AE was procured from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). TIE was sourced from Merck Sharp & Dohme (China). Additionally, Reverse Transcription Green qPCR SuperMix and bicinchoninic acid kits for protein concentration determination were procured from TransGen (Beijing, China). Enzyme-linked immunosorbent assay (ELISA) kits from R&D Systems (Minneapolis, MN, USA) were also used. Table 1 comprehensively details the antibodies used in this study, along with information on their providers.

Pathogen-free C57BL/6 mice, aged 8–10 weeks, were acquired from Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China), equally divided by sex, and housed at the Experimental Animal Center of Fujian Normal University. Mice were maintained under a controlled environment with temperature set between 23°C and 25°C, humidity levels between 40% and 60%, and a 12 h light/dark cycle. Male mice weighing 20–22 g and female mice weighing 18–20 g were provided ad libitum access to food and water, and were randomly assigned to various experimental groups. A total of 160 mice were used in this study, with an equal distribution of females (n = 80) and males (n = 80). These mice were randomly assigned to 8 groups, each consisting of 20 mice: sham, sham + AE, sham + TIE, sham + AE + TIE, CLP (sepsis model), CLP + AE, CLP + TIE, and CLP + AE + TIE. The treatments with AE (10 mg/kg/day), TIE (50 mg/kg/day), and the combination of AE + TIE were administered 1 h after sepsis induction.

The CLP procedure used to induce sepsis followed a well-documented protocol (Alverdy et al., 2020; Su et al., 2023b). Mice were anesthetized via intraperitoneal injection of 0.3% pentobarbital sodium. Subsequently, mice were placed on the surgical platform, and the lower left quadrant of their abdomen was shaved using an electric razor. The area was initially sterilized using a betadine solution, followed by 75% alcohol. An incision was made along the length of the specific peritoneal region, followed by a 1 cm cut through the outer layers to expose the cecum. Approximately two thirds of the cecum were tightly ligated near its base using a sterile 4.0 surgical thread. A small hole was then made in this area using a clean 5 mL syringe needle, allowing a small amount of fecal matter to extrude. After ligation and perforation, the cecum was gently placed back into the peritoneal cavity, and the incisions in the muscle and skin were sutured using a sterile 4.0 suture. To replenish fluids, a subcutaneous injection of 1 mL of pre-warmed 0.9% saline solution was administered on the back. Following surgery, mice were placed on a heated mat to ensure full recovery from anesthesia. In CLP model, the locus of sepsis is centered on the peritoneal cavity, specifically targeting the cecum as the primary site of infection. This site allows for controlled release of fecal matter into the peritoneal space, mimicking the polymicrobial nature of human intra-abdominal sepsis. The sham control group underwent only a laparotomy, without cecal ligation or puncture.

Mice were allocated into five experimental groups: sham (control), CLP (sepsis model), AE treatment (10 mg/kg/day), TIE treatment (50 mg/kg/day), and a combination of AE and TIE treatments. Treatments with AE and TIE commenced 1 h post-sepsis induction. Survival rates were observed at 6-h intervals post-administration. The severity of sepsis was assessed using the murine sepsis score (MSS) (Shrum et al., 2014). This scoring system assesses factors such as body temperature, activity levels, appearance, and breathing patterns. Each parameter is scored on a scale from 0 to 4, with higher scores indicating more severe symptoms of sepsis. The overall MSS is the sum of the individual scores, providing a quantitative measure of sepsis progression in the animals. The mice were monitored for signs consistent with systemic inflammatory response syndrome (SIRS), including elevated heart rate, respiratory rate, temperature changes, and leukocytosis, following the established clinical criteria adapted for animal models (Serafim et al., 2018). Following anesthesia with sodium pentobarbital, the systematic collection of blood, fecal matter, ascitic fluid, and tissue samples was performed according to established protocols (Wang P. et al., 2021).

Total RNA was isolated using TRIzol reagent (Takara, Tokyo, Japan), following the RT-qPCR methodologies outlined in a previous study (Wang Y. et al., 2021). mRNA levels were quantified utilizing specific primers detailed in Table 2 and were normalized against GAPDH as a reference gene. The expression changes relative to the expression in the phosphate-buffered saline (PBS) control were calculated employing the ΔΔCt method. This analysis was repeated thrice for each gene to confirm the precision and consistency of the data.

The samples were obtained and analyzed as described by Prakash et al. (2024). The concentrations of cytokines, including IFN-β (424001, Thermo Fisher Scientific), interleukin (IL)-1β (SMLB00C, R&D Systems), IL-6 (SM6000B, R&D Systems), and tumor necrosis factor (TNF)-α (SMTA00B, R&D Systems), were quantified using ELISA kits according to the manufacturer's instructions. Briefly, serum samples were collected from the mice at designated time points and diluted as necessary. Each sample was added to the pre-coated wells, followed by the addition of detection antibodies and subsequent color development. The absorbance was measured using a microplate reader, and cytokine concentrations were calculated based on a standard curve (Prakash et al., 2024).

Tissue samples, including lung, liver, heart, spleen, and kidney, were collected for histological analysis. The samples were rinsed with sterile saline and fixed in 4% paraformaldehyde for 24 h. These specimens were then processed following the hematoxylin and eosin (H&E) staining protocol (Su et al., 2024b). After undergoing dehydration, the tissues were sectioned into 4-μm slices and mounted on slides. These slides were then heated to 60°C for deparaffinization, rehydrated, and stained with H&E (Sigma-Aldrich, Burlington, MA, USA). Subsequently, excess dye was removed using ethanol and xylene, and coverslips were applied. For tissue, a semi-quantitative scoring system was used to assess inflammation, edema, hemorrhage, and alveolar septal thickening (Palumbo et al., 2017). Each parameter was scored on a 0- to 4-point scale, based on the severity of the injury observed in H&E-stained tissue sections. For inflammation, inflammatory cells were counted per field, and the other parameters were categorized based on the percentage of alveoli involved or septal thickness. The scores for each parameter were averaged to provide an overall injury score for each group.

Flow cytometry was conducted in line with the established guidelines (Su et al., 2024a), utilizing the antibodies specified in Table 1. This included the process of erythrocyte lysis, followed by the resuspension of cells in 2% HyClone PBS solution for counting. Cell surface markers were labeled through a 30-min incubation with targeted antibodies at 4°C. Peripheral blood mononuclear cells were identified using Fixable Viability Stain 780 (FVS780) and the viability of cells was assessed. The analysis was executed using a FACSymphony™ A5 cytometer (BD Biosciences, CA, USA) and the data were analyzed using FlowJo software (v10.5.3; Ashland, OR, USA).

To evaluate the impact of AE and TIE on the peritoneal microbiome in septic mice, animals were anesthetized and then euthanized by cervical dislocation. The procedural area was disinfected using 75% alcohol. In a sterile setting, 1 mL of saline was infused into the peritoneum, followed by gentle peritoneal massage to enhance fluid recovery. Ascitic fluid was collected in Eppendorf centrifuge tubes for further analysis.

Serial dilution was performed on the ascitics fluid samples, with 100 μL from each dilution plated on agar and incubated at 37°C for 24 h. Finally, colony counting was performed based on the following formula:

The number of colonies counted served as an indication of bacterial load.

Western blotting procedures were conducted according to established methods (Su et al., 2024c), which included steps such as protein extraction, electrophoresis, transfer to membranes, incubation with primary and secondary antibodies, washing, and chemiluminescent detection. Protein bands were quantitatively analyzed and their expression levels were presented as fold changes relative to control samples. The specific antibodies utilized are detailed in Table 1.

16S rRNA in ascitic fluid samples was sequenced and analyzed at Beijing BioMarker Bioinformatics Technology Co., Ltd. (Beijing, China). DNA was extracted, and 16S rRNA sequencing was conducted according to established procedures outlined in a previous study (Su et al., 2024e). The Magnetic Soil and Stool DNA Kit (Cat. 4992738, Tiangen Biotech Co., Ltd.) was utilized for DNA extraction following the manufacturer's instructions. The complete 16S rRNA gene was amplified using the general primers 27F (5′-AGRGTTTGATYNTGGCTCAG-3′) and 1492R (5′-TASGGHTACCTTGTTASGACTT-3′), with the addition of PacBio barcode sequences unique to each sample. Library construction, sequencing, and data analysis were carried out using the PacBio Sequel II platform (Beijing BioMarker Bioinformatics Technology Co., Ltd.). The bioinformatics analysis was conducted using BMK Cloud (https://www.biocloud.net) following a published protocol. PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) was employed to predict the functional traits of the microbiome by using a phylogenetic approach (Su et al., 2024c).

Beijing BioMarker Bioinformatics Technology Co., Ltd. conducted RNA sequencing (RNA-seq) on the Illumina NovaSeq 6000 platform (Illumina, Inc., San Diego, CA, USA). The procedure followed an established protocol for comprehensive genomic sequencing analysis (Zhang et al., 2021).

Photoshop and Illustrator 2020 (Adobe, San Jose, CA, USA) along with ImageJ v1.8.0 (NIH, MD, USA) were used for preparing images. According to the guidelines provided on the GraphPad website (https://www.graphpad.com/guides/prism/latest/user-guide/citing_graphpad_prism.htm), GraphPad Prism (v8.0) was used to conduct statistical analyses using t-tests, one- and two-way ANOVA, and the Mantel–Cox (log-rank) test. Data are reported as mean ± SD, with statistical significance defined at P < 0.05.

The therapeutic potential of the combination of AE (10 mg/kg/day) and TIE (50 mg/kg/day) was evaluated using a CLP-induced sepsis mouse model (Figure 1A). Control groups received an equivalent volume of vehicle (corn oil) to match the treatment groups. The dosages were selected based on prior studies that demonstrated significant therapeutic effects at these levels. The rationale behind the choice of these specific concentrations and the administration route has been detailed in our previous publications (Su et al., 2023a, 2024d). Compared to the control group solely treated with CLP, the administration of AE, TIE, or both in significantly improved the health of the mice. The AE + TIE group exhibited notably enhanced outcomes, including active feeding, improved fur sheen, and no fur clumping, illustrating the most substantial benefits (Figure 1B). Post-treatment, there was a significant decrease in the MSS, with the AE + TIE group showing a remarkable reduction compared to the CLP control group (P < 0.0001; Figure 1C), underscoring the enhanced therapeutic impact of the combination therapy. In contrast, neither AE nor TIE alone significantly altered these health indicators in the sham-operated group (P > 0.05; Figure 1D). Furthermore, the survival rate in the AE + TIE group was significantly higher than in the CLP control group, with a survival rate of 75% vs. 10% (P < 0.0001; Figure 1E), highlighting the critical effectiveness of AE and TIE in managing CLP-induced sepsis in mice.

The regulatory effect of AE and TIE on inflammatory cytokine production was assessed using a mouse model of CLP-induced sepsis. This investigation concentrated on how AE and TIE influence cytokine concentrations in various organs such as the serum, lungs, liver, and kidneys. The administration of AE and TIE did not yield statistically significant effects on these parameters in the sham group (P > 0.05, Supplementary Figure 1). However, the combined treatment notably decreased the heightened serum concentrations of IL-1β, IL-6, and TNF-β, demonstrating significant reductions (P < 0.01, Figures 2A–C). Cytokine levels also consistently declined in the tissues of the lungs, liver, and kidneys, validating the broad and effective nature of the treatment (Figures 2D–L).

Subsequent RT-qPCR analysis after CLP induction showed a marked decrease in IL-6, IL-1β, and TNF-α expression in the lung, liver, and kidney tissues of the AE + TIE group, markedly exceeding the results seen in the CLP-only group (P < 0.05, Figure 3). These results strongly highlight the potent anti-inflammatory effects of the AE and TIE combination, emphasizing its potential as an effective therapeutic approach for combating CLP-induced sepsis in mice.

Following the administration of the treatment, tissue samples were analyzed using H&E staining at 12 h post-treatment to assess injury severity. The combination therapy of AE and TIE notably diminished lung injury, as evidenced by the reduced alveolar wall thickening and fewer occurrences of hemorrhages (Figures 4A–E). In the liver, the AE + TIE group displayed significant injury reduction, characterized by diminished inflammatory cell infiltration, decreased swelling, and better hepatocyte organization, compared to the CLP control group (Figures 4F–J). Pathological evaluations further confirmed the significant effectiveness of AE and TIE combination treatment in reducing tissue damage across critical organ systems (P < 0.01; Figures 4K, L). Altogether, these findings underscore the considerable potential of AE and TIE combination therapy in alleviating the adverse effects of CLP-induced sepsis on respiratory, hepatic, and cardiovascular functions.

Using flow cytometry to evaluate immune responses 12 h post-treatment, we observed a notable decrease in B cell activation in the AE + TIE group compared to the TIE-only cohort (P < 0.05, Figures 5A, E). Additionally, CD8⁺ T cell activation was significantly increased after treatment with combined AE + TIE (P < 0.05, Figures 5B, F). Although treatments with AE and TIE alone did not significantly alter the NK cell count relative to that in the CLP model, AE + TIE treatment significantly decreased it (P < 0.05, Figures 5C, G). Notably, TNFR2⁺ T_reg cell counts were considerably elevated in the combination therapy group, suggesting an important immunomodulatory role (P < 0.05, Figures 5D, H). Despite these notable changes, no significant differences were observed in the CD4⁺ T cell and activated NK cell counts under septic conditions (P > 0.05, Supplementary Figure 2). These results highlight the selective effects of AE and TIE treatments on specific immune cell activations, demonstrating their targeted modulation of immune responses in sepsis-induced mice.

Our evaluation focused on the capability of AE and TIE to diminish bacterial colonies. According to Figure 6, although all groups showed some level of bacterial presence, the CLP group presented a notably higher colony count (P < 0.0001). In contrast, treatments with AE, TIE, and especially the combined AE + TIE therapy exhibited substantial reductions in bacterial growth, with TIE alone demonstrating the most significant decrease (P < 0.01).

Employing advanced 16S RNA sequencing, our study assessed changes in the microbial landscape post-treatment. Analysis using a Venn diagram and subsequent alpha- and beta-diversity assessments indicated unique microbial community structures among the groups (Figure 7A). To investigate the impact of AE + TIE on intestinal flora diversity and abundance in septic mice, we analyzed the alpha diversity at the operational taxonomic unit (OTU) level. This included the Ace, Chao1, Shannon, and Simpson indices, where Ace and Chao1 reflect species richness, and Shannon and Simpson indicate species diversity. The results showed that both the Ace and Chao1 indices significantly increased in the CLP group compared with those in the blank group (P < 0.01), with similar trends for the Shannon and Simpson indices (P < 0.01). In the CLP + AE + TIE group, the Ace and Chao1 indices further increased compared with those in the CLP group (P < 0.01) (Figures 7B–F). These findings suggest that CLP enhances intestinal flora diversity and abundance, and the combination of 10 mg/kg AE with 50 mg/kg TIE effectively boosts these metrics in CLP septic mice. At the genus level, Acinetobacter, Mucispirillum, Helicobacter, Escherichia, and Bacteroides were predominant. Acinetobacter levels were lower in the CLP model group compared with those in the Sham group, while Bacteroides levels were higher. However, compared with the CLP model groups, Acinetobacter levels were higher and Bacteroides levels were lower in the CLP +AE +TIE groups. Therefore, the AE and TIE treatments significantly modified the levels of pathogenic and beneficial bacteria, enhancing microbial homeostasis in the peritoneal cavity (Figures 7G–J).

Furthermore, using linear discriminant analysis effect size and predictive functional profiling, we elucidated the broad-spectrum impacts of these treatments on microbial gene function (Supplementary Figure 3). Using reconstruction of unobserved states analysis, the impact of combined AE + TIE therapy was also assessed on gut microbiome function. These analyses highlighted pathways involved in environmental adaptation, immune diseases, and signal transduction (Supplementary Figure 4), offering insights into the complex mechanisms through which AE and TIE therapies recalibrate microbial dynamics and community functions, thereby enhancing their therapeutic efficacy against sepsis-induced dysbiosis.

This comprehensive evaluation confirms the multifaceted role of AE and TIE therapy in restoring microbial balance, curtailing pathogenic bacterial overgrowth, and fostering beneficial bacterial proliferation. This positions the combined therapy as a potent therapeutic strategy against sepsis-induced microbial dysregulation.

Comprehensive RNA-seq of the entire genome elucidated the molecular pathways responsible for the protective effects of combined AE + TIE treatments in sepsis. Heatmaps in Figures 8A, B illustrate changes in gene expression, allowing the identification of differentially expressed genes. The combination therapy treatments attenuated the expression of key genes across multiple pathways, including the NF-κB (IL-1β and Nfkb1), Toll-like receptor (Mapk1 and IFNAR1), PI3K-Akt (Aft4 and Syk), and JAK-STAT (Fhl1 and Jak2) pathways (Figures 8A, B). Enrichment analyses of differentially expressed genes through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes highlighted the significant impact of AE + TIE treatment. CLP significantly activated the TNF, NF-κB, Toll-like receptor, PI3K-Akt, JAK-STAT, and TNF signaling pathways. In contrast, AE + TIE treatment effectively reduced inflammatory responses (Figures 8C–E). The volcano plot and Venn diagram revealed the induction of Retnig, Isg20, Tfdp2, Nptn, and Tpm4 by CLP and the suppression of Saa2, Lrf1, Coro1a, Upp1, and Emp3 by AE+TIE treatment (Figure 8F). STRING analysis displayed the functional interaction between TRP53, NF-κB, MAPK, JAK2, IFNAR1, IRF3, TNF, and CDK4 proteins (Figure 8G).

Moreover, western blot analysis of lung tissue was performed to evaluate the impact of AE + TIE therapy on sepsis induced by CLP, indicated notable inflammation. The combination treatment significantly decreased NF-κB activation, as evidenced by the reduced NF-κB (p65) and IKBα phosphorylation relative to that under CLP conditions (P < 0.001, Figures 8H–J). Overall, the combination treatment mitigates inflammatory responses triggered by CLP, thus having potential as a novel therapeutic approach for managing sepsis.