2.1 Genetically predicted serum VB₁₂ levels, but not folate or homocysteine show a protective effect against the risk of acute pancreatitis in patients

To test for the causal effect of key circulating one-carbon metabolism nutrients on the risk of acute pancreatitis, 3, 13, and 12 SNPs robustly associated with serum folate, homocysteine, and VB₁₂ levels, respectively, were used as instrumental variables (Table S2) to determine their effects on patients with acute pancreatitis. The MR results showed that the effect of homocysteine on acute pancreatitis was inconsistent in the FinnGen consortium (odds ratio [OR] = 0.8295, 95% confidence interval [CI] = 0.6904‒0.9966, p = 0.0458) and the UK Biobank cohort (OR = 1.2162, 95% CI = 0.8393‒1.7623, p = 0.3010) (Figure 1). The association between folate and acute pancreatitis was not observed (Figure 1). In contrast, the genetic prediction of the serum concentration of VB₁₂ was causally associated with decreased risk of acute pancreatitis in the UK Biobank cohorts (OR = 0.8130, 95% CI = 0.6618‒0.9987, p = 0.0485) and consistently observed in the FinnGen consortium (OR = 0.9484, 95% CI = 0.8256‒1.0895, p = 0.4540). To derive comprehensive genome-wide association study (GWAS) statistics for acute pancreatitis outcomes, we conducted a meta-analysis using data from both the FinnGen consortium and the UK Biobank cohorts. This combined analysis revealed no significant correlations between the risks of acute pancreatitis and the genetically predicted serum levels of homocysteine or folate (Figure 1). A suggestively negative causal association between serum VB₁₂ levels and the risks of acute pancreatitis in the human population was detected (Figure 1). For one-unit or one-standard deviation (SD) increase in the prevalence of these traits, the combined OR and corresponding 95% CI were 0.8806 (95% CI = 0.8095 ‒0.9579, p = 0.0031) for acute pancreatitis. The results remained directionally consistent in MR-Egger and weighted median model (Table 1).

Small heterogeneity was noticed in the analyses of acute pancreatitis (Figure S1). In the sensitivity analysis with each SNP that genetically predicted VB₁₂ levels, the associations remained (Table 1). Additionally, leave-one-out analyses showed that no single SNP drove the estimates (Figure S2), confirming the reliability of the causal effects of VB₁₂ levels in serum on acute pancreatitis in humans. Meanwhile, MR-Egger and MR-PRESSO tests showed no evidence of horizontal pleiotropy (Table 1). These results demonstrated that genetically predicted serum VB₁₂ levels are negatively associated with the risks of acute pancreatitis in patients.

2.2 Acute pancreatitis is aggravated in CD320-ablation mouse models

Although the MR analysis provided an associated relationship between serum VB₁₂ levels and acute pancreatitis, the roles of VB₁₂ in the progress of acute pancreatitis remained to be directly elucidated. Therefore, a genetically modified mouse model involving the ablation of the CD320 gene was used to evaluate the effects of cobalamin homeostasis on acute pancreatitis (Figure S3A‒D). The ablation of the CD320 gene causes the defects of cellular VB₁₂ uptake in cells in mouse tissues,¹⁷ resulting in reduced VB₁₂ levels in pancreatic tissues of mice (Figure S3E). Intraperitoneal L-arginine (L-Arg) administration is widely used to induce necrotizing pancreatitis in animal models. This non-invasive model leads to targeted, dose-responsive necrosis in acinar cell, making it an effective model for exploring the pathological mechanisms of acute necrotizing pancreatitis as well as the temporal evolution of the disease. Thus, an acute pancreatitis mouse model was generated by administering with L-Arg (Figure 2A). Seventy-two hours after L-Arg treatment, the mice were examined. The level of VB₁₂ in serum and pancreatic tissue significantly decreased in wild-type mice, but not CD320-ablation mice after administering with L-Arg (Figure S4A). In the wild-type mice, majorities of the mouse pancreatic lobules turned loose, pancreatic edema, necrosis, inflammatory cell infiltrations, and acinar atrophy were also observed (Figure 2B,C). Levels of amylase and lipase in serum were significantly increased at the 72 h stage of the disease (Figure 2D). The ductal tubular complex is an important pathological manifestation of acinar cell death in acute pancreatitis and can be recognized by CK-19, a sensitive duct marker.¹⁸ As acute pancreatitis progressed, many pancreatic epithelial cells changed to a flat shape and formed tube-like structures with strong staining of CK-19 at the 72 h stage of the disease (Figure 2E,F). In CD320-ablation mice, acute pancreatitis induced with L-Arg produced much severer pancreatic histopathological changes (Figure 2B,C). The CD320-ablation greatly increased the levels of amylase and lipase in serum (Figure 2D) and CD3⁺ T lymphocyte infiltration (Figure 2E,G), as well as the formation of ductal tubular complexes (Figure 2E,F) in the pancreas in mice with L-Arg-induced acute pancreatitis. CD320-ablation did not increase or slightly increased the level of activity of trypsin, CD11b⁺ macrophages, and MPO⁺ neutrophil infiltrations in pancreatic tissues (Figure S4B‒F). The results suggest that the VB₁₂ deficiency in cells can enhance acinar cell injuries, promote the formation of ductal tubular complexes, and trigger T lymphocyte infiltration in L-Arg-induced acute pancreatitis in mice.

Acute pancreatitis induced by cerulein (CER), a cholecystokinin analog, was the most well-characterized and widely used experimental model for acute edematous pancreatitis.^{19, 20} During the initial stages of pancreatitis triggered by CER, the fusion of zymogen granules within acinar cells gives rise to prominent autophagic vacuoles. This process is coupled with a surge in lysosomal enzyme activity and activation of trypsinogen, ultimately culminating in cellular necrosis.^{19, 21} In the CER model, the pancreas begins to recover 12 h after induction of pancreatitis.²² This model of CER-induced pancreatitis favors the analysis of intracellular events in the early phase of pancreatitis and can be used to define the early phenotypes occurring in pancreatic tissues in mice. After administering CER in wild-type and CD320-ablation mice (Figure 3A), the level of VB₁₂ in serum and pancreatic tissues significantly decreased in wild-type mice, but not in CD320-ablation mice (Figure S5A), which is consistent with the L-Arg-induced acute pancreatitis models. CER induced much severer pancreatic histopathological changes with increased necrosis and ductal tubular complexes in CD320-ablation mice compared to wild-type mice (Figure 3B‒E). The levels of amylase and lipase in serum (Figure S5B), the trypsin activities (Figure S5C), and MPO⁺ neutrophil infiltration (Figure S5D) in pancreatic tissues did not change or slightly increased in CD320-ablation mice, in comparison to the phenotypes in wild-type mice. In addition, significant CD3⁺ T lymphocyte infiltration was not detected in pancreatic tissues in the early phase of the CER-induced acute pancreatitis models (data not shown). The experiments of CER-induced acute pancreatitis models indicate that VB₁₂ directly protects necrosis of acinar cells during the early stage of acute pancreatitis. T lymphocyte infiltration most likely was secondary effects after the necrosis of acinar cells in L-Arg-induced acute pancreatitis in mice.

2.3 Vitamin B₁₂ displays therapeutic effects on acute pancreatitis in mouse models

We used CER-induced pancreatitis to test the effects of VB₁₂ on the intracellular events in pancreatic tissues in wild-type mice (Figure 4A). After treatment with VB₁₂, the level of VB₁₂ in pancreatic tissues significantly increased (Figure S6A). The pancreatic necrosis (Figure 4B,C), the level of serum amylase (Figure 4D), the formation of ductal tubular complexes (Figure 4B,E), and the trypsin activities (Figure 4F) in pancreatic tissues in mice with CER-induced acute pancreatitis significantly reduced. In contrast, the VB₁₂ treatment was unable to reduce the level of serum lipase (Figure S6B), the MPO⁺ neutrophil infiltration (Figure S6C) and CD11b⁺ macrophage (Figure S6D,E) in pancreatic tissues in mice with CER-induced acute pancreatitis. In addition, we did not detect significant CD3⁺ T lymphocyte infiltration (Figure S6D,F) in pancreatic tissues in the early phase of the CER-induced acute pancreatitis models. The results confirm that VB₁₂ can restore the pancreatic structures and prevent acinar cell necrosis in the early phase of pancreatitis in mouse models with CER-induced pancreatitis and is unable to prevent the innate immune responses in pancreatic tissues with induced acute pancreatitis in mice.

Next, we tested the effect of VB₁₂ in L-Arg-induced acute pancreatitis model (Figure 5A). VB₁₂ treatment greatly reduced the pancreatic histopathological changes of acute pancreatitis in mouse models (Figure 5B,C). The levels of amylase and lipase in serum (Figure 5D), trypsin activities (Figure 5E), ductal tubular complexes (Figure 5F,G), CD3⁺ T lymphocyte infiltration (Figure 5F,H), and CD11b⁺ macrophage (Figure S7B,C) in pancreatic tissue with L-Arg-induced acute pancreatitis were dramatically reduced by VB₁₂ treatment. Furthermore, a notable decrease in the infiltration of CD8⁺ T lymphocyte, rather than CD4⁺ cell, was observed, accompanied by an increase in Treg cell in CD4⁺ T lymphocyte (Figure S7E). The results show that VB₁₂ plays a protective role in L-Arg-induced pancreatitis in mice.

To test whether VB₁₂ conducts prevention of occurrences of acute pancreatitis (Figure 6A). VB₁₂ pretreatment prevented the gross loss of pancreatic structures, ameliorated pancreatic histopathological changes (Figure 6B,C), and dramatically reduced the elevation of amylase levels in serum (Figure 6D), the formation of ductal tubular complexes (Figure 6E,F), CD3⁺ T lymphocyte infiltration (Figure 6E,G), the CD45⁺ immune cell (Figure S6C,E), and the activity of trypsin in the pancreatic tissues (Figure S8A) in mice with induced acute pancreatitis. Furthermore, there was a notable increase in the proportion of Treg cell in CD4⁺ T lymphocyte (Figure S8F). CD11b⁺ macrophage and MPO⁺ neutrophil infiltrations were not altered in pancreatic tissues (Figure S8B‒D). These results demonstrate that VB₁₂ pretreatment can prevent injuries of acinar cells and reduces inflammatory reactions of T lymphocytes in pancreatic tissues in mice with L-Arg-induced acute pancreatitis.

2.4 Vitamin B₁₂ protects the acinar cells via ATP in pancreatic tissues in mouse models

VB₁₂ participates in two biochemical reactions: the conversion of homocysteine to methionine, and N⁵-methyl-tetrahydrofolate to tetrahydrofolate by the cytosolic methionine synthase, which requires methyl-Cbl as a cofactor and the conversion of methyl malonyl-coenzyme A to succinyl-CoA, the molecule is mostly involved in tricarboxylic acid cycle for ATP production.²³ Genetic analysis showed that a high level of circulation VB₁₂, but not folate and homocysteine, reduced the risk of acute pancreatitis, excluding the possibility that VB₁₂ displayed a protective role in the progress of acute pancreatitis via homocysteine or folate pathways. Homocysteine is involved in oxidative stress pathways and VB₁₂ has been shown to conserve GSH to reduce oxidative stress in pancreatic tissues in a mouse model with acute pancreatitis. We subjected GSH into the CD320-ablation mice and induced acute pancreatitis by CER (Figure S9A). The results showed that GSH enhanced the necrosis in pancreatic tissues in the mouse models (Figure S9B‒D). The tests exclude that VB₁₂ clears reactive oxygen species by conserving GSH to protect pancreatic damage in pancreatic tissues of acute pancreatitis.

Most likely, VB₁₂ restores the pancreatic structures and protects the acinar cells in the early phase of pancreatitis in mouse models via succinyl-CoA-based energy metabolism, which is involved in ATP production. Indeed, the level of ATP was greatly reduced in pancreatic tissues in CD320-ablation mice (Figure S10A). After administering L-Arg or CER, the ATP greatly lose in pancreatic tissues in wild-type mice and CD320-ablation mice (Figure S10A). However, the decreased degree of ATP levels was smaller in the pancreatic tissues of CD320-ablation mice, in comparison to the ones in wild-type mice (Figure S10B). The phenotypes are consistent with reduced VB₁₂ in pancreatic tissues in wild-type mice and CD320-ablation mice after induction of acute pancreatitis. After pretreatment or treatment with VB₁₂ in mouse models, the level of ATP significantly increased in pancreatic tissues (Figure S10C).

To further verify the results, ATP was used to treat CD320-ablation mice with CER-induced acute pancreatitis (Figure 7A). As expected, the ATP treatment greatly reduced the pancreatic histopathological changes of acute pancreatitis in CD320-ablation mouse models (Figure 7B,C). ATP treatment dramatically reduced the cell necrosis, the ductal tubular complexes, trypsin activities in the pancreatic tissue, and the serum amylase levels in the CD320-ablation mouse models with CER-induced acute pancreatitis (Figure 7B‒F). In contrast, the serum lipase levels and the MPO⁺ neutrophil infiltration in pancreatic tissues did not change after ATP treatment in the CD320-ablation mouse models with CER-induced acute pancreatitis (Figure 7D,G). The results that ATP treatment was able significantly to reduce the cell necrosis, but not the ductal tubular complexes, trypsin activities in the pancreatic tissues and the serum amylase levels in the wild-type mouse models with CER-induced acute pancreatitis further verified that ATP protected the cellular necrosis in pancreatic tissues with acute pancreatitis (Figure S11A‒G). Thus, the results indicate that VB₁₂ restores the pancreatic structures and protects the acinar cells in the early phase of pancreatitis in mouse models via the ATP production pathway.

4.1 Genetic instrument selection for exposure

The independent SNPs of folate,⁴⁰ homocysteine,⁴¹ and VB₁₂⁴⁰ were identified from corresponding meta-analyses of GWAS including 37,465, 44,147, and 45,576 individuals of European ancestries, respectively (Table S1). At the genome-wide significance threshold of p < 5 × 10⁻⁸, we extracted those SNPs significantly associated with serum folate, homocysteine, and VB₁₂ levels. To refine our selection, we excluded any SNPs that exhibited linkage disequilibrium, specifically those with an r² value below 0.001 within a 10-Mb region. The R² and F statistic of each SNP was calculated using the formula: R² = 2 × MAF × (1 − MAF) × β², and F statistic = R² × (N − 2)/(1 − R²). Subsequently, we excluded any SNPs with an F statistic fell below 10. We extracted 3, 14, and 14 SNPs robustly associated with serum folate, homocysteine, and VB₁₂ as instrumental variables (Table S2). Three of these SNPs (rs7788053, rs1141321, and rs2251468) were absent from at least one of our acute pancreatitis cohorts. The final set of instrumental variables comprised 3, 13, and 12 genetic variants for serum folate, homocysteine, and VB₁₂, respectively (Table S2).

4.2 Outcome data sources and meta-analysis

We analyzed the GWAS summary statistics pertaining to acute pancreatitis, harnessing data from two prominent, publicly available biobank-scale cohorts: the FinnGen consortium and the UK Biobank cohort (Table S1). FinnGen represents a comprehensive study that integrates both population-based and disease-oriented cohorts while the UK Biobank stands as a prospective, population-centered initiative, focusing primarily on UK citizens between the ages of 40 and 69 years.⁴² Pre-computed acute pancreatitis summary statistics of the FinnGen released 7 data was collected from the project website (https://r7.finngen.fi/). Acute pancreatitis GWAS summary statistics from fastGWA analysis of the UK Biobank imputed data were obtained from the website (https://yanglab.westlake.edu.cn/data/ukb_fastgwa/imp_binary/). The FinnGen consortium included 4648 cases of acute pancreatitis and 273,442 controls, while the UK Biobank cohort included 1748 cases of acute pancreatitis and 454,600 controls (Table S1). In this study, summary-level data for acute pancreatitis was included from two independent cohorts. To make the UK Biobank summary statistics compatible with the GRCh38 reference assembly used by the FinnGen cohort, MungeSumstats tool⁴³ was used to convert the UK Biobank data from the GRCh37 assembly to the GRCh38 assembly. We eliminated gene variants with low confidence and focused solely on those with a minor allele frequency greater than 0.01 for our meta-analysis. A z-score meta-analysis of acute pancreatitis summary statistics was conducted between the FinnGen and the UK Biobank samples using METAL.⁴⁴ The meta-cohort for acute pancreatitis ultimately included a total of 6396 cases and 728,042 controls. After rigorous quality control, we narrowed down the variants from both cohorts to only those that met our standards. The resulting set of 14,410,399 variants was then used for the meta-analysis in this study.

4.3 Mendelian randomization primary and sensitivity analyses

The effect of serum folate, homocysteine, and VB₁₂ on acute pancreatitis outcomes was estimated using the random-effects inverse variance weighted (IVW), MR-Egger, and weighted median. The IVW method was applied to obtain the primary outcome. MR-Egger and weighted median methods were performed to verify the causal effect. We harmonized SNPs to ensure that the effect estimates of each SNP on each trait and the risk of acute pancreatitis corresponded to the same allele. To assess the heterogeneity across each SNP, we employed Cochran's Q statistic and leave-one-out analysis. Additionally, we utilized the MR-Egger intercept test and the MR-PRESSO global test to detect potential horizontal pleiotropy. All these analyses were carried out using TwoSample MR (version 0.5.6)⁴⁵ and MRPRESSO (version 1.0) in R (version 4.2.0).

4.4 Mice strains

Mice were group-housed in individually ventilated plastic cages, maintained on a 14 h light and 10 h dark cycles, and provided with a standard chow diet (Chengdu Dashuo Experimental Animal Ltd.) as well as autoclaved water, both with ad libitum access. C57BL/6J wild-type mice and the CD320-ablation mice (C57BL/6J-CD320^em1C/Cya) were purchased from GemPharmatech and Cyagen Bioscience, respectively. For genotyping, genomic DNA obtained from the tails of 3-week-old offspring (CD320^−/−) mice and subjected to genotyping by polymerase chain reaction (PCR) and sequencing.

The total RNA was extracted from mouse pancreatic tissues with an RNA Extract Kit (TIANGEN, DP422). cDNA was generated using Evo M-NLV RT-PCR Kit (Accurate Biology, AG11601), and PCR analysis was performed. All primer sequences are listed in Table S3.

4.5 Induction of acute pancreatitis and treatment

The C57BL/6J wild-type and CD320^−/− mice, aged 6−8 weeks, were used and randomly divided into groups with each group containing 5−10 mice. Acute pancreatitis was induced by intraperitoneal injections of 20% L-arginine solution (L-Arg, Sigma‒Aldrich, A1513) at 3.3 g/kg per hour for 2 h. L-Arg solution was prepared in saline (0.9% sodium chloride) and filter sterilized. Control mice were injected with sterile saline. For VB₁₂ (Sigma‒Aldrich, 68-19-9) pretreatment, C57BL/6J wild-type mice received four intraperitoneal injections of VB₁₂ (500 µg/kg) 72, 48, and 24 h before L-Arg injection. For VB₁₂ treatment after L-Arg-induced pancreatitis, C57BL/6J wild-type mice received three intraperitoneal injections of VB₁₂ (500 µg/kg) 1, 3, and 5 h after L-Arg injection. All mice were anesthetized with 1% isoflurane and euthanized 72 h after the administration of L-Arg injection. For CER (TOCRIS, 6264) induced acute pancreatitis (CER-AP) in wild-type and CD320^−/− mice, the CER-AP group received seven intraperitoneal injections of 100 µg/kg CER hourly, with saline as controls. The CER + VB₁₂, CER + GSH (Sigma‒Aldrich, G6013), and CER + ATP groups received three intraperitoneal injections of VB₁₂ (500 µg/kg), GSH (500 mg/kg), and ATP (200 mg/kg, Sigma‒Aldrich, A6419-1G), respectively, which were administered at 2, 4, and 6 h after the first injection of CER. All mice were anesthetized with 1% isoflurane and euthanized 12 h after the first CER injection. Blood serum and the pancreatic tissues were collected for further experiments.

4.6 Measurement of amylase and lipase in the serum

Serum amylase (C016) and lipase (A054) levels were assayed precisely following the manufacturer's instructions from Nanjing Jiancheng Bioengineering Institute. In brief, serum samples were collected by centrifugation, and diluted with saline for drawing the dose‒response curve. After incubated with substrate reaction solution, absorbance at 660 and 420 nm were measured for serum amylase and lipase, respectively.

4.7 Pancreatic trypsin activity

A fluorometric assay was performed to detect pancreatic trypsin activity as previously described.⁴⁶ Briefly, pancreas tissues were homogenized in ice-cold pH 6.5 buffer (5 mM MOP, 1 mM MgSO₄, and 250 mM sucrose), centrifuged, and the supernatant was collected. The sample supernatant was detected with reaction buffer (50 mM Tris‒HCl, pH 8.0, 150 mM NaCl, 1 mM CaCl, 0.1 mg/mL bovine serum albumin [BSA]) containing Boc-Gln-Ala-Arg-MCA (Peptide, 3135-v) at excitation wavelength of 380 nm and emission wavelength of 440 nm using a plate reader (CLARIOstar, BMG Labtech) for 10 min. A standard curve was obtained from purified human trypsin. The pancreatic protein concentration was detected using a BCA protein analysis (P0012, Beyotime).

4.8 Pancreatic MPO activity

Pancreatic MPO activity was measured using the substrate 3,3′,5,5′-tetramethylbenzidine (TMB, Sigma‒Aldrich, 54827-17-7) as previously described.⁴⁶ Briefly, the pancreas tissues were homogenized in 20 mM phosphate buffer (pH 7.4), centrifuged, and the supernatant was collected. The supernatant was mixed with a phosphate buffer (100 mM, pH 5.4) containing 0.5% HETAB and 20 mM TMB. The mixture was incubated at 37°C for 3 min, followed by the addition 0.01% H₂O₂, and further incubated for another 3 min. The difference in absorbance between 0 and 3 min at 655 nm was detected using a plate reader. A standard curve was obtained from purified human MPO.

4.9 Detection of VB₁₂ in the serum and pancreatic tissues

The concentrations of VB₁₂ in serum and pancreatic tissues were detected using an ELISA (Shanghai Enzyme-Linked Biotechnology Co., Ltd., YJ057867). According to the manufacturer's instructions, standard solution and samples were added into the wells. Subsequently, the biotinylated VB₁₂ detection antibody was added and the wells were incubated at 37°C for 1 h. After washing four times, and the wells were then sequentially incubated with streptavidin‒ Horseradish peroxidase (HRP) at 37°C for 40 min, washed four times, incubated with TMB substrate at 37°C for 30 min. A stop solution was added and detected immediately at 450 nm using a plate reader (SYNERGY H1, BioTek). The concentrations of VB₁₂ were calculated based on standard curves. Additionally, the pancreatic tissue protein was determined by a BCA assay.

4.10 Measurement of ATP level in pancreatic tissues

For the detection of ATP levels, pancreatic tissues were collected and analyzed according to the Enhanced ATP Assay Kit protocol (Beyotime Biotechnology, S0027). In brief, pancreatic tissues were lysed using ATP detection lysis solution and centrifugated. The supernatant was collected and detected ATP content that calculated based on a standard ATP curve normalized to total protein according to the manufacturer's instruction.

4.11 Hematoxylin and eosin, immunohistochemistry staining, and immunofluorescence

For paraffin-embedded sections, the fresh pancreatic specimens from mice were collected, fixed with 10% formalin for 72 h, and embedded in paraffin. Hematoxylin and eosin (H&E) and immunohistochemistry staining were performed as previously described.⁴⁷ Sections (5 µm thick) were dewaxed, rehydrated, and stained with H&E. For the immunohistochemistry staining, tissue antigen retrieval was conducted by boiling in 10 mM (pH 6.0) citrate buffer (Sigma) for 45 min. Sections were allowed to cool to room temperature before treatment of 3% H₂O₂ and then hybridized with primary antibodies of CD45R (1:100, Santa Cruz, c-19597), cytokeratin 19 (CK-19) (1:100, Santa Cruz, sc-376126), CD11b (1:100, eBioscience, M1/70), and CD3 (1:100, Beyotime Biotechnology, AF1480) overnight at 4°C. Sections were then stained with secondary antibodies (ZSGB-Bio) for 1 h at room temperature. Negative and positive controls were applied in each staining run. CK-19-positive metaplasia and the infiltration of CD45R⁺, CD11b⁺, and CD3⁺ cells were manually quantified from five random and non-overlapping fields and averaged for each mouse as described previously.⁴⁸

For frozen sections, the fresh pancreatic specimens from mice were collected, fixed with 4% polyformaldehyde (PFA) at RT for 1 h, dehydrated with 30% sucrose overnight, and embedded in optimal cutting temperature compound (Sakura, 4583). For immunofluorescence, slides (5 µm thick) were fixed with 4% PFA, washed with phosphate-buffered saline (PBS), and blocking for 1 h with 4% BSA containing 0.5% Triton X-100. Primary antibodies were diluted in 4% BSA containing 0.5% Triton X-100 and incubated overnight at 4°C or 1 h at RT. Secondary antibodies were diluted in PBS with 0.1% Tween-20 and incubated for 1 h at RT. Slides were mounted in Antifade Mounting Medium with 4'6-diamidino-2-phenylindole (Beyotime Biotechnology, P0131).

4.12 Flow cytometry

The fresh pancreatic tissue was digested with collagenase to form a single-cell suspension and then cryopreserved in liquid nitrogen. The cells were resuscitated at 37°C and washed with Dulbecco's phosphate-buffered saline (DPBS). Subsequently, the cells were labeled with Fixable viability stain 700 (FVS700) and stained with fluorescent antibodies (BioLegend, CD3-BV421, CD4-FITC, CD8-APC, CD45-BV510, and CD25-PE). After washing with DPBS again, the cells were labeled with Foxp3 (BioLegend, PerCP-Cy5.5) using the Foxp3 staining kit (eBioscience, 00-5523-00). Finally, the cells were analyzed by flow cytometer (BD Celesta).

4.13 Histological scoring for acute pancreatitis

Histopathology scores for acute pancreatitis were evaluated basic on edema, inflammatory cell infiltration, and necrosis as previously reported.^{49, 50} In each experimental group, two separate and blinded evaluators assessed 10 randomly selected areas on each pancreatic slide, utilizing a 200× magnification. The degree of pancreatic damage was quantified by evaluating the presence and extent of edema, inflammatory cell infiltration, necrosis, and then calculating a comprehensive histopathology score that is the aggregate of individual scores for these features.⁴⁹

4.14 Data and statistical analysis

The methodology for data collection and statistical analysis in our study adhered to the guidelines established in pharmacology for experimental design and analysis.⁵¹ All experiments were repeated at least three times, and all data are expressed as mean ± SD. We compared the results of the treatment groups and controls in mice using one-way analysis of variance (ANOVA) and Tukey's multiple-comparison posttest or an unpaired t-test. Differences between groups were significant at a p-value of <0.05. Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software, Inc.).