Activation of the Cerebrospinal Fluid-Contacting Nucleus Mitigates Systemic Inflammation Induced by Sepsis

Abstract

Reciprocal interactions between the central nervous system and immune are now recognized as critical components of the host response in sepsis. However, the precise mechanism by which the central nervous system modulates immune function remains largely elusive. The cerebrospinal fluid (CSF)-contacting nucleus (CSF-contacting nucleus) is a unique nucleus, with its neurons located in the brain parenchyma and processes extending into the CSF. Here, we demonstrate that the CSF-contacting nucleus plays a protective role in sepsis. We observed a significant Fos high expression within the CSF-contacting nucleus in response to sepsis with immunofluorescence assays. Ablation of the CSF-contacting nucleus exacerbated sepsis severity, result in levels of proinflammatory cytokines (IL-1β, IL-6, and TNFα) elevated. Conversely, chemogenetic activation of the CSF-contacting nucleus resulted in a reduction in proinflammatory cytokine (IL-1β, IL-6, and TNFα) levels while simultaneously increasing the levels of IL-10. Inhibition of the CSF-contacting nucleus specifically elevated IL-6 levels. Notably, single-cell RNA sequencing was employed to identify key components within the CSF-contacting nucleus that regulate systemic inflammation in septic mice. Our findings indicate an upregulation of immune-related genes in the CSF-contacting nucleus during sepsis. Collectively, our study underscores a significant role of the CSF-contacting nucleus in neuro-immune interactions and suggests its potential as a novel therapeutic target for immune-mediated diseases.

INTRODUCTION

Sepsis is considered as a life-threatening syndrome characterized by organ dysfunction resulting from a dysregulated host response to infection (1–3). Inflammation and immunosuppression may occur either sequentially or simultaneously during sepsis. Despite extensive efforts to elucidate the pathophysiology and mechanisms underlying sepsis-associated immunosuppression (2,4,5), current therapeutic options for addressing the imbalance in inflammatory responses induced by sepsis remain limited. Recently, increasing attention has been directed toward the interplay between brain function and peripheral inflammation (6–8). Studies examining the brains of patients who succumbed to sepsis have revealed significant neuronal activation and apoptosis in specific regions (9). The vagus nerve has been reported to be critical in neural-inflammatory reflex, which controls innate immune responses and inflammation during pathogen invasion and tissue injury (10–14).

The cerebrospinal fluid contacting neurons (CSF-CNs) constitute a unique population of neurons whose cell bodies or axons are directly exposed to the CSF (15,16). These neurons exhibit polymodal interoceptive properties and have been shown to respond to bacterial metabolites present in the CSF. Several studies have demonstrated that activation of CSF-CNs during CNS infections leads to the secretion of compounds that enhance host survival (17). For instance, research on larval zebrafish has indicated that blocking the secretory activity of CSF-CNs reduces host survival during Streptococcus pneumoniae infection (18), suggesting that factors such as neuropeptides, neuromodulators, and secretory proteins released by these neurons play a crucial role in host defense mechanisms. However, the precise anatomical distribution of these CSF-CNs remains poorly defined, and their synaptic connections with other brain nuclei have yet to be fully characterized.

Notably, a distinct cluster of neurons known as the CSF-contacting nucleus is situated within the upper portion of the fourth ventricle and the ventral gray matter of the lower portion of the aqueduct consistently. This structure was first identified and named by the Zhang laboratory in their pioneering studies (19). Extensive projections throughout the brain were demonstrated between various functional nuclei and the CSF-contacting nucleus, suggesting that CSF-contacting nucleus plays a unique role in the crosstalk between the brain and CSF (20,21). However, it remains largely unexplored that the role of the CSF-contacting nucleus in the course of sepsis.

To explore the involvement of the CSF-contacting nucleus in sepsis, this study examined the expression of Fos within the CSF-contacting nucleus during sepsis. Subsequently, we employed chemogenetic techniques to modulate the activity of the CSF-contacting nucleus and evaluated its effects on the serum cytokine levels in septic mice. Additionally, we conducted single-cell RNA sequencing (Smart-seq2) to identify key genes within the CSF-contacting nucleus that regulate sepsis-induced systemic inflammation, thereby providing evidence and potential therapeutic targets for neuro-immune interactions and immune-related diseases.

MATERIALS AND METHODS

Ethics statement and animals

Male wild-type C57BL/6J mice, weighing 25 g ± 5 g, were obtained from the Experimental Animal Centre of Xuzhou Medical University (no. SCXK [Jiangsu] 2015–0009). All animal procedures adhered to the Guide for the Care and Use of Laboratory Animals and were approved by the Committee for the Ethical Use of Laboratory Animals, Xuzhou Medical University (ethics approval no. L20211001001). Animals were raised on a 12-h light/dark cycle (lights on 7:00 am) with ad libitum food and water in a controlled environment with temperatures of 24 ± 2°C.

General surgical procedures and viral delivery

Mice were anesthetized with 1.5% isoflurane (AbbVie Pharmaceutical) at an oxygen flow rate of 1 L/min. Then they were placed on the stereotaxic frame (RWD Life Science, 68513) with a heating pad to keep warm. The rAAV2/R-hSyn-CRE-WPRE-hGHpolyA (500–600 nL, titer: 3.5 × 10¹² v.g./mL, Braincase) was injected into the LV (lateral ventricle) anterior-posterior (AP): −0.60 ± 0.02 mm, dorsal-ventral (DV): −2.20 ± 0.02 mm, medial-lateral (ML): −1.4 ± 0.02 mm) with a glass pipette (tip diameter: 10–20 mm) at rate of 50 nL/min with a microinjection pump (RWD Life Science, r480). The AAV2/8-hSyn-DIO-hM3Dq-EGFP/AAV2/8-hSyn-DIO-hM4Di-mCherry (150 nL, titer: 4.5 × 10¹² v.g./mL, Braincase) was injected into the CSF-contacting nucleus (AP = 5 ± 0.02 mm, DV = −3.25 ± 0.02 mm, ML = 0 ± 0.02 mm). In addition, the AAV2/8-hSyn-DIO-EGFP was used as control (150 nL, titer: 4.27 × 10¹² v.g./mL, Braincase). After the injection, the glass pipettes were left in place for another 10 min. The animals were allowed to recover on a heating blanket. Cecum ligation and puncture (CLP) surgery was performed 4 weeks after the virus injection. Only animals with correct virus expression areas were included for analysis.

Chemogenetic manipulations

For in vivo chemogenetic manipulation of CSF-contacting nucleus, four weeks following the administration of the viral injection to the mice, saline or clozapine N-oxide (CNO, i.p., 1 mg/kg, Sigma) (22) was administrated intraperitoneally for 7 days consistently. Two hours following the final intraperitoneal injection of either saline or CNO, all mice were subjected to CLP surgery. The animals were euthanized 6 h after the CLP surgery to collect the serum, brain tissue and small intestine for further experiments.

Cecum ligation and puncture

CLP was applied to construct the sepsis model in mice as before (²³). Briefly, anesthetized mouse was supine-fixed. An abdominal incision was implemented after sterilized the abdominal skin in the midline of the abdomen. Expose, ligate, and perforate the cecum with the 21-gauge needle to establish the sepsis-model. Then replace the cecum to enterocoelia before the incision. Complementally, 1 mL of normal saline injected subcutaneously was needed. 6 h later, the standard septic mice would behave lethargy, piloerection, and diarrhea. Mice in the control group underwent the same procedure except for puncture and ligation.

Immunofluorescence

Mice were anesthetized with tribromoethanol (0.5 mg/g, i.p.). Saline and 4% paraformaldehyde (PFA) were used for perfuse procedure. Brains were carefully extracted and postfixed in 4% PFA for 6–8 h, Then, the postfixed brains were cryoprotected in 30% sucrose at 4°C. Free-floating brain sections were obtained at a thickness of 40 μm using a Cryostat (Leica CM 1950). The sections were blocked for 30 min at room temperature in 5% normal donkey serum solution. Sections were incubated with primary antibodies in PBST (with 1% normal donkey serum) at 4°C overnight followed by secondary antibodies in PBS at room temperature for 2 h with shaking. The primary antibody used was rabbit anti-Fos (1:2,000, Abcam, ab190289) while the secondary antibody employed was donkey anti-rabbit IgG-Alexa 488 (1:400, Jackson ImmunoResearch Laboratories, 2668665).

Hematoxylin-eosin staining

The intestines were fixed in 4% paraformaldehyde, embedded in paraffin wax, and subsequently processed into sections with a thickness of 5 μm. The sections were stained with hematoxylin and eosin and subsequently captured under a confocal microscope (Olympus, Tokyo, Japan) at ×20 magnification.

Survival rate

For the survival study, 52 mice were used to investigate the ablation of the CSF-contacting nucleus on the survival rate of sepsis induced by CLP surgery. In brief, they were randomly divided into the following four groups: control group (the mice only received PBS injection into the LV, n = 12), CLP group (mice only received CLP surgery, n = 12), SAP group (mice received CB-SAP injection into LV, n = 12), and SAP + CLP group (the mice receive CB-SAP injection into LV, 7 days later, CLP surgery was performed, n = 16). Mortality was recorded for up to 72 h after the CLP surgery.

Measurement of cytokine levels

The serum levels of IL-1β, IL-6, IL-10, and TNF-α were quantified using ELISA kits obtained from Proteintech Co., Ltd. (Wuhan, China). All procedures were meticulously carried out in full accordance with the manufacturer’s instructions.

Collection of CSF-contacting nucleus with laser capture microdissection

Animal model and tissue dissection

The CTB-594 was injected into the LV to label the CSF-contacting nucleus specifically. Five days later, these mice randomly received CLP operation or control operation respectively (n = 4 for the CLP group, n = 3 for the control group). Mice were anesthetized with 2,2,2-tribromoethanol (Sigma-Aldrich) and the brains were dissected 6 h after the surgery. We cleaned all tools and work surfaces with RNaseZAP wipes (Ambion). Tissues were carefully dissected and immediately snap-frozen in 2-methylbutane (Sigma-Aldrich) on dry ice within 15 min after decapitation. The tissues were subsequently stored at −80°C until further processing.

Tissue sectioning

To minimize tissue thawing, mouse brains were embedded in precooled OCT (Histolab) on dry ice. Thin coronal sections (14 μm) were collected in a cryostat and placed onto PEN membrane glass slides (Zeiss) at −20°C during the sectioning and subsequently stored at −80°C for further processing. Nuclease-free water and phosphate-buffered saline (PBS) (LifeTechnologies) were used, with reagents and tubes being of molecular biology/PCR grade. All following steps were performed on work surfaces and with equipment cleaned with RNaseZAP.

Laser capture microscopy

Slides were inserted into the slide holder of the Zeiss PALM MicroBeam microscope. To minimize potential contamination of adjacent cells, the cutting outline was meticulously delineated around individual neurons (Fig. 1C). Only neurons labeled with CTB-594 were selected. CSF-contacting neurons were excised at 20× magnification with laser power set to its minimal level with CloseCut + AutoPLC mode. The relative humidity was maintained from 17% to 65% and the temperature range was from 21°C to 27°C. A ribozyme-free 500-μL EP tube was positioned directly above the slide, with the open side down. The selected CSF-contacting neurons were automatically cut and inserted into the EP tube cap with the laser pulse. A 10-μm volume of lysis buffer was added and stored in liquid nitrogen for further processing after neurons were collected in the cap.

Fig. 1.

Chemogenetic manipulation of the CSF-contacting nucleus did not affect the locomotor activity. A–B, Chemogenetic activation of VB did not affect locomotor activity tested by open field test. Total distance (A) and velocity (B) showed no significant difference after CNO injection (1 mg/kg, i.p.) between AAV-DIO-EGFP group and AAV-DIO-Hm3Dq-EGFP group (n = 6 mice for each group). C–D, Chemogenetic inhibition of VB did not affect locomotor activity tested by open field test. Total distance (C) and velocity (D) showed no significant difference after CNO injection (1 mg/kg, i.p.) between AAV-DIO-EGFP group and AAV-DIO-Hm3Dq-EGFP group (n = 6 mice for each group). Two-way ANOVA was used followed by Bonferroni post hoc test. ANOVA, analysis of variance.

Single-cell RNA sequencing (Smart-seq2)

Total RNA was extracted from the tissue using TRIzol Reagent according to the manufacturer’s instructions, with 3 biological replicates for the control group and 4 biological replicates for the CLP group. RNA purification, reverse transcription, library construction, and sequencing were performed by Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China). The RNA-seq transcriptome library was prepared using the Illumina Stranded mRNA Prep, Ligation kit (San Diego, CA) with 1 μg of total RNA. Messenger RNA was isolated through polyA selection using oligo (dT) beads and then fragmentation. The SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA) was used to synthesize double-stranded cDNA with random hexamer primers. Then the cDNA underwent end-repairing, phosphorylation, and adaptor ligation. For cDNA target, libraries were size-selected for 300 bp on a 2% Low Range Ultra Agarose gel, followed by PCR amplification for 15 cycles with using of Phusion DNA polymerase (NEB). Using the NovaSeq Reagent Kit to sequence library on a NovaSeq X Plus platform (PE150) following the quantification with the Qubit 4.0. According to default parameters, trim the raw paired-end reads and the quality-controlled was using fastp. Using the HISAT2 software in a directed manner to align clean reads to the reference genome. Mapped reads for each sample were assembled via a reference-based approach using StringTie, and transcript expression levels were calculated using the transcripts-per-million (TPM) method to identify differentially expressed genes (DEGs). Gene abundances were quantified with RSEM, and differential expression analysis was carried out using DESeq2 or DEGseq. DEGs were defined as transcripts with |log2FC| ≥ 1 and an FDR < 0.05 (for DESeq2) or FDR < 0.001 (for DEGseq). Significant enrichment of the identified DEGs in GO terms and metabolic pathways was determined using GO and KEGG analyses with a Bonferroni-corrected P value <0.05 compared to the whole transcriptome background. Using Goatools and Python script software to conduct GO and KEGG pathway analyses. The PPI network was constructed using the website http://STRING-db.org.

Statistics

All experiments were conducted using a double-blind design. Data were presented as mean ± standard error (SEM) unless stated. The statistical analyses were performed using Prism 6 (GraphPad Software). Unpaired two-sided t test was applied for analyzing Fos expression. Log-rank test for trend was used to analyze the data obtained from the survival test. Data from the ELISA and HE assays were analyzed via one-way or two-way ANOVA analysis followed by Bonferroni post hoc tests. Statistical significance was set at P < 0.05.

RESULTS

Ablation of CSF-contacting nucleus aggravated the systemic inflammatory response in septic mice

We used cecum ligation and puncture (CLP) to elicit sepsis as before. To determine whether the CSF-contacting nucleus was involved in the sepsis, we performed the immunolabeling of Fos, a marker for neural activity, in the CSF-contacting nucleus. As shown in Figure 2A, Fos⁺ neurons in the CSF-contacting nucleus were elevated significantly 6 h later after the CLP-induced sepsis (Fig. 2B). Subsequently, we ablated the CSF-contacting nucleus specifically by injection of CB-SAP into the lateral ventricle (24) to observe the effect on the survival rate of the septic mice. Interestingly, we found that ablation of the CSF-contacting nucleus alone did not cause death in mice. However, the 72-h survival rate of septic mice with CSF-contacting nucleus ablation was significantly lower compared to mice that underwent CLP surgery alone (Fig. 2C, chi-square = 43.42, P < 0.001).

Fig. 2.

The CSF-contacting nucleus was involved in the CLP-induced sepsis. A, The expression of Fos in the CSF-contacting nucleus of the mice subjected to sepsis induced by CLP. The red represents the CSF-contacting nucleus, and the green is for the Fos. The scale bar is 50 μm. B, The analysis of the Fos expression between the CLP group and the control group as shown in (A), n = 4 for each group, ***P < 0.001, Student’s t test was used. (C) The 72-h survival rate of the mice in the control, CLP, SAP, and CLP + SAP groups, chi-square = 43.42, ***P < 0.001, log-rank test for trend was used. D–G, Quantitative analysis of the IL-1β (D), IL-6 (E), TNF-α (F), and IL-10 (G) in the serum of mice from the control, SAP, CLP, and SAP+CLP groups, n = 6 mice for control group, CLP group, and SAP+CLP group, n = 5 for SAP group. One-way ANOVA followed by Bonferroni post hoc test was used, *P < 0.05, **P < 0.01, ***P < 0.001. H–K, Representative histological images from small intestine sections stained with hematoxylin & eosin in control (H), SAP (I), CLP (J), and SAP+CLP (K) groups, scale bar = 50 μm. L, Chou’s score was used to assess the severity of inflammation and tissue damage in the small intestine of mice from the control, SAP, CLP, and SAP+CLP groups (n = 5 per group). One-way ANOVA followed by Bonferroni post hoc test was used, *P < 0.05, **P < 0.01, ***P < 0.001. ANOVA, analysis of variance; CSF, cerebrospinal fluid.

We further examined the serum levels of cytokines including IL-1β, IL-6, IL-10, and TNF-α. Ablation of the CSF-contacting nucleus alone elevated the serum levels of the IL-1β, IL-6, and TNF-α compared to the control mice, which underwent PBS injection into the lateral ventricle (LV) (Fig. 2, D–F). Moreover, ablation of the CSF-contacting nucleus also resulted in an increase of the concentration of IL-1β, IL-6, and TNF-α in septic mice (Fig. 2, D–F). The changes in cytokine levels between SAP group and CLP + SAP group appear as an additive effect. And IL-1β, IL-6, and TNF-α are usually considered as the proinflammatory factors (²⁵). Meanwhile, IL-10 is usually considered as an anti-inflammatory factor and is largely used in clinical (²⁵). In our work, compared with mice in control group, we found that the serum level of the IL-10 was decreased when the CSF-contacting nucleus was ablated (Fig. 2G).

Besides, in our work, we found that the concentration of the cytokines (IL-1β, IL-6, and TNF-α) in SAP group seems higher than the cytokines in CLP group. However, no statistically significant difference was observed between the two groups, although an increasing trend was noted. This phenomenon might due to the small n value and the discrete dataset in SAP group.

To further characterize the effects of the CSF-contacting nucleus on septic inflammation, we scored the grade of intestine inflammation using standard histopathology (26). The small intestine from the septic mice showed increased inflammatory cells between the crypts, and the inflammatory infiltrate extended into the villi even throughout the lamina propria (Fig. 2J). Surprisingly, we observed that inflammatory cells increased between the intestine crypts in the mice with CSF-contacting nucleus ablation as compared with the mice in the control group (Fig. 2, H–L), and the chou’s score was significantly increased when the CSF-contacting nucleus was ablated. In addition, Chou’s score was even higher in the septic mice with CSF-contacting nucleus ablation compared to mice with CSF-contacting nucleus ablation alone (Fig. 2L). Our work indicated CSF-contacting nucleus was closely involved in sepsis and ablation of the CSF-contacting nucleus aggravated the sepsis-induced inflammation.

Activation of the CSF-contacting nucleus attenuated the systemic inflammatory response in septic mice

Glial cells around the CSF-contacting nucleus may be activated in response to the ablation of the CSF-contacting nucleus, potentially leading to the ingestion of neuronal fragments, thus resulting in secreting proinflammatory cytokines (27). Hence, we utilized chemogenetics to manipulate the activity of the CSF-contacting nucleus specifically. We used the Cre/LoxP system to specifically infect the CSF-contacting nucleus with Designer Receptor Exclusively Activated by a Designer Drug (DREADD) by injecting the AAV2/R-hSyn-Cre into the LV and the AAV-DIO-hM3Dq-EGFP into the CSF-contacting nucleus (Fig. 3, A–C). Mice in control group were injected AAV-DIO-EGFP. Then, we administrated clozapine N-oxide (CNO, 1 mg/kg) or saline intraperitoneally for 7 consecutive days. The CLP surgery was performed 2 h later after the final CNO injection. Then we examined the serum levels of cytokines 6 h later. We found that activation of the CSF-contacting nucleus resulted in a significant reduction in the levels of cytokines IL-1β, IL-6, and TNF-α in CLP-induced septic mice (Fig. 3, D–F). In contrast, the serum concentration of IL-10 was increased in septic mice following excitation of the CSF-contacting nucleus (Fig. 3G).

Fig. 3.

Activation of the CSF-contacting nucleus alleviated the inflammation induced by the sepsis. A, Schematic representation of timeline and experimental design. B, Schematic showing virus injection in the LV and CSF-contacting nucleus of the wide-type mice. C, Representation of virus AAV-DIO-EGFP (left) and virus AAV-DIO-hM3Dq-EGFP (right) infection in the CSF-contacting nucleus, Aq: aqueduct, scale bar = 50 μm. D–G, The concentration of the inflammatory cytokines in the serum after activation of the CSF-contacting nucleus with chemogenetics. (D) is for IL-1β; (E) is for IL-6; (F) is for TNF-α; (G) is for IL-10. Two-way ANOVA followed by Bonferroni post hoc test was used, *P < 0.05, **P < 0.01, ***P < 0.001. H–K, Representative histological images from HE-stained small intestine sections of mice. L, Chou’s score measuring the severity of inflammation and tissue damage in the small intestine after activation of the CSF-contacting nucleus. n = 5 for each group, one-way ANOVA was used followed by Bonferroni post hoc test, *P < 0.05, **P < 0.01. ANOVA, analysis of variance; CSF, cerebrospinal fluid; LV, lateral ventricle.

Next, we assessed the effect of CSF-contacting nucleus activation on intestinal inflammation. We observed that septic mice exhibited crypt abscesses, increased immune cell infiltration, and epithelial hyperproliferation (Fig. 3, H–J). Activation of the CSF-contacting nucleus reduced immune cell infiltration and crypt abscess formation in septic mice (Fig. 3, K–L). Activation of the CSF-contacting nucleus decreased Chou’s score in the septic mice (Fig. 3L). These results demonstrated that the activation of the CSF-contacting nucleus alleviated the systemic inflammation induced by sepsis.

Inhibition of the CSF-contacting nucleus exacerbated the inflammatory response in septic mice

We next investigated whether chemogenetic inhibition of the CSF-contacting nucleus could alter CLP-induced inflammatory responses. We injected AAV viruses carrying a CRE into the LV and a Cre-dependent inhibitory DREADD into the CSF-contacting nucleus (Fig. 4, A–C). We found that inhibition of the CSF-contacting nucleus increased the level of the proinflammatory cytokine IL-6 in septic mice induced by CLP surgery (Fig. 4E). In addition, we also observed that inhibition of the CSF-contacting nucleus aggravated chorioedema of the intestinal mucosa, as well as crypt abscess formation (Fig. 4, H–K). Our work suggested that inhibition of the CSF-contacting nucleus exacerbated the inflammatory response in septic mice.

Fig. 4.

Inhibition of the CSF-contacting nucleus aggravated the inflammation induced by the sepsis. A, Schematic representation of timeline and experimental design. B, Schematic showing virus injection into the LV and the CSF-contacting nucleus of the wide type mice. C, Representation of virus AAV-DIO-EGFP (left) and virus AAV-DIO-hM4Di-EGFP (right) infection in the CSF-contacting nucleus, Aq: aqueduct, scale bar = 50 μm. D–G, The concentration of the inflammatory cytokines in the serum following activation of the CSF-contacting nucleus with chemogenetics, (D) was for IL-1β, (E) was for IL-6, (F) was TNF-α, and (G) was for IL-10, n = 5 for each group. *P < 0.05, **P < 0.01, ***P < 0.001. Two-way ANOVA was used followed by Bonferroni post hoc test. (H-K) Representative histological images from HE-stained small intestine sections from mice after inhibition of the CSF-contacting nucleus. (L) Chou’s score measuring the severity of inflammation and tissue damage in the small intestine after inhibition of the CSF-contacting nucleus. *P < 0.05, n = 4 for each group, one-way ANOVA was used followed by Bonferroni post hoc test. ANOVA, analysis of variance; CSF, cerebrospinal fluid; LV, lateral ventricle.

Besides, activation of the CSF-contacting nucleus decreased the level of the proinflammatory cytokines (IL-1β, IL-6, and TNF-α) and increased the level of anti-inflammatory cytokine IL-10. While, inhibition of the CSF-contacting nucleus only increased the level of IL-6 without affecting IL-1β, IL-6, and TNF-α. This discrepancy may potentially be attributed to the following factors. Firstly, our work reported that the activity of CSF-contacting nucleus is significantly elevated during the severe systemic inflammatory response (Fig. 2A). We think that chemogenetic inhibition of already heightened activity of CSF-contacting nucleus demonstrates limited potential to exacerbate the elevated proinflammatory cytokine levels given preexisting severe systemic inflammatory states. Second, several studies have also demonstrated that IL-6 has a physiological significance distinct from IL-1 and TNF-α and that these inflammatory factors are interrelated, but not always fully consistent (28,29). Third, there are many other inflammatory markers, for example: C-reactive protein (CRP), procalcitonin (PCT), angiopoietins, DAMPs/cell surface receptors, and so on (30). Meanwhile, studies preferred to select IL-6 as a biomarker to assess the severity of the inflammatory response (31,32) and guide the antibiotic therapy in sepsis (³³). Thus, the cytokine IL-6 might be a more sensitive biomarker during the inflammation induced by the CLP surgery. In conclusion, that is why inhibition of the CSF-contacting nucleus only increased the level of IL-6.

The cerebrospinal fluid-contacting neurons around the central canal of the spinal cord have been reported to respond to mechanical stimulation induced by fluid movements and induce spinal movement (34,35). Given that, we assessed whether activation or inhibition of the CSF-contacting nucleus influenced the mobility of the naive mice. Interestingly, as shown in the Figure 1, the activation of the CSF-contacting nucleus did not affect the locomotor activity of the mice (Fig. 1, A and B). Likewise, the inhibition of the CSF-contacting nucleus did not affect the distance and velocity of the mice in the open field test (Fig. 1, C and D). Thus, our results suggested that the CSF-contacting nucleus played an important role in the sepsis-induced systemic inflammation without affecting the locomotor activity.

Smart-seq of the CSF-contacting nucleus of the septic mice

To further identify specific and meaningful genetic targets in the CSF-contacting nucleus during the systemic inflammatory response in septic mice, we carried out single-cell RNA sequencing of the CSF-contacting nucleus. Briefly, we first injected CTB-594 into the LV to specifically label the CSF-contacting nucleus (21). Five days later, mice were randomly allocated into two groups. Mice in CLP group underwent CLP surgery, while the control group underwent sham surgery. Subsequently, the CSF-contacting nucleus tissue was precisely collected with laser capture microdissection 6 h later (Fig. 5, A–C). The tissues were collected and stored in strict accordance with the requirements for further single-cell sequencing. The coverage of sequencing showed the sequences were uniformly distributed in genes without significant bias peaks (Fig. 5D). The results of the quantitative analysis showed that the level of gene expression in all samples was consistent (Fig. 5E). The overview of the gene expression in the CSF-contacting nucleus was shown in Figure 5F. A total of 13,685 genes were expressed in CSF-contacting nucleus of the septic mice and 13,931 genes were in the control group. The CSF-contacting nucleus in two groups shared 12,803 common genes, accounting for 86.43% of the total. The overview heatmap of the DEGs was shown in Figure 5G. Principal component analysis (PCA) demonstrated that the samples within each group exhibited high reproducibility (Fig. 6A). A total of 262 DEGs were identified between the control and CLP group, including 218 up-regulated genes and 44 down-regulated genes (Fig. 6, B and C). We further employed RT-PCR assay to verify the reliability of sequencing results. We selected 5 top-up-regulated genes and 2 genes showing no difference for validation. Our results showed that the relative expression of Cdnk1α, Mt2, Phactr3, Mt1, and Baiap3 in CLP group was significantly increased compared with those in the control group. Additionally, no significant differences were observed in the relative expression levels of Cox5a and Xpo6 between the control and CLP groups (Fig. 6, D–J, and Table 1). Those results confirmed the reliability of our sequencing results. The top 20 histograms from KEGG and GO enrichment analyses revealed that the CSF-contacting nucleus is involved in immune system regulation (Fig. 7, A and C). The top 50 DEGs in the CSF-contacting nucleus during sepsis were shown in Figure 7B. To determine the interaction relationship between the proteins expressed by these DEGs in the systemic inflammation of sepsis, the PPI network was constructed using STRING V4.10 (Fig. 7D). Combined with the top 50 DEGs and PPI network analysis, we found several DEGs (somatostatin [Sst], early growth response protein 1 [Egr1], and calmodulin-like protein 4 [Calml4])) have high connectivity with other genes (Fig. 7D). Existing scholarly evidence demonstrates that Sst and Egr1 exhibit pivotal immunomodulatory functions, as substantiated by systematic analysis of peer-reviewed research (^36–38). Considering our work and previous studies, we suggested that the Sst and Egr1 genes were activated in the CSF-contacting nucleus during the systemic inflammation in septic mice, which might be a promising candidate in regulation of the sepsis.

Fig. 5.

Smart-Seq RNA of the CSF-contacting nucleus of mice receiving CLP surgery or the sham CLP surgery. A, Experimental timeline: mice were injected with CTB-594 retrograde tracer 5 days before CLP surgery. CSF-contacting nucleus was collected at H6 (6 h after CLP) using LCM for Smart-Seq2 RNA sequencing. B, Schematic for collecting CSF-contacting nucleus with the LCM. C, Representative images of the section that the CSF-contacting nucleus was intact (red, left) and the section that the CSF-contacting nucleus was cutting off with the LCM (right). Scale bar = 50 μm. D, Sequencing coverage across gene bodies (5′ → 3′), demonstrating uniform read distribution. E, Violin plots of gene expression distribution (log10(TPM)) across control (n = 3) and CLP (n = 4) groups. F, Venn diagram comparing DEGs between CLP and control groups. CLP-unique DEGs (882, 5.95%), control-unique DEGs (1,128, 7.61%), and shared genes (12,803, 86.43%) are shown. (G) Heatmap of gene expression patterns (rows: genes, columns: samples). Red indicates high expression; blue indicates low expression. Hierarchical clustering (top) reveals sample-group relationships. DEGs, differentially expressed gene; LCM, laser capture microdissection.

Fig. 6.

The analysis and validation of Smart-Seq RNA results. A, PCA analysis of the CSF-contacting nucleus samples from septic mice and control mice (n = 4 for CLP group, n = 3 for control group). B, Volcano plot showing the differential genes in CSF-contacting nucleus from mice in CLP and control group. Red dots represent the up-regulated genes, blue dots represent down-regulated genes, and gray dots represent genes no significantly changed. C, The histogram of differentially expressed genes. The red represented the up-regulated genes (n = 218), and the blue denoted the down-regulated genes (n = 44). D–I, The PCR test validated statistical diagrams of 5 top-up-regulated gene expressions in the CSF-contacting nucleus. (D) is for Cdnk1α, (E) is for Mt2, (F) is for Phactr3, (G) is for Mt, (H) is for Baiap3, (I) is for Cox5, (J) is for Xpo6. The data are presented as Mean ± SEM. Student t test was used *P < 0.05, **P < 0.01, n = 4 for each group. PCA, principal component analysis.

Table 1.

The primer sequences for the 5 top-up-regulated genes

Gene	Primers	Sequence (from 5′ to 3′)
Cdnk1α	Forward primer	CAGGCACCATGTCCAATCCT
	Reverse primer	AAAGTTCCACCGTTCTCGGG
Mt2	Forward primer	ATGCAAATGTACTTCCTGCAAGA
	Reverse primer	CTGGGAGCACTTCGCACAG
Phactr3	Forward primer	CCCACTGTTGATGAATTGAGAGA
	Reverse primer	TGTGCTCTTGCTACTTCCACATA
Mt1	Forward primer	AAGAGTGAGTTGGGACACCTT
	Reverse primer	CGAGACAATACAATGGCCTCC
Baiap3	Forward primer	CTTCAGTGACCCGTACTGTATGC
	Reverse primer	TTGACCTCCGTGACCTGGATA
Cox5a	Forward primer	GCCGCTGTCTGTTCCATTC
	Reverse primer	GCATCAATGTCTGGCTTGTTGAA
Xpo6	Forward primer	TAGGAGCTTGGAGATTCTGCT
	Reverse primer	ATCTTATCCTGAGATGGGACCC
Gapdh	Forward primer	AGGTCGGTGTGAACGGATTTG
	Reverse primer	GGGGTCGTTGATGGCAACA

Fig. 7.

The KEGG, GO enrichment, and PPI network analysis of the DEGs. A, The KEGG pathway analysis of the top 20 DEGs. B, The heatmap of the top 50 DEGs. C, The GO enrichment analysis of the DEGs. D, The PPI network of the DEGs. DEGs, differentially expressed genes. KEGG, Kyoto Encyclopedia of Gene and Genome, GO, gene ontology, PPI, protein-protein interaction.

DISCUSSION

In this study, we established the CSF-contacting nucleus as a critical regulator of inflammation in sepsis. The CSF-contacting nucleus was markedly activated in response to sepsis. Manipulation activity of the CSF-contacting nucleus modulated systemic inflammation induced by sepsis.

The body-brain circuit regulating the immune system

Interaction and communication between the immune system and the nervous system has long been identified (9,10,39–42). The vagus nerve has been shown to play a critical role in suppressing inflammation and modulating immune responses in various pathological conditions (43). Several studies have revealed the dopaminergic neurons in the vagus as a critical component to sense and transmit the peripheral immune signal (23,41,44). Their work reported that the caudal nucleus of the solitary tract receives peripheral inflammation signals from the vagus nerve and, in turn, modulates the immune response. Further study demonstrated that vagal sensory neurons encoded real-time, dynamic information of specific cytokines before they were transmitted to the brain (45).

Interestingly, retrograde tracing of the CSF-contacting nucleus demonstrated that the nucleus received the projections from the dorsal motor nucleus of the vagus (46). Our present study demonstrated that activation of the CSF-contacting nucleus reduced the levels of proinflammatory cytokines, elevated the levels of the anti-inflammatory cytokine IL-10, and mitigated intestinal inflammatory responses. Conversely, inhibition of the CSF-contacting nucleus increased the levels of IL-6, exacerbated chorioedema in the intestinal mucosa, and promoted the formation of crypt abscesses. By integrating the findings from previous studies with our results, we hypothesized that the vagus nerve perceives and processes peripheral inflammation, subsequently relaying this information to the CSF-contacting nucleus. Further studies are required to prove this hypothesis. Our work enriches the theory of neuro-immune crosstalk.

The brain-cerebrospinal fluid pathway regulating the immune system

The cerebrospinal fluid (CSF) has been considered to contain a tightly regulated immune system (17,47). Single-cell RNA sequencing on CSF revealed that CXCL16-CXCR6 signaling might account for antigen-specific T-cells infiltrate the brain, resulting in immune dysregulation (47). The CSF-contacting nucleus is defined by its unique architecture, wherein neuronal cell bodies are situated within the brain parenchyma while their processes project into the CSF. Thus, it can communicate with other brain structures as well as with the CSF. Our work indicated a role of the CSF-contacting nucleus which regulated the systemic inflammation response during sepsis, which seems to offer a reasonable explanation for the immunity regulation of the CSF. It appears that the CSF-contacting nucleus and CSF exhibit reciprocal regulation in immune modulation. Several studies have reported that the excitation of CSF-contacting neurons within the central canal of the spinal cord could modulate locomotion and posture (17,34). In essence, our work revealed that excitation or inhibition of the CSF-contacting nucleus does not influence the mice’s motor function. This indicated that the CSF-contacting nucleus may functionally differ from CSF-contacting neurons within the central canal of the spinal cord. Our work first validated that the CSF-contacting nucleus played a crucial role in the modulation of systemic inflammation, thereby extending the regulatory function of the brain-cerebrospinal fluid pathway within immune systems.

The potential targets within the CSF-contacting nucleus regulating the immune system

Cholinergic neurons in basal forebrain and have been reported to attenuate polymicrobial sepsis-induced inflammatory response (23). Hypothalamic AgRP neurons are a functional component of the immunoregulatory effects during endotoxemia (48). To explore the critical component within the CSF-contacting nucleus regulating the inflammatory response, we performed the single-cell RNA sequencing assay. The single-cell RNA sequencing on the CSF-contacting nucleus during the sepsis indicated that several immune-related genes were up-regulated. Somatostatin (Sst), for instance, a growth hormone inhibitory peptide, expressed in endocrine and nonendocrine tissues, immune cells, and the central nervous system (36), was significantly increased in CSF-contacting nucleus in response to sepsis. A study has reported that maternal immune activation could induce an increase in cortical Sst gene expression in adult rat offspring (49,50). The density of SST-positive neurons in the intestine exhibits age-dependent variation, accompanied by changes in SST colocalization with choline acetyltransferase (ChAT) (51). In addition, choline acetyltransferase receptor, the nicotinic acetylcholine receptor α7 subunit, has been identified as a critical regulator of inflammation (52). Further study has revealed that the cortical Sst gene plays a regulatory role in mediating immune response activation processes (53). Besides that, Egr1 has been reported to be critical for the rapid pathogen-specific recruitment of immune effector cells and modulated inflammatory cytokines expression (54–56). These data suggested that the Sst and Egr1 genes might be the potential candidates in the CS-contacting nucleus regulating the inflammation response in septic mice.

In conclusion, our work first reported the interference of CSF-contacting nucleus in systemic inflammation in sepsis, and activation of the CSF-contacting nucleus alleviated the systemic inflammation in septic mice. The single-cell RNA sequencing indicated that Egr1 and Sst might be the potential candidates in the CSF-contacting nucleus regulating the inflammation induced by sepsis, which still needs further research to confirm. These data suggest that clinically relevant technology development targeting the CSF-contacting nucleus may be potentially significant for sepsis and other immune-related diseases.