This study was designed to explore the changes of fecal and serum metabolites and metabolic pathways in AP patients with different etiologies. This prospective study, which was approved by the Institutional Review Board of Shanghai General Hospital, collected samples from July 2022 to July 2023. Written informed consent was obtained from all individuals before enrollment. All the authors had access to the study data and had reviewed and approved the final manuscript.

Patients with AP were recruited in inpatient department of the Shanghai General Hospital. Only patients who met the Revised Atlanta Classification (RAC) criteria were included in this study. The diagnosis of AP met two of the following three features: (1) abdominal pain consistent with acute pancreatitis (acute onset of a persistent, severe, epigastric pain often radiating to the back); (2) serum lipase activity (or amylase activity) at least three times greater than the upper limit of normal; and (3) characteristic findings of acute pancreatitis on contrast-enhanced computed tomography (CECT) and less commonly magnetic resonance imaging (MRI) or transabdominal ultrasonography (Banks et al., 2013).

Inclusion criteria for BAP met the following features: meeting the diagnostic criteria for AP; (4) gallbladder stones confirmed by ultrasound or MRCP (van Geenen et al., 2010).

Inclusion criteria for HAP met the following features: meeting the diagnostic criteria for AP; (4) the level of triglyceride (TG) in serum reached or exceeded 11.3 mmol/L; (5) excluding other causes, such as alcoholic, gallbladder, etc. and the TG was between 5.65 and 11.3 mmol/L (Banks et al., 2013).

Healthy controls were recruited from health management center of Jiading branch of Shanghai General Hospital. They had never been diagnosed with severe digestive disorders, such as acute or chronic pancreatitis, cholecystitis or cholangitis, and hepatic insufficiency.

Fecal samples were collected in sterile fecal collection containers from each participant. The samples were immediately frozen at −80°C until LC-MS analysis.

All the AP patients took blood on the second day of admission to the hospital. Healthy controls were fasted for 12 h before blood collection to rule out the effects of food or alcohol. Blood was collected in 1.5 mL EP tubes without anticoagulant. After centrifugation, serum was subsampled (300 uL) into prepared sterile tubes and frozen at −80°C until LC-MS analysis.

The LC-MS/MS analysis of sample was conducted on a Thermo UHPLC-Q Exactive system equipped with an ACQUITY HSS T3 column (100 mm × 2.1 mm i.d., 1.8 μm; Waters, United States) at Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). The mobile phases consisted of 0.1% formic acid in water: acetonitrile (95:5, v/v) (solvent A) and 0.1% formic acid inacetonitrile: isopropanol: water (47.5:47.5, v/v) (solvent B). The flow rate was 0.40 mL/min and the column temperature was 40°C.

The UPLC system was coupled to a Thermo UHPLC-Q Exactive Mass Spectrometer equipped with an electrospray ionization (ESI) source operating in positive mode and negative mode. The optimal conditions were set as followed: source temperature at 400°C; sheath gas flow rate at 40 arb; Aux gas flow rate at 10 arb; ion-spray voltage floating (ISVF) at −2800 V in negative mode and 3500 V in positive mode, respectively; Normalized collision energy, 20-40-60 V rolling for MS/MS. Full MS resolution was 70,000, and MS/MS resolution was 17,500. Data acquisition was performed with the Data Dependent Acquisition (DDA) mode. The detection was carried out over a mass range of 70–1,050 m/z.

The pretreatment of LC/MS raw data was performed by Progenesis QI (Waters Corporation, Milford, United States) software, and a three-dimensional data matrix in CSV format was exported. The information in this three-dimensional matrix included: sample information, metabolite name and mass spectral response intensity. Internal standard peaks, as well as any known false positive peaks (including noise, column bleed, and derivatized reagent peaks), were removed from the data matrix, deredundant and peak pooled. At the same time, the metabolites were identified by searching database, and the main databases were the HMDB (http://www.hmdb.ca/), Metlin (https://metlin.scripps.edu/) and Majorbio Database.

The data were analyzed through the free online platform of majorbio choud platform (cloud.majorbio.com). Metabolic features detected at least 80% in any set of samples were retained. After filtering, minimum metabolite values were imputed for specific samples in which the metabolite levels fell below the lower limit of quantitation, and each Metabolic features were normalized by sum. To reduce the errors caused by sample preparation and instrument instability, the response intensity of the sample mass spectrum peaks was normalized by the sum normalization method, and then the normalized data matrix was obtained. Meanwhile, variables with relative standard deviation (RSD) > 30% of QC samples were removed, and log10 processing was performed to obtain the final data matrix for subsequent analysis.

The 266-6 (CTCC-007-0526) cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 5% heat-inactivated fetal bovine serum (FBS), 1% penicillin and streptomycin and 0.1% Mycoplasma Removal Agent (immocell, China) at 37°C, 95% humidity, and 5% CO₂. Both DMEM and FBS were obtained from Bioagrio, china.

The effects of LPS (Sigma-Aldrich, United States), Gibberellin A4 and Catechin (MedChemExpress, China) on the cell viability of 266–6 cells were measured using CCK8 (Topscience, China) according to the manufacturer’s instructions. 266-6 cells were incubated at different concentrations of Gibberellin A4 (1 μM, 5 μM, 10 μM, 20 μM, 30 μM) or Catechin (0.1 μM, 1 μM, 10 μM, 50 μM, 100 μM) for 24 h. Another group of cells were incubated at 1 μg/mL LPS (Sigma-Aldrich, United States) and Gibberellin A4 or Catechin. Then 10 μL CCK eight reagent was added to the cells, followed by another 2 h of incubation. Finally, the absorbance at 450 nm was measured.

The inflammatory cytokines were quantified by real-time PCR. Reverse transcription (RT) was conducted using HyperScript III RT SuperMix for qPCR in conjunction with the gDNA Remover Kit (both from EnzyArtisan, Shanghai, China). All primers, which were customsynthesized by EnzyArtisan (China), are listed in Supplementary Table S1. Real-time PCR was performed with 2xS6 Universal SYBR qPCR Mix (EnzyArtisan, Shanghai, China), adhering strictly to the manufacturer’s guidelines. The relative expression levels of the genes were determined using the 2^−ΔΔCT method.

Perform variance analysis on the matrix file after data preprocessing. The R package “ropls” (Version 1.6.2) was used to perform principal component analysis (PCA) and orthogonal least partial squares discriminant analysis (OPLS-DA), and 7-cycle interactive validation evaluating the stability of the model. The metabolites with VIP > 1, p < 0.05 were determined as significantly different metabolites based on the Variable importance in the projeciton (VIP) obtained by the OPLS-DA model and the p-value generated by Student’s t-test.

Differential metabolites among two groups were mapped into their biochemical pathways through metabolic enrichment and pathway analysis based on KEGG database (http://www.genome.jp/kegg/). These metabolites could be classified according to the pathways they involved or the functions they performed. Enrichment analysis was used to analyze a group of metabolites in a function node whether appears or not. The principle was that the annotation analysis of a single metabolite develops into an annotation analysis of a group of metabolites. Python packages “scipy.stats” (https://docs.scipy.org/doc/scipy/) was used to perform enrichment analysis to obtain the most relevant biological pathways.

Pearson correlation analysis was performed between differential metabolites and clinical indicators. Receiver operating characteristic (ROC) was performed for differential metabolites associated with clinical indicators.

The demographic and clinical characteristics for 30 AP participants and 20 controls were presented in Table 1. No statistically significant difference in age, gender, and the proportion of hypertension was found (P > 0.05) between the AP group and the control group. Body mass index (BMI) and diabetes prevalence were found to differ between the AP and control groups (0.05 < P < 0.1).

The demographic and clinical characteristics for 15 BAP participants and 15 HAP participants were presented in Table 2. There was no statistically significant difference observed in age, gender, onset time, hospital stay, CTSI score, white blood cell count, and neutrophil ratio between the two groups (P > 0.05). Nevertheless, the level of the inflammatory indicators, such as C-reaction protein (CRP) and IL-6, were higher in the HAP group (P < 0.1). Additionally, it was shown that the onset age of HAP was younger than that of BAP (P < 0.05). The daily habits and customs of patients diagnosed with HAP were found to be comparatively inferior to those of patients diagnosed with BAP. This discrepancy was shown in terms of a higher rate of smoking (P < 0.05) and drinking (0.05 < P < 0.1) among the HAP patients. The rate of diabetes was also higher in the HAP group (P < 0.05).

Pearson correlation analysis showed that the fecal samples of AP group and control group were significantly separated (Figure 1a). As shown in Venn plot, 495 fecal metabolites were unique in AP patients, and 635 fecal metabolites were found in healthy people but not detected in AP patients (Figure 1b).

Partial least squares discriminant analysis (PLS-DA) recognition model was generated for the whole dataset to evaluate the clustering trend of the samples. The separation between AP and control group could be clearly observed in the PLS-DA patterns in both negative- and positive-ion modes, which demonstrated that there was a severe metabolic disorder in the patients diagnosed with AP. Besides, the BAP group and the HAP group were also significantly different from the control group. Nevertheless, there was a considerable degree of overlap in the fecal metabolites between the BAP group and HAP groups, suggesting that the distinction was not significant between the two groups. (Figures 1c,d). All of the analyses in the OPLS-DA pattern were consistent with those in the PLS-DA pattern (Supplementary Figure S1).

Differential metabolites were identified by metabolites-expression-level analysis with P values ≤0.05 and fold change values ≥1.5, and VIP pred OPLS-DA>1. As shown in Figure 2a, 77 (22 upregulated and 55 downregulated) fecal differential metabolites were identified between the AP and control groups. Further screening for the metabolites with fold change values ≥2 and excluding related drug metabolism factors found that two metabolites were upregulated, and eight metabolites were downregulated in the AP group (Table 3). To further investigate the metabolic changes in the patients diagnosed with AP, metabolic pathway analysis was performed. As shown in Figure 2e, Cyclo-dopa 5-O-glucoside, significantly declined in the AP group, is a constituent of the betalain biosynthetic pathway (P < 0.05). Similarly, the decreased 3-Methylxanthine in the AP group is an essential component in caffeine metabolism (0.05 < P < 0.1). Sinapoyl Malate, which was considerably higher in the AP group, is a component of the phenylpropanoid biosynthesis pathway (0.05 < P < 0.1).

As shown in Figure 2b, 95 (60 upregulated and 35 downregulated) fecal differential metabolites were identified between the BAP and control groups. Further screening out the metabolites of fold change values ≥ 2 and excluding related drug metabolism factors found that seven metabolites were upregulated and seven metabolites were downregulated in the BAP group (Table 4). Subsequent metabolic pathway analysis revealed that 3-Methylxanthine and theobromine are constituents of the caffeine metabolism, with statistical significance (P < 0.05). Theobromine is a component found within the biosynthesis of alkaloids derived from histidine and purine (0.05 < P < 0.1). Keto-3-deoxy-D-manno-octulosonic acid, which was decreased in the BAP group, is a constituent of lipopolysaccharide biosynthesis (0.05 < P < 0.1). And sinapoyl malate, which was considerably higher in the BAP group, is a component of the phenylpropanoid biosynthesis (0.05 < P < 0.1) (Figure 2f).

As shown in Figure 2c, 101 (26 upregulated and 75 downregulated) fecal differential metabolites were identified between the HAP and control groups. Further screening out the metabolites of fold change values ≥ 2 and excluding related drug metabolism factors found that four metabolites were upregulated and 10 metabolites were downregulated in the HAP group (Table 5). One of the differential metabolites, N-arachidonylethanolamine, which was increased in the HAP group, is involved in various metabolic processes, such as inflammatory mediator regulation of TRP channels, thermogenesis, axon regeneration, neuroactive ligand-receptor interaction, and retrograde endocannabinoid signaling (0.05 < P < 0.1). Besides, S-adenosyl-L-methioninamine, exhibited an elevated presence in the HAP group, is involved in the metabolism of cysteine, methionine, arginine, and proline (0.05 < P < 0.1). In addition, to the same as above, 3-Methylxanthine is a component of the metabolism of caffeine. It has been determined that cyclo-dopa 5-O-glucoside is a component of the betalain biosynthesis pathway (Figure 2g).

As shown in Figure 2d, 18 (13 upregulated and five downregulated) fecal differences metabolites were identified between the BAP and HAP groups. Further screening out the metabolites of fold change values ≥ 2 found that phenylalanylhistidine was upregulated and benzyl sulfate was downregulated in the BAP group (P < 0.05).

Differential metabolites were identified by metabolites-expression-level analysis with P values ≤ 0.05 and fold change values ≥ 2, and VIP pred OPLS-DA>1. A total of 10 metabolites were included in the person correlation analysis. As shown in Figure 3a, significant correlations were observed between some fecal differential metabolites and clinical parameters including gender, age, BMI, CTSI score, and so on. A significant correlation of 3-(3,5-dihydroxyphenyl)-1-propanoic acid sulphate with the overall condition of the AP patients were observed, such as gender (r = −0.371, P = 0.057), age (r = −0.385, P = 0.048), history of smoking (r = 0.445, P = 0.020), history of drinking (r = 0.581, P = 0.001), hypertension (r = −0.514, P = 0.006). Besides, the degree of expression of 3-(3,5-dihydroxyphenyl)-1-propanoic acid sulphate was also involved with onset time (r = 0.513, P = 0.006). A significant positive correlation of N1-(5-Phospho-a-D-ribosyl)-5,6-dimethylbenzimidazole with alanine transaminase (r = 0.405, P = 0.036) and γ-Glutamyl Transferase (r = 0.500, P = 0.008) were observed. Besides, 3-Methylxanthine was involved with aspartate Transaminase (r = 0.616, P < 0.001). Cyclo-dopa 5-O-glucoside related to leukocyte (r = −0.628, P < 0.001), neutrophil ratio (r = −0.670, P < 0.001), and C-reactive protein (r = −0.383, P = 0.049). Catechin related to C-reactive protein (r = −0.455, P = 0.017) and CTSI score (r = −0.557, P = 0.003). The correlation tables were shown in Supplementary Tables S1, S2.

Compared with the control group, the expression level of Cyclo-dopa 5-O-glucoside related to inflammation index was significantly decreased in the AP group (Figure 3b). Receiver operating characteristic analysis found that Cyclo-dopa 5-O-glucoside was able to discriminate the AP patients from the controls with an AUC of 0.798 in these 50 subjects (Figure 3c). The expression level of catechin related to CTSI score was also significantly decreased in the AP group (Figure 3d). Receiver operating characteristic analysis found that catechin was able to discriminate the AP patients from the controls with an AUC of 0.889 in these 50 subjects (Figure 3e). Incorporating the above two metabolites into receiver operating characteristic conjoint analysis, it was found that Cyclo-dopa 5-O-glucoside and catechin were able to discriminate the AP patients from the controls with an AUC of 0.936 in these 50 subjects (Figure 3h).

As shown in Figure 3f, phenylalanylhistidine exhibited a considerable upregulation in the BAP group. Receiver operating characteristic analysis found that phenylalanylhistidine was able to discriminate BAP from HAP with an AUC of 0.783 in these 30 subjects (Figure 3g).

Pearson correlation analysis showed that the serum samples of AP group and control group were significantly separated (Figure 4a). As shown in Venn plot, 70 serum metabolites were unique in AP patients, and 52 serum metabolites were found in healthy people but not detected in AP patients (Figure 4b).

The separation between the AP and control groups could be clearly observed in the PLS-DA patterns in both negative- and positive-ion modes, which demonstrated that there was a severe metabolic disorder in the patients diagnosed with AP. Besides, the BAP group and the HAP group were also significantly different from the control group (Figures 4c,d). As shown in Supplementary Figure S2, all of the analyses in the OPLS-DA pattern were consistent with those in the PLS-DA pattern.

Differential metabolites were identified by metabolites-expression-level analysis with P values of ≤0.05 and fold change values ≥1.5, and VIP pred OPLS-DA>1. As shown in Table 6, 5 (3 upregulated and two downregulated) serum differential metabolites were identified between the AP and control groups. In order to conduct a more comprehensive analysis of the metabolic alterations in patients diagnosed with AP, an investigation of metabolic pathways was conducted. Among the five differential metabolites, the amount of L-Histidinol was much higher in the AP group, which is mainly involved in histidine metabolism (P < 0.05). The amount of stearidonic acid was substantially lower in the AP group, which is predominantly involved in alpha-Linolenic acid metabolism (P < 0.05).

As shown in Table 7, 8 (4 upregulated and four downregulated) serum differential metabolites were identified between the BAP and control groups. The amount of caffeine was lower in the BAP group, which is involved in caffeine metabolism (0.05 < P < 0.1). In addition, same as the above, the amount of L-Histidinol was much higher in the BAP group, which is mainly involved in histidine metabolism (0.05 < P < 0.1). The amount of stearidonic acid was substantially lower in the BAP group, which is predominantly involved in alpha-Linolenic acid metabolism (0.05 < P < 0.1).

As shown in Table 8, 6 (4 upregulated and two downregulated) fecal differential metabolites were identified between the HAP and control groups. The amount of stearidonic acid was substantially lower in the HAP group, which is predominantly involved in alpha-Linolenic acid metabolism (P < 0.05).

Compared with the HAP group, aspartyl-Arginine was significantly decreased in the BAP group.

Differential metabolites were identified by metabolites-expression-level analysis with P values of ≤0.05 and fold change values ≥ 1.5, and VIP pred OPLS-DA>1. A total of five metabolites were included in the person correlation analysis. As shown in Figure 5a, noteworthy correlations were observed between some serum differential metabolites and clinical indicators including age, BMI, CTSI score, and so on. A significant correlation of 3-[4-(sulfooxy)phenyl]propanoic acid (r = −0.551, P = 0.002) and L-Histidinol (r = −0.569, P = 0.001) with BMI were observed. Both of 3-[4-(sulfooxy)phenyl]propanoic acid (r = −0.605, P < 0.001) and L-Histidinol (r = −0.535, P = 0.002) were also related with IL-6, which reflected to systemic inflammation of patients. Besides, the degree of expression of (2R,3R,4R,5S,6R)-2-(4-Chloro-3-(4-ethoxybenzyl)phenyl)-6-(hydroxymethyl)tetrahydro-2H-pyran-3,4,5-triol related with CTSI score (r = −0.371, P = 0.044) and IL-6 (r = −0.368, P = 0.046). The correlation tables were shown in Supplementary Tables S3, S4.

Compared with the control group, all of the expression level of 3-[4-(sulfooxy)phenyl]propanoic acid, L-Histidinol and (2R,3R,4R,5S,6R)-2-(4-Chloro-3-(4-ethoxybenzyl)phenyl)-6-(hydroxymethyl)tetrahydro-2H-pyran-3,4,5-triol were significantly increased in patients diagnosed with AP (Figures 5b,d,f). Receiver operating characteristic analysis found that 3-[4-(sulfooxy)phenyl]propanoic acid was able to discriminate the AP patients from the controls with an AUC of 0.975 in these 50 subjects (Figure 5c). L-Histidinol was able to discriminate the AP patients from the controls with an AUC of 0.968 in these 50 subjects (Figure 5e). (2R,3R,4R,5S,6R)-2-(4-Chloro-3-(4-ethoxybenzyl)phenyl)-6-(hydroxymethyl)tetrahydro-2H-pyran-3,4,5-triol was able to discriminate the AP patients from controls with an AUC of 0.928 in these 50 subjects (Figure 5g). Incorporating the above three metabolites into receiver operating characteristic conjoint analysis found that 3-[4-(sulfooxy)phenyl]propanoic acid, L-Histidinol and (2R,3R,4R,5S,6R)-2-(4-Chloro-3-(4-ethoxybenzyl)phenyl)-6-(hydroxymethyl)tetrahydro-2H-pyran-3,4,5-triol were able to discriminate the AP patients from the controls with an AUC of 0.999 in these 50 subjects (Figure 5j).

As shown in Figure 5h, Aspartyl-Arginine was downregulated significantly in the BAP group. Receiver operating characteristic analysis found that aspartyl-arginine in serum was able to discriminate the BAP patients from the HAP patients with an AUC of 0.724 in these 30 subjects (Figure 5i).

Differential metabolites were identified by metabolites-expression-level analysis with P values of ≤0.05 and fold change values ≥1, and VIP pred OPLS-DA>1. As shown in Figure 6a, a total of 126 differential metabolites in feces were identified between the AP and control groups. And there were 157 divergent metabolites in serum between the AP and control groups. A comprehensive analysis revealed the presence of 47 differential metabolites in both feces and serum samples. Among these metabolites, 26 exhibited an upregulation, while 21 displayed a downregulation. The 47 metabolites had significant associations with many biological processes, including caffeine metabolism, neuroactive ligand-receptor interaction, regulation of lipolysis in adipocytes, renin secretion, and cAMP signaling pathway (P < 0.05) (Figure 6b). The primary metabolites implicated in the metabolism of caffeine were found to be caffeine (decreased), paraxanthine (decreased), and theobromine (decreased). The differential metabolites that primarily participate in the neuroactive ligand-receptor interaction are adenosine (decreased), epinephrine (increased), and endomorphin-1 (increased). Besides, adenosine and epinephrine have been found to be involved in several metabolic pathways, such as renin secretion, the cAMP signaling system, and the regulation of lipolysis in adipocytes.

Subsequent characterization of the 47 candidate metabolites revealed distinct pathophysiological patterns. Specifically, fecal catechin concentrations in AP patients exhibited significant negative correlations with CTSI and CRP levels, paralleled by diminished serum catechin metabolites. Similarly, fecal Gibberellin A4 content demonstrated an inverse relationship with WBC and neutrophil percentage, accompanied by concurrent reduction of GA4-derived metabolites in systemic circulation. Consequently, Gibberellin A4 and catechin were specifically chosen to systematically assess their anti-inflammatory properties in pancreatic acinar cells.

To assess potential cytotoxic effects, in vitro models were exposed to physiologically relevant concentrations of GA4 and catechin. Cytotoxicity assessment through CCK-8 assays revealed no statistically significant differences in cellular viability between treatment groups and vehicle controls across tested dosages (Figures 7a,b), confirming the biocompatibility of these compounds at investigated concentrations.

To systematically evaluate the cytoprotective and anti-inflammatory properties of GA4 and catechin, we subjected 266–6 pancreatic acinar cells to LPS (1 μg/mL, 24 h)-induced injury modeling acute pancreatitis (AP), with concomitant administration of escalating concentrations (GA4: 1–30 μM; catechin: 0.1–100 μM). Quantitative analysis through CCK-8 assays demonstrated concentration-dependent amelioration of cellular viability. GA4 exhibited maximal protection at 10 μM, while catechin showed optimal efficacy at 1 μM, establishing distinct effective concentration thresholds for the two compounds (Figures 7c,d).

Mechanistically, co-treatment with GA4 or catechin during LPS challenge resulted in marked attenuation of pro-inflammatory cytokine secretion. PCR quantification revealed that GA4 suppressed IL-1β, TNF-α and IL-6 production by 75.7%, 67.6% and 51.7% at 100 μM respectively (vs. LPS, P < 0.01). Catechin displayed dose-responsive inhibition patterns, achieving maximal anti-inflammatory effects at 100 μM with 76.2% (IL-1β), 87.5% (TNF-α) and 78% (IL-6) reductions (vs. LPS, P < 0.01) (Figure 8).