Small intestinal γδ T17 cells promote C1q-mediated SAE by synaptic pruning in mice

doi:10.21203/rs.3.rs-5152680/v1

Download PDF

Biological Sciences - Article

Small intestinal γδ T17 cells promote C1q-mediated SAE by synaptic pruning in mice

https://doi.org/10.21203/rs.3.rs-5152680/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Sepsis is a severe global health issue with high mortality rates, and sepsis-associated encephalopathy (SAE) further exacerbates this risk. While recent studies have shown the migration of gut immune cells to the lungs after sepsis, their impact on the central nervous system remains unclear. Our research demonstrates that sepsis could induce the migration of IL-7R^high CD8^low γδ T17 cells from the small intestine to the meninges, where they secrete IL-17A, impairing mitochondrial function in microglia and activating the cGAS-STING-C1q pathway. This process is accompanied by inhibited ubiquitination of STING at the K150 site, resulting in STING accumulation and increased release of C1q-tagged hippocampal synapses, which are subsequently pruned by activated microglia. Importantly, 4-Octyl itaconate mitigates the excessive synaptic pruning by inhibiting γδ T17 cell migration and promoting STING ubiquitination, thereby alleviating SAE. Our findings reveal a novel mechanism of synaptic pruning by microglia via the cGAS-STING-C1q pathway, emphasize the critical role of gut-derived γδ T17 cell migration to the meninges in SAE, and highlight the importance of STING ubiquitination in modulating C1q-mediated excessive synaptic pruning.

Health sciences/Diseases/Infectious diseases/Bacterial infection

Biological sciences/Immunology/Infectious diseases/Bacterial infection

Sepsis-associated encephalopathy

Gut-brain axis

Synaptic pruning

Microglia

cGAS-STING

Sepsis-associated encephalopathy (SAE) is a severe complication of sepsis, characterized by diffuse or multifocal neurological dysfunction, and is a leading cause of increased mortality among ICU patients. Although SAE is often considered reversible, approximately 40% of patients experience long-term neurological symptoms, including memory impairment, depression, anxiety, and cognitive dysfunction, especially in cases of severe sepsis¹. The lack of in-depth understanding of the mechanisms and effective therapy for SAE remains a big challenge in clinical practice.

The gut is considered the “engine” driving sepsis and multiple organ dysfunction, a concept known as “gut-origin sepsis”². As a critical immune organ, the gut contains diverse lymphocyte subsets, including γδ T cells and Th17 cells, primarily located in the small intestinal lamina propria (SI-LP). These cells are crucial for maintaining mucosal immune balance and intestinal barrier integrity³. Although γδ T cells comprise a small fraction of total T cells, they are essential for epithelial integrity, tissue repair, host homeostasis, and pathogen defense³. When the intestinal barrier is compromised, leading to bacterial invasion and inflammation, γδ T cells could differentiate into interleukin-17A (IL-17A)-producing γδ T17 cells, which exacerbates the inflammatory response⁴.

T cells in the meninges can secrete IL-17A, which could affect border-associated macrophages, reduces cerebral blood flow, and impairs cognition⁵. Similarly, γδ T17 cells can migrate from the SI-LP to the meninges, worsening the acute ischemic brain injury⁶. IL-17A secreted from γδ T17 cells could activate microglia and contribute to lipopolysaccharide (LPS)-induced neuroinflammation and cognitive deficits⁷. We recently showed that IL-7R^high CD8^low γδ T17 cells could migrate from the gut to the lungs and exacerbate cecal ligation and puncture (CLP)-induced acute lung injury and mortality via secreting IL-17A⁸. However, it remains unclear whether gut-derived IL-7R^high CD8^low γδ T17 could migrate to the central nervous system (CNS) and exacerbate SAE via IL-17A.

Synaptic pruning, a natural process of synapse elimination, is crucial for proper neural circuit formation, but aberrant pruning can lead to CNS disorders such as schizophrenia, anxiety, and autism⁹. Aberrant synaptic pruning affects neural signal transmission and network functionality, which is primarily mediated by microglial phagocytosis¹⁰. Complement-mediated synaptic pruning by microglia is closely linked to the progressive synaptic loss in LPS-induced depression¹¹. Furthermore, sepsis-activated microglia could release C1q, which tags synapses for abnormal pruning and thereby exacerbates peritoneal contamination and infection-induced SAE¹². However, the connection between C1q-mediated synaptic pruning and IL-17A and its potential mechanisms remain to be clarified.

This study demonstrated that the migration of IL-7R^high CD8^low γδ T17 cells from the small intestine to the meninges during sepsis was a critical factor contributing to SAE. IL-17A mediated mitochondrial damage in microglia and activated the cGAS-STING pathway, which was accompanied by inhibited ubiquitination of STING at the K150 site, resulting in STING accumulation. Increased levels of STING promoted the release of C1q and thus enhanced C1q-tagged hippocampal synapses, which exacerbated microglia-mediated synaptic pruning and SAE. 4-Octyl itaconate (4-OI), an itaconate derivative with potent anti-inflammatory properties^13,14, inhibited IL-7R^high CD8^low γδ T17 cell migration, increased the ubiquitination of STING at the K150 site in microglia, reduced STING expression, and subsequently decreases C1q-tagged synapses and excessive synaptic pruning, thereby ameliorating SAE.

Gut-derived IL-7R ^high CD8^low γδ T17 cells migrate to the meninges and exacerbate SAE

First, we investigated whether small intestinal γδ T cells could migrate to CNS after sepsis, we performed flow cytometry (FCM) analysis following CLP. We observed significant reduction of γδ T cells in the small intestine, with a concurrent increase in the meninges (Fig. 1a, b), while γδ T cells in the brain parenchyma remained low (Extended Data Fig. 1a). These findings suggest that the meninges serve as the primary CNS destination for migrating small intestinal γδ T cells after sepsis. Notably, CLP induced elevated production of IL-17A by γδ T cells in the meninges, with over 50% of these cells identified as γδ T17 cells (Fig. 1c–e). Building on previous findings that IL-7R^high CD8^low γδ T17 cells in small intestine could migrate to the lungs after sepsis⁸, we observed a similar elevation of this subset in the meninges after CLP (Fig. 1f, g). Therefore, we hypothesized that small intestinal IL-7R^high CD8^low γδ T17 cells could directly migrate to the meninges following sepsis. To investigate this, the small intestine of Kaede-tg mice underwent localized irradiation to track the migration of small intestinal lymphocytes (Fig. 1h). We found that the most IL-7R^high CD8^low γδ T17 cells in the meninges originated from the small intestine (Fig. 1i).

We further investigate the impact of IL-7R^high CD8^low γδ T17 cells on SAE, γδ T cell-specific knockout mouse models targeting Cd8a and Il7r were used. Specifically, mice were developed by crossing Trdc^CreERT2 mice with Il7r^flox/flox and Cd8a^flox/flox mice, leading to a conditional knockout of Il7r and Cd8a in γδ T cells, respectively. In particular, Trdc^CreERT2 Il7r^flox/flox mice exhibited a significant reduction in γδ T17 cells in the meninges compared to wild-type (WT) and Trdc^CreERT2 Cd8a^flox/flox mice (Extended Data Fig. 1c). Additionally, Golgi staining showed that CLP-induced dendritic spine loss and disorganization were reversed in Trdc^CreERT2 Il7r^flox/flox mice (Fig. 1j, k). Furthermore, transmission electron microscopy (TEM) revealed attenuation of synaptic vesicle loss, postsynaptic density thinning, and mitochondrial damage in Trdc^CreERT2 Il7r^flox/flox mice (Fig. 1l, m).

Behavioral assessments highlighted significant cognitive and memory impairments in CLP-induced SAE, including reduced exploration in Y-maze (Fig. 1n, o) and Morris-water maze (MWM) tests (Fig. 1p–r), increased anxiety-like behavior in open field tests (OFT) (Extended Data Fig. 1f, g), and decreased time spent exploring the novel object in the novel object recognition test (NORT) (Extended Data Fig. 1h). However, these behavioral deficits were markedly alleviated in Trdc^CreERT2 Il7r^flox/flox mice, suggesting that IL-7R^high CD8^low γδ T17 cells could exacerbate cognitive dysfunction in SAE.

γδT cells enhance microglial cGAS-STING expression and exacerbate synaptic pruning via C1q release

Having established that the migration of IL-7R^high CD8^low γδ T17 cells from the intestine to the meninges could exacerbate SAE, we next explored the molecular mechanisms involved. We found that more than 50% of IL-17A in the meninges was produced by γδ T17 cells (Fig. 1d, e). IL-17A from γδ T17 cells is known to activate microglia, contributing to neuroinflammation and cognitive dysfunction⁷. Although IL-17A has been implicated in mitochondrial degradation via enhanced mitophagy in bronchial fibroblasts in severe asthma¹⁵, its effect on microglial mitochondrial integrity remains unclear. To evaluate the impact of IL-17A, produced by γδ T cells, on microglial function, we treated BV2 cells with recombinant IL-17A (100 ng ml^–1) and LPS (1000 ng ml^–1). This treatment exhibited decreased mitochondrial membrane potential and mitochondrial swelling, as indicated by JC-1 staining and TEM (Fig. 2a–c), suggesting mitochondrial damage. Mitochondrial dysfunction can release mitochondrial DNA (mtDNA), triggering the cGAS-STING pathway, a key intracellular DNA-sensing mechanism involved in host defense and inflammation¹⁶. Activation of the cGAS-STING pathway not only triggers inflammatory responses but also exacerbates various CNS injuries, including traumatic brain injury, spinal cord injury, subarachnoid hemorrhage, and hypoxic-ischemic encephalopathy¹⁷.We confirmed elevated cGAS and STING expression in BV2 cells following IL-17A and LPS treatment (Fig. 2d). Importantly, we found that mitochondrial damage was necessary for cGAS-STING upregulation, as mitochondria-depleted BV2^ρ0 cells did not exhibit increased cGAS and STING expression (Extended Data Fig. 1i–k). Consistent with these findings, silencing IL-17A in primary γδ T cells using siRNA reduced mitochondrial damage and decreased cGAS, STING, and C1q levels in primary microglia (Fig. 2f–k).

In vivo, hippocampal overexpression of IL-17A increased cGAS and STING expression (Fig. 2l–n) and impaired cognitive function (Extended Data Fig. 2a–g). Conversely, neutralization of IL-17A improved cognition in septic mice (Extended Data Fig. 2h–n). As lowering IL-17A levels reduced C1q expression in microglia (Fig. 2j, k). Given that C1q can bind to synapses and mediate synaptic pruning in a murine polymicrobial sepsis model¹², we examined synaptic proteins in the hippocampus. IL-17A overexpression elevated C1q and reduced the levels of postsynaptic density protein 95 (PSD95) and synaptophysin (SYN), indicating that IL-17A could promote increased synaptic pruning by increasing C1q levels (Fig. 2o–r). However, this synaptic damage was reversed in Trdc^CreERT2 Il7r^flox/flox mice (Fig. 2s, t), which also exhibited lower expression of cGAS and STING in the hippocampus (Fig. 2u, v). Confocal imaging and 3D reconstruction using Imaris software showed that microglia in the hippocampus of septic mice displayed an activated phenotype, with increased phagocytosis of PSD95, linking microglial activation to synaptic protein loss (Fig. 2w, x).

Considering that STING promotes neuroinflammation¹⁸ and contributes to synaptic impairments, we investigated whether pharmacological inhibition of STING could reverse the associated neuroinflammatory responses and synaptic deficits (Fig. 3a–c). Treatment with the STING inhibitor H151 (750 nM) reduced C1q, increased the expression of PSD95 and SYN (Fig. 3d, e), and mitigated synaptic pruning (Fig. 3f, g), which led to improved cognitive outcomes (Fig. 3h, i and Extended Data Fig. 3a–e). H151 also reversed synaptic impairments caused by cGAS overexpression in microglia (Fig. 3j–s, Extended Data Fig. 3f–l). Additionally, direct C1q neutralization reduced microglial phagocytosis of synaptic proteins, resulting in improved cognition and restoring synaptic integrity (Fig. 3t–z). In line with this, microglial depletion with PLX3397 similarly reduced synaptic pruning and improved behavior deficts in septic mice (Extended Data Fig. 4e–k). These findings demonstrate that synaptic pruning after sepsis is closely related to the phagocytosis of C1q-labeled synapses by microglias. IL-17A produced by γδ T cells could drive C1q release via the cGAS-STING pathway, representing a key therapeutic target to reduce cognitive impairment in sepsis.

K150-mediated STING ubiquitination prevents cognitive dysfunction

Given that sepsis could increase STING expression and contributes to C1q-mediated synaptic pruning in our findings, we investigated whether the ubiquitin-proteasome system (UPS), a key pathway for protein degradation and maintenance of protein homeostasis¹⁹, could regulate STING accumulation. We hypothesized that the ubiquitination of STING could play an important role in the pathogenesis of SAE. As expected, we found that exposure of primary microglia to LPS resulted in a significant reduction in STING ubiquitination, which was also observed in the hippocampus of CLP mice (Fig. 4a, b). Mass spectrometry identified several potential ubiquitination sites on STING, with lysine 150 (K150) emerging as critical, mutation of K150 markedly reduced STING ubiquitination in primary microglia (Fig. 4c, d). A specific antibody targeting ubiquitinated K150 confirmed a reduction in the hippocampus of septic mice (Fig. 4e, f).

To further explore the role of K150, we generated an adeno-associated virus (AAV) vector carrying the K150R mutation and injected it into the hippocampus of Cx3cr1^Cre mice to mutate K150 site specifically in microglia. This mutation led to reduced STING ubiquitination (Fig. 4g), increased microglial phagocytosis of PSD95 (Fig. 4h, i), and elevated STING protein levels (Fig. 4j, k). Consequently, mice with the K150R mutation in microglia showed enhanced synaptic pruning, evidenced by increased C1q and decreased levels of PSD95 and SYN (Fig. 4l, m). Furthermore, microglial K150R-mutant mice exhibited severe synaptic and mitochondrial damage (Fig. 4n, o), along with pronounced learning and memory deficits in behavioral tests (Fig. 4p–v). These findings demonstrate that K150-mediated ubiquitination in microglia is essential for STING degradation and prevention of cognitive dysfunction.

We further explored the ubiquitination process of STING. We used UbiBrowser to predict potential E3 ubiquitin ligases for STING and identified RNF5 as a key candidate (Extended Data Fig. 5a). Co-immunoprecipitation confirmed a direct interaction between RNF5 and STING, which decreased following LPS stimulation or CLP-induced sepsis (Extended Data Fig. 5b–f). Furthermore, RNF5 knockdown in microglia significantly reduced STING ubiquitination (Extended Data Fig. 5g), while RNF5 overexpression increased STING ubiquitination, reduced C1q levels, increased synaptic proteins, and improved cognitive function in septic mice (Extended Data Fig. 5h–t). Importantly, the K150 mutation disrupted the interaction between STING and RNF5, highlighting the dependence of RNF5-mediated STING ubiquitination on K150 (Extended Data Fig. 5u). Together, these findings underscore the importance of K150-mediated STING ubiquitination in mcroglia in preventing excessive synaptic pruning and cognitive dysfunction, with RNF5 serving as a key regulator of this process.

4-OI Reduces γδ T17 Cell Migration and Promotes K150-Mediated STING Ubiquitination

To explore potential therapeutic strategies for SAE, we considered the role of small intestinal γδ T17 cell migration and the strong correlation between sepsis, gut barrier dysfunction, and systemic inflammatory response²⁰. Proteomic analysis of the small intestine from septic mice revealed significant upregulation of aconitate decarboxylase 1 (ACOD1), which produces itaconate, known for its anti-inflammatory effects¹³ (Fig. 5a–c). Hence we further investigated the protective effects of 4-octyl-itaconate (4-OI), a blood-brain barrier-permeable itaconate derivative, on sepsis-induced systemic inflammation.

Notably, administration of 4-OI led to a marked reduction in pro-inflammatory cytokines, as demonstrated by multiplex cytokine analysis (Extended Data Fig. 6a, b), with improved survival rates and enhanced intestinal barrier integrity (Fig. 5d–h). Interestingly, 4-OI significantly reduced the number of IL-7R^high CD8^low γδ T17 cells in the meninges, correlating with decreased migration from the small intestine (Fig. 5i, j). This reduction in migration was accompanied by mitigation of hippocampal synaptic and mitochondrial damage (Fig. 5k), reduced reactive oxygen species (ROS) levels, and enhanced antioxidant activity, as confirmed by dihydroethidium (DHE), malondialdehyde (MDA), and superoxide dismutase (SOD) assays, respectively (Extended Data Fig. 7a–d). Additionally, positron emission tomography-computed tomography (PET-CT) and laser speckle imaging demonstrated improved hippocampal glucose metabolism and restored sepsis-induced decrease in the cerebral blood flow after 4-OI treatment (Extended Data Fig. 8a, b).

Moreover, 4-OI reduced microglial phagocytosis of PSD95 (Fig. 5l), increased dendritic spine density (Fig. 5m), and improved cognitive performance in the MWM, OFT, Y-maze and NORT (Fig. 5n; Extended Data Fig. 7e–i). Mechanistically, 4-OI reduced the expression of cGAS and STING, leading to decreased C1q and increased levels of PSD95 and SYN, indicating that 4-OI could attenuate C1q-mediated synaptic pruning (Fig. 5o–q). In vitro, 4-OI inhibited IL-17A release and mitigated mitochondrial dysfunction in BV2 cells (Fig. 5r, s). 4-OI also promoted microglial M2 polarization, enhancing the M2/M1 ratio in both in vitro and in vivo models while reducing the migration and phagocytic activity of LPS-stimulated microglia (Extended Data Fig. 9a–i). Furthermore, 4-OI inhibited STING activity in microglia (Extended Data Fig. 9g).

Interestingly, transcriptomic analysis of the hippocampus from septic mice also showed upregulation of Acod1 (Fig. 5x; Extended Data Fig. 10a). In the hippocampus, 4-OI enhanced K150-mediated STING ubiquitination and upregulated RNF5 expression (Fig. 5y, z; Extended Data Fig. 5v, w). These results confirm that 4-OI could protect against SAE not only by inhibiting γδ T17 cell migration from small intestine into the meninges, but also by promoting K150-mediated STING ubiquitination in the hippocampus.

Silencing Acod1 in Microglia Aggravates Cognitive Dysfunction

Given the increased expression of ACOD1 in both the intestine and hippocampus after sepsis, we further investigated its role in microglia and its impact on cognitive impairment in SAE. We silenced Acod1 using Acod1– -AAV in Cx3cr1^Cre mice (Fig. 6a). This suppression significantly upregulated cGAS and STING expression in the hippocampus, leading to synaptic and mitochondrial damage (Fig. 6b–e). Silencing Acod1 also elevated C1q levels and reduced the levels of PSD95 and SYN, indicating that Acod1 suppression enhanced complement-mediated synaptic pruning (Fig. 6f, g). Correspondingly, microglial phagocytosis of PSD95 was notably increased following Acod1 suppression, further confirming its involvement in excessive synaptic pruning (Fig. 6h, i).

Consistent with the promotion of K150-mediated STING ubiquitination by 4-OI, Acod1 suppression led to a significant reduction in K150-mediated STING ubiquitination (Fig. 6j–l), promoting cGAS-STING pathway activation and exacerbating synaptic injury (Fig. 6b–i). Taken together, these findings suggest that ACOD1 could regulate microglial function by promoting K150-mediated STING ubiquitination, thereby limiting cGAS-STING activation and preventing excessive synaptic pruning. Therefore, Acod1 suppression in microglia could worsen hippocampal synaptic and mitochondrial dysfunction by impairing K150-mediated STING ubiquitination, activating the cGAS-STING-C1q pathway, and enhancing synaptic pruning, ultimately leading to cognitive dysfunction after sepsis.

Our latest study reveals that IL-7R^high CD8^low memory γδT17 cells, originating from the small intestine, could migrate to the lungs and contribute to acute lung injury following sepsis⁸. Interestingly, this present research further demonstrates that the migration of small intestine-derived IL-7R^high CD8^low γδ T17 cells to the meninges plays a pivotal role in exacerbating SAE, establishing a crucial link between intestinal γδT17 cells and central nervous system pathology. Our findings underscore the importance of gut-centric inter-organ interactions.

Prior research has predominantly focused on gut dysbiosis and intestinal barrier dysfunction in SAE^21,22. However, our study provides the first direct evidence of IL-7R^high CD8^low γδ T17 cell migration playing a pathogenic role in SAE. Notably, similar mechanisms have been observed in hypertension, where IL-17-secreting T cells in meninges could contribute to CNS injury⁵. Consistent with these findings, we observed a significant increase in γδ T17 cells in the meninges after CLP, along with an upregulation of IL-7R^high CD8^low γδ T17 cells. However, the number of γδ T cells in the brain parenchyma remained low in both CLP and sham-operated mice, suggesting a specific migration pattern to the meninges rather than the brain tissue.

IL-17A, a pro-inflammatory cytokine known to exacerbates immune responses under inflammatory conditions²³, appears to be a key mediator in this “small intestine-derived γδ T17 cells induced SAE”. We found that over 50% of IL-17A-secreting immune cells in the meninges following sepsis were γδ T cells, significantly higher than that of 30% in the sham group. We further confirmed that IL-7R^high CD8^low γδ T17 cells are the primary IL-17A source in the meninges after sepsis, as shown using Trdc^CreERT2 Il7r^flox/flox and Trdc^CreERT2 Cd8a^flox/flox transgenic mice. The IL-17A secreted by IL-7R^high CD8^low γδ T17 cells activates microglia, triggering CNS inflammation and exacerbating SAE.

Microglia, as CNS resident immune cells, can be neuroprotective or neurotoxic²⁴. In the CLP model of sepsis, activated microglia could exacerbate SAE by upregulation neuronal NAT10 expression²⁵, induction of hippocampal neuronal ferroptosis via the CXCL2/CXCR2 pathway²⁶, and aberrant synaptic pruning²⁷. In the peritoneal contamination and infection (PCI)-induced SAE model, microglial complement C1q-dependent synaptic pruning is shown to be able to worsen SAE¹², although the mechanisms underlying sepsis-induced C1q release remains unclear. Our study reveals that IL-17A-stimulated microglia upregulate C1q-dependent synaptic pruning through activation of the cGAS-STING pathway. C1q, a key initiator of the classical complement pathway, mediates synaptic pruning during both development and disease. We found that inhibiting STING with H151 or depleting microglia with PLX3397 significantly reduced C1q levels and synaptic pruning, suggesting that STING-mediated microglial C1q release contributes to the worsening SAE. Traditionally associated with antiviral immunity through the detection of cytosolic DNA and activation of type Ⅰ interferon responses, the cGAS-STING pathway has recently been implicated in neuroinflammatory diseases¹⁸. In multiple sclerosis, STING activation enhances autophagy in neurons, increasing their susceptibility to glutamate-induced excitotoxicity, while STING inhibitors such as C176 or H151 have been shown to reduce neuroinflammatory damage²⁸. Furthermore, reducing cGAS-STING signaling has been found to inhibit glial cell activation in aging models²⁹.

In exploring potential therapeutic interventions, we focused on itaconate, an immunomodulatory metabolite produced by the enzyme ACOD1. Elevated itaconate levels, dependent on ACOD1, have been observed in the blood during Plasmodium infection³⁰. Macrophages and myeloid cells produce itaconate under inflammatory stimulation, and microglia also express ACOD1 highly under pro-inflammatory conditions³¹, which aligns with our findings. Our proteomic and transcriptomic analyses revealed significant upregulation of ACOD1 in both the small intestine and hippocampus after sepsis. Exogenous supplementation with itaconate and its derivative 4-OI has been shown to mitigate systemic inflammation and exert anti-inflammatory effects in the small intestine during CLP-induced sepsis¹⁴. Given that itaconate can be released into the bloodstream³², we propose that 4-OI inhibits the migration of IL-7R^high CD8^low γδ T17 cells from the small intestine to the meninges, offering a theoretical strategy for SAE. Moreover, our data indicate that 4-OI enhances the expression of RNF5, an E3 ubiquitin ligase that promotes ubiquitination and degradation of STING at the K150 residue. RNF5 has been shown to target STING at K150 for ubiquitination and degradation following viral infection³³. The K150 site is crucial for recruiting the deubiquitinating enzyme VP1-2 to STING, influencing its ubiquitination by other E3 ligases such as TRIM32³⁴. These findings suggest that STING ubiquitination may differ across diseases, but the role of STING ubiquitination in CNS disorders remains unclear, our study is the first to demonstrate its relevance in SAE. By enhancing STING ubiquitination, 4-OI effectively dampens neuroinflammatory processes mediated by the cGAS-STING pathway. Although other E3 ligases such as TRIM32 have been shown to promote STING ubiquitination in different contexts like in herpes simplex virus encephalitis³⁴, the role of these ligases in CNS disorders remains to be elucidated. Future research should explore the involvement of additional ubiquitin ligases and the precise molecular mechanisms governing STING regulation in neuroinflammation.

In summary, our study highlights a novel mechanism wherein in the migration of small intestine-derived IL-7R^high CD8^low γδ T17 cells to the meninges leads to IL-17A secretion, microglia activation, and inhibition of K150-dependent STING ubiquitination. And promotes the cGAS-STING-C1q pathway, playing a key role in cognitive impairment during SAE. By enhancing STING ubiquitination and controlling γδT17 cell migration, 4-OI emerges as a promising therapeutic agent to alleviate the neurological consequences of sepsis. This study lays the foundation for “gut-brain” axis-targeted therapies in inflammatory CNS diseases, offering new hope for patients with SAE and other neuroinflammatory conditions.

Animals

Male WT C57BL/6 J mice, Trdc^CreERT2 mice, Il7r^flox/flox mice, Cd8a^flox/flox mice, and Cx3cr1^Cre mice (all on a C57BL/6 background), weighing between 23.0 and 25.0 grams and aged between 8 and 10 weeks, were obtained from Vital River Laboratory Animal Technology Co Ltd., Beijing, China. Trdc^CreERT2 mice were crossed with Il7r^flox/flox mice and Cd8a^flox/flox mice to generate Trdc^CreERT2Il7r^flox/flox and Trdc^CreERT2Cd8a^flox/flox offspring. To induce Cre activity in γδ T cells, tamoxifen (150 mg kg^− 1) was administered intraperitoneally to Trdc^CreERT2Il7r^flox/flox mice and Trdc^CreERT2Cd8a^flox/flox mice for five consecutive days. The CLP model was performed one week after tamoxifen induction. Kaede-transgenic (Kaede-Tg) mice (B6. Cg-Gt (ROSA)26Sor < tm1.1(CAG-kikGR) Kgwa>) were generously provided by M. Tomura of Kyoto University. The mice were housed in a controlled, specific pathogen-free environment. They were exposed to a 12:12 light/dark cycle, maintained at a regulated temperature and humidity, and provided with unrestricted access to food and water. Ethical approval for all experiments was obtained from the Experimental Animals Committee of Tongji Medical College (permission number: 4028), in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study adhered to ARRIVE guidelines (Animals in Research: Reporting In Vivo Experiments).

Animal model

Mice were anesthetized with sodium pentobarbital (0.3% solution) at a dose of 40 mg kg^–1 body weight. A midline laparotomy was performed to conduct a central lymphadenectomy, following which the abdominal region was shaved and sterilized.

A 1 cm incision was made along the midline to expose the cecum, which was ligated 1 cm from the distal end using a 4 − 0 silk suture. A 20-gauge needle was then used to puncture the cecum, allowing a small amount of cecal content to extrude from both openings. After the procedure, the ligated cecum was returned to the abdominal cavity, and the incision was closed in multiple layers. Either 50 mg kg^–1 of 4-OI or 10 ml kg^–1 of 0.9% saline was administered intraperitoneally immediately after CLP or sham-operation. Inhibitor H151 was administered at a concentration of 750 nM via intraperitoneal injection, three times per week for three weeks prior to the CLP procedure. Survival rates were monitored and assessed 7 days post-CLP.

To increase IL-17A expression in the brain, mice received adenoviruses (Adv) via lateral ventricle injection three weeks before sham surgery or CLP. Each Adv injection contained 1 × 10¹² plaque-forming units (PFU) of IL-17A-expressing recombinant AAV (IL-17A⁺-Adv). Control mice received an equivalent dose of Adv expressing GFP. In a separate experimental group, IL-17A neutralizing antibody was administered intraperitoneally at 100 µg per day for 5 days, with four doses given prior to the CLP challenge and one dose immediately after. Control mice received equivalent doses of normal hamster serum IgG.

To specifically interfere with gene expression in microglia, we injected Cx3cr1^Cre mice with DIO-AAV vectors. The AAVs were obtained from Brainvta (Wuhan, China) and included the following constructs: rAAV-SFFV-DIO-cGAS-2a-EGFP-WPREs, rAAV-SFFV-DIO-RNF5-His-2a-EGFP-WPREs, rAAV-CWV-DIO-(EGFP-U6)-shRNA1(Acod1), and rAAV-SFFV-DIO-STING1 (K150R)-2a-EGFP-W.

Kaede photoconversion

To track immune cells in the small intestines in vivo, we performed photoconversion following the method described in a previous study⁸. Briefly, following CLP or sham surgery, the small intestine of Kaede-transgenic mice was exposed to a 405-nm laser for 10 min, while the surrounding tissue was shielded from light with aluminum foil.

Brain stereotaxic injection

A previous study³⁵ detailed the stereotaxic injection of viruses and neutralizing antibodies into the hippocampus. Briefly, mice were anesthetized with 1–2% isoflurane. Openings were made at specific coordinates targeting the unilateral hippocampal CA1 region (x: ± 2.15 mm; y: − 2.5 mm; z: − 2.25 mm). A volume of 400 nanoliters (nL) was injected into each hippocampus at a rate of 26.67 nL min^–1. Following the injection, the needle was left in place for 5 min to ensure proper diffusion before removal. The CLP model was performed 21 days post-viral injection and 1 day after C1q neutralizing antibody administration.

Morris water maze

Experimental data were collected in a 120 cm circular pool filled with opaque water maintained at 20–22 ℃. Mice were trained four times daily for four days to locate a hidden platform submerged 1 cm below the water surface. The average latency to find the platform was calculated from four trials. Each mouse was allowed 60 seconds per trial to search for the platform. Mice remained on the platform for 15 seconds if they found it, or were placed on it for 10 seconds if they failed to locate it. A probe trial without the platform was conducted 24 h after the final hidden platform test. Mice were tested for 60 seconds to locate the platform. The number of times the mice crossed the platform target and the total time spent in the target quadrant were recorded. Mice were monitored via video cameras throughout training and probe trials. Data were analyzed using TopScan Lite (Clever Sys. Inc.).

Y-maze test

Three identical arms (30 cm long, 5 cm wide, 20 cm high) were positioned at 120° angles in the Y-maze (YM). In the initial 10-minute training session, mice explored two arms while the third arm was blocked. One hour later, mice were given free access to all arms during the retention test. Mice were recorded exploring the novel arm for 5 min. Each trial was separated by cleaning the Y-maze arms with 75% ethanol. Arm entries and time spent in the novel arm were recorded, and short-term memory was calculated as the ratio of novel arm time to total exploration time. Data were examined using TopScan Lite (Clever Sys. Inc.).

Open filed test

The testing area was an acrylic box measuring 40 cm long, 40 cm wide, and 30 cm high. The center of the base contained a 20 x 20 cm square core region. Each mouse was gently placed in the center of a dimly lit open-field arena and allowed to explore for 5 min. A mobile camera automatically recorded and tracked their movement. The time spent in the central region was used to assess exploratory behavior. To eliminate odors, the chamber floor was wiped with 75% ethanol after each session. Data were examined using TopScan Lite (Clever Sys. Inc.).

Novel objection recognition

Two similar objects were placed at the edges of a 40 × 40 × 30 cm box. Mice were then gently placed in the box and allowed to explore the objects freely for 5 min, with the time spent on each object recorded. Exploration was defined as sniffing or touching an object within 0–2 cm with the nose. One hour later, one object was replaced with a novel one, and mice were given another 5 min to explore both objects, with the time spent on each recorded. The novel object recognition rate was calculated as (time spent on the novel object / total time spent on both objects) × 100%.

Tissue preparation

Cold PBS was perfused transcardially into sedated mice. Brains and small intestines were then collected. For immunofluorescent labeling, tissues were fixed overnight at 4 ℃ in 4% paraformaldehyde and cryoprotected in 30% sucrose for at least 2 days. Brains were sectioned at 30 µm thickness using a Leica CM1950 cryostat. Small intestines were paraffin-embedded and sectioned at 4 µm thickness. Blood was collected from sedated mice via cardiac puncture into anticoagulant tubes for biochemical analysis. Plasma was obtained and stored at − 80 ℃. Small intestines were stored at − 80 ℃ after a gentle flush with cold PBS. After dissociation, the hippocampus was snap-frozen in liquid nitrogen and stored at − 80 ℃ for protein extraction.

Immunofluorescence

Brain slices were washed in PBS and blocked for 2 h at room temperature with 5% BSA (Biofroxx, Germany) and 0.3% Triton X-100 in PBS. The sections were incubated overnight at 4 ℃ with mouse anti-Iba1 (1:100, Abcam) and rabbit anti-PSD95 (1:250, CST) primary antibodies. After PBS washing, the sections were incubated with Alexa Fluor 488-conjugated goat anti-rabbit (1:1000, Abcam) and Alexa Fluor 549-conjugated goat anti-mouse (1:1000, Abcam) for 1 h at room temperature the following day. The slices were mounted with SouthernBiotech DAPI Fluoromount-G after a final wash.

After deparaffinization in xylene, intestinal sections were rehydrated through a graded ethanol series to water. Sections were heated in citrate buffer (pH 6.0) for 20 min for antigen retrieval. After cooling to room temperature, the sections were blocked for 1 h with 5% BSA and 0.3% Triton X-100 in PBS. Sections were incubated overnight at 4 ℃ with rabbit anti-ZO-1, rabbit anti-MUC2, and rabbit anti-Occludin primary antibodies. The following day, sections were washed in PBS and incubated with Alexa Fluor 449-conjugated goat anti-rabbit (1:1000, Abcam) for 1 h at room temperature. The slices were mounted with SouthernBiotech DAPI Fluoromount-G after a final wash.

Microscopy and analysis

Immunofluorescent-stained brain slices were imaged using a Dragonfly spinning disk confocal microscope (Andor Technology) equipped with 405, 488, and 561 nm laser lines. Z-stack images were captured using a 60x oil immersion objective. Image processing and analysis were performed using Imaris 10.2 (Bitplane). Z-stack images were generated using Imaris 10.2 for 3D reconstruction and quantification of Iba1 and PSD95 expression. Immunofluorescent signal intensity in ROIs was compared across experimental groups.

Immunofluorescent-stained small intestine sections were scanned at 20x magnification using an Olympus VS200 slide scanner. High-resolution images were analyzed using Image J (1.8.0, NIH, USA). The fluorescence intensity of ZO-1, MUC2, and Occludin was quantified in randomly selected fields of view.

Primary microglia culture

Based on a recent study³⁶, we made slight modifications to the primary microglia culture procedure. Cortical brain tissues from 1-day-old C57BL/6J mice were dissected in cold HBSS (Servicebio, China, cat. no. G4203). After removing the meninges, the brain tissues were minced and rinsed three times with HBSS. The tissues were then digested in 0.25% trypsin-EDTA (ThermoFisher, cat. no. 25200056) for 20 min at 37 ℃ and triturated into a single-cell suspension. The primary cells were plated on poly-L-lysine-coated plates in Neurobasal medium supplemented with 10% FBS (Gibco, USA. cat. no. 10099-141), 1% GlutaMAX (Gibco, USA. cat. no. 35050061), and 1% penicillin-streptomycin (P.S.) (Gibco, USA. cat. no. 15140122). After 10 days of culture, the cell culture flasks were shaken at 250 rpm at 37 ℃ for 2 h. The collected culture media were centrifuged at 1000 rpm for 10 min, and the cells were resuspended in DMEM (Gibco, USA. cat. no. 11995073) with 10% FBS for inoculation

Primary γδ T culture

Primary γδ T cells were isolated from mouse spleens. Briefly, T25 cell culture flasks were coated with 5 µg mL^–1 TCR γδantibody one day prior. After anesthesia, mouse spleens were extracted under sterile conditions, homogenized with 2 mL of precooled mouse lymphatic separation solution, filtered through a 70-mesh filter, gently layered over 5 mL of precooled DMEM medium, and centrifuged at 2000g for 20 min. The intermediate layer was carefully aspirated, washed with PBS, and centrifuged at 420 × g for 6 min at 4 ℃. Cells were resuspended in inoculation medium (DMEM + 10% FBS + 1% P.S + 0.1 mM β-mercaptoethanol + 5 µM Zoledronic acid monohydrate + 1000 IU IL-2 + 20 ng mL^–1 IL-7) at 1 × 10⁵ cells mL^–1 and seeded into six-well plates for growth. The culture medium was replaced every 3 days with DMEM containing 10% FBS, 1% P.S, 0.1 mM β-mercaptoethanol, 1000 IU IL-2, and 20 ng mL^–1 IL-7. On day 12, primary cells were transfected and stimulated.

Primary γδ T cells-microglia co-culture

Primary γδ T cells were isolated from mouse spleens, and microglia were obtained from the brains of postnatal day 1 (P1) mice. For co-culture experiments, microglia were seeded in the bottom chamber of a transwell device with 0.4 µm pore-size polycarbonate membrane inserts, while γδ T cells were seeded in the top chamber. To investigate the impact of IL-17A, γδ T cells were transfected with siIL-17A or siCtrl using Lipofectamine RNAiMAX. Additionally, γδ T cells were treated with recombinant mouse IL-17A (100 ng mL^–1) and LPS (1000 ng mL^–1) for 6 h, followed by a medium change. The transwell inserts containing γδ T cells and microglia were co-cultured for 24 h post-treatment. After co-culture, microglia were harvested for analysis.

Cell Culture and treatment

BV2 murine microglial cells were obtained from Punosai Life Science and Technology Co., Ltd. and cultured in RPMI-1640 medium (Gibco, USA. cat. no. 11875119) supplemented with 10% FBS and 1% P.S. at 37 ℃ in 5% CO₂. To model SAE, BV2 cells were treated with 1000 µg mL^–1 LPS for 6 h. After incubation, Cells were collected and processed for further assays. To block the cGAS-STING pathway, BV2 cells were treated with H151 (0.75 µM) for 2 h prior to LPS stimulation. To generate BV2^ρ0 cells, which are devoid of mitochondria, BV2 cells were treated with ethidium bromide (50 ng mL^–1) for 4 weeks³⁷.

Cell Transfection

For gene silencing, primary γδ T cells were transfected with IL-17A siRNA, and BV2 microglial cells transfected with RNF5 siRNA. Additionally, BV2 cells were transfected with STING protein plasmids containing site-directed mutations for ubiquitination site analysis. All siRNAs and plasmids were obtained from Obio Technology Co., Ltd., Shanghai. Following the manufacturer’s instructions. Lipofectamine RNAiMAX was used for siRNA transfections, and Lipofectamine 3000 was used for plasmid transfections. Transfection efficiency was confirmed by Western blot analysis.

Western blot

Total proteins were extracted from cells and tissues, and protein concentrations were measured using the BCA assay. Equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes (Millipore). Membranes were incubated overnight at 4 ℃ with primary antibodies against ZO-1, MUC-2, Occludin, cGAS, STING, C1q, PSD95, Synaptophysin (SYN), ACOD1, RNF5, Ubiquitin (UB), β-actin,GAPDH and a custom-made antibody for the K150 ubiquitination site on STING (Abclonal). After incubation, membranes were treated with HRP-conjugated anti-mouse or anti-rabbit IgG. Protein bands were visualized using ECL Western blot Detection Reagents (Beyotime) and captured with a UVP gel documentation system (UVP, LLC, Phoenix). Band intensity was quantified using ImageJ (1.8.0, NIH, USA).

Synaptosomal proteins were extracted from mouse hippocampi using the Syn-PER Synaptic Protein Extraction Reagent (Thermofisher) following the manufacturer’s instructions. Western blot analysis was then performed on the isolated synaptosomal proteins.

FCM analysis

As previously dicribed⁸, cell suspensions from the meninges and small intestine were prepared. After counting 1 × 10⁶ cells, they were resuspended in 100 µL of 1% BSA in PBS, blocked with 5 ng µL^–1 anti-CD16/CD32 antibody for 5 min at 4 ℃, and then stained with the desired antibodies. Antibodies used for extracellular staining included fixable viability dye, CD45, CD3, TCR γδ, IL7R, CD8, F4/80, CD206, CD86, and CD11B. For intracellular staining, cells were first labeled with surface markers, then frozen and permeabilized before being labeled with IL-17A antibody (4 ng µL^–1). Samples were analyzed using a Beckman CytoFLEX or BD LSRFortessa X-20 cytometer (BD Biosciences, San Jose, CA). Data were analyzed using FlowJo 10.0 (FlowJo, Oregon, USA).

JC-1 Assay

Mitochondrial membrane potential in BV2 microglial cells was measured using the JC-1 detection kit (Elabscience, China) according to the manufacturer’s instructions. BV2 cells were seeded in 6-well plates following experimental criteria. After treatment, cells were stained with JC-1 at 37 ℃ for 20 min. The cells were rinsed with buffer solution and analyzed using a fluorescence microscope or flow cytometer after incubation. The mitochondrial membrane potential was assessed by the red-to-green fluorescence ratio.

Ubiquitination level detection

MG132 (100 µM) was added to cells 6 h before collection to inhibit proteasome activity and accumulate ubiquitinated proteins in cell culture studies. Hippocampus tissues were lysed in IP lysis buffer containing protease and phosphatase inhibitors to extract proteins. The lysates were incubated overnight at 4 ℃ with an anti-STING antibody and protein A/G magnetic beads for co-immunoprecipitation. After washing, the co-immunoprecipitated proteins were eluted from the beads. Eluted proteins were analyzed by Western blot using an anti-UB antibody (1:1000, Proteintech) to measure ubiquitination.

Detection of ubiquitination modification sites

Immunoprecipitation (IP) was performed to pull down STING from BV2 microglial cells for the identification of ubiquitination modification sites. Ubiquitination sites were identified by mass spectrometry following STING isolation. The identified modification sites were used to create mutant plasmids for each ubiquitination site using site-directed mutagenesis. These mutant plasmids were then transfected into BV2 cells to specifically alter STING ubiquitination sites. Ubiquitination levels of WT and mutant STING proteins were compared in subsequent assays to identify key ubiquitination sites.

Microglia depletion

PLX3397 (Selleck, China) was used to deplete brain microglia. As previously reported³⁸, PLX3397 was incorporated into AIN-76A chow at a concentration of 300 mg kg^–1. The PLX3397-supplemented chow was prepared by Jiangsu Xietong, Inc., Nanjing. Mice were fed AIN-76A chow or PLX3397-supplemented chow (300 mg kg^–1 in AIN-76A) for 21 days. Microglia depletion was evaluated by quantifying microglia numbers and assessing Iba1 expression after the 21-day feeding period.

Golgi staining

After anesthesia, mouse brains were immediately placed in Golgi stain fixing solution and immersed in dye solution for 14 days at room temperature in the dark. The treatment solution was changed after 1 h and then stored at 4 ℃ in the dark for 3 days. Brain samples were sectioned into 60 µm coronal slices using a vibrating microtome. Dendritic spine morphology was examined under a microscope.

Reactive Oxygen Species (ROS) Detection by DHE Staining

Oxidative stress in the hippocampal CA1 region was measured using dihydroethidium (DHE) staining. Brain slices from treated animals were processed as previously described. Sections were treated with Beyotime DHE staining reagent according to the manufacturer’s instructions. F Slices were incubated with DHE (10 µM) for 30 min in a light-protected, humidified chamber at 37 ℃. Superoxide anions oxidize DHE to ethidium, which intercalates with DNA and fluoresces red, indicating tissue ROS levels.

After incubation, sections were washed in PBS and mounted with anti-fade media. Fluorescent signals were captured using an Olympus fluorescence microscope (Olympus, Japan) with DHE filters. Red fluorescence intensity in the hippocampal CA1 region was quantified using ImageJ (NIH, USA) to assess oxidative stress. Oxidative stress levels between experimental groups were assessed by statistical analysis of mean fluorescence intensity.

TEM

Mice were perfused transcardially with 2.5% glutaraldehyde. 1-mm coronal brain slices were collected immediately after perfusion. The CA1 region of the hippocampus was meticulously microdissected and post-fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer at 4 ℃ for 24 h. After treatment, BV2 and primary microglia were centrifuged and fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer at 4 ℃ for 24 h.

Both tissue samples and cells were processed similarly after fixation. After three washes with 0.1 M sodium cacodylate buffer, the samples were post-fixed in 1% OsO₄ in the same buffer at 4 ℃ for 1.5 h. After dehydration in graded ethanol (50%, 70%, 90%, 100%), the samples were embedded in epoxy resin. Thin sections (60–70 nm) were cut using a Leica EM UC7 ultramicrotome and placed on copper grids. For contrast enhancement, sections were stained with 2% uranyl acetate for 15 min and lead citrate for 10 min. Electron micrographs were captured using an 80 kV Hitachi HT7800 transmission electron microscope (HITACHI, Japan). Ultrastructural characteristics of hippocampal CA1 synapses, BV2 cells, and primary microglia were identified and analyzed.

MDA and SOD activity determination

MDA and SOD activity were measured using kits (Beyotime, China) according to the manufacturer’s instructions. The hippocampus was excised, weighed, and homogenized in MDA and SOD assay solutions. Plasma and homogenate supernatants were added to the reaction system. MDA activity was measured at 532 nm by absorbance and reported as µmol mg^–1 or µM. SOD activity in the samples was measured by absorbance at 450 nm and expressed as U mg^–1 or U mL^–1 of total protein.

Determination of cytokine levels

Cytokine levels in mouse plasma and hippocampus were measured using the ABplex Mouse Cytokine 8-Plex Assay Kit (Abclonal). The analyzed cytokines included IL-2, IL-4, IL-12p70, IL-1β, IL-6, IFN-γ, TNF-α, and IL-17A. The samples were clarified by centrifugation and analyzed by FCM.

Scratch Assay

BV2 microglial cell migration was assessed using a scratch assay. BV2 cells were seeded into 12-well plates and cultured until 90% confluence. A linear scratch was made in the cell monolayer using a sterile 200 µL pipette tip. After scratching, wells were gently rinsed with PBS to remove floating cells and debris before fresh media was added. Cell migration into the scratch region was tracked for 24 h using the Opera Phenix Plus high-content imaging system (Revvity). Images were frequently captured to monitor cell migration. The migration rate was calculated as the proportion of the scratch area covered by migrating cells relative to the control group. ImageJ (NIH, USA) was used for image analysis to compare cell migration rates between experimental conditions.

Phagocytosis Assay

The phagocytic capability of primary microglial cells was assessed using the Cell Meter™ Fluorimetric Phagocytosis Assay Kit (Red Fluorescence, Cat#21225). I Primary microglial cells were allowed to adhere overnight in a 96-well confocal cell culture plate. After experimental treatments, cells were incubated with Protonex™ 600-labeled beads according to the manufacturer’s instructions. After incubation, non-internalized beads were washed away, and fluorescence from ingested beads was detected using a Zeiss LSM 880 confocal laser scanning microscope. Phagocytic activity was quantified as bead engulfment (%), calculated as the ratio of red fluorescent area (engulfed beads) to total cell area. ImageJ software (NIH, USA) was used to calculate this ratio and statistically compare microglial phagocytic capability between experimental groups.

PET-CT Imaging for Brain Metabolism

Brain metabolic activity was assessed via PET-CT imaging on mice 24 h after CLP. To minimize glucose fluctuations that could affect 18F-fluorodeoxyglucose (18F-FDG) uptake, mice were fasted for 4–6 h with access to water before imaging. After fasting, mice were intravenously administered 18F-FDG at a dose of 15.6 ± 1.7 MBq in 0.5 mL saline. Following injection, mice were housed in a warm, calm environment for 30–45 min to enhance tracer uptake and minimize stress and muscular activity that could affect glucose distribution.

Small-animal PET-CT scanners with high resolution and sensitivity were utilized under isoflurane anesthesia. Optimized scanning covered the entire brain, with a focus on the hippocampi. CT scans provided anatomical reference, followed by PET scans to assess glucose metabolism.

The SUVmean, which represents the average 18F-FDG uptake adjusted for body weight and injected dose, was used to measure brain metabolic activity, particularly in the hippocampal region. SUVmean data were statistically compared between experimental groups to evaluate post-CLP brain metabolism.

Laser Speckle Contrast Imaging (LSCI) for Cerebral Blood Flow

Cerebral blood flow dynamics in mice were assessed by LSCI 24 h post-CLP. After anesthesia, the scalp was gently removed to expose the skull. LSCI uses coherent laser light to illuminate cortical tissue. Red blood cells in cerebral vessels alter the speckle pattern captured by a high-resolution camera. These oscillations are used to map blood flow across the cortical surface in real time. Data were collected using a laser-synchronized high-sensitivity CCD camera. Relative blood flow changes were computed based on speckle contrast. To evaluate the impact of CLP on cerebral perfusion, blood flow was quantified and statistically compared between experimental groups.

TMT-based quantitative proteomics analysis

Sample preparation

Proteomic analysis was conducted on CLP and sham-operated mice (n = 3 per group). After anesthesia, intestinal tissues were promptly removed and snap-frozen in liquid nitrogen. Weighed protein samples were added to SDS L3-EDTA lysis buffer. The samples were homogenized (25,000 g, 4 ℃, 5 min). The homogenate was then treated with 10 mM DTT and 45 mM IAM and incubated in the dark for 45 min. Proteins were precipitated with cold acetone, followed by centrifugation (25,000g, 4 ℃, 15 min). After air-drying the pellet, SDS L3-free lysis solution was added to fully solubilize the proteins. Following another centrifugation (25,000 g, 4 ℃, 15 min), protein concentration was measured using the Bradford method. Proteins were desalted after trypsin digestion. The freeze-dried peptides were separated using a Shimadzu LC-20AB liquid chromatography system.

High-performance liquid chromatography and mass spectrometry

Peptides were separated using gradient elution on a self-packed C18 column on an Easy nLC 1200 system. The isolated peptides were analyzed using Data Dependent Acquisition (DDA) tandem mass spectrometry. TMT-proteomic analysis was performed by BGI-Shenzhen, China.

Bioinformatics analysis

The UniProtKB database (Release 2016_10) was used to obtain FASTA protein sequences of differentially expressed proteins for Gene Ontology (GO) mapping and annotation. KEGG Orthology (KO) identities and pathways were determined by aligning the FASTA sequences of significantly altered proteins with the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://geneontology.org/). Protein expression data from relevant KEGG pathways were visualized using a hierarchical clustering heat map.

Statistical analysis

Data reported as mean ± standard deviation (Mean ± SD) were analyzed using GraphPad Prism 10.0 software. Data normality was assessed using the D’Agostino & Pearson test or Shapiro-Wilk test. One-way analysis of variance (ANOVA) and two way ANOVA with Tukey’s or Šídák’s multiple comparisons test was used for multiple-group comparisons. Unpaired Student’s t-test or Mann-Whitney test was used for group comparisons, and Log-rank (Mantel–Cox) test was performed for survival rate analysis. P values < 0.05 were considered statistically significant. All data analyses and statistical figures were generated using GraphPad Prism 10.0.

Reporting summary

Additional details on study design are available in the Nature Portfolio Reporting Summary associated with this article.

Acknowledgements

The study was funded by the National Natural Science Foundation of China (Grant No. 82472223 and 82071480 [to JCZ]; Grant No. 82272231 [to SYY]; Grant No. 82302471 [to YJZ]; Grant No. 82402568 [to BX]), and the Science foundation of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology (2022xhyn054 [to YY]).

Author contributions

J.-c.Zhang, K. Hashimoto and Y.-m.Wu conceived and designed experiments. Y.-m.Wu and Y.-j.Zhang carried out most of the experiments, analyzed the data and prepared figures. Y.Yu, Z.-f.Zhen and X.W carried out the TMT-based quantitative proteomics analysis. B.Xie and M.-q.Han conducted CLP surgery. X.-y.Zhang, X.-q.Sun, and M.-y.Wang carried out the Kaede photoconversion. X.-y.Weng conducted cell culture. J.-c.Zhang, S.-y.Yuan and Y.Shang contributed to experimental design, data analysis, discussions and advice. Y.-m.Wu, J.-c.Zhang and Y.-j.Zhang wrote the manuscript.

Competing interests

The authors declare no competing interests.

Robba C, Crippa IA, Taccone FS (2018) Septic Encephalopathy. Curr Neurol Neurosci Rep 18:82. 10.1007/s11910-018-0895–6
Assimakopoulos SF et al (2018) Gut-origin sepsis in the critically ill patient: pathophysiology and treatment. Infection 46:751–760. 10.1007/s15010-018-1178–5
Nielsen MM, Witherden DA, Havran WL (2017) gammadelta T cells in homeostasis and host defence of epithelial barrier tissues. Nat Rev Immunol 17:733–745. 10.1038/nri.2017.101
Wu P et al (2014) gammadeltaT17 cells promote the accumulation and expansion of myeloid-derived suppressor cells in human colorectal cancer. Immunity 40:785–800. 10.1016/j.immuni.2014.03.013
Santisteban MM et al (2024) Meningeal interleukin–17-producing T cells mediate cognitive impairment in a mouse model of salt-sensitive hypertension. Nat Neurosci 27:63–77. 10.1038/s41593-023-01497-z
Benakis C et al (2016) Commensal microbiota affects ischemic stroke outcome by regulating intestinal gammadelta T cells. Nat Med 22:516–523. 10.1038/nm.4068
Sun J et al (2015) IL–17A is implicated in lipopolysaccharide-induced neuroinflammation and cognitive impairment in aged rats via microglial activation. J Neuroinflamm 12:165. 10.1186/s12974-015-0394–5
Xie B et al (2024) Gut-derived memory gammadelta T17 cells exacerbate sepsis-induced acute lung injury in mice. Nat Commun 15:6737. 10.1038/s41467-024-51209–9
Zhang YQ et al (2021) Dopamine D2 receptor regulates cortical synaptic pruning in rodents. Nat Commun 12:6444. 10.1038/s41467-021-26769–9
Dejanovic B et al (2022) Complement C1q-dependent excitatory and inhibitory synapse elimination by astrocytes and microglia in Alzheimer's disease mouse models. Nat Aging 2:837–850. 10.1038/s43587-022-00281–1
Hao W et al (2024) Gut dysbiosis induces the development of depression-like behavior through abnormal synapse pruning in microglia-mediated by complement C3. Microbiome 12, 34 10.1186/s40168-024-01756–6
Chung HY et al (2023) Microglia mediate neurocognitive deficits by eliminating C1q-tagged synapses in sepsis-associated encephalopathy. Sci Adv 9:eabq7806. 10.1126/sciadv.abq7806
Mills EL et al (2018) Itaconate is an anti-inflammatory metabolite that activates Nrf2 via alkylation of KEAP1. Nature 556:113–117. 10.1038/nature25986
Li W et al (2023) 4-octyl itaconate as a metabolite derivative inhibits inflammation via alkylation of STING. Cell Rep 42:112145. 10.1016/j.celrep.2023.112145
Ramakrishnan RK et al (2020) IL–17 Induced Autophagy Regulates Mitochondrial Dysfunction and Fibrosis in Severe Asthmatic Bronchial Fibroblasts. Front Immunol 11:1002. 10.3389/fimmu.2020.01002
Ghosh M, Saha S, Li J, Montrose DC, Martinez LA (2023) p53 engages the cGAS/STING cytosolic DNA sensing pathway for tumor suppression. Mol Cell 83:266–280. 10.1016/j.molcel.2022.12.023
Hu X et al (2022) Emerging role of STING signalling in CNS injury: inflammation, autophagy, necroptosis, ferroptosis and pyroptosis. J Neuroinflamm 19:242. 10.1186/s12974-022-02602-y
Ferecsko AS et al (2023) STING-Triggered CNS Inflammation in Human Neurodegenerative Diseases. Biomedicines 11:1375. 10.3390/biomedicines11051375
Meyer-Schwesinger C (2019) The ubiquitin-proteasome system in kidney physiology and disease. Nat Rev Nephrol 15:393–411. 10.1038/s41581-019-0148–1
Chancharoenthana W, Kamolratanakul S, Schultz MJ, Leelahavanichkul A (2023) The leaky gut and the gut microbiome in sepsis - targets in research and treatment. Clin Sci 137:645–662. 10.1042/CS20220777
Xu K et al (2024) Phosphatidylserine improves aging sepsis survival, modulates gut microbiome, and prevents sepsis-associated encephalopathy. Biomed Pharmacother 178:117200. 10.1016/j.biopha.2024.117200
Xu Y et al (2024) Palmatine ameliorated lipopolysaccharide-induced sepsis-associated encephalopathy mice by regulating the microbiota-gut-brain axis. Phytomedicine 124:155307. 10.1016/j.phymed.2023.155307
Mills K (2023) IL–17 and IL–17-producing cells in protection versus pathology. Nat Rev Immunol 23:38–54. 10.1038/s41577-022-00746-9
Li J et al (2021) Microglial Phenotypic Transition: Signaling Pathways and Influencing Modulators Involved in Regulation in Central Nervous System Diseases. Front Cell Neurosci 15:736310. 10.3389/fncel.2021.736310
Gao S et al (2024) N-acetyltransferase 10 mediates cognitive dysfunction through the acetylation of GABA(B)R1 mRNA in sepsis-associated encephalopathy. Proc. Natl. Acad. Sci. U. S. A. 121, e1884403175 10.1073/pnas.2410564121
Yang YS et al (2024) Quercetin protects against sepsis-associated encephalopathy by inhibiting microglia-neuron crosstalk via the CXCL2/CXCR2 signaling pathway. Phytomedicine 134:155987. 10.1016/j.phymed.2024.155987
Yin XY et al (2023) HMGB1 mediates synaptic loss and cognitive impairment in an animal model of sepsis-associated encephalopathy. J Neuroinflamm 20:69. 10.1186/s12974-023-02756-3
Woo MS et al (2024) STING orchestrates the neuronal inflammatory stress response in multiple sclerosis. Cell 187:4043–4060. 10.1016/j.cell.2024.05.031
Jimenez-Loygorri JI et al (2024) Mitophagy curtails cytosolic mtDNA-dependent activation of cGAS/STING inflammation during aging. Nat Commun 15:830. 10.1038/s41467-024-45044-1
Ramalho T et al (2024) Itaconate impairs immune control of Plasmodium by enhancing mtDNA-mediated PD-L1 expression in monocyte-derived dendritic cells. Cell Metab 36:484–497. 10.1016/j.cmet.2024.01.008
Chen YJ et al (2023) Targeting IRG1 reverses the immunosuppressive function of tumor-associated macrophages and enhances cancer immunotherapy. Sci Adv 9:eadg654. 10.1126/sciadv.adg0654
Gu X et al (2023) Itaconate promotes hepatocellular carcinoma progression by epigenetic induction of CD8(+) T-cell exhaustion. Nat Commun 14:8154. 10.1038/s41467-023-43988-4
Zhong B et al (2009) The ubiquitin ligase RNF5 regulates antiviral responses by mediating degradation of the adaptor protein MITA. Immunity 30:397–407. 10.1016/j.immuni.2009.01.008
Bodda C et al (2020) HSV1 VP1–2 deubiquitinates STING to block type I interferon expression and promote brain infection. J Exp Med 217:e20191422. 10.1084/jem.20191422
Wu X et al (2024) Foxq1 activates CB2R with oleamide to alleviate POCD. Brain Pathol. e13289 10.1111/bpa.13289
Liu Y et al (2024) NG2 glia protect against prion neurotoxicity by inhibiting microglia-to-neuron prostaglandin E2 signaling. Nat Neurosci 27:1534–1544. 10.1038/s41593-024-01663-x
Zecchini V et al (2023) Fumarate induces vesicular release of mtDNA to drive innate immunity. Nature 615:499–506. 10.1038/s41586-023-05770-w
Shi Y et al (2019) Microglia drive APOE-dependent neurodegeneration in a tauopathy mouse model. J Exp Med 216:2546–2561. 10.1084/jem.20190980

There is NO Competing Interest.

ExtendeddataTable.docx
Extended data table 1
SupplementaryFigures.pdf
Supplementary Figures

Download PDF

Version 1

posted

You are reading this latest preprint version

Small intestinal γδ T17 cells promote C1q-mediated SAE by synaptic pruning in mice

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Materials and Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1