed spatial learning and memory ability following anesthesia/surgery6,29,55. Specifically, the Y maze, made of gray polyvinyl, was placed in a quiet and illuminated room and consisted of three arms extending from a triangular central platform, with an angle of 120 degrees between each arm. To help the mice establish their spatial memories, the walls of arms were marked with colored geometric shapes. The three arms included the start arm, in which the mouse started to explore (always open); the novel arm, which was blocked at the first trial but opened at the second trial; and the other arm (always open). The start arm and other arm were designed randomly to avoid spatial memory error. The Y-maze test consisted of two trials separated by a 2-hour inter-trial interval. The first trial (training) was 10 min in duration and allowed the mouse to explore the start arm and the other arm of the maze with the novel arm being blocked. 2 h after the first trial, the second trial was conducted. For the second trial, the mouse was placed back in the maze in the same start arm with free access to all three arms for 5 min. A video camera, which was linked to the Any-Maze animal tracking system software (Noldus Ethovision), was installed to monitor and analyze the number of entries and the time spent in each arm. The time spent in and entries into the novel arm indicated the spatial recognition memory (learned behavior). Each of the arms of the Y maze was cleaned with 75% ethanol solution between trials. Behavioral composite Z-scores The composite Z-scores, which present the postoperative delirium-like behaviour of each mouse, were calculated using the methods described in previous studies29,53,54. The formula to calculate Z-scores was Z= [ΔXtrail– MEAN(ΔX)control]/SD(ΔX)control. Briefly, (ΔX)control was the change score of mice in the control group at 6 h or 24 h after the control condition minus the score at the baseline; ΔXtrail was the change score of mice in the A/S, A/S + MB or MB group at 6h or 24 h after the three groups condition minus the score at baseline; MEAN(ΔX)control was the mean of ΔXcontrol, and SD(ΔX)control was the standard deviation of ΔXcontrol. The composite Z score for the mouse was calculated as the sum of the values of 6 Z scores (latency to eat food, time spent in the center, latency to the center, freezing time, entries in novel arm, and duration in novel arm) normalized with the SD for that sum in the control conditions. Given that the reduction (rather than increase) in time spent in the center and freezing time (open field test) and the reduction in duration and entries in the novel arm (Y maze test) indicate impairment of the behavior, we multiplied the Z score values representing these behaviors by − 1 prior to calculating the composite Z score using these values. Western blot analysis Hippocampus and prefrontal cortex of the mouse were harvested 6 h after surgery. The total protein content of these tissues was isolated using radioimmunoprecipitation assay (RIPA) lysis buffer containing protease and phosphatase inhibitors on ice. BCA kit was used to determine the protein concentration. After mixing with 5 × loading buffer, the protein was boiled for 10 min. Then equal amounts of protein were separated by 7.5%−12.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels, and the separated proteins were transferred onto a polyvinylidene difluoride membrane, followed by blocking in skim milk for one-two hours at room temperature. The membranes were incubated with various specific primary antibodies at 4 °C overnight and then incubated with the appropriate HRP-conjugated secondary antibodies for 2 h at room temperature. Anti-β-actin was used to normalize and control for loading differences in protein levels. The bands were measured using image analysis software (v1.37, NIH, USA), and changes in protein levels were presented as folds of those in the control group. Real-time quantitative polymerase chain reaction (RT-qPCR) Total RNA was extracted from tissue samples using lysis and wash buffers according to the manufacturer’s instructions (RNOO1A, EZBioscience, USA), and transcribed to cDNA by using 4 × Reverse Transcription Master Mix (A0010GQ, EZBioscience, USA). QPCR was performed using SYBR Green Master Mix (A0012-R2, EZBioscience, USA) on real-time qPCR instrument (Applied Biosystems, QuantStudio 7 Flex). β-actin was selected as internal reference gene to normalized, data was treated by comparative cycle threshold method. The primers used for qPCR are listed in Supplementary Table S1 and were purchased from and validated by Sangon Biotech (Shanghai, China). Immunofluorescence Six hours after surgery, mice were anesthetized with Tribromoethanol and perfused transcranially with ice-cold normal saline followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate-buffered saline (PBS) at PH 7.4. The brains were harvested and fixed in 4% PFA in 0.1 M PBS at 4 °C, cryoprotected in 30% sucrose for 72 h, frozen in TissueTek OCT (Feica), and cut sequentially into 20-µm-thick slices. For brain section staining, sections were rinsed with PBS for 10 min, permeabilized for 20 min with PBS containing 0.2% Triton X-100 and blocked with 5% bovine serum albumin (BSA) at room temperature to block non-specific binding. Then, the sections were incubated overnight with rabbit anti-Iba-1 primary antibody (1:300) at 4 °C. After washing, the sections were incubated with secondary antibodies CoraLite488-conjugated Donkey Anti-Rabbit IgG(H + L) (1:500) for 1 h at room temperature in dark. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) and the sections were mounted in DAPI Fluoromount-G. All images were taken with EVOS M700. Analysis was conducted with Image J software (v1.37, NIH, USA). Dihydroethidium (DHE) staining Brain sections were rinsed with PBS for 10 min, then the sections from the four groups blocked with 5% BSA for 30 min at 37 °C. After washing, the sections were incubated with PBS diluted DHE (10µM) at room temperature for 30 min in light-protected conditions. The images were observed under EVOS M700, in which the red fluorescence reflected the ROS content, and the fluorescent intensity was conducted with Image J software. Evaluation of blood-brain barrier permeability Immunoglobulin G (IgG) leakage was performed to access BBB permeability. Brain sections were baked for 20 min at 37 °C, followed by staining with Cy-3-conjugated goat anti-mouse IgG (1:200) for 2 h. After being mounted with a glass coverslip, the slides were scanned in a fluorescence microscope. Dextran was used to measure BBB permeability as described in previous studies with modifications9,15. Spectrophotometric quantification of 10-kDa dextran tracer (4 mg/ml) from hippocampus and prefrontal cortex extracts was used as described in previous study15. Specifically, each mouse was injected intravenously with 100 µl dextran tracer at 6 h after surgery. Fifteen minutes after injection of the tracer, mouse received thoracotomy under brief anesthesia (1.4% isoflurane for 5 min) and perfused with transcranial phosphate-buffered saline (150 ml for 5 min). Then, the mice were sacrificed, and the hippocampus and prefrontal cortex were harvested. Then, the brain tissue was homogenized and tissue lysates were centrifuged at 16,000 rpm for 20 min. The fluorescence of the supernatant was measured on a Synergy H1 Multi-Mode Microplate Reader (BioTek) (ex/em 492/520 nm). Statistical analysis Data are presented as mean ± standard error of the mean (SEM), with the error bars indicating the SEM. Shapiro-Wilk test was used to analyze the normal distribution. Multiple group comparisons were performed by one-way ANOVA with Bonferroni post hoc test. Statistical analysis was performed using GraphPad Prism 7.0 (GraphPad, New York, NY, United States). Differences were considered statistically significant at P < 0.05. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1 Supplementary Material 2 Supplementary Material 2 Supplementary Material 3 Supplementary Material 3 Supplementary Material 4 Supplementary Material 4 Supplementary Material 5 Supplementary Material 5 Supplementary Material 6 Supplementary Material 6 Supplementary Material 7 Supplementary Material 7