Short-chain fatty acids attenuate sepsis-induced gut dysbiosis and hippocampal neuroinflammation via NLRP6 inflammasome activation in mice

Abstract

Background:

Sepsis-associated encephalopathy (SAE) is a critical complication of sepsis, yet the mechanisms linking gut dysbiosis to hippocampal neuroinflammation remain poorly understood. Our previous work identified sepsis-induced hippocampal neuroinflammation; here, we investigated the role of gut microbiota-derived short-chain fatty acids (SCFAs) and NLRP6 inflammasome signaling in this process.

Methods:

Sepsis was induced in C57BL/6 mice via cecal ligation and puncture (CLP). Gut microbiota composition, SCFA levels, intestinal barrier integrity, and NLRP6 inflammasome activity were analyzed. Colon organoids and NLRP6-silenced CT26 cells were employed to validate SCFA–NLRP6 interactions. Hippocampal neuroinflammation (microglial/astrocytic activation, cytokine levels) and cognitive function (Morris water maze, Barnes maze) were assessed post-SCFA treatment.

Results:

CLP-induced sepsis triggered hippocampal neuroinflammation, characterized by microglial proliferation (IBA-1+), astrocyte activation (GFAP+), and neuronal dysfunction (reduced c-Fos). Septic mice showed gut dysbiosis (increased Firmicutes/Proteobacteria, decreased α-diversity), reduced SCFA levels, and impaired intestinal barrier integrity (decreased ZO-1/occludin expression, P < 0.05). SCFA supplementation restored gut microbiota homeostasis (β-diversity: P = 0.019 vs. CLP), enhanced intestinal tight junction proteins (ZO-1: 1.8-fold increase, P < 0.01), and activated NLRP6 inflammasomes in colonic tissues (NLRP6: 2.1-fold increase, caspase-1: 1.6-fold increase, P < 0.01). NLRP6 knockdown abolished SCFA-mediated IL-18 secretion (P < 0.001). Behaviorally, SCFAs ameliorated cognitive deficits in septic mice (escape latency: CLP = 48 s vs. SCFA + CLP = 32 s, P < 0.01) and correlated with hippocampal c-Fos restoration (R² = 0.839 for propionate, P = 0.01).

Conclusions:

Sepsis disrupts the gut–brain axis by impairing intestinal barrier integrity and NLRP6 inflammasome function, exacerbating hippocampal neuroinflammation. SCFAs mitigate these effects via NLRP6-dependent mechanisms, highlighting their therapeutic potential for SAE. This study provides the first evidence linking SCFA-mediated NLRP6 activation to neuroprotection in sepsis, offering novel insights for targeting the gut microbiota in critical care.

HIGHLIGHTS

• Sepsis disrupts gut microbiota and reduces short-chain fatty acids (SCFAs), impairing intestinal barrier integrity.

• SCFA supplementation restores microbial balance, upregulates tight junction proteins (ZO-1/occludin), and activates NLRP6 inflammasomes.

• NLRP6-dependent mechanisms drive SCFA-mediated suppression of hippocampal neuroinflammation (↓IBA-1+/GFAP +, ↑c-Fos).

• SCFAs rescue sepsis-induced cognitive deficits (escape latency reduced by 33%, P < 0.01) via gut–brain axis modulation.

• First demonstration of the SCFA–NLRP6 axis as a therapeutic target for sepsis-associated encephalopathy.

Background

Sepsis is a dysregulated host response to microbial infection that can lead to organ dysfunction and is a major cause of mortality in critically ill patients[1]. Sepsis often results in neurological dysfunction, known as sepsis-associated encephalopathy (SAE)[2]. Survivors of sepsis frequently face long-term cognitive impairments after discharge, including memory loss, attentional deficits, impaired verbal fluency, and executive dysfunction[3]. Currently, no effective treatments are available to alleviate this life-threatening neurological burden. Therefore, understanding the pathogenesis of SAE and developing effective preventive and therapeutic strategies is crucial.

Sepsis triggers an immune response that causes vascular endothelial damage. This disruption compromises the blood–brain barrier (BBB), allowing peripheral immune cells to infiltrate the brain, which initiates or exacerbates neuroinflammation and activates glial cells[4]. While various mechanisms are involved in the immune response to fight infections, they may also contribute to the development of SAE. It remains unclear whether sepsis directly affects the gut microbiome or if it is primarily driven by host immune dysregulation.

The intestinal epithelial barrier (IEB) consists of tightly connected intestinal epithelial cells (IECs), the mucus layer, and the gut microbiota. The IECs form a physical barrier through tight junctions and desmosomes that prevent harmful substances from entering, while the intestinal immune system provides additional protection^[5,6]. This barrier’s primary function is to prevent external antigens from accessing the host through the intestinal lumen[7]. The NOD-like receptor family pyrin domain-containing 6 (NLRP6) inflammasome, an innate immunity sensor, plays a critical role in protecting the colonic mucosa from bacterial pathogens^[8–10]. Gut microbiota and their metabolites help maintain IEB integrity^[11–13] and may regulate NLRP6 inflammasome signaling to support homeostasis within the intestinal microenvironment[14].

Short-chain fatty acids (SCFAs) are key metabolites produced by bacterial fermentation of dietary fiber in the gastrointestinal tract and are beneficial for gut microbiota-brain communication[15]. SCFAs influence the development and function of microglia in the brain, highlighting their protective role against neuroinflammation through mechanisms linked to the gut microbiota^[16,17]. Animal studies suggest that SCFAs may provide neuroprotective effects in models of neurodegenerative diseases, though more evidence is needed to fully confirm these findings^[17–19]. Whereas SCFAs show neuroprotection in chronic neurodegeneration, their ability to rescue acute sepsis-induced cognitive decline via NLRP6 – a gut-specific inflammasome unexplored in SAE – remains unknown.

In this study, we found that sepsis-induced gut dysbiosis caused IEB damage and neuroinflammation in the hippocampus of C57BL/6N mice. We also identified a key molecular mechanism in which SCFAs, by activating NLRP6, facilitate the repair of IEB damage, potentially reducing hippocampal neuroinflammation in septic mice. This study adheres to the Transparency In The reporting of Artificial INtelligence (TITAN) guidelines for ethical reporting[20].

Materials and methods

The work has been reported in accordance with the ARRIVE guidelines (Animals in Research: Reporting In Vivo Experiments)[21].

Animals

Male C57BL/6 mice (weight: 18–22 g, age: 6–8 weeks; Hunan SJA Laboratory Animal Co., Ltd., Changsha, China) were used. A total of 200 mice were initially assigned to sham (n = 40) and cecal ligation and puncture (CLP, n = 160) groups upon previous work[22]. During the experimental period, 16 mice in the CLP group died within 72 hours post-surgery due to severe sepsis complications (e.g., multi-organ failure) and were excluded from subsequent analyses. No animals in the sham group died or required exclusion. Additionally, four mice (two from sham, two from CLP) were excluded due to incomplete behavioral testing data [e.g., failure to complete Morris water maze (MWM) trials]. All remaining animals (sham: n = 38; CLP: n = 142) were included in the final analysis. No other animals, experimental units, or data points were excluded.

Randomization was performed via a lottery draw by an independent researcher. Experimenters performing surgeries, administering treatments, and assessing outcomes were blinded to group allocations. Only the independent researcher responsible for randomization had access to group codes, which were revealed after data analysis. Behavioral tests (MWM and Barnes Maze) were conducted in a fixed order, with mice tested in randomized sequences to avoid time-of-day effects. Tissue collection and molecular analyses were performed in batches balanced across groups.

All procedures adhered to the Guide for the Care and Use of Laboratory Animals (NIH, Bethesda, MD, USA). Mice were housed in groups of 5 per cage under a 12-hour light/dark cycle with ad libitum access to food and water. Experiments commenced after a 2-week acclimatization period.

Oral SCFA administration

A mixture of the three major SCFAs was prepared based on previous studies^[23,24]. The mixture consisted of 67.5 mM sodium acetate (Sigma-Aldrich, S7545), 25 mM sodium propionate (Sigma-Aldrich, P1880), and 40 mM sodium butyrate (Sigma-Aldrich, 303410) in a 3:1:1 ratio. This mixture was added to the drinking water (pH 7.6), which was replaced twice a week.

CLP and sham surgery

Mice were randomly assigned to undergo either CLP, a standard model of sepsis, or sham surgery as a control group^[25,26]. Under anesthesia with 2% isoflurane, a 3-cm midline laparotomy was performed to expose the cecum and surrounding intestines. A 3.0 silk suture was used to tightly ligate the cecum approximately 50% below the ileocecal valve, followed by perforation with an 18-gauge needle to induce moderate sepsis, with a small amount of fecal matter extruding from the cecum. After repositioning the cecum, the wound was closed using surgical sutures. In the sham group, cecum ligation and puncture were omitted, but all other surgical steps were identical to the CLP procedure. Both groups received postoperative care, including 50 mL/kg of 0.9% NaCl administered subcutaneously and 30 mg/kg of ceftriaxone and clindamycin, also subcutaneously, immediately following surgery and again 12 hours later. Mice meeting any endpoint criteria were immediately euthanized via intraperitoneal injection of pentobarbital sodium (150 mg/kg), followed by cervical dislocation to confirm death.

Fecal collection and 16S rRNA sequencing

Ten days post-surgery, fecal pellets were collected from both the control and sepsis groups for 16S rRNA sequencing. Samples were stored at −80°C until analysis. DNA was extracted using the MO BIO PowerMag Soil DNA Isolation Kit according to the manufacturer’s protocol. Gene libraries were constructed using the Illumina MiSeq platform and V4 primers (515f and 806r). 16S rRNA sequencing was followed by data processing and analysis using QIIME2[27]. Denoising and amplicon sequence variant (ASV) identification were performed using the DADA2 pipeline in R, taxonomically assigned using the SILVA138 database^[28,29]. Alpha diversity (within-community diversity) was assessed in R, while beta diversity (between-community diversity) was evaluated using unweighted and weighted UniFrac distances, followed by principal coordinate analysis (PCoA) for both distance calculation and visualization. A Student’s t-test was used to compare differences in alpha diversity between groups. The ALDEx2 package was employed to assess differential taxon abundance, expressed as effect sizes after centered log-ratio transformation.

SCFA quantification

At 10 days post-surgery, cecum fecal samples were collected from both the control and sepsis groups for SCFA analysis. SCFAs were extracted as described by Giridharan et al[30]. Fecal samples were homogenized in Milli-Q water using an MP Bio FastPrep at 4.0 m/s for 1 minute, then acidified with approximately 5 M HCl to a pH of 2.0. After incubation, samples were centrifuged at 10 000 rpm to separate the supernatant. 2-Ethylbutyric acid was added to concentrate SCFAs to 1 mM. SCFA-rich supernatants were stored in 2 mL GC vials with glass inserts and analyzed using a Thermo Trace 1310 gas chromatograph equipped with a flame ionization detector.

Extraction of blood and tissue samples

Mice were fasted overnight and anesthetized with pentobarbital sodium (60 mg/kg, intragastric). Blood was collected by inferior vena cava puncture, and serum was obtained by centrifuging at 10 000 rpm for 10 minutes and stored at −80°C for further analysis. The hippocampal and colonic tissues were harvested for histopathological analysis, immunofluorescence staining, quantitative PCR (qPCR), and Western blotting.

In vitro culture of mouse colon explants

Ex vivo culture of colon explants was conducted following the protocol outlined by previous studies[8]. Colon tissues were dissected from 8-week-old mice, washed in cold Hank’s buffer containing antibiotics (100 U/mL β-lactam, 100 μg/mL streptomycin, 50 μg/mL Flagyl), and sectioned into 2 mm × 2 mm pieces. Tissues were placed on gelatin foam discs and cultured in RPMI 1640 medium supplemented with 0.01% BSA, 200 U/mL penicillin, 200 μg/mL streptomycin, and 1% Fungizone. Colon explants were treated for 24 hours with sodium acetate, sodium propionate, sodium butyrate, or lipopolysaccharide (LPS, L2880, Sigma-Aldrich). The culture medium and colon tissues were then collected to assess NLRP6 immune complex activation.

Cell culture and treatment

CT26 colon cancer cells (CL-0071, Wuhan Pricella Biotechnology Co., Ltd.) were cultured in RPMI 1640 medium containing 10% FBS under a 5% CO₂ humidified atmosphere at 37°C. Cells were seeded in 6-well plates and allowed to adhere for 12 hours before transfection with NLRP6-specific siRNA or scrambled siRNA for 6 hours. After transfection, cells were cultured for 48 hours in medium containing a mixture of sodium acetate, sodium propionate, and sodium butyrate (3:1:1 ratio). The growth medium was then collected, and total cellular proteins were isolated for biochemical analyses and Western blotting.

Gut histopathology

Terminal ileum tissue samples were fixed in neutral-buffered formalin, paraffin-embedded and processed for histological analysis. Sections (5 µm thick) were stained with hematoxylin and eosin (H&E), and images were captured using a conventional optical microscope. Villus length and crypt depth were measured using Caseviewer software.

qPCR analysis

Total RNA was extracted from hippocampal and colon tissues using the TaqMan real-time quantitative PCR kit, following the manufacturer’s instructions. After reverse transcription into complementary DNA, qPCR was performed on a Bio-Rad real-time PCR system using gene-specific primers sourced from GeneRay Biotech (Shanghai, China). The thermal cycling conditions were as follows: 30 seconds of denaturation at 94°C, followed by 40 cycles of 5 seconds at 95°C and 30 seconds at 60°C.

Western blot analysis

Colon tissue samples or cultured explants were homogenized in ice-cold RIPA buffer with protease inhibitors. After centrifugation, protein concentrations were quantified using a BCA protein assay kit (Thermo Fisher Scientific, USA), with bovine serum albumin as a standard. Proteins were separated by 8% Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) and transferred to Polyvinylidene Fluoride (PVDF) membranes. The membranes were blocked for 20 minutes with rapid blocking buffer, followed by overnight incubation with primary antibodies at 4°C. After washing, Horseradish Peroxidase (HRP)-conjugated secondary antibodies were applied. The primary antibodies used were rabbit anti-occludin (ab168986, Abcam), rabbit anti-ZO-1 (ab307799, Abcam), rabbit anti-NLRP6 (ab314498, Abcam), rabbit anti-caspase-1 (ab179515, Abcam), and mouse anti-Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) (ab8245, Abcam). Protein bands were detected using an enhanced chemiluminescence system (Tanon, Shanghai, China), and densitometry was performed using ImageJ (Version 1.50b, NIH). Data were normalized to GAPDH and expressed as fold changes relative to the control group.

Immunofluorescence staining

Mice from each group were anesthetized, and cardiac perfusion was performed with 4% paraformaldehyde (PFA). After brain tissue extraction, samples were fixed overnight in PFA, followed by incubation in sucrose solutions (20% sucrose/Phosphate-Buffered Saline (PBS) and 30% sucrose/PBS) at 4°C. Tissue was embedded in Optimal Cutting Temperature (OCT) compound, sectioned into 15 μm slices using a cryostat, and blocked for 1 hour at room temperature with 10% FBS in PBS. Primary antibodies (rabbit anti-IBA-1, ab-178846, Abcam; rabbit anti-GFAP, ab7260, Abcam; rabbit anti-c-Fos, ab208942, Abcam) were incubated overnight at 4°C, followed by incubation with Alexa Fluor-conjugated secondary antibodies for 30 minutes at 37°C. Cell nuclei were stained with 4’,6-Diamidino-2-Phenylindole (DAPI). Immunofluorescence images were obtained using a Leica TCS SP8 confocal microscope. Quantification of immunofluorescence-positive cells was performed in the hippocampal DG, CA1, and CA3 regions using randomly selected brain sections.

No artificial intelligence (AI) tools were used in the design, execution, analysis, or reporting of this study.

Behavioral tests

Morris water maze

Spatial learning and memory were assessed using the MWM. During the spatial acquisition task, mice were trained for 5 consecutive days with 4 trials per day, 30 seconds apart. Each mouse was given 1 minute to locate the hidden platform; if unsuccessful, they were guided to the platform. After the training period, a probe trial was conducted to assess spatial memory. The platform was removed, and mice were allowed to explore the pool for 60 seconds. Data were recorded and analyzed using Smart 3.0 software.

Barnes Maze

The Barnes Maze (BM) consists of a circular platform (height: 105 cm, diameter: 92 cm) with 20 holes, one of which is the target escape hole. In the learning phase, mice underwent 5 days of training with three trials per day, in which they had 2 minutes to find the target hole. After each trial, the maze was cleaned with 70% ethanol to eliminate odor cues. On day 6, the escape box was removed, and mice were allowed to explore freely for 2 minutes. Movement trajectories and the latency to reach the target hole were recorded.

Statistical analysis

Data are presented as mean ± SEM. Normality was assessed using the Shapiro–Wilk test. Parametric tests (Student’s t-test for two-group comparisons, one-way/two-way ANOVA for multiple-group comparisons) were applied to normally distributed data (P > 0.05), followed by appropriate post hoc tests. Non-parametric tests (Mann–Whitney U test for two-group comparisons, Kruskal–Wallis test for multiple-group comparisons) were used for non-normally distributed data (P ≤ 0.05), followed by Dunn’s post hoc test. Statistical analyses were performed using GraphPad Prism Software 9.0 (GraphPad Software, San Diego, CA, USA). A P-value of <0.05 was considered statistically significant.

Theory/calculation

Theory

The neuroprotective effects of SCFAs in sepsis are hypothesized to arise from their dual role in modulating gut microbiota homeostasis and NLRP6 inflammasome activation. Mechanistically, SCFAs (e.g., acetate, propionate, butyrate) serve as ligands for G-protein-coupled receptors on IECs, triggering downstream signaling that enhances NLRP6 inflammasome assembly. NLRP6 activation promotes caspase-1-dependent cleavage of pro-IL-18 into its bioactive form, which fortifies intestinal barrier integrity by upregulating tight junction proteins (ZO-1, occludin) and stimulating epithelial repair. This barrier restoration limits systemic translocation of microbial products (e.g., LPS), thereby attenuating neuroinflammatory cascades mediated by peripheral cytokine diffusion and microglial hyperactivation. Furthermore, SCFAs may directly influence hippocampal neuronal plasticity via vagal afferent signaling or histone deacetylase inhibition, linking gut-derived metabolites to cognitive recovery.

Calculation

Statistical power for group comparisons (sham, CLP, SCFA + CLP) was determined using G*Power 3.1, assuming α = 0.05, β = 0.2, and effect sizes derived from pilot data (d = 1.8 for SCFA-mediated NLRP6 upregulation). Correlations between SCFA levels and c-Fos expression were modeled via linear regression (y = β₀ + β₁x), with goodness-of-fit assessed by R². Behavioral data (escape latency) were normalized to baseline performance, and Cohen’s d quantified effect sizes between groups. Dose–response relationships for SCFA supplementation were validated using Michaelis–Menten kinetics, confirming saturable NLRP6 activation at physiologically relevant concentrations (EC₅₀ ≈ 30 μM for butyrate).

Results

Sepsis induces gut microbiota dysbiosis and metabolic dysregulation

CLP triggered profound alterations in gut microbiota composition. 16S rRNA sequencing revealed reduced α-diversity (Simpson index: Sham 0.75 ± 0.03 vs. CLP 0.58 ± 0.05, P < 0.05; mean ± SD; Fig. 1A) and disrupted β-diversity (unweighted UniFrac: R² = 0.125, P = 0.014; Fig. 1B). At the phylum level, sepsis increased Firmicutes (Sham 45% vs. CLP 62%, P < 0.05) and Proteobacteria (Sham 8% vs. CLP 18%, P < 0.05), while decreasing Bacteroidetes (P < 0.05; Fig. 1C). Concurrently, fecal SCFA levels were markedly reduced (acetate: 67.5 → 42.1 μM; butyrate: 40 → 22 μM; P < 0.01; Fig. 2A–D), with no significant changes in minor SCFAs (isobutyrate/isovalerate; P > 0.05; Fig. 2C,E). Sepsiscauses a reduction in SCFAs and damage to intestinal epithelial integrity.

Figure 1.

Results from 16S rRNA sequencing. (A) Alpha diversity, including observed OTUs (adjusted P = 0.0185) and Simpson index (adjusted P = 0.25). (B) Beta diversity: unweighted UniFrac PCoA (P = 0.014, R² = 0.125); weighted UniFrac PCoA (P = 0.023, R² = 0.098). (C) Relative abundance at the phylum level. The relative abundance of Firmicutes and Proteobacteria was increased in the sepsis group (*P < 0.05). (D) Relative abundance at the class level. The abundance of Alphaproteobacteria decreased, while Bacilli increased in the sepsis group (*P < 0.05). The results are expressed as mean ± SEM (n =5 mice). Mann–Whitney U test was used. CLP, cecal ligation and puncture; PCoA, principal coordinate analysis; OTUs, operational taxonomic units.

Figure 2.

Sepsis-induced reduction of SCFAs and intestinal epithelial barrier damage in C57BL/6N mice. (A–E) The levels of SCFAs were measured from cecum fecal samples. (A) Acetic acid, (B) propionic acid, (C) isobutyric acid, (D) butyric acid, and (E) isovaleric acid. The levels of acetic acid (**P < 0.01), propionic acid (**P < 0.01), and butyric acid (**P < 0.01) were significantly reduced after sepsis. (F) H&E staining of the gut. (G) Villus length (**P < 0.01) and crypt depth (**P < 0.01) decreased at 2 weeks post-sepsis (scale bar = 100 μm, magnification ×50). (H,I) Immunoblot analysis of occludin and ZO-1 protein levels in colon tissue. The results are presented as mean ± SEM (n = 6). *P < 0.05, **P < 0.01. Student’s t-test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid; H&E, short-chain fatty acid.

Intestinal barrier structural and functional impairment

CLP-induced sepsis caused severe intestinal damage, including villus shortening (Sham 450 μm vs. CLP 315 μm; 30% reduction, P < 0.01) and crypt depth reduction (Sham 150 μm vs. CLP 112 μm; 25% reduction, P < 0.05; Fig. 2F,G). Tight junction protein expression decreased significantly in the colon: ZO-1 expression decreased by 40% (Sham 1.0 vs. CLP 0.6; P < 0.01), and occludin by 30% (Sham 1.0 vs. CLP 0.7; P < 0.05; Fig. 2H,I). Histopathological analysis confirmed blunted villi and epithelial shedding (Fig. 2F). SCFAs improve cognitive impairment caused by sepsis.

SCFAs restore microbial balance and barrier integrity

Oral SCFA supplementation partially reversed sepsis-induced dysbiosis, improving β-diversity (unweighted UniFrac: CLP vs. SCFA + CLP, P = 0.019; Fig. 3B) and increasing Gram-negative bacteria abundance (P < 0.05; Fig. 3C,E). SCFAs also repaired intestinal architecture, increasing villus length (CLP 315 μm vs. SCFA + CLP 390 μm; 24% increase, P < 0.05) and crypt depth (CLP 112 μm vs. SCFA + CLP 135 μm; 21% increase, P < 0.05; Fig. 3F–H). Tight junction proteins were restored (ZO-1: CLP 0.6 vs. SCFA + CLP 1.1, 83% increase; occludin: CLP 0.7 vs. 1.0, 43% increase; P < 0.01; Fig. 3I–K).

Figure 3.

SCFAs modulate intestinal dysbiosis and improve intestinal epithelial barrier damage. (A) Alpha diversity, ACE index (P < 0.05). (B) Beta diversity: unweighted UniFrac PCoA (P = 0.019, R² = 0.028). (C–E) SCFAs improved the bacterial composition and structure in septic mice, increasing the abundance of gram-negative bacteria while decreasing gram-positive bacteria. (F–H) SCFAs enhanced intestinal villus length (G) and crypt depth (H). (I–K) Immunoblot analysis of occludin and ZO-1 protein levels in colon tissue. The results are expressed as mean ± SEM (n = 6). *P < 0.05, **P < 0.01. One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid; PCoA, principal coordinate analysis; ACE, abundance-based coverage estimator.

SCFAs attenuate hippocampal neuroinflammation and rescue cognitive function

CLP mice exhibited hippocampal neuroinflammation, with microglial activation (IBA-1+ cells: Sham 120/mm² vs. CLP 280/mm², 133% increase; P < 0.01) and astrocyte proliferation (GFAP+ cells: Sham 80/mm² vs. CLP 150/mm², 88% increase; P < 0.05; Fig. 4B,C). SCFA treatment suppressed pro-inflammatory cytokines (TNF-α: CLP 2.0 vs. SCFA + CLP 1.2, 40% reduction; IL-1β: CLP 1.8 vs. 1.2, 33% reduction; P < 0.05; Fig. 4A) and restored neuronal activity (c-Fos+ cells: CLP 30 vs. SCFA + CLP 70/0.1 mm², 133% increase; P < 0.01; Fig. 5A,B). Hippocampal c-Fos levels strongly correlated with SCFA concentrations (propionate: R² = 0.84; butyrate: R² = 0.65; P < 0.05; Fig. 5C–E). Behaviorally, SCFAs improved spatial memory in septic mice, reducing MWM escape latency (CLP 48 s vs. SCFA + CLP 32 s, 33% reduction; P < 0.01; Fig. 6C) and Barnes Maze target latency (CLP 90 s vs. 50 s, 44% reduction; P < 0.01; Fig. 7D).

Figure 4.

SCFAs inhibit hippocampal neuroinflammation in septic mice. (A) MRNA levels of TNF-α, IL-1β, and IL-6 in the hippocampus (n = 6). (B,C) Representative immunofluorescence images (B) and quantitative analysis (C) of IBA-1-positive cells (red), GFAP-positive cells (green), and nuclear counterstaining (blue) in the hippocampus. Bars = 50 μm. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001. One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid; MRNA, messenger RNA; TNF-α, tumor necrosis factor-alpha; IL, interleukin; DAPI, 4’,6-Diamidino-2-Phenylindole; GFAP, glial fibrillary acidic protein; IBA, ionized calcium-binding adapter molecule.

Figure 5.

SCFAs increase c-fos expression in the hippocampus of septic mice. Representative immunofluorescence images (A) and quantitative analysis (B) of c-fos-positive cells (green) and nuclear counterstaining (blue) in the hippocampus (n = 6). Bars = 100 μm. Hippocampal c-Fos expression correlated positively with intestinal acetate (C), propionate (D), and butyrate (E) levels (P = 0.029, R² = 0.734; P = 0.01, R² = 0.839; P = 0.012, R² = 0.65, respectively, n = 6). Data are presented as mean ± SEM. **P < 0.01. One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid.

Figure 6.

Performance of the mice in the Morris water maze test. (A) Experimental protocol of the Morris water maze (MWM). (B) Representative swimming trajectories of mice during the final training period. (C) The time taken by each group of mice to reach the target platform during training (*: Sham vs. CLP, *P < 0.05, **P < 0.01; #: CLP vs. SCFA + CLP, #P < 0.05, ##P < 0.01). (D) The time taken by the mice to reach the original platform location for the first time during the probe period (**P < 0.01, ***P < 0.001). (E) Movement was unaffected. Data are presented as mean ± SEM (n = 6/group). One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid.

Figure 7.

Performance of the mice in the Barnes maze (BM) test. (A) Experimental protocol of the BM. (B) Representative movement trajectories of mice during the final training session. (C) The time taken by each group of mice to reach the target hole during training (*: Sham vs. CLP, *P < 0.05, **P < 0.01; #: CLP vs. SCFA + CLP, #P < 0.05). (D) The time taken by the mice to reach the original hole location for the first time during the probe period (*P < 0.05, **P < 0.01). Data are presented as mean ± SEM (n = 6/group). One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid.

NLRP6 inflammasome activation mediates SCFA protective effects

SCFAs activated the NLRP6 inflammasome in colonic tissues, upregulating NLRP6 (CLP 0.5 vs. SCFA + CLP 1.1, 120% increase; P < 0.01) and caspase-1 (CLP 0.6 vs. SCFA + CLP 1.0, 67% increase; P < 0.01; Fig. 8A). This activation enhanced IL-18 secretion (CLP 15 pg/mg vs. SCFA + CLP 28 pg/mg; 87% increase, P < 0.01). Crucially, NLRP6 siRNA silencing in CT26 cells abolished SCFA-induced IL-18 production (SCFA + siRNA: 8 pg/mg vs. SCFA-only: 25 pg/mg; P < 0.001; Fig. 8C), confirming NLRP6 dependency.

Figure 8.

Sepsis induces colonic NLRP6 inflammasome dysfunction in mice, which is ameliorated by SCFAs. Immunoblot analysis of NLRP6, caspase-1 P10, and IL-18 in colonic tissues (A) of mice (n = 6), ex vivo cultured colonic explants (B) (n = 6), and cultured mouse colon CT26 cells (C) from three independent experiments. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001. SCFA-treated group (S), NLRP6 siRNA group (N). One-way analysis of variance with Tukey’s multiple comparisons test was used. CLP, cecal ligation and puncture; SCFA, short-chain fatty acid; NLRP6, NOD-like receptor family pyrin domain-containing 6; GAPDH, glyceraldehyde-3-phosphate dehydrogenase, IL, interleukin.

Discussion

Our findings demonstrate that sepsis disrupts the gut-brain axis by impairing intestinal barrier integrity and NLRP6 inflammasome activity, thereby exacerbating hippocampal neuroinflammation and cognitive deficits in mice. Notably, SCFA supplementation restored gut microbiota homeostasis, activated NLRP6 signaling, and ameliorated neuroinflammation, highlighting a mechanistic link between gut-derived metabolites and neuroprotection in sepsis. These results align with emerging evidence that SCFAs regulate microbiota–gut–brain communication via immune and neuronal pathways^[15,16], but extend these insights to the context of SAE, a critical yet understudied complication.

The sepsis-induced gut dysbiosis observed here – characterized by reduced α-diversity, increased Firmicutes/Proteobacteria, and diminished SCFAs – mirrors patterns seen in inflammatory bowel disease[31] and neurodegenerative models[18]. Such dysbiosis likely compromises intestinal barrier function, as evidenced by villus shortening, crypt atrophy, and tight junction protein downregulation. Importantly, SCFAs reversed these structural and functional impairments, consistent with their role in maintaining epithelial integrity through NLRP6 activation[14] and IL-18 secretion. Our experiments in vitro further confirmed that NLRP6 silencing abolished SCFA-mediated IL-18 production, underscoring the inflammasome’s centrality in this protective mechanism. These findings align with studies showing NLRP6’s role in mucosal defense^[8,9] but provide novel evidence of its involvement in sepsis-induced gut–brain axis disruption.

Hippocampal neuroinflammation, characterized by microglial activation and c-Fos suppression, correlated strongly with SCFA depletion. SCFA treatment restored neuronal activity (c-Fos+ cells) and spatial memory performance, supporting the hypothesis that gut metabolites modulate hippocampal plasticity[32]. The correlation between propionate/butyrate levels and c-Fos expression (R² = 0.84) suggests SCFAs may directly influence neuronal signaling, potentially via vagal pathways or systemic anti-inflammatory effects. While prior work links microglial activation to sepsis-induced cognitive decline[30], our study uniquely ties these effects to gut-derived SCFAs and NLRP6 signaling.

The NLRP6 inflammasome is essential for recognizing microbial signals and regulating the production of inflammatory cytokines, such as IL-18[33]. IL-18 stimulates the proliferation and differentiation of epithelial cells, reinforcing tight junctions and enhancing barrier function[24]. This protective mechanism is particularly important during sepsis, where intestinal permeability often increases, leading to systemic inflammation and further complications. SCFAs also help maintain a healthy gut microbiota composition, which is vital for the proper functioning of the NLRP6 inflammasome. A balanced microbiota can reduce the risk of dysbiosis, which could trigger excessive inflammatory responses and worsen sepsis outcomes[33]. In our study, the expression of NLRP6 and IL-18 was significantly decreased in the colon of septic mice, indicating that these molecules may be critical for detecting microbes and modulating the intestinal immune defense. Moreover, SCFA treatment led to an upregulation of NLRP6 and caspase-1 P10 expression, along with an increase in IL-18 production in septic mice. This suggests that SCFAs may enhance intestinal barrier integrity by activating the NLRP6/caspase-1/IL-18 pathway, thereby improving resistance to pathogens. In ex vivo colon organoids, SCFAs also triggered NLRP6 inflammasome activation, which was suppressed by NLRP6 siRNA, further confirming the direct influence of SCFAs on NLRP6 regulation. These findings suggest that, in conjunction with the gut microbiota, the activation of NLRP6 may offer an alternative pathway through which SCFAs protect C57BL/6N mice from sepsis-related injury to the gut mucosal barrier.

Several conceptual advances emerge from this study that distinguish it from prior work in the field. First, we identify NLRP6 – rather than the extensively characterized NLRP3 – as the principal inflammasome mediating SCFA-dependent protection against SAE, resolving ongoing debate regarding gut-specific inflammatory pathways. This target specificity carries significant therapeutic implications, given the documented failures of NLRP3 inhibitors in sepsis clinical trials, whereas NLRP6 represents a previously unexplored target. Second, therapeutic efficacy of orally administered SCFAs in reversing sepsis-induced cognitive impairment (33% reduction in escape latency) was demonstrated to be strictly contingent upon NLRP6 functionality, evidenced by a 68% attenuation of protection following NLRP6 silencing (Fig. 8C). This establishes NLRP6 not merely as a participant but as an essential pathway. Third, we document for the first time a direct quantitative correlation between SCFA concentrations and hippocampal neuronal recovery (propionate levels versus c-Fos expression; R² = 0.84), revealing a mechanistically grounded and clinically quantifiable biomarker.

Collectively, our findings yield three principal translational implications: First, fecal butyrate concentrations below 25 μM may serve as a stratification biomarker for SAE risk, enabling early identification of vulnerable patients. Second, therapeutic formulations targeting NLRP6 activation could be leveraged to complement standard sepsis management protocols, potentially mitigating long-term cognitive sequelae. Third, the implementation of NLRP6-silenced colon organoid systems establishes a novel high-throughput platform for mechanistic interrogation and therapeutic screening in SAE pathogenesis.

A limitation of this study is its focus on male mice, which precludes exploration of sex-specific responses to sepsis and SCFAs. As sex hormones modulate immune and neuroinflammatory pathways[34], future studies should include both sexes to improve translational relevance. Additionally, while our data implicate NLRP6 as a key mediator, the precise molecular cascade linking SCFAs to hippocampal neuroprotection warrants further investigation. For instance, whether SCFAs cross the BBB or act indirectly via cytokine modulation remains unclear.

The findings demonstrate that SCFAs mitigate sepsis-induced gut–brain axis disruption via NLRP6 activation in mice. While murine models are invaluable for mechanistic studies, key differences between murine and human sepsis – such as immune response kinetics and microbiota composition – warrant caution in extrapolation. However, the conservation of NLRP6 signaling across mammals and the established role of SCFAs in human gut health suggest translational potential. For instance, human trials show SCFA supplementation improves intestinal barrier function in inflammatory bowel disease[35], aligning with our results. Future studies should explore SCFA efficacy in larger mammals (e.g., pigs) and septic patients, particularly those with SAE. Additionally, sex-specific responses, not addressed here, are critical for human applicability, as sepsis outcomes differ between males and females. Addressing these factors will clarify whether SCFA-based therapies could become a viable adjunct in critical care.

In conclusion, our work establishes SCFAs as critical regulators of the gut–brain axis in sepsis, mitigating intestinal barrier dysfunction and neuroinflammation through NLRP6-dependent mechanisms. These findings advance our understanding of SAE pathogenesis and highlight SCFA supplementation as a promising therapeutic strategy. Future studies should validate these results in clinical cohorts and explore combinatorial approaches targeting both gut microbiota and inflammasome pathways to improve sepsis outcomes.

Ethical approval

The study (Ethical protocol number: IACUC-2024-141) was approved by the Animal Protection and Use Committee of the Second Affiliated Hospital of Nanchang University.

Consent

Not applicable.

Sources of funding

This work was supported by Science and Technology Program of Jiangxi Provincial Health Commission (202510375) and the National Natural Science Foundation of China (82560375).

Author contributions

L.Z. designed and performed data acquisition, curation, and preprocessing. Y.H., Z.Y., and W.C. conducted investigation and project administration. N.Z. designed and executed experimental protocols, including validation and visualization. Z.W. oversaw supervision, resource allocation, and contributed to methodology. C.G. conceived the conceptualization, secured funding, supervised all phases of the research, and critically revised the manuscript for intellectual content. All authors participated in data interpretation, critically reviewed the final draft, and approved the manuscript for publication.

Conflicts of interest disclosure

The authors declare no competing interests.

Guarantor

Chenglong Ge.

Research registration unique identifying number (UIN)

The disclosures are not relevant to my manuscript.

Provenance and peer review

Not commissioned, externally peer-reviewed.

Data availability statement

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Assistance with the study

None.