LP18 was cultured in MRS (De Man, Rogosa, and Sharpe Broth) medium and incubated at 37 °C for 12 h. The LP18 strain was isolated via centrifugation at 4,000 rpm for 5 min at 4 °C, yielding a bacterial suspension of 1 × 10⁸ CFU/mL prepared with phosphate-buffered saline (PBS).

Twenty-four male C57BL/6J rodents, 6 weeks old and weighing 16–18 g, were obtained from GemPharmatech Co., Ltd. The animal study was approved by the Animal Ethics Committee of The First Affiliated Hospital of Zhengzhou University (Approved No. ZZU-LAC2025032509). During the feeding phase, standard illumination conditions (12-h light-dark cycle), temperature (22 ± 2 °C), and sufficient food and water were provided daily. After a 7-day adaptation period, the rodents were arbitrarily allocated into three groups (n = 8): the control group (CON), the SAP group (SAP), and the LP18 group (LP18). The CON group and SAP group were administered 200 μL of PBS daily via gavage. The LP18 group was administered an intragastric dose of 200 μL of LP18 at a concentration of 1 × 10⁸ CFU/mL daily. Starting from day 28, mice in the SAP and LP18 groups were granted unrestricted access to injections of caerulein (Cae) and lipopolysaccharide (LPS). The SAP model was developed through intraperitoneal administration of Cae (300 μg/kg, NJPeptide, China, Cat. Pep03263) and LPS (10 μg/kg, MCE, USA, Cat. HY-D1056) in accordance with body weight. There were seven total administrations of Cae, with one injection given every hour. One hour after the final Cae injection, a single dose of LPS was administered. The control group was administered an identical volume of PBS injection. The initial administration of Cae was performed at 0 h. On the 29th day of the experiment, after a 10-h fast, the rodents were anesthetized with 3% isoflurane administered via inhalation for a duration of 2 min.

Pancreatic tissues were collected immediately after euthanasia. The abdominal cavity was rapidly opened under sterile conditions, and the pancreas was carefully dissected from surrounding tissues, including the stomach, spleen, duodenum, and mesentery, using fine forceps to avoid mechanical damage. The excised pancreas was gently rinsed in ice-cold phosphate-buffered saline (PBS) to remove blood and debris, blotted dry, and divided into portions for downstream analyses. For histological examination, pancreatic tissues were fixed in 4% paraformaldehyde at room temperature for 24 h, followed by paraffin embedding. For molecular analyses, tissues were snap-frozen in liquid nitrogen and stored at −80 °C until use. Blood was drawn from the cornea, centrifuged at 4 °C at 4,000 rpm for 15 min, and the serum was subsequently collected and stored at −80 °C until needed. The cecum contents and jejunum were collected in a sterile environment for subsequent microbiome sequencing. A section of the jejunum was immersed in 4% paraformaldehyde for subsequent histopathological examination, while the remaining specimens were cryopreserved in liquid nitrogen and stored at −80 °C.

Caco-2, a human colonic epithelial cell line, was cultured at 37 °C in a 5% (v/v) CO₂ atmosphere using Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin/streptomycin. A controlled environment with elevated humidity, maintained at 37 °C with a 5% carbon dioxide concentration, was employed for cell culture. Once the cells achieved 80% confluence, the culture medium was replaced every 2 days or the cells were harvested. The Caco2 cells were subsequently divided into three distinct groups: CON (treated solely with PBS), TNF-α (treated exclusively with TNF-α), and LP18 (treated with a combination of LP18 and TNF-α). The Caco2 cells in the control group were treated with phosphate-buffered saline (PBS), whereas the cells in the TNF-α and LP18 groups were exposed to 100 ng/mL TNF-α or 1 × 10⁸ CFU/mL LP18, respectively. In the experimental group designated as LP18, the Caco-2 cells were pretreated with a concentration of 1 × 10⁸ CFU/mL LP18 for a period of 24 h. Subsequently, the cells were subjected to a treatment with 100 ng/mL of TNF-α for a duration of 24 h. Prior to the administration of LP18 or TNF-α, each experimental group underwent two flushes with PBS.

Intestinal morphological analysis entailed dehydrating jejunum tissue specimens through a sequential series of ethanol and xylene solutions, then fixing them in paraformaldehyde solution for 24 h at room temperature. The materials were subsequently processed into paraffin slabs. A 5 μm-thick cross-section was extracted from each specimen and stained with hematoxylin and eosin (Sigma-Aldrich, St. Louis, MO, USA). Following the administration of cedar oil, the tissues were examined and documented using a microscope (BX63, Olympus, Tokyo, Japan).

The concentrations ofIFN-γ, ALT, AST, D-LA, and DAO in the serum were quantified in accordance with the manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) using a spectrophotometer (SpectraMax M5, Molecular Devices, Sunnyvale, California, USA) in the laboratory.

Mouse serum samples were collected, and concentrations of tumor necrosis factor-alpha (TNF-α), interleukin-1β (IL-1β), IL-2, IL-4, IL-6, and LPS were quantified using the same ELISA kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The absorbance was measured at 450 nm using a microplate reader.

As previously reported, the Ussing chamber experiment was conducted to assess jejunum permeability (Hu et al., 2013). The proximal segment of the jejunum was excised, treated with Ringer's solution, and secured using a Ussing chamber system (VCC MC6, San Diego, CA). Following equilibration, the transendothelial electrical resistance (TER, Ω·cm²) was assessed at 15-min intervals over a duration of 2 h. The TER value of a caco2 cells was determined by calculating the mean of the data. On the mucosal surface, FITC-labeled dextran (FD4, Sigma-Aldrich, Saint Louis, MO, USA) was administered. FD4 flux from the mucosal to the serosal side was assessed at 30-min intervals over a 90-min period utilizing a multifunction microplate reader (Bio-Tek, Flx800, USA).

Total RNA was isolated from jejunum and Caco2 tissues employing TRIZOL reagent (TaKaRa, Japan). RNA purity and integrity were evaluated using spectrophotometry (SpectraMax M5, Molecular Devices, USA), with A260/A280 ratios maintained within the range of 1.8–2.0. cDNA synthesis was performed using the PrimeScript II First Strand cDNA Synthesis Kit (Vazyme, Nanjing, Jiangsu, China). For RT-qPCR analysis, reactions were conducted in triplicate on 96-well plates utilizing the StepOne Plus Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) and the HiScript II One Step RT-qPCR SYBR Green Kit (Vazyme, Nanjing, Jiangsu, China). Primers designed using NCBI Primer-Blast are provided in Table S3. Gene expression levels were normalized to β-actin and quantified utilizing the 2–ΔΔCt method.

Jejunum tissue lysates prepared with RIPA buffer were subjected to BCA protein quantitation (Sigma, Saint Louis, MO, USA). The protease inhibitor PMSF was freshly added to the RIPA buffer immediately before use according to the manufacturer's instructions. Samples were incubated on ice for 30 min with intermittent vortexing and subsequently centrifuged at 12,000 × g for 15 min at 4 °C. The supernatants were collected, and protein concentrations were determined using a BCA assay prior to downstream analyses. Equal protein concentrations were subjected to 12% SDS-PAGE and subsequently deposited onto PVDF membranes (Millipore, MA, USA) through electroblotting. Membranes were blocked with 5% skim milk for 2 h at 26 °C and subsequently incubated overnight at 4 °C with primary antibodies targeting claudin-1, Occludin, p-NF-κB p65, NF-κB p65 and TRAF6 and β-actin (Cell Signaling Technology, MA, USA). Following five TBST washes, HRP-conjugated secondary antibodies were applied for 2 h at 26 °C. Protein bands were detected using a Millipore chemiluminescent HRP substrate kit (MA, USA) and imaged with a Tanon system (China). Band intensities quantified using ImageJ software, with protein levels normalized to β-actin.

Genomic DNA from microbes was isolated from rodent cecum contents under sterile conditions employing the TIANamp Stool DNA Kit (Tiangen, Beijing, China). DNA integrity was confirmed through spectrophotometric analysis and agarose gel electrophoresis. Amplification of the 16S rRNA V3–V4 region was performed using primers 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′), with nuclease-free water serving as a negative control. PCR products were subjected to electrophoresis in a 2% agarose gel. DADA2 (version 1.8) and QIIME2 (https://qiime2.org/) were utilized to denoise the data, followed by the excision of chimeric sequences using VSEARCH (version 2.15.0). Operational taxonomic unit (OTU) tables were generated from quality-filtered readings, employing SILVA-based taxonomy assignment via the RDP classifier (version 2.2). Alpha and beta diversity assessments employed unweighted UniFrac distance metrics. Differential abundance analysis at the phylum/genus level and functional predictions based on PICRUSt2 were conducted.

Serum samples (100 μL) were processed according to established protocols (Gao et al., 2016): combined with 400 μL of 80% methanol in EP tubes, vortexed, and centrifuged at 4 °C at 15,000 g for 20 min. Supernatants were diluted with mass spectrometry-grade water to achieve a methanol concentration of 53%. Pooled quality control samples derived from equal aliquots ensured analytical consistency. Chromatographic separation was conducted using a Vanquish UHPLC system (Thermo Fisher, MA, USA) equipped with a Hypesil Gold column (100 × 2.1 mm, 1.9 μm) in both positive and negative ion modes. Positive mode conditions: mobile phase A = 0.1% formic acid, B = methanol; negative mode conditions: both phases = methanol (5 mM ammonium acetate in A, pH 9.0). Column parameters: discharge rate of 0.2 mL/min, temperature set to 40 °C. Mass spectrometry was conducted using a Q Exactive™ HF instrument (Thermo Fisher, MA, USA) with a scan range of 100–1,500 m/z. ESI settings: spray voltage 3.5 kV, sheath gas at 35 kPa, auxiliary gas at 10 L/min, capillary temperature 320 °C, S-lens RF 60, auxiliary heater at 350 °C.

Previous reports (Zhang et al., 2024) were used to prepare the sample. The following parameters were employed to quantify the levels of SCFA: a flow rate of 2 mL/min, an initial column temperature of 90 °C, a duration of 6 min, a temperature ramp of 10 °C/min to 200 °C, followed by an additional 6-min hold. The gas chromatograph employed was an Agilent 7890, located in Palo Alto, California, USA. The carrier gas has a flow rate of 2 mL/min. With a hydrogen flow rate of 30 mL/min and an air flow rate of 300 mL/min, the FID detector is maintained at a temperature of 250 °C.

Statistical analyses were conducted using SPSS 20.0 (IBM, USA) for one-way ANOVA with Tukey's post hoc test to perform multiple comparisons. Data presented as mean ± SEM; P < 0.05 denotes statistical significance. Graphs produced using GraphPad Prism 8.4.2.

Comparing with the normal pancreatic morphology in the control group, the pancreas tissues showed typical damage in the SAP group, including extensive edema, inflammatory infiltration and vacuolation of acinar cells. In contrast to the morphological disruption in the SAP group, LP18 administration markedly alleviated the histological injury in the pancreas, as reflected by reduced stromal edema and preservation of acinar integrity (Figure 1A). In addition, the LP18 supplementation also significantly (P < 0.05) restored the pancreatic exocrine function, with a notable (P < 0.05) decrease in the levels of enzyme, lipase and triglyceride comparing with the SAP group (Figure 1B).

Histological analysis of the small intestine revealed marked villus atrophy and mucosal injury in the SAP group comparing with the control group (Figure 2A). However, the LP18 supplementation ameliorated the villus structure and improved overall mucosal morphology. Quantitative measurements demonstrated a substantial reduction in villus height and villus height-to-crypt depth ratio in SAP mice (Figure 2B), whereas LP18 treatment significantly (P < 0.05) increased both parameters relative to the SAP group, indicating improved epithelial architecture. To further evaluate intestinal barrier permeability, the circulating levels of D-lactic acid (D-LA), diamine oxidase (DAO), and lipopolysaccharide (LPS) were measured (Figure 2C). Consistent with impaired barrier function, these molecules were significantly increased in the SAP group, while LP18 treatment effectively mitigated their elevation (P < 0.05), suggesting partial restoration of barrier integrity. Protein expression of the tight junction markers claudin-1 and occludin was examined by Western blotting (Figure 2D). SAP mice exhibited notable (P < 0.05) reductions in both proteins, whereas LP18 treatment significantly (P < 0.05) increased their expression relative to SAP, indicating improved tight junction maintenance.

To evaluate the effect of LP18 on the systemic inflammation, the levels of inflammatory factors in serum were assessed (Figures 3A, B). Compared with the SAP group, the levels of liver enzymes were significantly decreased in the LP18 group, indicating that LP18 exerts a protective effect on the liver (Figure 3A). The concentrations of key proinflammatory cytokines, including tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β) and interleukin-6 (IL-6), were significantly (P < 0.05) increased in the SAP group. In contrast, LP18 treatment significantly (P < 0.05) reduced the upregulated levels of these cytokines compared with the SAP group (Figure 3B).

The overall distribution of microbial abundance across ranked species is shown in Figure 4A. The relative abundance curve was altered in the SAP group compared with the control group, whereas the LP18 supplementation led to a significant change in the distribution pattern. Venn analysis demonstrated both shared and unique operational taxonomic units (OTUs) among the three groups (Figure 4B). α-diversity indices, including Feature count, ACE, Simpson, and Chao1, were subsequently assessed (Figure 4C). SAP mice displayed reduced Shannon and PD-whole tree indices relative to the control group, indicating decreased microbial richness and phylogenetic diversity. LP18 supplementation increased Shannon and PD-whole tree values compared with the SAP group. Similarly, Feature count, ACE, and Chao1 indices showed higher values in LP18-treated mice relative to SAP, whereas Simpson index remained relatively stable across groups. β-diversity analyses illustrated clear separation of microbial community structures among groups. Principal coordinate analysis (PCoA) plots showed distinct clustering of CON, SAP, and LP18 groups based on PC1, PC2, and PC3 (Figure 4D). Partial least squares discriminant analysis (PLS-DA) further emphasized these separations, demonstrating distinct microbial community profiles in each group (Figure 4E).

To further identify specific taxa contributing to microbial differences among groups, LEfSe analysis was performed. The cladogram (Figure 5A) illustrated distinct bacterial enrichments across the CON, SAP, and LP18 groups. SAP mice exhibited strong enrichment of taxa belonging to Actinobacteriota, including the class Actinobacteria, order Bifidobacteriales, family Bifidobacteriaceae, and genus Bifidobacterium. LP18 treatment was associated with increased abundance of several taxa primarily within Firmicutes, including Allobaculum, Dubosiella, Fournierella, Christensenella, and members of Ruminococcaceae. In contrast, the control group displayed enrichment across a broader set of bacterial lineages, such as Turicibacter, Lachnoclostridium, Parasutterella, Desulfovibrio, and taxa associated with Clostridia- and Desulfobacterota-related families.

Consistent with the cladogram, LDA score analysis revealed clear group-specific discriminatory taxa (Figure 5B). SAP mice showed high LDA scores for Actinobacteriota-associated taxa, whereas the LP18 group was characterized by differential enrichment of multiple Firmicutes-derived genera. Control mice demonstrated enrichment of several genera from Firmicutes, Proteobacteria, and Desulfobacterota.

Relative abundance plots of representative differential genera are presented in Figure 5C. These taxa exhibited distinct distribution patterns among the three groups, reflecting the group-specific enrichment identified by LEfSe analysis. Genera enriched in SAP mice, such as Bifidobacterium, Gordonibacter, and members of Bifidobacteriales, were lower or absent in LP18-treated mice. Conversely, LP18 supplementation increased the abundance of several Firmicutes-associated genera that were less represented in SAP mice.

Principal component analysis (PCA) showed clear separation between CON and SAP groups, with LP18 displaying a distinct clustering pattern (Figure 6A). The PCA scatter plots and pairwise component distributions demonstrated that there were marked alterations in global metabolic profiles between the SAP and CON groups. LP18-treated mice exhibited a shift away from the SAP cluster, indicating a partial restoration of metabolic patterns. Orthogonal partial least squares discriminant analysis (OPLS-DA) further revealed distinct metabolic signatures among groups (Figure 6B). Clear group separation was observed between CON and SAP, as well as between LP18 and SAP groups, confirming systematic metabolic differences occurred in the SAP mice but modified by LP18.

A heatmap of significantly altered metabolites across samples (Figure 6C) demonstrated marked clustering of metabolic features by group. CON and SAP groups exhibited divergent profiles, while LP18 samples showed an intermediate pattern, characterized by distinct changes in multiple metabolite classes, including lipids, organic acids, and bioactive small molecules. Volcano plot analysis identified numerous differentially expressed metabolites (Figure 6D). A total of 251 metabolites were upregulated and 230 were downregulated in SAP compared with CON. In contrast, comparison between LP18 and SAP identified 22 upregulated and 34 downregulated metabolites, indicating a relatively moderate but targeted metabolic shift following LP18 treatment.

KEGG pathway enrichment analysis was performed to identify metabolic pathways associated with differential metabolites. As shown in Figure 7A, comparison between CON and SAP groups revealed significant enrichment of multiple pathways, including inositol phosphate metabolism, glycerolipid metabolism, glycerophospholipid metabolism, phenylalanine metabolism, and β-alanine metabolism. Bubble plots further indicated that several of these pathways exhibited high pathway impact values, reflecting substantial metabolic perturbation induced by SAP.

Comparison of LP18 and SAP groups identified a distinct set of enriched pathways (Figure 7B). Metabolic pathways such as pyruvate metabolism, pyrimidine metabolism, porphyrin metabolism, arginine and proline metabolism, and alanine, aspartate, and glutamate metabolism were significantly enriched. Pathways with high impact scores included pyruvate metabolism, arginine biosynthesis, and amino acid-related metabolic modules.

Short-chain fatty acids (SCFAs) were measured to evaluate microbial metabolic output among groups (Figure 8). SAP mice exhibited a reduction in total SCFA levels compared with the control group. LP18 supplementation increased total SCFAs relative to SAP group, resulting in levels comparable to controls. Among individual SCFAs, acetic acid and propionic acid were both decreased (P<0.05) in the SAP mice relative to controls, while LP18 treatment increased their concentrations compared with the SAP group. Butyric acid was markedly (P<0.05) reduced in the SAP mice but partially restored in the LP18 group. Isobutyric acid and valeric acid showed no significant differences among groups. Levels of isovaleric acid were higher in SAP and LP18 groups than in controls. Caproic acid increased in both SAP and LP18 groups relative to controls.

Expression levels of SCFA receptors GRP41, GRP43, and GRP109A were further evaluated among these three groups (Figure 9A). The relative mRNA expression levels of GRP41, GRP43, and GRP109A were significantly reduced in the SAP group compared with the control group, whereas levels of GRP43 and GRP109A were markedly restored after the LP18 treatment. However, the expression level of GRP41 still remained lower in both SAP and LP18 groups than in controls. To further examine the downstream inflammatory signaling, protein levels of phosphorylated NF-κB p65 (p-NF-κB p65), total NF-κB p65 (t-NF-κB p65), and TRAF6 were analyzed (Figure 9B). Compared with the control group, the ratio of p-NF-κB p65 to t-NF-κB p65 and the relative expression level of TRAF6 protein were increased in the SAP mice. However, both these two levels were significantly reduced after LP18 treatment, indicating partial attenuation of the activation of NF-κB pathway induced by SAP.

Cell viability, membrane damage, epithelial permeability, and tight junction-related gene expression were evaluated in TNF-αstimulated epithelial cells (Figure 10). TNF-α markedly reduced cell viability compared with the control group, whereas LP18 treatment partially restored viability (Figure 10A). LDH release was significantly increased in TNF-α-treated cells, indicating enhanced cellular damage; LP18 reduced LDH release relative to TNF-α (Figure 10B). Permeability measurements showed that TNF-α increased FD4 flux while decreasing TEER values, demonstrating impaired barrier integrity (Figures 10C, D). LP18 treatment decreased FD4 flux and partially restored TEER compared with TNF-α, indicating improved epithelial barrier function. Expression of tight junction-related genes, including Claudin-1, Claudin-5, Occludin, and ZO-1, was significantly downregulated in TNF-α-treated cells (Figure 10E). LP18 supplementation upregulated all four genes relative to TNF-α, although levels remained lower than those in untreated control cells.

Pro-inflammatory cytokine expression was assessed in TNF-α treated epithelial cells (Figure 11A). TNF-α markedly increased the mRNA levels of IL-1β, IL-6, and IFN-γ compared with the control group. LP18 supplementation significantly reduced these cytokines relative to TNF-α treatment, although values remained above baseline. IL-10 expression was elevated in TNF-α treated cells and remained similarly high in the LP18 group.

To further examine inflammatory signaling, mRNA expression of iNOS, MyD88, TRAF6, and NF-κB p65 was measured (Figure 11B). All four markers were significantly upregulated in response to TNF-α stimulation. LP18 treatment reduced the expression of iNOS, MyD88, TRAF6, and NF-κB p65 compared with TNF-α, indicating partial suppression of TNF-α induced inflammatory pathway activation.