scRNA-seq of the intestine reveals the key role of mast cells in early gut dysfunction associated with acute pancreatitis

Abstract

BACKGROUND

Intestinal barrier dysfunction is a prevalent and varied manifestation of acute pancreatitis (AP). Molecular mechanisms underlying the early intestinal barrier in AP remain poorly understood.

AIM

To explore the biological processes and mechanisms of intestinal injury associated with AP, and to find potential targets for early prevention or treatment of intestinal barrier injury.

METHODS

This study utilized single-cell RNA sequencing of the small intestine, alongside in vitro and in vivo experiments, to examine intestinal barrier function homeostasis during the early stages of AP and explore involved biological processes and potential mechanisms.

RESULTS

Seventeen major cell types and 33232 cells were identified across all samples, including normal, AP1 (4x caerulein injections, animals sacrificed 2 h after the last injection), and AP2 (8x caerulein injections, animals sacrificed 4 h after the last injection). An average of 980 genes per cell was found in the normal intestine, compared to 927 in the AP1 intestine and 1382 in the AP2 intestine. B cells, dendritic cells, mast cells (MCs), and monocytes in AP1 and AP2 showed reduced numbers compared to the normal intestine. Enterocytes, brush cells, enteroendocrine cells, and goblet cells maintained numbers similar to the normal intestine, while cytotoxic T cells and natural killer (NK) cells increased. Enterocytes in early AP exhibited elevated programmed cell death and intestinal barrier dysfunction but retained absorption capabilities. Cytotoxic T cells and NK cells showed enhanced pathogen-fighting abilities. Activated MCs, secreted chemokine (C-C motif) ligand 5 (CCL5), promoted neutrophil and macrophage infiltration and contributed to barrier dysfunction.

CONCLUSION

These findings enrich our understanding of biological processes and mechanisms in AP-associated intestinal injury, suggesting that CCL5 from MCs is a potential target for addressing dysfunction.

Key Words: Single-cell RNA sequencing; Acute pancreatitis; Mast cell; CCL5; Intestinal barrier function

Core Tip: Our study provides a comprehensive picture of the transcriptome of small intestine cells during the early stage of acute pancreatitis, revealing a total of 33232 cells across all samples and 17 main clusters. Through our investigation, we established that mast cells (MCs) were promptly activated in the intestine, and we identified CCL5 derived from MCs as an indispensable factor contributing to the infiltration of inflammatory cells and the progression of gut barrier dysfunction.

INTRODUCTION

Acute pancreatitis (AP) is the most common acute abdominal disease, with an annual incidence of 34 per 100000 person-years in developed countries[1]. Most AP patients present with limited clinical symptoms and recover within 1 week, but approximately 20% of patients deteriorate to severe AP (SAP), with a substantial mortality rate of 20%-40%[2]. According to a meta-analysis of 18 studies, the occurrence rate of intestinal barrier dysfunction in AP is as high as 59%[3]. Intestine barrier dysfunction is a high-risk factor for sepsis and necrotic infectious pancreatitis[4]. Clinical trials and clinical practices underscore the importance of restoring the function of the intestinal barrier in AP patients, but efficient treatments are currently limited[5]. On the other hand, various pathological events are thought to be involved in intestinal barrier dysfunction during AP, including the acute inflammatory response, edema, microcirculation disorders, the oxidative stress response, and apoptosis[6]. The complexity and interconnection among these events impede the understanding of the development of AP-associated intestinal barrier dysfunction. The molecular mechanisms involved in the early stage of AP-associated intestinal barrier dysfunction remain largely unknown and are possibly neglected. Therefore, identifying key events and the underlying molecular mechanisms would be beneficial in understanding AP-associated intestinal barrier dysfunction.

Recent advances in the single-cell transcriptome have provided valuable techniques for revealing genetic and functional heterogeneity, detecting sealed cell subpopulations, and characterizing biological mechanisms at the cellular level[7,8]. Previous studies have successfully applied single-cell transcriptomes to explore the mechanism of intestinal injury or regeneration. For example, the single-cell transcriptome revealed a distinct damage-induced quiescent cell type responsible for resolving the homeostatic stem cell compartment and regenerating the intestinal epithelium[9]. A recent study revealed that intestinal regeneration and repair require the intimate cooperation of several types of epithelial cells[10]. In addition, single-cell RNA sequencing revealed that oligosaccharides could preserve small intestinal function by increasing the number of various types of enterocytes[11]. Nevertheless, considerable efforts are still required to elucidate the pathological process of AP-associated intestinal barrier dysfunction, especially when emerging technologies are used. Currently, no studies using single-cell transcriptomes to explore the heterogeneity of genetic, functional, or cell subpopulations have been reported.

A deeper study and understanding of intestinal barrier function have revealed that dysregulation of various cell types in the intestine can potentially lead to intestinal barrier dysfunction. For example, the epithelial cells lining the intestine form a physical barrier, and disruptions in the integrity of these cells can lead to increased permeability and impaired barrier function during AP[12]. Disruption of mucus production by goblet cells or alterations in the mucus composition can impair various site-specific protective functions and increase susceptibility to infections. The inhibition of MUC2, a mucus protein, leads to goblet cell dysfunction, gut dysbiosis, and bacterial translocation, worsening AP[13]. Intestinal stem cells (ISCs) are essential for restoring barrier homeostasis and maintaining the epithelial cell pool upon injury[14,15]. Previous works have shown that cholic acid or drugs potentiate gut barrier dysfunction by suppressing ISC renewal by inhibiting the expression of peroxisome proliferator-activated receptor alpha[16]. Natural killer (NK) cells can secrete interferon-γ (INF-γ), thereby activating the intestinal epithelium, leading to restoration of the mucosal barrier and preservation of immunotolerance[17]. In the context of gut barrier dysfunction associated with AP, prior studies have focused primarily on intestinal epithelial cells, the gut microbiota, and macrophages[18,19]; further investigations are needed to explore the potential involvement of other cell types, such as goblet cells, neutrophils and mast cells (MCs) in gut barrier dysfunction.

In this study, we employed a single-cell transcriptome approach, along with in vitro and in vivo experiments, to examine the balance of intestinal barrier function and explore the underlying biological processes and potential mechanisms involved. We identified 17 main types of cells in the intestine. Enterocytes exhibited elevated programmed cell death while maintaining their functional capabilities. Additionally, cytotoxic T cells and NK cells in the intestine showed enhanced anti-pathogen abilities during the early stage of AP. Furthermore, intestinal MCs were activated early in AP and secreted chemokine (C-C motif) ligand 5 (CCL5) to promote the infiltration of neutrophils and macrophages in the intestine. Thus, our findings enhance the current understanding of the key biological processes and mechanisms involved in AP-associated intestinal injury and may provide new pathways for preventing intestinal barrier dysfunction related to AP.

MATERIALS AND METHODS

Single-cell RNA sequencing

The collected mouse small intestine was cut into 1 mm³ pieces and enzymatically digested with the MACS Tumour Dissociation Kit Mouse (Miltenyi Biotec, Germany) for 30 minutes at 37 °C. This was followed by filtration of the cells through 70 μm and 40 μm cell filters (BD, United States). The filtered cells were subsequently centrifuged at 300 g/minute for 10 minutes, after which the supernatant was removed. This was followed by the addition of erythrocyte lysate (Thermo Fisher Scientific, United States) for a period of 2 minutes, resulting in cell lysis. The cell precipitates were then resuspended in PBS (0.04% BSA) following two washes with PBS. cDNA amplification and library construction were performed according to the instructions of the MGI DNBelab C series reagent kit (MGI, China). In summary, single-cell suspensions were converted into barcoded scRNA-seq libraries through a series of steps, including droplet encapsulation, emulsion breakage, mRNA-captured bead collection, reverse transcription, cDNA amplification, and purification. The resulting libraries were then identified via a Qubit ssDNA assay kit (Thermo Fisher Scientific, United States) and an Agilent Bioanalyzer 2100. All the libraries were subsequently subjected to further sequencing on the DIPSEQ T1 platform. DNBelab C series scRNA analysis software was used to filter the resulting sequencing data and create gene expression matrices.

Immunofluorescence staining

The collected small intestines were paraffin-embedded and cut into 4 μm thick sections. The sections were heated in an oven at 60 °C for 60 minutes, washed and dewaxed in xylene, and hydrated through a graded series of ethanol. The epitopes were then repaired with sodium citrate buffer (10 mmol/L sodium citrate and 0.05% Tween 20, pH 6.0) for a period of 20 minutes at 96 °C. The sections were subsequently blocked with 5% goat serum (16210064, Fisher Scientific, United States), and the following primary antibodies were added dropwise: Anti-CD44 (A12410, ABclonal, 1:100), anti-CD56 (A7913, ABclonal, 1:100), anti-c-kit (sc-365504, Santa Cruz, 1:100), anti-CD68 (97778S, CST, 1:100), and anti-ly6 g (65140-1-Ig, Wuhan Sanying, 1:50) antibodies. The secondary antibodies coupled with Alexa Fluor 488 or 594 (Invitrogen, United States) were added the next day and incubated for 60 minutes at room temperature. Then, the samples were stained dropwise with DAPI (Invitrogen, United States) containing an anti-fluorescence quencher. Images were captured using a Zeiss microscope (ZEN, Germany).

Immunohistochemical staining

For the immunohistochemical (IHC) staining assay, 4 μm paraffin sections were deparaffinized and placed in sodium citrate buffer (10 mmol/L sodium citrate and 0.05% Tween 20, pH 6.0) at 96 °C for 20 minutes to repair the epitopes. Thereafter, the sections were processed following the steps of the IHC kit (ab64261, Abcam), with sufficient drops of peroxidase added to the tissues for 10 min and washed with PBS twice. Subsequently, the slides were placed in a wet box, and the following primary antibodies were added: Anti-Ki-67 (9449S, 1:100, CST); anti-cleaved caspase 3 (abs132005, 1:100, Absin); anti-phospho-MLKL (af7420, 1:100, Affinity), and then the slides were incubated at 4 °C overnight. The next day, the primary antibody was removed by washing, and the secondary antibody was introduced in accordance with the species of the primary antibody. Subsequently, the tissue was washed and a drop of prepared DAB chromogenic solution was applied, and then rinsed under running water for 10 minutes until the color developed completely. The nuclei were then stained with hematoxylin for 10 minutes, after which the hematoxylin was removed under running water. Finally, the samples were sealed in neutral resin after dehydration in anhydrous ethanol and xylene. Images were obtained using a Zeiss microscope.

Genotype identification

c-Kit^w-sh/w-sh mice were purchased from Cyagen (China). For genotype identification, the tails of two-week-old mice were clipped, and DNA extraction was performed according to the instructions provided with the kit (Beyotime Biotech Inc.). The mouse tail samples were combined with lysate and subjected to a water bath at 55 °C for six hours. Subsequently, the mixture was transferred to a DNA purification column and centrifuged at 12000 rpm for one minute to remove the lower layer of liquid. Next, wash solution 1 was added, and the centrifugation process was repeated under the same conditions. After that, wash solution 2 was added, and the centrifugation was repeated twice. Finally, the eluent was collected, and the DNA was separated into the lower layer of the liquid. The following primers were employed to identify the mouse genes: Region: P1: 5’-CCTGACCCTGTCTCACAAGAG-3’; P2: 5’-AGGATTCATAGTTGTTCAATGTCC-3’; Region 2: P3: 5’- AGTGGTGAGAGCCAGCAACT-3’; P2: 5’- AGGATTCATAGTTGTTCAATGTCC-3’.

HE staining

The collected small intestine and pancreatic tissues were paraffin-embedded and cut into 4 μm thick sections. The sections were heated in an oven at 60 °C for 60 minutes, washed and dewaxed in xylene and hydrated through a graded series of ethanol. Then, the sections were immersed in a hematoxylin solution until the nuclei exhibited bluish-purple coloration. The sections were subsequently rinsed with tap water for 5 minutes and differentiated with 1% acidic alcohol to adjust the depth of staining. The sections were immersed in eosin solution until the cytoplasm and other structures exhibited pink coloration. The sections were then cleared in xylene and dehydrated in a graded series of ethanol, followed by sealing with neutral resin. Images were captured using a Zeiss microscope (ZEN, Germany).

Quantitative reverse transcription-PCR

The tissues or cells were treated with TRIzol reagent (Invitrogen, United States) for a period of 10 minutes, after which they were subjected to centrifugation at 12000 rpm/minute for 15 minutes at 4 °C. Then, the supernatant was collected and mixed with isopropanol to isolate the RNA, after which it was centrifuged at 12000 rpm/ minute for 10 minutes at 4 °C. After the RNA was obtained, the purity and concentration of the RNA were analyzed via a Nanodrop 1000 spectrophotometer (Thermo Fisher, United States). cDNA was synthesized using a High Capacity cDNA Reverse Transcription Kit (Life Tec, United States). Quantitative reverse transcription-PCR (qRT-PCR) was conducted using 2 × Universal SYBR Green Rapid qRT-PCR Mix (ABclonal, China). The following primer sequences were used: interleukin (IL)-1β, forward, TGGACCTTCCAGGATGAGGACA, reverse, GTTCATCTCGGAGCCTGTAGTG; IL-6, forward, TACCACTTCACAAGTCGGAGGC, reverse, CTGCAAGTGCATCATCGTTGTTC; CCL5, forward, CCTGCTGCTTTGCCTACCTCTC, reverse, ACACACTTGGCGGTTCCTTCGA; β-actin, forward, CATTGCTGACAGGATGCAGAAGG, reverse, TGCTGGAAGGTGGACAGTGAGG; 16s rDNA, forward, TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3', reverse, 5'-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAA-TCC-3'.

Toluidine blue staining

The collected small intestine tissues were paraffin-embedded and cut into 4 μm thick sections. The sections were heated in an oven at 60 °C for 60 minutes, washed and dewaxed in xylene and hydrated in a graded series of ethanol. Then, the sections were treated with 0.5% toluidine blue staining solution (Solarbio, China) for 30 minutes and rinsed with running water for 2 minutes. Then, 1% hydrochloric acid alcohol was added for a few seconds, and the samples were rinsed under running water for 2 minutes. The samples were subsequently cleared in xylene and dehydrated through a graded series of ethanol. Then, they were sealed with neutral resin. Images were captured using a Zeiss microscope (ZEN, Germany).

Detection of CCL5 expression in MCs by flow cytometry

The small intestines were cut into approximately 1 cm pieces and incubated in PBS buffer with 0.4 mg/mL collagenase P (Roche, Switzerland) at 37 °C for 30 minutes. Then, the small intestines were triturated slightly through a 1 mL pipette tip 10 times and filtered through a 70 μm filter. After centrifugation at 1000 rpm/minute for 5 minutes, the MCs were obtained and resuspended in PBS. The number of MCs was counted. A total of 1 × 10⁶ cells were transferred to a new 1.5 mL tube, and antibodies against CCL5 (149104, Biolegand, United States) and CD117 (553356, BD, United States) were added and incubated for 45 minutes at room temperature. The cells were subsequently centrifuged at 1000 rpm/min for 5 min and fixed in 4% paraformaldehyde for 10 minutes. Analysis was subsequently performed via flow cytometry (BD Biosciences, United States). The data were analyzed with FlowJo software (version 11, United States).

Measurement of amylase and lipase activity in plasma

Plasma was obtained by centrifuging the blood with 20 μl of 0.5 M EDTA at 1000 rpm/minute for 10 minutes. Trypsin and lipase activities in the plasma were determined via a clinical analysis system. The activity of amylase was determined using a kit that employed the substrate EPS, whereas the activity of lipase was determined using a kit that utilized the substrate methyl triazine. The signals were determined via the AU1000 system (Beckman, United States).

Animal models

C57BL/6J mice were purchased from Hunan SJA Laboratory Animal Co. The mice were kept under pathogen-free conditions. This study was ethically approved by the Second Xiangya Hospital of Central South University and complied with the guidelines for animal care and use.

For the mild AP model, 6-8-week-old male mice were fasted for 12 hours and then intraperitoneally injected with caerulein (50 μg/kg body weight) 4, or 8 times at 1 hour intervals, with 8 mice per group (randomly assigned). Mice were sacrificed (euthanasia was performed by injection of pentobarbital sodium) 6 hours or 12 hours after the first injection, and blood, pancreas, and small intestine were collected for further analysis.

Statistical analysis

All the data are expressed as the mean ± SE and were analyzed using GraphPad Prism 8. The significance of the differences between the two groups was analyzed using unpaired Student's t tests (parametric tests were employed when the data were normally distributed and homogeneous, and nonparametric tests were used otherwise), and one-way analysis of variance was used to determine differences between multiple groups. A P value of less than 0.05 was considered to indicate statistical significance.

RESULTS

Single-cell expression atlas and cell typing in small intestines with or without AP

To explore the cellular diversity of the intestine in the early stage of AP, we established early AP models via the intraperitoneal injection of caerulein 4 times and sacrificed the mice 2 hours after the last injection (namely, AP1), followed by the intraperitoneal injection of caerulein 8 times and sacrificed 4 hours after the last injection (namely, AP2; Figure 1A). Histology of the pancreas is shown in Figure 1B. The small intestines were collected and dissociated using enzymatic digestion. Single cells were subsequently captured via droplet-based systems. RNA extraction and library preparation were performed on the captured single cells, followed by sequencing of the prepared libraries (Figure 1C). After initial quality control, we detected 33232 cells across all samples and identified 980 genes per cell in the normal intestine, 927 genes per cell in the AP1 intestine, and 1382 genes per cell in the AP2 intestine (Supplementary Table 1). The transcriptomes of individual intestinal cells from all of the samples revealed 17 main clusters, including enterocyte 0, enterocyte 1, enterocyte 2, enterocyte 3, B cell, tuft cell, cytotoxic T cell, dendritic cell, enteroendocrine cell, goblet cell, M1 macrophage, MC1, MC2, monocyte, NK cell, neutrophil, and stem cell (Figure 1D). The distribution of cell types in the intestine is illustrated in Figure 1E. Gene expression signatures were generated for each cell cluster via the generation of cluster-specific marker genes by comparing gene expression differences, and representative genes were used to identify each cluster (Figure 1F). In many cells, well-known cell type markers are involved in cell type identification, such as C1qa for M1 macrophages, Napsa for neutrophils, S100a6 for MCs, Myo15b for enterocytes, Cd79b for dendritic cells, Dcn for tuft cells, Cd44 for stem cells, Chgb for enteroendocrine cells, and Igha for B cells (Figure 1F). We also identified various additional markers for each cell type, such as Taco1, Zfp791, Peak1 and Gm²6917 for enterocytes; Defa24, Tff3, Zg16, and Spink4 for goblet cells; Trbc2, Gzma, and Cd3 g for NK cells; Gzmb and Rgs1 for cytotoxic T cells; Ms4a1 and Cd74 for dendritic cells; Mgp and Sparc for tuft cells; Sct and Cck for enteroendocrine cells; and Cdk8, Cmss1, and Lars2 for stem cells (Figure 1G). Within the main clusters, 4 types of enterocytes were identified and labeled enterocyte 1, enterocyte 2, enterocyte 3, and enterocyte 4. Additionally, two types of MCs were identified and labeled MC1 and MC2.

Figure 1 Single-cell expression atlas of the small intestine in mice with acute pancreatitis. A: Modeling procedure for AP1 and AP2 in mice; B: HE staining of the pancreas in the AP1, AP2 and normal group; C: Workflow for the collection and processing of specimens from AP1, AP2, and control small intestine samples for scRNA-seq; D: Cell type and number ratio identified by single-cell sequencing; E: UMAP plot illustrating major cell types in the small intestine of mice; F: Heatmap showing the expression levels of specific markers in each cell type of the small intestine; G: Violin plots displaying the expression of representative well-known markers across identified cell types in the small intestine.

The proportions of these 17 clusters of cells in each sample were calculated using the Seurat program, and the stratification and cell type identification of each cluster in the normal, AP1 and AP2 groups are illustrated in Figure 2A. The number of cytotoxic T cells increased from 6.05% of the total cells in the normal group to 13.05% in the AP1 group and 13.67% in the AP2 group. MCs decreased from 9.07% of total cells in the normal group to 2.2% in the AP1 group and 2.37% in the AP2 group, and stem cells increased from 4.27% of total cells in the normal group to 9.74% in the AP1 group and 11.37% in the AP2 group (Figure 2B and C). The numbers of brush cells, enteroendocrine cells, and goblet cells were comparable among the groups. Although early infiltration of immune cells, such as neutrophils and macrophages, is typically considered a key characteristic of intestinal barrier damage[20-22], our analysis of single-cell sequencing data did not reveal significant infiltration of these cells. This discrepancy could be partially attributed to the unique characteristics of neutrophils and macrophages, including their high granularity, propensity to undergo spontaneous apoptosis, and considerable heterogeneity in activation states[23]. These features can introduce potential biases during the isolation and preparation steps of single-cell sequencing, leading to the results we observed. Overall, a relative decrease in the number of B cells, dendritic cells, MCs, and monocytes was observed, and the number of cytotoxic T cells, NK cells, and stem cells was greatest in the intestine of the AP2 group.

Figure 2 Proportions of these 17 clusters of cells in each group. A: UMAP plot illustrating major cell types in the small intestine of each group; B: The cell proportions of each cluster in each group; C: The cell count for each cluster in each group.

Enterocytes exhibited increased programmed cell death but maintainable function in the early stage of AP

The number of enterocytes in each group was comparable among the normal, AP1, and AP2 groups (accounting for 39.68%, 41.60%, and 41.85% of all identified cells in each group, respectively; Figure 2B and C). Given that many studies have suggested that the gut barrier is dysregulated in AP[24], we first examined the integrity of the gut barrier by immunofluorescence staining for ZO-1, a marker of the integrity of the gut. The results revealed that both the AP1 and AP2 groups presented decreased expression of ZO-1 (Figure 3A). Furthermore, the expression of ZO-1 was lower in the AP2 group than in the AP1 group, indicating that even in the early stage of AP, the gut barrier was disturbed. The differentially expressed genes (DEGs) associated with enterocytes in each group are shown in Supplementary Table 2. To understand the key signaling pathways of enterocytes that were dysregulated in AP, we performed Gene Ontology (GO) enrichment analysis for each group and found that enterocytes in AP1 were significantly related to cell stress, mitochondrial function, and catabolism (Figure 3B), which was further supported by the dysregulated expression of gene markers for these biological processes, such as Hsp90ab1, Hspa1b, Hspa1a, Hspd1, and Hspe1 (markers for cell stress); Tfam, Ucp2, Prdx1 and Park7 (markers for mitochondrial function); and Acly, Cpt1a, Cyp2d26, Fasn, and Ldha (markers for catabolism; Figure 3C, Supplementary Table 1). On the other hand, programmed cell death was observed more often in the enterocytes of the AP2 group compared with those of the normal group, as indicated by apoptosis-associated terms, and as the disease progressed, programmed cell death became more obvious, characterized by enrichment of necroptosis-, autophagy- and apoptosis-associated terms (Figure 3D and E). This observation was further supported by the dysregulation of several genes associated with programmed cell death, such as Casp3, Casp6, Casp7, Casp8, Apaf1, Xiap, and Ripk3 (Figure 3C). We also observed that complicated changes in protein metabolism occurred in the AP2 group. We subsequently used IHC assays to examine the expression of cleaved casp3, a marker of apoptosis, and p-MLKL, a marker of necroptosis, in the AP groups, and the results revealed a gradual increase in apoptosis and necroptosis in the AP groups (Figure 3F). These results indicated an obvious increase in cell stress and programmed cell death among enterocytes during the early stages of AP.

Figure 3 Enterocytes exhibit increased programmed cell death while maintaining function in the early stages of acute pancreatitis. A: Protein expression of ZO-1 in the AP1, AP, and normal small intestine; B: Representative enriched gene ontology (GO) terms for differentially expressed genes (DEGs) in enterocytes from the AP1 and normal groups on the basis of scRNA-seq data; C: Selected dysregulated genes associated with cell stress, mitochondrial function, and catabolism (P < 0.05). "Percent expressed" indicates the percentage of cells among groups, whereas "average expression" reflects changes in mean expression; D: Representative enriched GO terms for DEGs in enterocytes from the AP2 and normal groups on the basis of scRNA-seq data; E: Representative enriched GO terms for DEGs in enterocytes from the AP2 and AP1 groups on the basis of scRNA-seq data; F: Protein expression of cleaved caspase-3, Ki-67, and p-MLKL in the AP1, AP2, and normal groups in the small intestine; G: Selected DEGs associated with SLC family genes, nutrient digestion and absorption, and organic acid metabolism in enterocyte subcluster 2 (P < 0.05). "Percent expressed" indicates the percentage of cells among subclusters in enterocytes, whereas "average expression" reflects changes in mean expression. GO: Gene ontology. ^aP < 0.001.

To characterize each enterocyte subgroup comprehensively, we conducted a functional enrichment analysis for each subcluster. Enterocytes, the most abundant cell type in the intestine, are involved in essential functions, such as the production of digestive enzymes and the absorption and processing of nutrients, including sugars, lipids, and vitamins[10]. Notably, Subcluster 2 was significantly enriched in the intracellular translocation or localization of nutrient substances (Figure 3G). This subgroup exhibited high expression of SLC family genes responsible for transporting a wide range of nutrients and metabolites. Specifically, transporters such as SLC51A and SLC51B (bile salt transporters); RBP2 and SLC23A2 (vitamin transporters); SLC7A7, SLC38A2, and SLC7A9 (amino acid transporters); as well as SLC2A2 and SLC2A5 (sugar transporters), were prominently expressed in Subcluster 2. Furthermore, Subcluster 2 presented elevated expression of digestive enzymes, such as LCT, MGAM, ANPEP, ENPEP, and XPNPEP1, which are involved primarily in nutrient digestion and absorption in the small intestine. Consequently, Subcluster 2 emerged as the main subset of enterocytes responsible for nutrient digestion and absorption. Furthermore, Subcluster 2 also demonstrated high expression of transporters responsible for organic solute transportation (Slc22a5, Slc25a37, Slc47a1, Abcb1, Abcg2, Slc26a3) and nucleotide transportation (SLC46A1, and SLC22A5; Supplementary Figure 1A). GO term analysis further revealed that Subcluster 2 was enriched in various terms associated with digestion and substance transportation (Supplementary Figure 1B). Furthermore, Subcluster 2 displayed a notable association with organic acid metabolism (Supplementary Figure 1C). This observation was supported by the prominent expression of key genes involved in organic acid metabolism, such as CBR1, SULT1B1, MDH1, FABP1, and GSTA1 (Supplementary Figure 1D).

Notably, while the overall cell count across groups remained relatively stable, Subcluster 0 exhibited a discernible decrease, and Subcluster 1 showed an increase. Subcluster 2 remained unchanged, suggesting a primarily preserved absorption function of the intestine in the early stage of AP. These novel insights contribute to our understanding of AP pathogenesis and may have implications for targeting AP-related intestinal barrier dysfunction. Compared with that in the normal group, the number of cells in Subcluster 1 showed an increase in AP1 and AP2 compared to the normal group (accounting for 11.23%, 41.60%, and 41.85% of the total number of enterocytes, respectively). Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that Subcluster 1 was enriched in several programmed cell death pathways (apoptosis, necroptosis, and ferroptosis; Supplementary Figure 1E). Thus, Subcluster 1 represented the cells that died in the intestine during the early stage of AP.

The renewal of the intestinal epithelium is pivotal for maintaining the epithelial balance and is driven primarily by two cell populations: ISCs and transient amplifying (TA) cells[25]. We found that Subcluster 0 exhibited significant enrichment in cell cycle-associated functions (cell cycle, DNA replication) and substance metabolism-related functions, such as mRNA metabolism, ribonucleoprotein biogenesis, peptide biosynthesis, and amine metabolism, indicating its association with anabolism (Supplementary Figure 1F). Upon further examination of characteristic gene expression within Subcluster 0, we identified similarities to TA cells. This was evident through the high expression of genes such as Mki67, Ccnd1, Ube2s, Rack1, Cdk4, and Cbx3 (Supplementary Figure 1G). Interestingly, the number of cells in Subcluster 0 was significantly lower in the AP group than in the normal group. These findings suggest that intestinal barrier dysfunction in AP may be linked to decreased proliferation of TA cells. This previously overlooked aspect could hold potential value in addressing AP-related impairments to the intestinal barrier. Overall, our results showed that enterocytes were challenged by programmed cell death and impaired proliferation of TA cells during the early stage of AP.

The number of ISCs increased and are responsible for the regeneration of epithelial cells in the early stage of AP

Stem cells play crucial roles in self-renewal and the generation of daughter cells that differentiate into various epithelial cell types[25,26]. The amplification of stem cells may be necessary to maintain gut barrier function in the early stage of AP. ISCs highly express several well-known markers, such as Lgr5, Cd44, Cdk8, and Sox9 (Figure 4A). The number of ISCs was greater in the AP groups than in the normal group (accounting for 9.74%, 11.37%, and 4.27% of all identified cells in each group, respectively; Figure 2B and C). We verified the number of stem cells in various groups via immunofluorescence staining for Cd44 in the intestine (Figure 4B). TA cells are highly proliferative cells that are responsible for fast renewal of the intestinal epithelium[27] and are characterized by high expression of Mki67, Cdk1, Cdk4, Cdk2ap2, Cdk16, Ccnd1, and Pcna. We observed that the expression of these genes was synergistically decreased in the AP groups, indicating that the function of TA cells might be impaired in the early stage of AP (Figure 4C, Supplementary Tables 2 and 3). Although ISCs are believed to be cells that respond to damage by proliferating, our cell cycle analysis revealed a distinct blockade in the S stage rather than an increase in the cell cycle in the AP group (Figure 4D). Actually, the intricate process of intestinal restoration necessitates the coordination of cell proliferation and dedifferentiation. Research has demonstrated that differentiated villus cells and proliferative intervillus cells have the capacity to revert to adult stem cells[28], potentially contributing to the expansion of the ISCs population upon intestine injury. Interestingly, the genes essential for the dedifferentiation of enterocytes into ISCs, including Ascl2 and Il11ra1, were upregulated in the AP groups compared to the normal group (Supplementary Tables 2 and 3). These findings suggest that during the initial stages of AP, the expansion of ISCs primarily occurs through dedifferentiation rather than the proliferation of ISCs.

Figure 4 Increased number of intestinal stem cells responsible for regeneration of enterocytes in the early stages of acute pancreatitis. A: Representative genes in intestinal stem cells (ISCs), including Lgr5, Cd44, Cdk8, and Sox9, as demonstrated by the scRNA-seq data; B: Protein expression of CD44 in the AP1, AP2, and normal groups in the small intestine; C: Selected dysregulated genes in ISCs, including Mki67, Cdk1, Cdk4, Cdk2ap2, Cdk16, Ccnd1, and Pcna; D: Cell cycle analysis of ISCs across each group; E: Representative enriched gene ontology (GO) terms for differentially expressed genes (DEGs) in ISCs from the AP1 and normal groups on the basis of scRNA-seq data; F: Representative enriched GO terms for DEGs in ISCs from the AP2 and AP1 groups on the basis of scRNA-seq data; G: UMAP plot showing four ISC subclusters in the intestine; H: Selected genes in ISC subclusters, including Mki67, Cdk1, Cdk4, Cdk16, Ccnd1, Ascl2, Il11ra1, and Slc2a5. GO: Gene ontology; ISC: Intestinal stem cell. ^aP < 0.001.

To investigate the transcriptional changes in the stem cell compartment at the early stage of AP, DEGs were analyzed, and the list of upregulated and downregulated genes was used for GO enrichment analysis to observe functional changes across all groups. Compared with those in the normal group, ISCs in both the AP1 and AP2 groups were enriched in terms of cell proliferation (such as the cell cycle and DNA replication) and DNA repair-associated functions (such as DNA repair and nucleotide excision repair) (Figure 4E). This observation was further verified by the downregulation of associated genes (Dbb1, Cops5, Rbx1, and Ubc; as shown in Supplementary Table 4), indicating that stem cells in the intestine potentially experienced replication impairment and DNA repair impairment during AP. With the development of AP, ISCs exhibited increased programmed cell death and obvious responses to damage stimuli (Figure 4F). Thus, ISCs exhibit impaired DNA replication and repair in the early stages of AP, and as AP progresses, fragile ISCs might collapse, potentially leading to SAP-associated intestinal barrier dysfunction.

To delineate the diversity of ISCs in AP, ISCs were divided into subclusters, and 4 unique stem cell populations were revealed on the basis of transcriptional profiles (Figure 4G). The number of Subclusters 1 and 3 remained largely unchanged, the number of Subclusters 2 decreased, and the number of Subcluster 0 obviously increased in the AP groups compared with the normal group (Figure 4G). Mki67 and the CDK family (Cdk1, Cdk4, and Cdk16) were highly expressed in Subcluster 2, collectively indicating that the cycling population of ISCs was inhibited (Figure 4H). Subcluster 1 was highly expressed in Ascl2, Il11ra1, and enterocyte marker Slc2a5 (Figure 4H). Thus, in response to AP, there is a notable expansion of the stem cell population from dedifferentiated villus cells but not from self-renewal. This process leads to the production of mature intestinal cells, contributing to restoration of the gut barrier.

Cytotoxic T cells in the intestine enhanced the antibacterial ability in AP

T cells are essential players in gut immunity[29], with cytotoxic T cells being crucial for the elimination of bacteria, particularly intracellular ones[30]. We explored the number and phenotype changes of them in the intestine during the early stage of AP. The number of cytotoxic T cells was greater in the AP1 and AP2 groups than in the normal group (13.05%, 13.67% and 6.05% of all identified cells in each group, respectively) (Figure 2B and C). We first explored the functional diversity of cytotoxic T cells in AP and normal intestines through GO functional analysis. Cytotoxic T cells in the AP group were enriched in the innate immune response, adaptive immune response, and phagocytosis (Figure 5A). Compared with AP1, AP2 amplified the antibacterial immune response, as evidenced by enrichment of pathways such as epithelial cell signaling in Helicobacter pylori infection and the cytosolic DNA-sensing pathway (Figure 5B). Consistently, cytotoxic T cells in AP highly expressed genes related to the phagocytosis process (Apoa1, Actg1, and Coro1A) and the lysis or antimicrobial activity of pathogens, such as Gzma, Gzmb and Reg3b (Figure 5C). During AP, cytotoxic T cells highly expressed genes associated with the electron transport chain and ATP production, such as MT-ND5, MT-CO1, ATP5b and Cox7a2 (Supplementary Tables 5 and 6). During the bactericidal process, cytotoxic cells utilize ATP to fuel various effector mechanisms. For example, ATP is crucial for the mobility and migration of cytotoxic cells toward infected cells[31]. These data demonstrated that cytotoxic T cells exhibited enhanced antibacterial ability in AP, which is beneficial for reducing the risk of bacterial translocation.

Figure 5 Enhanced antibacterial ability of intestinal cytotoxic T cells in acute pancreatitis. A: Representative enriched gene ontology terms for differentially expressed genes (DEGs) in cytotoxic T cells from the AP1 and normal groups based on scRNA-seq data; B: Representative enriched Kyoto Encyclopedia of Genes and Genomes pathways for DEGs in cytotoxic T cells from the AP2 and AP1 groups; C: Selected dysregulated genes, including Apoa1, Actg1, Coro1a, Gzma, Gzmb, and Reg3b, in cytotoxic T cells; D: UMAP plot showing 6 cytotoxic T-cell subclusters in the intestine; E: Expression of selected genes in cytotoxic T-cell subclusters, including Gzma, Gzmb, Ifng, Apoa1, Coro1a, Cd27, Cd8a, Cd69, Lag3, Cd244, Cd38, Itga4, Ccr9, Cxcr3, Cd3e, Cd3d, Il2, Mki67, Cdk1, Cdk4, and Cdk6; F: Percentage of cells in each cytotoxic T-cell subcluster. GO: Gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes.

The cytotoxic T cells were further divided into 6 subclusters (Figure 5D). Previous studies have shown that cytotoxic T cells can be subtyped by effector cytotoxic T cells, memory cytotoxic T cells, resident memory cytotoxic T cells, stem-like cytotoxic T cells, and exhausted cytotoxic T cells[32]. Effector cytotoxic T cells are a subset of T cells that play crucial roles in the immune response against pathogens and infected cells. The markers of effector cytotoxic T cells, such as Gzma, Gzmb, and Ifng, were expressed mainly in Subcluster 5 (Figure 5E). The genes associated with the phagocytosis process (Apoa1 and Coro1a) were expressed mainly in Subclusters 5. Memory cytotoxic T cells are a subset of T cells that have previously encountered an antigen and developed a memory response. These cells play crucial roles in providing long-term immunity against specific pathogens or antigens. Memory cytotoxic T cells (Cd27) were expressed in Subclusters 2. Resident memory cytotoxic T cells are a specialized subset of memory T cells that reside in peripheral tissues, such as the skin, mucosa, and other nonlymphoid organs[33]. These cells provide rapid and localized immune responses to reinfection at the site of initial infection with pathogens[33]. Genes specifically expressed in resident memory cytotoxic T cells (Cd8a, Cd69) were attributed to Subcluster 2. Exhausted cytotoxic T cells are a subset of T cells that have become dysfunctional and exhibit reduced effector functions owing to prolonged exposure to persistent antigen stimulation. The genes associated with exhausted cytotoxic T cells (Lag3, Cd244, and Cd38) were expressed mainly in Subclusters 4. On the other hand, cytotoxic T cells express adhesion molecules (such as ITGA4) and chemokine receptors (such as CCR9 and CXCR3) to facilitate their migration into the intestines. The main populations expressing these genes were Subclusters 2. Furthermore, essential genes involved in orchestrating the activation of cytotoxic T cells, such as Cd3e, Cd3d, and IL-2, were expressed in Subclusters 2 and 5. These subclusters encompass effector cytotoxic T cells, memory cytotoxic T cells, resident memory cytotoxic T cells, and exhausted cytotoxic T cells. Additionally, stem-like cytotoxic T cells, which possess self-renewal capacity and multipotency similar to that of stem cells (high expression of Mki67, Cdk1, Cdk4, and Cdk6), were mainly found in Subclusters 2, with a stable cell count in the AP group. The number of Subcluster 0, 1, 3, and 5 in the AP group was more than twice that of the normal group (Figure 5F). Thus, our data indicate an enhanced antibacterial ability of cytotoxic T cells in the AP group due to increased activation, infiltration, and likely proliferation.

The ability of NK cells in the intestine to defend against infected cells and pathogens is increased in AP

NK cells play a vital role in maintaining immune homeostasis and orchestrating effective inflammatory responses to stimuli in the intestine[34,35]. By doing so, they ensure immunity at environmental interfaces and contribute to the preservation of intestinal barrier integrity. Previous studies have demonstrated that NK cells are responsible for pathogen defense, aid in tissue repair, and act as regulators of mucosal inflammation[36]. The number of NK cells increased from 10.21% to 16.49% and 14.80% of all identified cells in the normal, AP1, and AP2 intestines, respectively (Figure 2B and C). The number of NK cells was then verified by IF staining for Cd56, a marker of NK cells (Figure 6A). The DEGs associated with each group of NK cells are shown in Supplementary Table 7 and 8. To understand the key signaling pathways of NK cells that are dysregulated in AP, we performed functional enrichment analysis. During AP, NK cells in the intestine are significantly associated with Th cell differentiation and programmed cell death (Figure 6B and C). Previous studies have shown that NK cells can modulate the differentiation and balance of Th cell subsets through the production of cytokines and direct cell-cell interactions[37]. This crosstalk between NK cells and Th cells contributes to the regulation of intestinal immunity, inflammation, and the preservation of intestinal barrier integrity. The expression of genes involved in Th1 differentiation, including Ifng, Stat1, and Ccr5, in NK cells was downregulated in the presence of AP (Figure 6D). Similarly, for Th17 differentiation, the expression of Stat3, Batf, and Lat was also suppressed (Figure 6D). Consequently, during the early stages of AP in the intestine, the differentiation process of Th cells, assisted by NK cells, is inhibited. However, in contrast to this inhibition, the expression of Gzma and Gzmb, two genes crucial for the NK cell-mediated targeted elimination of infected cells, was upregulated in AP. This observation indicates increased activation of a cellular elimination mechanism aimed at clearing potentially infected cells or pathogens from the intestine by NK cells during AP.

Figure 6 Enhanced ability of natural killer cells to defend against infected cells and pathogens in acute pancreatitis. A: Protein expression of CD56 in natural killer (NK) cells from the AP1, AP2, and normal groups in the small intestine (left) and semiquantitative analysis (right); B: Representative enriched gene ontology (GO) terms for differentially expressed genes (DEGs) in NK cells from the AP1 and normal groups based on scRNA-seq data; C: Representative enriched GO terms for DEGs in NK cells from the AP2 and normal groups on the basis of scRNA-seq data; D: Expression of selected genes, including Ifng, Stat1, Ccr5, Stat3, Batf, and Lat, in NK cells across each group; E: UMAP plot showing five NK cell subclusters in the intestine; F: Expression of selected genes, including Klrb1, Il2ra, Cd226, and Prf1, in NK cell subclusters; G: Percentage of cells in each NK cell subcluster. GO: Gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; NK: Natural killer. ^aP < 0.001.

We further divided the NK cells into 5 distinct subclusters on the basis of their transcriptional profiles (Figure 6E). The Cd56 high-expressing NK cell Subcluster is characterized by elevated levels of Cd56, lower cytotoxicity, and a predominant capacity for cytokine production, indicating its immunoregulatory functions. Notably, Klrb1 and Il2ra, which are markers for this specific NK cell subcluster, were expressed primarily in Subclusters 0 and 2 (Figure 6F). These findings suggest that Subclusters 0 and 2 may play crucial roles in the immunoregulatory functions of NK cells during AP. On the other hand, CD56dim NK cells exhibit increased cytotoxic activity and are involved primarily in target cell killing, resulting in more diverse functional responses. Cd226, a marker for CD56dim NK cells, was highly expressed in Subclusters 3, and Prf1, a gene involved in target cell lysis, was expressed mainly in Subcluster 4, suggesting that Subclusters 3 and 4 can be defined as CD56dim NK cells (Figure 6F). The overall number of Subclusters 0 and 2 did not significantly change, whereas the number of Subclusters 3 and 4 noticeably increased in the AP group compared with the normal group (Figure 6G). This finding indicates that enhanced elimination of infected cells or pathogens was predominantly observed in the AP group. In conclusion, our data suggest that NK cells contribute to maintaining gut barrier functions during AP through their cellular elimination mechanism.

Intestinal MCs are activated early in AP

MCs are involved in the response to both external and internal stimuli and the regulation of intestinal barrier function and integrity[38]. We observed a significant reduction in MC numbers as early as the initial stage of AP, as revealed by single-cell sequencing (Figure 2A). The percentage of MCs decreased from 9.07% in the normal group to 2.2% and 1.76% in the AP1 and AP2 groups, respectively. To validate these findings, we performed immunofluorescence staining for c-Kit to examine the number of MCs in each group. Surprisingly, we found that the number of MCs did not actually change in the AP1 and AP2 groups (Figure 7A). The reasons for this discrepancy are discussed in the following text. MCs are long-lived granulated immune cells that mature in tissues and can release various mediators or degranulate in response to stimuli[39]. GO term analysis revealed that MCs in the AP group were enriched in the protein process and response to cytokines (Figure 7B). KEGG signaling pathway analysis consistently revealed the enhancement of the protein process in the AP group (Figure 7C). Enhanced MCs activation in the AP group resulted in increased expression and release of mediators such as IL-10 and Cxcl10 (Figure 7D), which might affect intestinal permeability. Interestingly, the most significant difference between the AP1 and AP2 groups was that AP2 resulted in increased apoptosis and cell cycle progression (Figure 7E). This finding was supported by enrichment in apoptotic processes and elevated expression of apoptosis-associated genes such as Fos, Pecam1, and Hes1 (Supplementary Table 9). Hence, MCs during the early stages of AP demonstrate characteristics of cell activation and potential cell death as the disease progresses.

Figure 7 Early activation of mast cells in the intestine of acute pancreatitis. A: Protein expression of cKit in the small intestine from the AP1, AP2, and normal groups (upper left) and semiquantitative analysis (upper right). Toluidine blue staining of the small intestine (lower left) and semiquantitative analysis of activated mast cells (MCs; lower right); B and C: Representative enriched Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for differentially expressed genes (DEGs) in MCs from the AP1 and normal groups on the basis of scRNA-seq data; D: Violin plots showing the expression of IL-10 and Cxcl10 across each group in the small intestine; E: Representative enriched GO terms and KEGG pathways for DEGs in MCs from the AP2 and AP1 groups on the basis of scRNA-seq data; F: Selected dysregulated genes in MC subclusters, including Pclaf, Pcna, Top2a, Mki67, Il6, Tnf, Il10, Cxcl10, Ccl3, Ccl11, Ccl5, Ccl4, Ccl2, Alox12, Ltc4s, Ptgs1, Gata2, Hdc, and Cpa3. GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; MC: Mast cell; NS: No significance. ^aP < 0.001.

To further investigate the functional variability in MCs, the identified MCs were subclustered into 6 unique clusters. The mucosal MCs marker genes Cpa3 and Hdc were expressed mainly in Subclusters 0, 3, and 4 (Figure 7F). The transcription factor for the specification and differentiation of MCs, Gata2, was expressed mainly in Subclusters 4 and 5 (Figure 7F). The genes responsible for prostaglandin biosynthesis (Ptgs1) and histamine biosynthesis (HDC) were expressed in Subcluster 4, and those for leukotriene biosynthesis (Ltc4s and Alox12) were expressed in Subcluster 5. MCs can express genes encoding diverse cytokines, chemokines, and growth factors, which are essential for immune responses and inflammation. The expression profiles of these genes exhibited significant heterogeneity within the MCs. Specifically, Ccl2 was expressed primarily in Subcluster 5, whereas Ccl44 and CclL5 were expressed in Subclusters 2 and 3. Ccl11 expression was observed specifically in Subcluster 5, whereas Ccl3 and Cxcl10 were expressed in Subclusters 0, 2, and 3. Furthermore, IL-10 was detected in Subcluster 0, and tumor necrosis factor-α (TNF-α) was predominant in Subcluster 3, whereas IL-6 expression was prominent in Subcluster 3. The gene markers for MCs with high proliferation rates, including Mki67, Top2A, Pcna, and Pclaf, were expressed mainly in Subclusters 0 and 3 (Figure 7F). Overall, MCs in the intestine are highly heterogeneous during AP, and the activation of MCs might be a marker for AP.

CCL5 derived from MCs promoted the infiltration of neutrophils and macrophages

MCs, which are located in close proximity to epithelial cells, play crucial roles in the regulation of intestinal barrier function[38]. MCs can influence immune responses in the gut, regulating the activity of other immune cells, such as neutrophils and macrophages, and are involved in maintaining intestinal homeostasis and defending against pathogens[40-42]. The role of MCs in gut barrier dysfunction caused by AP has not been evaluated. Although single-cell sequencing revealed a noticeable decrease in the number of MCs in the AP groups, our immunofluorescence staining of c-Kit revealed that the MCs count remained unchanged following AP. This discrepancy can be attributed to the fact that most MCs were activated in the AP groups, as shown by toluidine blue staining, and exhibited a relatively fragile state during the dissociation process. As a result, fewer MCs were detected by single-cell sequencing because of their reduced stability and increased likelihood of damage. Next, we examined whether the activation of MCs could affect gut barrier function during AP. Cromolyn sodium, an MCs stabilizer, was used as a pretreatment before the induction of AP[43], and 12 hours after the first injection of caerulein, samples were collected (Figure 8A). Toluidine blue staining demonstrated that the administration of Cromolyn sodium protected against the activation of MCs in the gut (Figure 8B). The integrity of gut barrier function was assessed by measuring the expression of bacterial 16S rDNA in the bloodstream. The results revealed that treatment with Cromolyn sodium significantly reduced the expression of bacterial 16S rDNA (Figure 8C), indicating an improvement in gut barrier function. H&E staining revealed that the damage to intestinal epithelial cells was alleviated by Cromolyn sodium (Figure 8D). To visually investigate whether stabilization of MCs can improve intestinal barrier dysfunction during AP, we utilized transgenic mice with fluorescently labeled intestinal epithelial cells, allowing for the observation of intestinal epithelial continuity. These transgenic mice were generated by crossing mice expressing the reporter gene (Rosa26-CAG-LSL-tdTomato) with mice harboring intestinal epithelial cell-specific Cre recombinase (PVillin-Cre), named Villin-tdT mice (Figure 8E and F). At 6 weeks of age, these mice were given an intraperitoneal injection of tamoxifen to induce specific expression of the red fluorescent protein tdTomato in intestinal epithelial cells (Figure 8G). The results demonstrated that Cromolyn sodium significantly improved the integrity of intestinal epithelial cells (Figure 8H). Additionally, when proinflammatory factors (IL-1β and IL-6) in intestinal tissue were examined, inhibition of mast cell activation significantly reduced their expression levels (Figure 8I and J). The infiltration of neutrophils and macrophages was obviously inhibited by the administration of Cromolyn sodium (Figure 8K and L). In summary, these findings indicate that MCs play a crucial role in AP-associated dysfunction of the intestinal barrier and that MC activation contributes to intestinal barrier damage by induction of the pro-inflammation effect in the intestine.

Figure 8 Stabilization of mast cells alleviates intestinal damage and inflammation in the intestine of acute pancreatitis. A: Workflow schematic for mouse experiments; B: Representative images of toluidine blue staining of small intestine samples from the Cromolyn sodium-treated and control groups (left) and semiquantitative analysis of activated mast cells (right); C: Relative levels of 16S rDNA in plasma from the Cromolyn sodium-treated and control groups; D: HE staining of the small intestine in the Cromolyn sodium-treated and control groups; E: Gene targeting strategy for Rosa26-CAG-LSL-tdTomato mice; F: Genotype identification of Rosa26-CAG-LSL-tdTomato mice; G: Workflow schematic for mouse experiments of Villin-tdT mice; H: Immunofluorescence images of the small intestine in the Cromolyn sodium-treated and control groups; I and J: mRNA expression levels of IL-1β and IL-6 in the small intestine of the Cromolyn sodium-treated and control groups; K: CD68 staining of the small intestine in the Cromolyn sodium-treated and control groups (left) and semiquantitative analysis (right); L: Ly6G staining of the small intestine in the Cromolyn sodium-treated and control groups (left) and semiquantitative analysis (right). ^aP < 0.001; ^bP < 0.05; ^cP < 0.01.

Next, we further analyzed the molecular mechanisms underlying MC-mediated intestinal barrier dysfunction in AP. We conducted a differential gene analysis between MCs from AP and normal tissues and identified significant upregulation of CCL5, a chemokine known to recruit immune cells to sites of inflammation or injury[44]. The violin diagram illustrates the differential expression levels of CCL5 between AP and normal MCs (Figure 9A). We further demonstrated that, compared with that in the control intestine, the mRNA expression of CCL5 in the intestine of AP mice was significantly greater (Figure 9B). To demonstrate the expression of CCL5 derived from MCs, we utilized flow cytometry and observed a significant increase in CCL5 levels in MCs from the AP groups compared with those from the normal group (CCL5-positive rates were 12.6%, 24.9%, and 28.8% in the normal, AP1, and AP2 groups, respectively) (Figure 9C). Furthermore, in co-cultures of primary acinar cells with MC/9 cells, a type of mouse mast cell line, for 6 hours, a significant increase in CCL5 expression was observed (Figure 9D).

Figure 9 CCL5 derived from mast cells promotes infiltration of neutrophils and macrophages. A: Violin plots displaying the expression of CCL5 across each group in the small intestine; B: CCL5 mRNA expression levels in the small intestine of the AP1, AP2, and normal groups; C: Flow cytometry analysis of CCL5 expression in CD117-positive cells in the small intestine; D: CCL5 mRNA expression levels in MC/9 cells treated with varying amounts of primary acini; E: Identification of genotypes in c-Kit^W-sh/W-sh mice; F: Weight comparison between c-Kit^W-sh/W-sh mice and wild-type mice; G: Workflow schematic for mouse experiments; H: Immunofluorescence staining confirms the successful inoculation of MC/9 cells into the small intestine; I: HE staining of the pancreas in c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (left) and semiquantitative analysis (right); J: Amylase and lipase activity in plasma from c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (CCL5 shRNA or control); K: HE staining of the intestine of c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (CCL5 shRNA or control); L: 16S rDNA expression in plasma from c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (CCL5 shRNA or control); M: Immunofluorescence staining of CD68 in the small intestine of c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (CCL5 shRNA or control); N: Immunofluorescence staining of Ly6G in the small intestine of c-Kit^W-sh/W-sh mice inoculated with MC/9 cells (CCL5 shRNA or control). MC: Mast cell; NS: No significance. ^aP < 0.001; ^bP < 0.05; ^cP < 0.01.

To validate the role of MC-derived CCL5 in vivo, we introduced CCL5-knockdown MC/9 cells into MC-devoid mice (c-Kit^W-sh/W-sh) via tail vein injection. The genotype of the c-Kit^W-sh/W-sh mouse is shown in Figure 9E. The c-Kit^W-sh/W-sh mice exhibit similar weights compared to wild-type mice (Figure 9F). These MC/9 cells were infected with a lentivirus expressing GFP, resulting in the production of green fluorescence. Three days after injection, frozen sections of the small intestine were collected (Figure 9G), confirming the presence of GFP-expressing MC/9 cells entering the intestines (Figure 9H). Research has revealed the crucial role of MCs in the pancreas in initiating local and systemic inflammatory responses during the early stages of AP. Our study demonstrated that CCL5 in MCs is not the primary determinant in this process, and that no significant difference in pancreatic pathology or plasma amylase or lipase was observed (Figure 9I and J). Interestingly, in the intestine, MC/9 deficiency in CCL5 significantly enhanced the integrity of the gut barrier and reduced bacterial translocation compared to the control (Figure 9K and L). Moreover, compared with the control treatment, CCL5 knockdown resulted in a decrease in the number of neutrophils and macrophages in the intestine upon induction of AP (Figure 9M and N). In conclusion, our findings highlight the importance of CCL5 in MCs for facilitating the infiltration of neutrophils and macrophages during the early stages of AP. Intervening with or targeting CCL5 could offer therapeutic benefits by protecting the integrity of the gut barrier function in AP.

DISCUSSION

Elaborate maintenance of gut barrier function relies on a diverse array of genetic factors, including genes crucial for cell commitment, junctional complexes, mucus production and secretion, pathogen sensing, and elimination of reactive oxygen species and infected cells[45]. Moreover, specific environmental factors, such as bacterial infections, nutrient components, and trypsin secretion from the pancreas, also contribute to the regulation of gut barrier function[46,47].

The gut epithelium, lining the lumen of the intestine, serves a dual function: Acting as a selective filter that allows beneficial nutrients, electrolytes, and water to pass into the circulation while also functioning as a protective barrier that shields against harmful intraluminal substances such as bacteria and toxins[48]. Intestinal epithelial cells form a tightly interconnected barrier through the presence of desmosomes, adherens junctions (mediated by epithelial–catenin interactions), and tight junctions (consisting of occludins, claudins, junctional adhesion molecules, and tricellulin). The appropriate level of programmed cell death in epithelial cells is crucial for maintaining homeostasis[49]. For example, villus tip epithelial cells typically undergo anoikis, whereas infected epithelial cells undergo pyroptosis. However, excessive cell death through apoptosis, necroptosis, and pyroptosis can result in compromised epithelial integrity, erosion, dysbiosis, and the dissemination of pathogens and toxins throughout the system[49]. Excessive epithelial cell death and an acute inflammatory response are key features of intestinal barrier dysfunction in AP, resulting from pancreatic enzymes, microcirculatory disturbances, and a burst of inflammatory mediators[50]. On the other hand, the levels of Claudin-4, a vital protein involved in tight junction formation, and aquaporins, which are water channel proteins, are reduced in the intestine, thereby contributing to an improvement in intestinal permeability in AP[51]. In our analysis, we detected three types of programmed cell death, namely, apoptosis, necroptosis, and ferroptosis, in the intestine at the early stage of AP. Recent studies have suggested that various types of programmed cell death can converge, namely, PANoptosis, which involves pyroptosis, apoptosis, and necroptosis under certain conditions[52]. Further exploration is necessary to determine whether programmed cell death in the intestine is triggered by a single uniform molecular mechanism and to investigate whether inhibition of this mechanism would be beneficial for reducing intestinal barrier dysfunction during AP.

MCs are essential immune cells strategically located near blood vessels and nerve endings, particularly near the epithelial surface and within the lamina propria[53]. MCs participate in both innate and adaptive immunity in the intestine via the selective or nonselective release of various inflammatory mediators, such as histamine, cytokines (IL-4, IL-5, and IL-6), proteases (tryptase and chymase), and chemotactic factors (CCL2, 3, 5, and CXCL8)[54]. MCs can impair the gut barrier by various mechanisms, including the release of the abovementioned inflammatory mediators, activation of the immune response of macrophages and neutrophils, induction of epithelial cell death, and modulation of mucus production. For example, acute psychological stress can increase small intestinal permeability by activating MCs in the intestine, while the administration of an MC stabilizer has the potential to reverse this process[55]. MCs also directly influence epithelial cells via tryptase-mediated downregulation of junctional adhesion molecule-A expression in epithelial cells, and the induction of epithelial cell death via TNF-α contributes to intestinal barrier dysfunction[56]. Recent studies have investigated the therapeutic potential of targeting MC activation using drugs such as costunolide, metformin, and disodium cromoglycate to treat intestinal barrier dysfunction, which has beneficial effects in various disease models[38]. However, the role and mechanism of MCs in AP-associated gut barrier dysfunction are still unclear. A previous study indicated that severe necrotizing AP is associated with reduced villous height and mast cell numbers in the small intestine[57]. Subsequent investigations revealed that the administration of C48/80, an MC activator, resulted in significant intestinal barrier dysfunction in mice with AP. However, the exact role of MCs in AP, especially in the early stage of AP, remains unclear[58]. In our study, we observed a clear early activation of MCs in the gut in AP. In AP, trypsinogen is released early into pancreatic tissue or serum[59]. The premature activation of pancreatic enzyme can trigger the digestion of pancreatic tissue, leading to damage and an inflammatory response, with the release of numerous inflammatory mediators like IL-6, and TNF-α[60], which affect MC activity in a variety of ways[61]. In addition, early activation of lipase leads to the digestion of adipose tissue, releasing large amounts of free fatty acids[62]. Free fatty acids can induce MC activation by activating the toll-like receptor pathway[63]. Consequently, we hypothesize that the early activation of intestinal MCs in AP may be related to the premature activation of pancreatic enzymes and the release of inflammatory factors. We confirmed through further functional experiments that the application of MCs membrane stabilizer Cromolyn sodium can effectively improve intestinal barrier dysfunction during AP. Inhibiting MCs can effectively reduce the severity of intestinal barrier dysfunction by reducing the infiltration of inflammatory cells and epithelial cell death. However, MCs do not have a significant impact on pancreatic pathology.

CCL5, a crucial chemokine, plays an essential role in immune responses and inflammation by effectively recruiting diverse immune cells to sites of infection and inflammation[44]. CCL5 is produced by a variety of cell types, with MCs, T cells, and macrophages being the most prevalent types of cells that express CCL5[64]. CCL5 has the ability to bind to multiple receptors, including CCR1, CCR3, CCR4, and CCR5, with the highest affinity observed for CCR5[64]. A growing body of evidence indicates that CCL5 plays a crucial role not only in facilitating macrophage/monocyte infiltration into inflammatory sites but also in conferring antiapoptotic properties to inflammatory cells and inhibiting M2 macrophage polarization[65,66]. Our analysis revealed high expression of CCL5 in MCs, NK cells, and cytotoxic T cells within the intestine, with subsequent upregulation specifically observed in MCs but not in NK cells or cytotoxic T cells during the development of AP. We further demonstrated that introducing CCL5-knockdown MC/9 cells into MC-deficient mice resulted in impaired infiltration of neutrophils and macrophages. Given the limited evidence indicating a direct role of CCL5 in neutrophil recruitment to inflammatory sites, our speculation is that the effects observed following CCL5 knockdown in MC/9 cells could be attributed to the attenuation of intestinal damage or an indirect pathway.

The intestinal barrier maintains homeostasis through a highly self-renewing epithelium, facilitated by the presence of stem cells and their differentiated cells, which are regularly packed within crypts and villi[50]. Enhanced restoration of the damaged epithelium is crucial in addressing intestinal barrier dysfunction, highlighting the significant role and importance of ISCs, which could address multiple pathological defects. For example, maintaining an adequate number of goblet cells derived from ISCs through tightly regulated gene expression during terminal differentiation is crucial for ensuring the secretion of mucus, a gel-like substance in the intestine. Defects in the environmental sensor aryl hydrocarbon receptor result in uncontrolled proliferation of ISCs, impaired cell differentiation, a significant decrease in goblet cells, and reduced expression of Muc2[14]. In our work, in the early stage of AP, we observed a significant increase in the number of ISCs compared with that in the normal intestine. Additionally, putative markers for ISCs and differentiated cells, including Paneth cells and goblet cells, presented high expression levels. These findings suggest that mechanisms promoting ISC expansion and the subsequent maintenance of terminally differentiated cells are activated during AP in the intestine. On the other hand, we also observed that ISCs activated the apoptosis pathway in AP and significantly downregulated the expression of antiapoptotic genes (HSPA1A, HSPA1B, PPIA, PRDX1, and PGX1). As apoptosis plays a critical role in determining the cell fate of stem cells, the balance between apoptosis and cell survival signals is crucial for regulating stem cell homeostasis and controlling their differentiation and maintenance. The number of each enterocyte subpopulation differs; thus, we speculated that the activation of ISC apoptosis might be involved in epithelial homeostasis in AP.

In response to damage in the intestine, a subsequent phase of increased ISC proliferation occurs, but the duration and kinetics of this proliferative response can differ depending on various factors. Previous works have demonstrated that +4 cells (positioned above Paneth cells and responsible for injury-induced regeneration) are not damage sensitive, a property proposed to protect against damage and preserve the capacity for intestinal repair[25]. Intestinal restoration is a complicated process that requires the coordination of cell proliferation and differentiation. Studies have shown that differentiated villus cells, as well as proliferative intervillus cells, can revert to adult stem cells, which might also contribute to the expansion of the ISC population in the early stage of AP[67]. Thus, further research is needed to elucidate whether the increased number of ISCs is associated with the dedifferentiation of mature epithelial cells and whether enhanced differentiation of TA cells is an early event in maintaining intestinal epithelial homeostasis during AP.

Furthermore, we analyzed the intestinal cytotoxic T cells and NK cells in the early stages of AP. The study found an enrichment of these cells in the AP group, which enhanced the intestinal anti-infection and pathogen clearance capabilities during AP. The role of immune cells in the pathogenesis of AP is significant and should not be overlooked; potentially directly involved in AP progression and mediating damage to the pancreas and extrapancreatic organs[68]. Cytotoxic T cells are integral to the immune response in the gut[69]. In the inflammatory environment of AP, cytotoxic T cells decrease as the disease progresses, suggesting that they play a protective role in AP[70,71]. These cells exhibit increased cytotoxicity against infected and damaged cells, contributing to controlling pathogen infections and maintaining intestinal homeostasis[72]. The pancreas is recognized as a primary source of pro-inflammatory cytokines, capable of inducing systemic T cell activation in the early stages of AP[73]. Therefore, in this study, the upregulation of intestinal cytotoxic T-cell activation is closely linked to the increased serum release of AP cytokines. This enhanced activity is crucial for preventing secondary infections and maintaining the integrity of the intestinal mucosa. The interaction between cytotoxic T cells and other immune cells, such as dendritic cells and macrophages, is essential in modulating their antimicrobial functions. Dendritic cells present antigens to cytotoxic T cells, priming them for enhanced activity[74]. Macrophages stimulate CD4⁺ T cells to secrete IFN-γ and IL-2 through the secretion of cytokines like IL-12, where IFN-γ can activate other resting macrophages and IL-2, along with other cytokines, can promote CD8⁺ T cells to activate into cytotoxic effector T cells[75]. This synergistic interaction ensures that cytotoxic T cells are effectively primed and capable of responding to the diverse challenges presented during AP. The ability of these cells to adapt and enhance their antimicrobial functions is vital for protecting the gut during the early stages of AP, contributing significantly to the host's defense mechanisms. In addition, NK cells are critical components of the innate immune system. Our research shows that in the intestine, their role in combating infections is enhanced during AP. The increase in NK cell activity during AP is crucial for early immune response. In the early stages of AP, memory-like circulating NK cells migrate to the pancreas and trigger a rapid immune response to the same antigen encountered again, and the low number of NK cells is closely related to the development of secondary infections in AP patients[76]. NK cells also play an important role in the immune defense of the intestine and can interact with various cell types, including fibroblasts, macrophages, dendritic cells, and T lymphocytes, helping to maintain immune homeostasis and the development of effective immune responses[77]. This interaction ensures that NK cells are effectively activated and capable of responding to the pathogenic challenges presented during AP. The ability of NK cells to rapidly respond and eliminate infected cells is essential for controlling inflammation and preventing further tissue damage. Their enhanced role in early intestinal immunity during AP underscores their importance in maintaining gut homeostasis and preventing secondary infections.

CONCLUSION

In conclusion, our study provides a comprehensive picture of the transcriptome of small intestine cells during the early stage of AP, revealing a total of 33,232 cells across all samples and 17 main clusters. Through our investigation, we established that MCs were promptly activated in the intestine, and we identified CCL5 derived from MCs as an essential factor contributing to the infiltration of inflammatory cells and the progression of gut barrier dysfunction. The application of mast cell membrane stabilizers in the early stages of AP may be an effective measure to prevent intestinal barrier disruption. In addition, there are some interesting discoveries, such as the number of ISC expansions in the early stage of AP, which might be attributed to epithelial dedifferentiation but not stem cell proliferation. The decrease in TA cell proliferation might be explained by the promotion of cell differentiation but not replenishment by ISCs. The absorption function of the intestine was preserved during the early stage of AP, and the augmented activation of NK cells and cytotoxic T cells aimed at clearing potentially infected cells or pathogens from the intestine. By dissecting the early responses involved in the AP-associated intestine, we can gain valuable insights into the underlying mechanisms and develop novel strategies for addressing gut barrier dysfunction.