GPR43 deficiency aggravates sepsis by promoting gut microbiota–dependent barrier disruption and HIF-1α–ENO1 axis–mediated M1 polarization of macrophages

Abstract

Background

GPR43, a receptor for short-chain fatty acids (SCFAs), is broadly expressed in intestinal epithelial and immune cells and is essential for preserving barrier integrity and immune homeostasis. Nevertheless, how GPR43 influences gut microbiota composition and intestinal barrier integrity while also regulating macrophage immunometabolism in the context of sepsis remains poorly understood.

Methods

A cecal ligation and puncture model was used to induce sepsis in mice. Survival, histopathology, and immune responses were compared between Gpr43^−/− and wild-type mice; 16S ribosomal RNA (rRNA) sequencing and untargeted metabolomics were performed to evaluate gut microbiota composition and metabolic profiles. Antibiotic-mediated microbiota depletion and fecal microbiota transplantation were used to assess functional impacts. Bone marrow-derived macrophages were employed to investigate the effects of GPR43 deficiency on macrophage polarization. RNA sequencing, metabolic flux analysis, and Western blotting were conducted to explore the molecular mechanisms involved. Peripheral blood mononuclear cell samples from patients with sepsis were analyzed for clinical correlation.

Results

Gpr43^−/− mice exhibited significantly reduced survival following CLP, along with impaired intestinal barrier function and elevated proinflammatory cytokine levels. Microbiota diversity and SCFA-producing bacteria were markedly decreased, accompanied by reduced SCFA levels in fecal metabolites. Fecal microbiota transplantation (FMT) partially restored gut function and survival in Gpr43^−/− mice. GPR43-deficient macrophages displayed a strong M1-polarized phenotype with the upregulation of the glycolytic enzyme ENO1 and its upstream regulator HIF-1α. The inhibition of either ENO1 or HIF-1α reversed the proinflammatory phenotype. A clinical data analysis revealed that GPR43 expression was negatively correlated with IL-6, ENO1, and lactate levels.

Conclusions

GPR43 exerts a dual protective role in sepsis by maintaining gut microbiota homeostasis and barrier integrity and by modulating macrophage metabolism and polarization via the HIF-1α–ENO1 axis. This study provides novel insights into the GPR43 in pathogenesis of sepsis and suggests potential therapeutic targets for intervention.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s11658-025-00833-4.

Introduction

Sepsis is a life-threatening syndrome caused by infection, characterized by a dysregulated host response that leads to uncontrolled systemic inflammation, multiple organ dysfunction, and high mortality [1]. Despite significant advances in antimicrobial therapy and supportive care, sepsis remains a major global cause of death, with an estimated 49 million cases and 11 million deaths each year [2]. Currently, effective immunomodulatory strategies capable of restoring host immune balance are still lacking [3]. At the core of sepsis pathogenesis lies profound immune dysregulation, often accompanied by an imbalance in cytokine production [4]. Macrophages are central effector cells that produce large amounts of proinflammatory cytokines such as TNF-α, IL-6, and IL-1β, which, when excessively released, amplify systemic inflammation [5]. The cytokine profile of macrophages is tightly linked to their polarization and metabolic state: M1 macrophages rely predominantly on glycolysis and produce proinflammatory mediators that promote pathogen clearance, whereas M2 macrophages depend on oxidative phosphorylation and secrete anti-inflammatory cytokines that suppress excessive immune activation and support tissue repair [6–8]. During sepsis, incomplete pathogen clearance often results in sustained macrophage activation and skewing toward the M1 phenotype, thereby driving cytokine storms, vascular injury, and organ failure [9]. These observations underscore the urgent need to precisely regulate macrophage function and metabolism to restore immune balance and improve outcomes in sepsis.

Given the central role of macrophage metabolism in sepsis, increasing attention has been directed toward host–microbiota interactions and their metabolites as potential regulators of immune balance [10]. Short-chain fatty acids (SCFAs), including acetate, propionate, and butyrate, are the major products of dietary fiber fermentation by the gut microbiota [11, 12]. Beyond serving as energy substrates for intestinal epithelial cells, SCFAs act as signaling molecules that regulate immune responses and barrier integrity. Among their receptors, GPR43 (also known as FFAR2) is highly expressed in intestinal epithelial and immune cells and mediates SCFA-dependent regulation of both epithelial barrier function and host immunity [13, 14]. During infection, SCFA–GPR43 signaling enhances tight junction integrity, stimulates mucus secretion, and limits bacterial translocation, thereby protecting against systemic inflammation [15–19]. In parallel, it modulates immune responses by promoting neutrophil recruitment, enhancing antimicrobial activity, and restraining excessive macrophage activation [20–23]. Animal models of inflammatory bowel disease, bacterial infection, and acute lung injury consistently demonstrate protective roles of GPR43 [24–27], while clinical data show that higher GPR43 expression correlates with improved outcomes in patients with sepsis [28]. Collectively, these findings position GPR43 as a pivotal molecular link between the gut microbiota, host metabolism, and immune defense against infection, with its protective functions being most extensively characterized in most intestinal diseases.

However, the role of GPR43 in systemic infection remains incompletely defined. Most previous studies have focused on cellular responses to exogenous SCFA supplementation while largely neglecting the intrinsic functions of GPR43 itself, particularly in macrophages [29, 30]. Moreover, the literature reports conflicting findings regarding GPR43 in inflammatory regulation. For example, Maslowski et al. reported that GPR43 suppresses neutrophil recruitment and ameliorates colitis [31], while Macia et al. showed that GPR43 limits inflammasome activation [32], indicating a protective role in inflammation. By contrast, Sina et al. demonstrated that GPR43 deficiency exacerbates colitis by impairing neutrophil recruitment [33], and Kim et al. reported that GPR43 deficiency aggravates colitis via MAPK pathway activation [34]. These contradictory results raise the unresolved question of what role GPR43 plays in the context of sepsis. In addition, it remains unclear how GPR43 regulates gut microbiota composition, maintains intestinal barrier integrity during infection, and influences macrophage polarization and metabolic reprogramming.

In this study, we systematically investigated the role of GPR43 in a clinically relevant polymicrobial sepsis model. We hypothesized that GPR43 exerts dual protective effects by preserving the gut barrier integrity and intrinsically regulating macrophage function. By integrating multiomics analyses with in vivo and in vitro macrophage experiments, we found that GPR43 deficiency disrupts gut homeostasis through microbiota-dependent mechanisms and promotes M1 polarization via hypoxia-inducible factor 1 alpha (HIF-1α)–enolase 1 (ENO1)–driven glycolysis. Notably, pharmacological activation of GPR43 with the selective agonist TUG-1375 alleviated inflammation and improved survival in septic mice. Collectively, these findings provide new mechanistic insights into the immunometabolic functions of GPR43 and highlight its potential as a therapeutic target in sepsis.

Materials and methods

Animal models

Polymicrobial sepsis was induced in mice using the cecal ligation and puncture (CLP) procedure. Briefly, mice were anesthetized with isoflurane, and a midline abdominal incision was made to expose the cecum. The cecum was ligated at approximately 50% of its length and punctured once with a 21-gauge needle. A small amount of fecal content was gently extruded, and the cecum was returned to the peritoneal cavity. The incision was closed, and mice were resuscitated with subcutaneous sterile saline (1 mL per 20 g body weight). Peritoneal lavage fluid (PLF) was collected by instilling 5 mL of cold sterile PBS into the peritoneal cavity, gently massaging the abdomen, and aspirating the fluid for downstream analyses. Sepsis was alternatively induced by intraperitoneal injection of lipopolysaccharide (LPS, 20 mg/kg). For pharmacological interventions, mice were treated with the GPR43-specific agonist TUG-1375 (100 mg/kg, intraperitoneally) or the ENO1 inhibitor AP-III-a4 (10 mg/kg, intraperitoneally) 3 h prior to CLP surgery.

Bone marrow–derived macrophage (BMDM) culture and polarization

Bone marrow cells were isolated from the femurs and tibias of mice and cultured in complete Dulbecco’s modified Eagle medium (DMEM) supplemented with macrophage colony-stimulating factor (M-CSF) (25 ng/mL, Novoprotein, catalog number (cat. no.) CD34) at 37 °C with 5% CO₂. After 7 days of differentiation, Bone marrow–derived macrophages (BMDMs) were harvested for experiments. For polarization, cells were stimulated for 24 h with lipopolysaccharide (LPS) (100 ng/mL, Invivogen, cat. no. Tlrl-eklps) plus interferon gamma (IFN-γ) (20 ng/mL, Novoprotein, cat. no. C746) to induce M1 macrophages or with interleukin (IL)−4 (20 ng/mL, Novoprotein, cat. no. CK74) plus IL-13 (20 ng/mL, Novoprotein, cat. no. CH18) to induce M2 macrophages. For the flow cytometry, cells were washed with phosphate-buffered saline (PBS), blocked with anti-CD16/32, stained with CD45, CD11b, and F4/80, fixed/permeabilized (Cytofix/Cytoperm Kit, BD, cat. no. 554714), and subjected to intracellular staining for inducible nitric oxide synthase (iNOS) and arginase 1 (Arg1) before analysis (FACSCelesta, BD). For the inhibitor or agonist treatments, BMDMs were pretreated with glycolysis inhibitor 2-deoxyglucose (2-DG) (2 mM, 6 h, MCE, cat. no. HY-13966), ENO1 inhibitor AP-III-a4 (20 μM, 6 h, MCE, cat. no. HY-15858), HIF-1α inhibitor acriflavine (5 μM, 6 h, MCE, cat. no. HY-100575), or the GPR43 agonist TUG-1375 (20 μM, 6 h, MCE, cat. no. HY-112813) before M1 polarization.

Histological analysis and tissue injury scoring

Lung and intestinal tissues were collected 24 h after CLP, fixed in 4% paraformaldehyde for 24 h, paraffin-embedded, sectioned at 5 μm, and stained with hematoxylin and eosin (H&E) using standard protocols. Sections were evaluated by two blinded investigators under a light microscope. Lung injury was scored on the basis of five parameters [35]: alveolar structure destruction, inflammatory infiltration, edema, hemorrhage, and hyaline membrane formation, each graded 0–4, with a maximum total score of 20. Intestinal injury was scored on a 0–4 scale [36], with 0 meaning no inflammation; 1 meaning mild infiltration in lamina propria; 2 meaning moderate infiltration with crypt separation and mild hyperplasia; 3 meaning extensive infiltration with mucosal destruction, goblet cell loss, and epithelial thickening; and 4 meaning crypt abscesses or ulceration. Representative images were obtained with a digital slide scanner (Pannoramic MIDI, 3DHISTECH).

TUNEL staining of lung and intestinal tissues and quantification

Apoptotic cells in lung and intestinal tissues were detected using terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) Apoptosis Assay Kit (Beyotime, cat. no. C1086). Paraffin-embedded sections were deparaffinized, rehydrated, and permeabilized with 0.1% Triton X-100. Sections were incubated with TUNEL solution at 37 °C for 1 h in the dark, followed by 4′,6-diamidino-2-phenylindole (DAPI) counterstaining. Images were captured using a Zeiss LSM 980 confocal microscope. For each sample, five random high-power fields were analyzed. The apoptotic index was calculated as the percentage of TUNEL-positive nuclei among total DAPI-stained nuclei using Image J.

FITC–dextran permeability assay

To evaluate intestinal permeability, mice were subjected to CLP surgery and, 24 h post-operation, were fasted for 4 h. Subsequently, they were orally gavaged with 600 mg/kg fluorescein isothiocyanate (FITC)-dextran (Sigma-Aldrich, cat. no. 46944). After 4 h, blood was collected via retro-orbital puncture and allowed to clot at room temperature. The serum was isolated by centrifugation at 3000g for 10 min and analyzed for fluorescence using a microplate reader (Tecan Spark, excitation: 485 nm, emission: 528 nm).

Glucose uptake assay

BMDMs were treated with or without LPS + IFNγ, followed by incubation in glucose-free DMEM (Gibco, cat. no. 11966025) for 1 h. Cells were then incubated with 2-NBDG (Thermo Fisher, cat. no. N13195) at 100 μM for 45 min at 37 °C in the dark. After washing three times with PBS, 2-NBDG fluorescence was measured by flow cytometry (BD LSRFortessa) using the FITC channel.

Measurement of lactic acid and inflammatory cytokines

Lactic acid quantification in cell culture supernatant was performed using a Lactic Acid Assay Kit (Nanjing Jiancheng Bioengineering Institute, cat. no. A019-2-1) according to the manufacturer’s instructions. Cytokine concentrations in cell culture supernatants and PLF were quantified using enzyme-linked immunosorbent assay (ELISA) kits for tumor necrosis factor alpha (TNF-α) (Dakewe, cat. no. 1217202), IL-1β (Dakewe, cat. no. 1210122), IL-6 (Dakewe, cat. no. 1210602), and IL-10 (Dakewe, cat. no. 1211002) following the manufacturer’s protocols.

Flow cytometry analysis of PLF macrophages

Peritoneal lavage fluid (PLF) was centrifuged, and red blood cells were lysed with red blood cell (RBC) lysis buffer for 5 min. After washing, cells were divided into two panels. In panel 1 (cytokine production), cells were stimulated for 4 h at 37 °C in Roswell Park Memorial Institute 1640 (RPMI 1640) containing 10% fetal bovine serum (FBS), cell stimulation cocktail with protein transport inhibitor (eBioscience, cat. no. 00–4975-93), LPS (100 ng/mL), and antibiotics. After washing, cells were stained with Zombie dye, blocked with anti-CD16/32, and surface-stained for CD45, CD11b, F4/80, and lineage markers (CD19, Ly6G, CD3, NK1.1). Cells were then fixed, permeabilized, and intracellularly stained for IL-6, IL-1β, TNF-α, and IL-10 before flow cytometric analysis. In panel 2 (polarization markers), cells were surface-stained for CD45, CD11b, F4/80, MHCII, and CD86 followed by fixation, permeabilization, and intracellular staining for iNOS, CD206, and Arg1 before analysis. Antibodies are presented in Supplementary Table S1.

Flow cytometry analysis of splenic Th17 and T_reg cells

Single-cell suspensions were prepared from spleens by mechanical dissociation, followed by RBC lysis. After Fc receptor blocking, cells were surface-stained with CD45, CD3, CD4, and CD25, fixed, and permeabilized using the Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific, cat. no. 00–5523-00). Intracellular staining for Foxp3 and RORγt was performed at 4 °C for 4 h. T_reg cells were defined as CD45⁺CD3⁺CD4⁺CD25⁺Foxp3⁺, and Th17 cells as CD45⁺CD3⁺CD4⁺Foxp3⁻RORγt⁺. Data were analyzed using FlowJo software.

Western blotting

After treatment, cells were lysed in radioimmunoprecipitation assay (RIPA) buffer supplemented with protease and phosphatase inhibitors. Protein samples were denatured by boiling at 100 °C for 5 min and then separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The gel was run at 80 V for 30 min followed by 120 V for 60 min. Proteins were subsequently transferred onto polyvinylidene fluoride (PVDF) membranes at 90 V for 90 min in a cold transfer buffer. The membranes were blocked with 5% nonfat dry milk in Tris-buffered saline with Tween 20 (TBST) for 1 h at room temperature and then incubated overnight at 4 °C with the appropriate primary antibodies (Supplementary Table S1). After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence reagents and imaged with a chemiluminescence detection system.

RNA extraction, qRT–PCR, and RNA sequencing

Total RNA was extracted from BMDM and peripheral blood mononuclear cell (PBMC) cells treatment using TRIzol reagent (Thermo Fisher Scientific, cat. no. 15596026), following the manufacturer’s protocol. RNA isolation and quantitative real-time PCR (qRT–PCR) were conducted as described previously, and the primer sequences used are listed in Supplementary Table S2. The RNA sequencing was carried out by the Beijing Genomics Institute (BGI, Shenzhen, China) on the BGISEQ-500 platform, as detailed in reference [37].

OCR and ECAR measurements

BMDMs were seeded at 5 × 10⁴ cells per well in XFe96 plates and cultured with or without LPS (100 ng/mL) + IFN-γ (20 ng/mL) for 24 h. Before the assay, cells were switched to Seahorse XF base DMEM and equilibrated for 1 h at 37 °C in a CO₂-free incubator. For the glycolysis stress test, assay medium contained 2 mM glutamine (no glucose/pyruvate) and sequential injections were glucose (10 mM), rotenone/antimycin A (0.5 μM), and 2-deoxyglucose (50 mM). For the mitochondrial stress test, the assay medium contained 10 mM glucose, 1 mM pyruvate, and 2 mM glutamine with sequential injections of oligomycin (1.5 μM), FCCP (1 μM), and rotenone/antimycin A (0.5 μM). The oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were recorded on a Seahorse XFe96 analyzer and analyzed in the vendor’s software.

Immunofluorescence staining and confocal microscopy

Frozen tissue samples were embedded in OCT, rapidly frozen in liquid nitrogen, and sectioned at 10-μm thickness. Sections were mounted on slides, rehydrated in PBS, and blocked with 10% goat serum for 30 min at room temperature. After blocking, sections were incubated overnight at 4 °C with primary antibodies against ZO-1, MUC2 and F4/80 followed by incubation with appropriate fluorescent secondary antibodies for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI for 5 min. Slides were mounted with antifade medium and imaged using a Zeiss LSM980 confocal microscope.

Cecal microbiome and metabolome profiling via 16S rRNA and LC–MS

Fecal and cecal content samples were collected from WT and Gpr43^−/− mice and stored at −80 °C until analysis. For microbiome profiling, 16S ribosomal RNA (rRNA) sequencing targeting the V3–V4 region was performed by Shanghai OE Biotech using the primers 343 F (TACGGRAGGCAGCAG) and 798R (AGGGTATCTAATCCT), followed by microbial diversity analysis. For metabolomic profiling, 30 mg of cecal content was extracted with 400 μL methanol–water, homogenized, ultrasonicated, and centrifuged. The supernatants were analyzed using a Waters ACQUITY UPLC I-Class plus system coupled to a Thermo QE HF high-resolution mass spectrometer. Quality control samples were prepared by pooling equal volumes from all extracts. All raw sequencing and metabolomics data are available from the authors upon request.

Mouse fecal microbiota transplantation (FMT) procedure

Mice received an antibiotic cocktail in drinking water for 14 days (ampicillin 1 g/L, neomycin 1 g/L, metronidazole 1 g/L, vancomycin 0.5 g/L), followed by a 2-day washout. Donor feces were homogenized in prechilled PBS, briefly stored at −80 °C, and administered by oral gavage once daily for three consecutive days.

Transmission electron microscopy (TEM) of mitochondria in BMDMs

BMDMs were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer at 4 °C overnight. After washing with phosphate buffer, cells were postfixed in 1% osmium tetroxide for 1 h at room temperature, dehydrated through a graded ethanol series, and embedded in epoxy resin. Ultrathin sections were cut using an ultramicrotome, mounted on copper grids, and stained with uranyl acetate and lead citrate. Mitochondrial ultrastructure was observed and imaged using a transmission electron microscope (JEM-1400).

Clinical data and statistical analysis

Study participants: all patients were diagnosed with sepsis according to the guidelines of the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). Inclusion criteria were as follows: (1) confirmed or strongly suspected infection; (2) secondary organ dysfunction or acute exacerbation of preexisting organ dysfunction; and (3) a Sequential Organ Failure Assessment (SOFA) score ≥ 2. The clinical characteristics of all 36 enrolled sepsis patients were documented at the time of enrollment (Supplementary Table S3). A total of 15 healthy individuals were recruited as controls. Peripheral blood mononuclear cells (PBMCs) were collected using lymphocyte separation medium (Solarbio, cat. no. P8610), and serum lipopolysaccharide-binding protein (LBP) concentrations were determined with a Human LBP ELISA Kit (Invitrogen, cat. no. EH297RBX5). All procedures and sample collection were approved by the Clinical Ethics Committee of the First Affiliated Hospital of Anhui Medical University (approval no. PJ2024-10–50).

Statistical analysis

All experiments were independently performed in triplicate. Statistical analyses were conducted using GraphPad Prism 9. Spearman’s rank correlation was used to assess correlations, and survival curves were generated via the Kaplan–Meier method. For comparisons involving nonnormally distributed data, the Wilcoxon rank-sum test was applied, while normally distributed data were analyzed using unpaired Student’s t-tests. One-way or two-way analysis of variance (ANOVA) was used for comparisons among multiple groups, followed by Tukey’s multiple comparison test to calculate p-values. Experimental results are presented as the mean ± standard deviation (SD) and differences were considered statistically significant at p < 0.05.

Results

GPR43 deficiency increases mortality in CLP-induced sepsis, associated with enhanced inflammatory cytokine release and impaired intestinal epithelial barrier function

To investigate the role of GPR43 in sepsis, we first established a cecal ligation and puncture (CLP) model to mimic polymicrobial sepsis, with sham-operated mice serving as controls to exclude surgical trauma. In this model, GPR43 knockout (Gpr43^−/−) mice exhibited a markedly lower survival rate (0%) compared with wild-type (WT) controls (40%) (Fig. 1A), suggesting that GPR43 plays a critical protective role in host defense against infection.

Fig. 1.

GPR43 deficiency exacerbates mortality, tissue injury and intestinal barrier dysfunction in CLP-induced sepsis. A Survival curves of WT and Gpr43^−/− mice after sham or CLP surgery (n = 10). B Representative H&E staining of lung and intestinal tissues from WT and Gpr43^−/− mice following sham or CLP 24 h. C, D Quantification of intestinal (C) and lung (D) injury scores (n = 6). E Intestinal permeability assessed by FITC–dextran assay after CLP 24 h (n = 6). F Representative immunofluorescence staining of MUC2 (red) and ZO-1 (green) in intestine; DAPI (blue) for nuclei. G, H Quantification of mean fluorescence intensity of MUC2 (G) and ZO-1 (H) in intestinal tissues (n = 6). I Relative mRNA expression of tight junction–related genes in intestinal tissues determined by qRT–PCR (n = 6). Data are presented as means ± SD and analyzed by one-way ANOVA with Tukey’s post hoc test. Survival by Kaplan–Meier with log-rank test. Data are representative of three independent experiments. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

To further elucidate the underlying mechanisms of increased mortality, we performed a histological analysis of major target organs. GPR43-knockout (KO) mice displayed aggravated intestinal and pulmonary injury, characterized by epithelial disruption, extensive inflammatory cell infiltration, and higher histological damage scores (Fig. 1B–D). TUNEL staining further revealed significantly increased apoptosis in these tissues of KO mice (Supplementary Fig. S1A–C). Given that GPR43 is predominantly expressed in intestinal epithelial cells and immune cells, and the intestine is considered the central “driver organ” in CLP-induced sepsis [38], we next focused on intestinal barrier function. An in vivo FITC–dextran permeability assay demonstrated significantly elevated intestinal permeability in CLP-treated KO mice (Fig. 1E), indicating severe barrier disruption. The immunofluorescence analysis confirmed reduced expression and disorganized distribution of the tight junction protein ZO-1 and the goblet cell–derived mucin marker MUC2 in the intestinal epithelium of KO mice (Fig. 1F–H). ZO-1, a key scaffold protein in the tight junction complex, is essential for maintaining intercellular adhesion and preventing pathogens and toxins from crossing the intestinal barrier into circulation. Meanwhile, MUC2, secreted by goblet cells, constitutes the major component of the mucus layer that provides the first line of chemical defense in the intestine [39]. Their simultaneous downregulation suggests that GPR43 deficiency substantially impairs intestinal barrier integrity.

Consistently, quantitative polymerase chain reaction (qPCR) analysis revealed that mRNA levels of tight junction–related genes, including Tjp1, Ocln, and Cldn1, were significantly decreased in the intestines of knockout mice (Fig. 1I), further confirming barrier disruption. Concomitant with barrier dysfunction, proinflammatory cytokines such as TNF-α, IL-6, and IL-1β were markedly upregulated in the intestinal tissue (Supplementary Fig. S1D), and their serum concentrations were also elevated (Supplementary Fig. S1E), indicating that increased permeability amplified gut-derived inflammation. To validate the generalizability of these findings, we further employed a lipopolysaccharide (LPS)-induced sepsis model. Similar to the CLP model, GPR43 deficiency resulted in higher mortality (Supplementary Fig. S1E), aggravated intestinal pathology (Supplementary Fig. S1G,H), and exacerbated barrier injury (Supplementary Fig. S1I–K).

Taken together, these findings demonstrate that GPR43 deficiency disrupts intestinal barrier structure and function, leading to increased intestinal permeability, amplified inflammatory cytokine release, and aggravated tissue injury, thereby predisposing mice to heightened sepsis susceptibility and mortality. This highlights the essential role of GPR43 in maintaining intestinal barrier homeostasis and controlling systemic inflammation.

GPR43 deficiency impairs intestinal barrier integrity through gut microbiota and metabolic dysregulation

The gut microbiota is well known to play a pivotal role in maintaining intestinal barrier function, and the SCFA–GPR43 axis is thought to be a key mediator in this process [40]. However, whether GPR43 deficiency alters gut microbial composition during sepsis remains unclear. To address this question, we performed 16S rRNA sequencing of the gut microbiota in Gpr43^−/− and WT mice. Principal component analysis (PCA) revealed a clear separation between the two groups, indicating substantial differences in microbial composition (Fig. 2A). Alpha diversity, as measured by the Chao1, Shannon, and Simpson indices, was significantly reduced in Gpr43^−/− mice (Fig. 2B–D). The beta diversity analysis using Bray–Curtis distance and principal coordinate analysis (PCoA) further demonstrated a distinct clustering between WT and Gpr43^−/− groups, with permutational multivariate analysis of variance (PERMANOVA) showing p < 0.001 (Fig. 2E). Collectively, these data suggest that GPR43 deficiency profoundly alters gut microbial community structure.

Fig. 2.

Gut microbiota, metabolomic profiles, and fecal microbiota transplantation in WT and Gpr43^−/− mice. A PCA of gut microbiota composition in WT and Gpr43^−/− mice based on 16S rRNA sequencing. B–D Chao1 (B), Shannon (C), and Simpson (D) indices of microbial communities in WT and Gpr43^−/− mice. E A PCoA plot showing the beta diversity of gut microbial communities in WT and Gpr43^−/− mice. The analysis was based on Bray–Curtis distances. The statistical significance of the separation was determined by a PERMANOVA test. F A heat map of the different gut microbiota composition at the phylum level in WT and Gpr43^−/− mice. G A heat map of the different gut microbiota composition at the genus level in WT and Gpr43^−/− mice. H PCA of untargeted fecal metabolomic profiles from WT and Gpr43^−/− mice. I Volcano plot of differential metabolites identified in WT and Gpr43^−/− mice. J Lollipop plot showing the top ten significantly upregulated and downregulated metabolites in Gpr43^−/− compared with WT mice. K Survival curves of WT and Gpr43^−/− mice subjected to CLP with or without FMT. In the Gpr43^−/− + FMT group, WT microbiota were transplanted into Gpr43^−/−; in the WT + FMT group, Gpr43^−/− microbiota were transplanted into WT (n = 12). L–N Quantification of mean fluorescence intensity for MUC2 (L) and ZO-1 (M) in intestinal tissues of WT and Gpr43^−/− mice ± FMT (n = 6) and representative immunofluorescence images (N) showing MUC2 (red) and ZO-1 (green) with DAPI (blue) for nuclei. O FITC–dextran assay of intestinal permeability in WT and Gpr43^−/− mice ± FMT (n = 6). Data in B–D are shown as median ± interquartile range and analyzed using the Wilcoxon rank-sum test. Data in L–O are shown as means ± SD and analyzed using one-way ANOVA with Tukey’s post hoc test. Survival by Kaplan–Meier with log-rank test. Data in B–D were obtained from 16S rRNA sequencing analysis. Data in L–O are representative of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001

At the phylum level, GPR43 deficiency significantly reduced the abundance of Firmicutes and Bacteroidetes while increasing the abundance of Proteobacteria (Fig. 2F). At the genus level, several beneficial SCFA-producing taxa, including Parabacteroides, Lactobacillus, Roseburia, Odoribacter, and Bacteroides, were markedly decreased (Fig. 2G). To further explore the metabolic consequences of these microbial shifts, we conducted an untargeted metabolomic profiling of fecal samples. The PCoA of metabolomic data revealed distinct clustering between WT and Gpr43^−/− mice (Fig. 2H). In total, 330 metabolites were upregulated, and 350 were downregulated in Gpr43^−/− mice compared with WT controls (Fig. 2I). Among these, inflammation-associated metabolites such as PGG2 were markedly elevated, whereas lipid metabolism–related metabolites, including glycerophosphorylethanolamine and 3R-hydroxy-5Z-dodecenoic acid, were significantly reduced (Fig. 2J). Importantly, the levels of SCFAs such as acetate, propionate, and butyrate were also consistently decreased (Supplementary Fig. S2A–C). A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis further confirmed significant alterations in multiple pathways, including fatty acid metabolism, taurine metabolism, and steroid hormone biosynthesis (Supplementary Fig. S2D). Together, these results indicate that GPR43 deficiency reshapes gut microbiota and disrupts host metabolic homeostasis, particularly in lipid and inflammatory metabolism.

To determine whether GPR43-related barrier defects are microbiota-dependent, we treated mice with a broad-spectrum antibiotic cocktail for 2 weeks to deplete endogenous microbiota, followed by fecal microbiota transplantation (FMT). Specifically, microbiota from WT mice were transferred to Gpr43^−/− mice, while microbiota from Gpr43^−/− mice were transplanted into WT recipients. Following CLP-induced sepsis, survival analysis showed that Gpr43^−/− mice receiving WT microbiota exhibited a significantly improved survival rate. Conversely, WT mice transplanted with Gpr43^−/− microbiota showed a trend toward reduced survival, although this did not reach statistical significance (Fig. 2K). These findings suggest that gut microbiota contribute to the heightened susceptibility of Gpr43^−/− mice to sepsis.

To further evaluate whether this effect was linked to barrier dysfunction, we performed FITC–dextran permeability assays and immunofluorescence staining for ZO-1 and MUC2 in intestinal tissues. The WT microbiota transfer partially restored barrier integrity in Gpr43^−/− mice, while the transplantation of Gpr43^−/− microbiota exacerbated barrier disruption in WT mice (Fig. 2L–O). Taken together, these results provide strong evidence that gut microbiota play a causal role in mediating intestinal barrier dysfunction and sepsis susceptibility associated with GPR43 deficiency. However, it is noteworthy that the survival rate of Gpr43^−/− mice was not fully rescued by WT microbiota transplantation, and WT mice receiving Gpr43^−/− microbiota did not exhibit a statistically significant reduction in survival. These findings imply that additional GPR43-regulated mechanisms, beyond the microbiota, likely contribute to sepsis outcomes. Overall, our study demonstrates that GPR43 deficiency reshapes the gut microbiota, reduces SCFA-producing bacteria and their metabolites, disrupts intestinal homeostasis and barrier integrity, and ultimately increases susceptibility to sepsis.

Excessive inflammatory cytokine production in Gpr43^−/− mice during sepsis is associated with enhanced macrophage M1 polarization

Our previous findings indicated that Gpr43^−/− mice exhibited more severe intestinal barrier disruption following CLP. Given the pivotal role of the inflammatory response in the pathogenesis of sepsis, we next assessed the production of proinflammatory cytokines. An analysis of PLF showed that levels of TNF-α, IL-6, and IL-1β were markedly elevated in Gpr43^−/− mice compared with WT controls after CLP (Fig. 3A), suggesting an exaggerated systemic inflammatory response. Since macrophages are a major source of inflammatory cytokines during sepsis, we investigated whether GPR43 influences the host response to sepsis through macrophage regulation. The flow cytometric analysis of PLF 24 h after CLP (Supplementary Fig. S3A) revealed no significant differences in the total number of peritoneal cells or macrophages (CD11b⁺F4/80⁺) between WT and Gpr43^−/− mice (Fig. 3B–D). However, the intracellular cytokine staining revealed striking functional differences: compared with WT macrophages, Gpr43^−/− macrophages produced significantly higher levels of proinflammatory cytokines (TNF-α, IL-6, and IL-1β) and lower levels of the anti-inflammatory cytokine IL-10 (Fig. 3E, F). As cytokine production is closely linked to macrophage polarization, these results imply a shift toward a proinflammatory phenotype in GPR43-deficient macrophages.

Fig. 3.

GPR43 deficiency enhances proinflammatory cytokine production and skews macrophages toward an M1 phenotype during sepsis. A Concentrations of cytokines in PLF measured by ELISA from WT and Gpr43^−/− mice 24 h after CLP (n = 6). B Representative flow cytometry plots showing gating of peritoneal macrophages (CD11b⁺F4/80⁺). C, D Quantification of macrophage numbers (C) and total peritoneal cell numbers (D) in PLF. E, F Intracellular cytokine staining of PLF macrophages showing percentages of IL-1β⁺, IL-6⁺, IL-10⁺, and TNF-α⁺ cells (n = 6) (E) and representative flow-cytometry plots (F). G Representative histograms of M1 marker (CD86, MHCII, iNOS) and M2 marker (Arg1, CD206) markers in PLF macrophages. H–L Quantification of MFI of CD86 (H), MHCII (I), iNOS (J), Arg1 (K), and CD206 (L) (n = 6). Data are presented as means ± SD and analyzed by two-tailed unpaired Student’s t-test. Data are representative of three independent experiments. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001

To determine macrophage polarization status more precisely, we evaluated M1 and M2 markers in PLF macrophages by flow cytometry. Gpr43^−/− macrophages displayed an enhanced proinflammatory M1 phenotype, characterized by the increased expression of CD86, MHCII, and iNOS (Fig. 3G–J), whereas the expression of M2-associated markers such as CD206 and Arg1 was significantly reduced (Fig. 3K, L). Given prior reports that GPR43 deficiency may affect the balance of TH17 and T_reg cells [41, 42], we also examined splenic TH17 (CD3⁺CD4⁺Foxp3⁻RORγt⁺) and T_reg (CD3⁺CD4⁺CD25⁺Foxp3⁺) populations. However, no significant differences were observed between WT and Gpr43^−/− mice (Supplementary Fig. S3B-C). Collectively, these in vivo data demonstrate that GPR43 deficiency skews macrophages toward a proinflammatory M1 phenotype.

Importantly, circulating SCFA levels were markedly reduced in Gpr43^−/− mice (Supplementary Fig. S2A–C), and SCFAs are known to exert anti-inflammatory effects and promote M2 polarization. Thus, it remains to be determined whether the altered macrophage phenotype observed in Gpr43^−/− mice arises as an indirect consequence of reduced SCFA levels or whether it reflects an intrinsic regulatory role of GPR43 signaling within macrophages themselves.

GPR43 deficiency promotes LPS + IFN-γ–induced M1-like macrophage polarization in vitro

To directly determine the role of GPR43 in macrophage polarization, we performed in vitro experiments using BMDMs cultured under identical conditions without exogenous SCFAs. BMDMs were stimulated with either M1-polarizing stimuli (LPS + IFN-γ) or M2-polarizing stimuli (IL-4 + IL-13). Consistent with our hypothesis, Gpr43^−/− BMDMs exhibited a stronger proinflammatory M1 phenotype, as evidenced by markedly increased iNOS expression and elevated production of proinflammatory cytokines (TNF-α, IL-6, and IL-1β) under M1-polarizing conditions (Fig. 4A–C). By contrast, under M2-polarizing conditions, no significant differences were observed between groups in the expression of M2-associated markers (Arg1, Ym1, Fizz1) or the anti-inflammatory cytokine IL-10 (Fig. 4D–F), indicating that GPR43 deficiency specifically affects M1 but not M2 polarization.

Fig. 4.

GPR43 deficiency promotes M1 polarization of macrophages in vitro. A qRT–PCR analysis of M1 marker mRNA expression in WT and Gpr43^−/− BMDMs were polarized to M1 with LPS (100 ng/mL) and IFN-γ (20 ng/mL) for 24 h or mock control (n = 3). B, C Flow cytometry analysis of iNOS expression in WT and Gpr43^−/− BMDMs under M1 polarization (n = 3). Representative flow cytometry histograms (B) and quantification of iNOS MFI (C). D qRT–PCR analysis of M2 marker mRNA expression in WT and Gpr43^−/− BMDMs were polarized to M2 with IL-4 (20 ng/mL) + IL-13 (20 ng/mL) for 24 h, or mock control (n = 3). E, F Flow cytometry analysis of Arg1 expression in WT and Gpr43^−/− BMDMs under M2 polarization (n = 3). Representative flow cytometry histograms (E) and quantification of Arg1 MFI (F). G, H Flow cytometry analysis of iNOS and Arg1 expression in M1/M2-polarized WT and Gpr43^−/− BMDMs pretreated with the GPR43 agonist TUG-1375 (20 μM) for 6 h (n = 3). Representative iNOS (left) and Arg1 (right) flow cytometry histograms (G) and quantification of iNOS (left) and Arg1(right) MFI (H). Data are presented as means ± SD and analyzed using two-way ANOVA with Tukey’s multiple-comparisons test. Data were obtained from three independent experiments. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

We next investigated whether SCFAs themselves could modulate macrophage polarization. The treatment of BMDMs with SCFAs followed by M1/M2 stimulation reduced M1 polarization while enhancing M2 polarization (Supplementary Fig. S4B-C), suggesting a bidirectional regulatory role of SCFAs. However, given that SCFAs not only activate FFAR2 but also act on FFAR1 and HCAR2, function as histone deacetylase (HDAC) inhibitors (notably HDAC1, HDAC2, and HDAC3), and directly participate in cellular energy metabolism, their effects on macrophage polarization are likely mediated through multiple parallel mechanisms [43, 44]. To dissect the specific contribution of GPR43 from the pleiotropic effects of SCFAs, we employed the selective FFAR2 agonist TUG-1375 [45]. The treatment of WT BMDMs with TUG-1375 significantly suppressed the iNOS expression without affecting IL-10 or Arg1 (Fig. 4G, H). Consistently, ELISA assays revealed that TUG-1375 inhibited the production of IL-6 and TNF-α but did not alter IL-10 levels (Supplementary Fig. S4D). Importantly, the inhibitory effect of TUG-1375 on M1 polarization was completely abolished in Gpr43^−/− BMDMs, confirming that its action is strictly dependent on GPR43.

Taken together, these findings demonstrate that GPR43 functions as an intrinsic suppressor of M1 macrophage polarization. During septic inflammation, the loss of GPR43 disrupts this regulatory mechanism, leading to excessive macrophage activation upon LPS stimulation and the overproduction of inflammatory mediators, thereby exacerbating disease progression.

GPR43 deficiency promotes macrophage M1 polarization through glycolysis activation

To elucidate the molecular mechanisms by which GPR43 deficiency promotes M1 macrophage polarization, we performed transcriptomic profiling of WT and Gpr43^−/− BMDMs following LPS + IFN-γ stimulation. RNA sequencing (RNA-seq) identified 501 differentially expressed genes, including 72 upregulated and 429 downregulated genes in Gpr43^−/− macrophages compared with WT controls (p < 0.05, |log₂FC|> 1) (Supplementary Fig. S5A). The KEGG pathway enrichment analysis revealed that these DEGs were mainly associated with MAPK, PI3K–Akt, cytokine–cytokine receptor interaction, and glycolysis pathways (Supplementary Fig. S5B). Gene set enrichment analysis (GSEA) further confirmed glycolysis as one of the most significantly enriched metabolic pathways in Gpr43^−/− macrophages (Supplementary Fig. S5C). The heat map analysis illustrated robust upregulation of proinflammatory and glycolytic genes in Gpr43^−/− macrophages (Fig. 5A,B). On the basis of these transcriptomic findings, we next validated the expression of glycolysis-related genes by qPCR. Among them, the glycolytic enzyme ENO1 showed the most prominent increase in Gpr43^−/− BMDMs after LPS stimulation, accompanied by the altered expression of several other glycolytic enzymes, HK2 and LDHA (Fig. 5C). A western blot analysis further validated that ENO1 showed the strongest upregulation among glycolytic enzymes (Fig. 5D and Supplementary Fig. S5D). These results suggested that GPR43 deficiency promotes metabolic reprogramming, particularly glycolysis, to drive M1 polarization.

Fig. 5.

GPR43 deficiency promotes glycolysis-driven macrophage M1 polarization activation. A, B Heat map of proinflammatory cytokine gene (A) and glycolysis-related gene (B) expression in WT and Gpr43^−/− BMDMs ± LPS + IFN-γ. C qRT–PCR validation of glycolytic enzyme gene expression in WT and Gpr43^−/− BMDMs following M1 polarization (n = 3). D Representative immunoblots of glycolysis related proteins levels in M1-polarized WT and Gpr43^−/− BMDMs. The protein levels are shown in Supplementary Fig. S5D. E Lactate production in culture supernatants of M1-polarized WT and Gpr43^−/− BMDMs (n = 3). F, G Glucose uptake (2-NBDG) measured by flow cytometry in M1-polarized WT and Gpr43^−/− BMDMs (n = 3). Representative histograms (G) and quantification of 2-NBDG MFI (F). H–K Seahorse XF glycolysis stress test in M1-polarized WT and Gpr43^−/− BMDMs (n = 3). ECAR traces (H) with sequential injections: glucose (10 mM), rotenone/antimycin A (0.5 µM), and 2-deoxyglucose (50 mM). Quantification of glycolysis (I), glycolytic capacity (J), and glycolytic reserve (K) from ECAR data. L–O Seahorse XF mitochondrial stress test in M1-polarized WT and Gpr43^−/− BMDMs (n = 3). OCR traces (L) with sequential injections: oligomycin (1.5 µM), FCCP (1 µM), and rotenone/antimycin A (0.5 µM each). Quantification of basal respiration (M), maximal respiration (N), and spare respiratory capacity (O) from OCR data. Data in C, E, F, I–K and M–O were obtained from three independent experiments are presented as means ± SD. Statistical analysis by two-tailed unpaired Student’s t-test or one-way ANOVA with Tukey’s post hoc test, as appropriate. The data in A, B were obtained from RNA-seq analysis. ns, not significant; *p < 0.05, ****p < 0.0001

To functionally assess glycolytic activity, we measured lactate production and glucose uptake using the fluorescent analog 2-NBDG. Both assays indicated enhanced glycolysis in Gpr43^−/− BMDMs (Fig. 5E–G). Real-time extracellular flux analysis using the Seahorse XF system demonstrated that, while baseline extracellular acidification rate (ECAR) was comparable between groups, Gpr43^−/− BMDMs exhibited significantly higher ECAR, glycolytic capacity, and glycolytic reserve following LPS stimulation compared with WT controls (Fig. 5H–K). By contrast, oxygen consumption rate (OCR), including basal respiration, maximal respiration, and spare respiratory capacity, showed no significant differences between groups under either basal or LPS-stimulated conditions (Fig. 5L–O). Transmission electron microscopy further confirmed that mitochondrial morphology was not altered between groups (Supplementary Fig. S5E). Together, these data strongly demonstrate that GPR43 deficiency results in a marked increase in glycolytic flux in macrophages upon LPS activation, while mitochondrial metabolism remains largely unaffected. Given that M1 macrophage polarization is known to be driven by glycolytic metabolism [46], we further tested whether glycolysis is causally involved in the enhanced polarization phenotype. The treatment with the glycolysis inhibitor 2-deoxyglucose (2-DG) markedly suppressed the augmented M1 activation observed in Gpr43^−/− macrophages (Supplementary Fig. S5F).

In conclusion, our results provide compelling evidence that GPR43 deficiency promotes M1 macrophage polarization through glycolytic reprogramming. The hyperactivation of glycolysis is intrinsically linked with excessive activation of M1 signaling pathways, thereby driving macrophages toward an exaggerated proinflammatory phenotype. These findings highlight GPR43 as a critical regulator of macrophage immune responses during sepsis, through modulation of glycolytic metabolism.

GPR43 regulates ENO1 expression via HIF-1α signaling

In our earlier experiments, we found that glycolytic reprogramming in Gpr43^−/−macrophages was characterized by a pronounced upregulation of ENO1. We hypothesized that ENO1 is not simply a marker of enhanced glycolysis but a key driver of metabolic reprogramming. To test this, we performed functional rescue experiments using the ENO1-specific inhibitor AP-III-a4. The pretreatment with AP-III-a4 markedly reversed the excessive M1 polarization observed in Gpr43^−/− macrophages upon LPS stimulation, as evidenced by the reduced iNOS expression and decreased production of TNF-α and IL-6, restoring them to levels comparable to WT macrophages (Fig. 6A–D and S6A). This effect was accompanied by a significant suppression of glycolytic activity (Fig. 6E). In vivo, ENO1 inhibition improved the survival of septic Gpr43^−/− mice (Fig. 6F). Histological analyses of lung tissues further revealed that AP-III-a4 treatment attenuated M1 polarization, reduced tissue damage, and improved outcomes (Fig. 6G–I). These findings establish ENO1 as a pivotal mediator linking GPR43 deficiency to macrophage polarization and disease severity.

Fig. 6.

GPR43 deficiency promotes macrophage M1 polarization via the HIF-1α–ENO1 axis. A Representative immunoblots of iNOS and ENO1 in WT and Gpr43^−/− BMDMs ± AP-III-a4 (ENO1 inhibitor, 20 µM) for 6 h. The protein levels are shown in Supplementary Fig. S6A. B, C Flow cytometry analysis of iNOS expression in M1-polarized WT and Gpr43^−/− BMDMs ± AP-III-a4 (n = 3). Representative flow cytometry histogram (B) and the iNOS MFI quantification (C). D Lactate production in M1-polarized WT and Gpr43^−/− BMDMs ± AP-III-a4 (n = 3). E Supernatant cytokines measured by ELISA in M1-polarized WT and Gpr43^−/− BMDMs ± AP-III-a4 (n = 3). F Kaplan–Meier survival curves for WT and Gpr43^−/− mice after CLP ± AP-III-a4 (10 mg/kg) (n = 10). G–I Lung histopathology and immunofluorescence after CLP in WT and Gpr43^−/− mice ± AP-III-a4 (n = 6). Representative H&E and immunofluorescence images (G, iNOS, green; F4/80, red; nuclei, DAPI, blue). Lung injury scores (H) and quantification of iNOS mean fluorescence intensity (I). J Representative immunoblots of HIF-1α, c-Myc, p-AKT and p-mTOR expression in M1-polarized WT and Gpr43^−/− BMDMs. The protein levels are shown in Supplementary Fig. S6B. K, L Flow cytometry analysis of iNOS expression in M1-polarized WT and Gpr43^−/− BMDMs ± the HIF-1α inhibitor acriflavine (5 μM) for 6 h (n = 3). Representative histogram of iNOS expression (K) and quantification of the iNOS MFI (L). M Supernatant cytokines measured by ELISA in M1-polarized WT and Gpr43^−/− BMDMs ± acriflavine (n = 3). Data are presented as means ± SD. Two-group comparisons used two-tailed unpaired Student’s t-test; multiple groups used one-way ANOVA with Tukey’s post hoc test; survival by log-rank test. Data in C–E and L–M were obtained from three independent experiments. Data in H, I are representative of three independent experiments. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

To determine why ENO1 expression is elevated in the absence of GPR43, we examined upstream regulatory pathways, including HIF-1, c-Myc, and AKT/mTOR [47]. Notably, only the HIF-1α was significantly activated in Gpr43^−/− macrophages, whereas c-Myc and AKT/mTOR signaling remained unchanged (Fig. 6J and Supplementary Fig. S6B), implicating HIF-1α as a key upstream regulator of ENO1. Consistent with this, inhibition of HIF-1α using acriflavine effectively reversed the proinflammatory phenotype of Gpr43^−/− macrophages under LPS stimulation, as shown by decreased iNOS expression (Fig. 6K, L) and reduced production of TNF-α, IL-1β, and IL-6 (Fig. 6M). Together, these results support a model in which GPR43 deficiency promotes glycolytic reprogramming and M1 polarization through HIF-1α–driven ENO1 upregulation.

Furthermore, the activation of GPR43 with the selective agonist TUG-1375 in WT BMDMs significantly downregulated HIF-1α and ENO1 expression (Supplementary Fig. S6C,D), reduced glycolytic flux (Supplementary Fig. S6E), and suppressed M1 polarization and proinflammatory cytokine production (Supplementary Fig. S4D). Importantly, TUG-1375 administration markedly improved survival and mitigated lung injury in septic mice (Supplementary Fig. S6F–H) as well as reduced iNOS expression and systemic cytokine levels (Supplementary Fig. S6I, J). However, these protective effects were completely abolished in Gpr43^−/− mice, confirming receptor specificity.

In conclusion, our findings demonstrate that aberrant macrophage activation in the absence of GPR43 is driven by the HIF-1α–ENO1–glycolysis axis. Targeting this pathway—either by inhibiting ENO1 or HIF-1α or by pharmacologically activating GPR43—effectively reprograms macrophage metabolism, alleviates inflammatory responses, and improves sepsis outcomes. These results identify the GPR43–HIF-1α–ENO1 axis as a critical regulatory pathway in macrophage immune function and a promising therapeutic target in sepsis.

Downregulation of GPR43 is associated with metabolic reprogramming and inflammation in patients with sepsis

To assess whether the GPR43–ENO1–glycolysis axis identified in mice is relevant in human sepsis, we conducted a preliminary clinical study. Peripheral blood mononuclear cells (PBMCs) were collected from 36 patients with sepsis (clinical characteristics in Supplementary Table S3). Quantitative real-time PCR revealed a marked reduction of GPR43 mRNA expression in septic PBMCs compared with controls, which was further validated at the protein level by western blotting (Fig. 7A, B).

Fig. 7.

GPR43 expression in patients with sepsis and its correlation with clinical and metabolic markers. A Representative immunoblots (left) and quantification (right) of GPR43 expression in PBMCs from healthy controls and sepsis patients (n = 36 for sepsis, n = 15 for healthy controls). B qRT–PCR quantification showing GPR43 mRNA expression in PBMCs from sepsis patients compared with healthy controls. C–G Correlation analysis between GPR43 mRNA expression and clinical/metabolic parameters in septic patients, including serum LBP levels (C), blood lactate concentrations (D), ENO1 mRNA expression (E), CRP levels (F), and IL-6 mRNA expression (G). Data are presented as mean ± SD. Statistical significance in A, B was determined by unpaired Student’s t-test, and correlations in C–G were assessed using Spearman’s correlation test, ** p < 0.01

Correlation analyses showed that GPR43 expression was inversely associated with lipopolysaccharide-binding protein (LBP), a marker of intestinal barrier injury [48] (Fig. 7C), thereby linking GPR43 downregulation to barrier dysfunction and bacterial translocation, consistent with our murine findings. Moreover, GPR43 expression showed significant negative correlations with lactate concentration and ENO1 expression (Fig. 7D,E). Since lactate is the terminal product of glycolysis, these results suggest that reduced GPR43 expression may drive ENO1 upregulation and glycolytic overactivation, thereby worsening tissue hypoperfusion and metabolic imbalance. Additionally, GPR43 expression was negatively correlated with systemic inflammatory markers, including C-reactive protein (CRP) and IL-6 (Fig. 7F,G), further supporting its role in limiting excessive inflammation.

Collectively, these clinical observations mirror our experimental findings, demonstrating that reduced GPR43 expression in sepsis is closely associated with intestinal barrier disruption, glycolytic dysregulation, and systemic inflammatory amplification. Together, they highlight the GPR43–ENO1 axis as a critical regulator of immune–metabolic balance in sepsis.

Discussion

Despite significant advances in antimicrobial therapy and supportive care, sepsis remains one of the leading causes of death worldwide, with an estimated 49 million cases and 11 million deaths reported annually [49]. Effective immunomodulatory strategies to restore immune balance have proven elusive, largely owing to the complex interplay between initial, excessive inflammation and subsequent immunosuppression. Targeting macrophage polarization and metabolic reprogramming has emerged as a promising approach to address this challenge. Macrophage polarization is increasingly recognized as a central determinant of immune dysregulation in sepsis. Proinflammatory M1 macrophages produce high levels of TNF-α, IL-6, and IL-1β, which amplify cytokine storms, thereby propagating vascular injury and driving multiorgan failure. By contrast, anti-inflammatory M2 macrophages promote the resolution of inflammation and tissue repair. Growing evidence indicates that these functional states are tightly coupled to distinct metabolic programs. M1 macrophages rely predominantly on aerobic glycolysis to meet their energetic and biosynthetic demands for rapid inflammatory responses, while M2 macrophages depend on oxidative phosphorylation and fatty acid oxidation to sustain anti-inflammatory and reparative functions [50–52]. Thus, glycolytic rewiring is not merely a metabolic adaptation but a key driver of M1 polarization and inflammatory amplification. In the setting of sepsis, this glycolysis–M1 axis is particularly detrimental, as sustained M1 activation triggers cytokine storms, tissue damage, impaired microcirculation, and, ultimately, poor outcomes.

Our findings collectively establish GPR43 as a critical checkpoint that restrains glycolysis-driven M1 activation, thereby preventing excessive cytokine release and tissue injury. A clinical data analysis revealed that GPR43 expression in patients with sepsis was negatively correlated with systemic inflammation and lactate levels, and higher GPR43 expression was associated with better prognosis [28]. Mechanistically, GPR43 deficiency relieves the suppression of glycolytic pathways by activating HIF-1α and upregulating the key glycolytic enzyme ENO1. This metabolic reprogramming skews macrophages toward a sustained M1 phenotype, resulting in excessive cytokine secretion and subsequent tissue damage. We highlighted the pathogenic relevance of this mechanism in the murine CLP sepsis model. Conversely, pharmacological activation of GPR43 with the selective agonist TUG-1375 effectively inhibited this glycolytic program, alleviated systemic inflammation, and improved survival in septic mice, underscoring the therapeutic potential of receptor-specific targeting. A key advantage of this approach lies in its specificity: unlike natural SCFAs, which not only activate GPR43 but also exert pleiotropic effects through histone deacetylase inhibition, selective agonists such as TUG-1375 provide a more precise means of restoring immune balance. On the basis of these results, we propose that the GPR43–HIF-1α–ENO1 axis represents a novel metabolic checkpoint that intrinsically limits pathological macrophage activation. This positions GPR43 as an endogenous “brake” against maladaptive metabolic reprogramming in sepsis, offering a promising strategy to mitigate pathological inflammation and improve outcomes. Moreover, the observation that higher GPR43 expression in patients with sepsis correlates with improved prognosis further supports its therapeutic potential.

In addition, the modulation of the gut microbiota represents another promising therapeutic opportunity. We observed that FMT partially restored gut barrier function and improved survival in Gpr43^−/− mice, underscoring the therapeutic relevance of the microbiota–SCFA–GPR43 axis. This finding aligns with the growing clinical interest in FMT, which has demonstrated efficacy in recurrent Clostridioides difficile infection and is currently being explored in a range of inflammatory and metabolic disorders [53, 54]. Similarly, probiotic supplementation aimed at restoring SCFA-producing taxa may provide a feasible strategy to enhance gut barrier integrity and immune homeostasis in septic patients. Although still experimental, combining microbiota-based interventions with pharmacological activation of GPR43 could yield synergistic benefits.

Nonetheless, our study has several limitations that warrant acknowledgment. First, our patient cohort was modest in size, and only PBMCs were analyzed. This leaves the tissue-specific functions of GPR43 unexplored and calls for larger cohorts and analysis of tissue biopsies in the future to confirm our findings. Second, our in vivo work was largely restricted to the CLP model. Thus, whether similar mechanisms operate in other clinically relevant models of infection, such as pneumonia, remains uncertain. Third, while FMT provided partial functional rescue, our analysis was limited to whole-community transplantation and did not identify the specific bacterial taxa responsible for the protective effects. This highlights a need for future work to precisely delineate the key microbial players.

In conclusion, our study provides a comprehensive understanding of GPR43 as a dual regulator of host defense in sepsis. We reveal that GPR43 acts through two distinct yet complementary pathways: it maintains gut barrier integrity via microbiota-dependent mechanisms while also intrinsically suppressing glycolysis-driven macrophage polarization through the novel HIF-1α–ENO1 axis. This dual action effectively restrains systemic inflammation, prevents excessive tissue damage, and significantly improves survival in septic models. These findings not only provide mechanistic insight into the immunometabolic role of GPR43 but also reconcile its previously paradoxical roles in different inflammatory diseases. Our work establishes GPR43 as a protective immunometabolic checkpoint, resolving the ambiguity of its function.

Conclusions

GPR43 protects against sepsis by preserving gut barrier integrity through microbiota-dependent mechanisms and by limiting glycolysis-driven M1 macrophage polarization via the HIF-1α–ENO1 axis.

Abbreviations

2-DG

2-deoxyglucose

ANOVA

One-way analysis of variance

Arg1

Arginase 1

BMDM

Bone marrow–derived macrophage

CLP

Cecal ligation and puncture

CRP

C-reactive protein

ECAR

Extracellular acidification rate

ELISA

Enzyme-linked immunosorbent assay

ENO1

Enolase 1

FMT

Fecal microbiota transplantation

GPR43/FFAR2

G protein–coupled receptor 43 / Free fatty acid receptor 2

GSEA

Gene set enrichment analysis

H&E

Hematoxylin and eosin

HIF-1α

Hypoxia-inducible factor 1 alpha

IFN-γ

Interferon gamma

IL

Interleukin

iNOS

Inducible nitric oxide synthase

KEGG

Kyoto Encyclopedia of Genes and Genomes

LBP

Lipopolysaccharide-binding protein

LPS

Lipopolysaccharide

MFI

Mean fluorescence intensity

MHCII

Major histocompatibility complex class II

MUC2

Mucin 2

OCR

Oxygen consumption rate

PBMC

Peripheral blood mononuclear cell

PERMANOVA

Permutational multivariate analysis of variance

PLF

Peritoneal lavage fluid

qRT–PCR

Quantitative (reverse transcription) polymerase chain reaction

SCFA

Short-chain fatty acid

SD

Standard deviation

SOFA

Sequential Organ Failure Assessment

TUNEL

Terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling

WT

Wild type

Author contributions

Jiabin Li, Ting Wu, Hongru Li, and Cui Wang were responsible for the project design and reagent purchase. Mingyang Tang wrote the manuscript. Fei Tang, Bao Meng, Lifen Hu, Yanyan Liu, Jingjing Pan, and Ying Xu were responsible for data analysis and statistics. Mingyang Tang, Yuanlong Shu, Qingyue Zhang, Yuexin Xu, and Chengcheng Li performed the experiments. All authors contributed to the article and agreed to the submitted version.

Funding

This study was supported by the National Natural Science Foundation of China (grant nos. U24A20643, 82270015, 82400132, 82470108, 82100017, 82302577, 82304209, 82370016), Anhui Province Clinical Medical Research Transformation Special Project (grant nos. 202304295107020032, 202304295107020043), Top Talents Academic Funding Key Programs in Universities (grant no. gxbjZD08), Anhui Medical University Natural Science Research Project (grant nos. 2023AH010083, 2024AH010114, 2024AH050826, 2024AH050829), Anhui Province Postdoctoral Research Funding Project (grant nos. 2023B700, 2024C875), Anhui Provincial Natural Science Foundation (grant nos. 2208085MH264, 2308085QH284, 2308085MH243), Basic and Clinical Cooperative Research Program of Anhui Medical University (grant no. 2021xkjT021), Anhui Province Scientific Research Planning Project (grant nos. 2023AH01008, 2023AH053282), China Primary Health Care Foundation (grant no. MTP2022A015), Anhui Provincial Health Research Program (grant no. AHWJ2022b076), and the Youth Talent Support Program of the Anhui Province Association for Science and Technology (grant no. RCTJ202425).

Data availability

All raw experimental data generated or analyzed during this study are available from the corresponding author upon reasonable request. The RNA sequencing datasets generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under the BioProject accession number PRJNA1263183. The metabolomics datasets generated in this study have been deposited in the OMIX repository under the accession number OMIX011981.

Declarations

Ethics approval and consent to participate

The animal experiments in this study were reviewed and approved by the Animal Ethics Committee of Anhui Medical University (approval no. LLSC20240063, approved on 15 March 2024). This committee adheres to the principles and guidelines set forth by the International Council for Laboratory Animal Science (ICLAS), ensuring that our animal research meets international ethical standards. All experiments involving human samples and clinical data were reviewed and approved by the Clinical Ethics Committee of the First Affiliated Hospital of Anhui Medical University (approval no. PJ2024-10-50, approved on 10 October 2024). The committee conducted its review in strict accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.