Introduction

Acute pancreatitis (AP) is a common, life-threatening disease of the pancreas. Non-mild AP is characterized by acute onset, rapid progression, and a poor prognosis1,2. The course of AP often lasts from tens of days to months, depending on the severity of the disease. Based on clinical characteristics, AP is generally divided into three phases: acute reaction, progression, and recovery. While not all patients experience a three-phase course, every surviving patient undergoes the recovery phase3.

Pancreatic regeneration and functional repair may begin during the acute reaction phase. When inflammation occurs, organ repair and regeneration mechanisms are activated, which continue throughout the course of the disease. During the acute reaction and progression phases, the pathophysiological processes include injury and necrosis of pancreatic cells, as well as inflammatory responses. Upon entering the recovery phase, pancreatic regeneration and functional repair become the dominant pathophysiological processes, while inflammatory responses gradually weaken or attain a balance4,5. At this stage, relief of discomfort, regression of inflammation, and recovery of gastrointestinal function occur as responses to treatment, enhancing the quality of life of patients6. However, the risk of deterioration exists.

The human intestinal barrier system consists of the gut microbiota, the mucosal immune system, and the mucosal epithelium. The stability of gut microbiota depends on intestinal barrier function7. Gut microbiota-derived metabolites are crucial components involved in intestinal-pancreatic axis interactions8. Inflammatory responses, bacterial translocation, and impairment of the intestinal barrier are essential for AP progression and complications9,10. Increased intestinal permeability and subsequent translocation of gut-derived bacteria to pancreatic tissue may aggravate the inflammation in AP11. Escherichia coli, Enterococcus, and Enterobacteriaceae are the primary pathogens responsible for secondary infections in AP12,13. Additionally, dysbiosis of the intestinal microbiota reduces the levels of short-chain fatty acid-producing bacteria, which affects the integrity of the intestinal barrier and worsens clinical conditions14. Indeed, dysbiosis of the gut microbial community or mucosal damage in patients with AP could elevate intestinal permeability, leading to the translocation of bacteria from the gut to the blood or nearby tissues, such as the pancreas, thereby increasing the risk of pancreatic infection and exacerbating inflammation15. Most pancreatic and extra-pancreatic organ infections originate in the gut, thereby triggering inflammatory responses16. Although previous studies have reported potential routes and mechanisms underlying AP development, more information is required on intestinal dysfunction over the entire course of AP, especially during the recovery phase17.

In this study, we employed metagenomic sequencing technology to investigate composition and function of the gut microbiota in patients with AP of varying severities during the acute and recovery phases, revealing the complex interactions between gut microbes and the host.

Materials and methods

Patients

From January 2021 to September 2021, patients diagnosed based on the 2012 Revision of the Atlanta Classification and admitted to Peking Union Medical College Hospital, Beijing, China, were included in this study. Patients were enrolled within 48 h of disease onset and followed up for 30 d to observe their recovery. Patients were excluded if they had a history of immunodeficiency, inflammatory bowel disease, asthma, irritable bowel syndrome, celiac disease, gastroenteritis, colon cancer, human immunodeficiency virus infection, necrotizing enterocolitis, or arthritis. The criteria for individuals recovering from AP were as follows: (1) relief of abdominal pain in patients with mild AP (MAP); (2) decrease in the improved Marshall score to 0 points in patients with moderately severe AP (MSAP) or severe AP (SAP); and (3) tolerable enteral nutrition. The control group was defined as described in our previous study12. Four healthy individuals were recruited and matched for age, sex, and body mass index. Written informed consent was obtained from all the patients. Approval was obtained from the Institutional Review Board of Peking Union Medical College Hospital (No. JS1826).

Sample collection

A flowchart of the study is shown in Fig. 1. Rectal swabs were obtained from each patient with AP immediately after admission and during the recovery phase, owing to the difficulties in direct fecal collection due to gastrointestinal dysfunction18,19. The control group consisted of four healthy volunteers, with all individuals having empty stomachs before sample collection. All individuals were enrolled at Peking Union Medical College Hospital, Beijing, China. Fecal sample collection methods have been previously described in detail12. Soap, water, and 70% alcohol were used to clean around the anus. The disinfectant was allowed to evaporate to reduce commensal skin contamination. A sterile swab was soaked in normal saline for 2 min before being inserted into the anus to a depth of 4–5 cm and then rotated gently to obtain fecal samples. Rectal swabs with fecal samples were quickly placed into sterile tubes and immediately stored at −80℃ before shipping to the laboratory under cold chain conditions.

Fig. 1
figure 1

The flow chart of this study.

Full size image

Metagenomics sequencing, quality control, and genome assembly

DNA was extracted from the fecal samples using MP-soil FastDNA™ Spin Kit for Soil (#6560 − 200, MP Biomedicals, Southern California, USA) according to the manufacturer’s instructions. DNA purity, concentration, and quality were determined using NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA), TBS-380, and a 1% agarose gel electrophoresis system, respectively. Paired-end sequencing was performed on an Illumina HiSeq 4000 platform (Illumina, San Diego, CA, USA). The raw sequencing reads for each sample were processed for quality control using fastp (v 0.23.0). Firstly, sequencing adapters were removed from sequencing reads using fastp. Secondly, low-quality reads with an average quality score of less than 20 and sequences that fell below 50 bp after contaminant removal and trimming were removed. Thirdly, human genomic DNA reads were removed by mapping to the human genome using BWA (v 0.7.17)20. After quality control, the reads were mapped to representative genes with 95% identity using SOAPaligner (v 2.21)21, and gene abundance was assessed for each sample. The proportions of bacterial reads, viral reads, and archaeal reads were 97.28%, 1.26%, and 0.01%, respectively.

Alpha diversity and principal coordinates analysis

The α-diversity analysis was performed using the R packages “vegan” (v 4.0.5) and “reshape2” (v 4.0.5). Species diversity was assessed using the Shannon index, with a higher value indicating higher microbial diversity. In addition, principal coordinates analysis (PCoA), which reflects the differences and distances between samples by analysing the community composition of different samples, was performed using R (v 3.3.1). The relatively similar mean values of the samples from different groups indicated a similarity in species composition.

Analysis of enterotypes and dominant Microbiome composition

Enterotypes are effective in stratifying populations and providing a global overview of the inter-individual variations in gut microbial composition22. Enterotype analysis can group different samples with a similar structure of dominant microbiomes into one group using statistical clustering. Based on the relative abundance of the microbiome at the genus level, the Bray–Curtis distance was calculated for clustering. After calculating the optimal clustering K using the Calinski–Harabasz index22, the 28 samples were divided into different enterotypes. In patients with AP and those recovering from AP, histogram plots were generated using the Majorbio Cloud Platform (www.majorbio.com) to visualise the distribution of the dominant microflora among different samples. Circos diagrams were also generated using Circos (v 0.67–7) to show the microbiome composition. In this analysis, the dominant microflora was defined as the taxa with the highest relative abundance of microbial taxa in these samples23.

Differential Microbiome analysis and functional annotation

The sequencing depths ranged between 10 and 14 Gb in this study. In order to reduce the difference in microbiota abundance caused by varying sequencing depths, Reads Number_Relative abundances were generated to compare the proportion of different microbiota across samples as percentages. Based on the read abundance data, species with significant differences were identified using the Wilcoxon rank-sum test (P < 0.05). Nonredundant genes were annotated against the KEGG database (v 94.2)24 using Diamond (v 2.0.13)25 with an optimized e-value (cutoff of 1e − 5).

Statistical analyses

Statistical analyses and data visualisation were conducted using custom scripts that are publicly available (https://github.com/songxiao203/sx-project)26. The Strengthening The Organizing and Reporting of Microbiome Studies checklist was completed and uploaded to Zenodo (https://zenodo.org/records/10676912?token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6ImM1YTIzYmI0LTU0MmYtNDQ4NS05NzdkLTVmOTRkOWFmNGI0NCIsImRhdGEiOnt9LCJyYW5kb20iOiIxNjE4NmI1NWVmNjc1ZTY4YjYyMjcxOTRkNTI5NzRlMyJ9.T3S5fNFyjdfxPgK-BGoCYPaVx4H7gKLMLGcJ8cYuogIp9mIhJkdl1QUqH9vJm5qaEeZmEmue2Jqqc1XCFLo28A).

Results

Patient characteristics

Twelve patients with AP and four healthy controls were included in this study. Among the 12 patients with AP, 4 had MAP, 5 had MSAP, and 3 had SAP. After treatment, the clinical condition of all patients with AP initially improved and their serum lipase and amylase levels returned to normal or met the expected values. Meanwhile, Acute Physiology and Chronic Health Evaluation II (APACHE II) and Balthazar computed tomography (CT) scores decreased during the recovery phase, which also indicated the improvement of clinical manifestations. Detailed information on the patients with AP and healthy controls is presented in Table S1.

Gut microbiome diversity in patients with AP during acute and recovery phases

Compared to healthy controls, the α-diversity (Shannon index) in patients with AP was significantly reduced during the acute phase and further decreased during the recovery phase (Fig. 2A). A reduction in diversity was observed among different severities during the recovery phase (Fig. 2B). During the recovery period, clinical manifestations started to improve, but gut microbiota diversity did not recover concomitantly. Thus, the recovery of gut microbiota and intestinal function was slower than the overall clinical improvement. Category-based PCoA revealed differences in β-diversity between the control and AP groups, showing distinct clustering. However, no significant clustering was observed between the acute and recovery phases or among the MAP, MSAP, and SAP subgroups (Fig. 2C). The structural diversity of acute-phase samples differed only slightly across different severity levels (Figure S1). During the recovery phase, samples from different severities exhibited clear clustering.

Fig. 2
figure 2

Microbial diversity and principal coordinates analysis in the recovery phase of AP. A Alpha diversity analysis in the acute and recovery phases of AP patients; B Alpha diversity analysis in AP patients of different severities; C Principal coordinates analysis (PCoA) of Bray-Curtis distance in all samples at the species level. A_MAP acute phase of MAP, R_MAP recovery phase of MAP, A_MSAP acute phase of MSAP, R_MSAP recovery phase of MSAP, A_SAP acute phase of SAP, R_SAP recovery phase of SAP.

Full size image

Analysis of enterotypes and dominant Microbiome composition during the recovery phase of AP

The concept of “enterotypes”, which is stable, mainly affected by long-term dietary habits, and without direct relationship with gender, age, geography, or cultural background, was first introduced by Arumugam et al.22. The 28 samples were clustered into two distinct enterotypes, with those having a high abundance of the Bacteroides genus being classified as enterotype 1 and those having a high abundance of the Enterococcus genus classified as enterotype 2, for both the total group (Figure S2A) and the subgroups (Figure S2B). Enterococcus was the dominant genus in the gut microbiota during the acute and recovery phases of AP (Fig. 3A, S2C). At the species level, the abundances of unclassified g Enterococcus and Enterococcus faecium were high during both the acute and recovery phases of AP, consistent with the results observed at the genus level (Fig. 3B, S2D). Similarly, in this subgroup, Enterococcus was the dominant genus in the gut microbiota during the acute and recovery phases of AP (Fig. 3C, S2E). At the species level, the abundances of unclassified g Enterococcus and Enterococcus faecium were high during the acute and recovery phases of AP, consistent with the results observed at the genus level (Fig. 3D, S2F). Comparison of the acute and recovery phases of MAP showed that two patients had decreased and the other two had increased abundance of Enterococcus (without significant differences, Fig. 4A). The same comparison revealed an increase in the abundance of Bacteroidales (without significant differences, Fig. 4B). Conversely, a comparison of the acute and recovery phases of MSAP revealed an increase in the abundance of Enterococcus (Fig. 4C) and a decrease in that of Bacteroidales (Fig. 4D), without significant differences. Similarly, in SAP, the same comparison revealed an increase in the abundance of Enterococcus (Fig. 4E) and a decrease in that of Bacteroidales (Fig. 4F), with no significant differences. Thus, Enterococcus was the dominant genus in the gut microbiota during recovery.

Fig. 3
figure 3

Analysis of the dominant microbiome composition in the recovery phase of AP. Distribution of the predominant dominant gut microbiota species at the genus level in the total group (A), at the species level in the total group (B), at the genus level in the subgroup (C), and at the species level in subgroup (D). A_MAP acute phase of MAP, R_MAP recovery phase of MAP, A_MSAP acute phase of MSAP, R_MSAP recovery phase of MSAP, A_SAP acute phase of SAP, R_SAP recovery phase of SAP.

Full size image
Fig. 4
figure 4

Difference analysis of Enterococcus and Bacteroidales in the acute phase of MAP and recovery phase of MSAP. A Difference analysis of Enterococcus in MAP; B difference analysis of Bacteroidales in MAP. C difference analysis of Enterococcus in MSAP; D difference analysis of Bacteroidetes in MSAP; E difference analysis of Enterococcus in SAP; F difference analysis of Bacteroidetes in SAP. A_MAP acute phase of MAP, R_MAP recovery phase of MAP, A_MSAP acute phase of MSAP, R_MSAP recovery phase of MSAP, A_SAP acute phase of SAP, R_SAP recovery phase of SAP. *0.01 < p ≤ 0.05, ns: not significant.

Full size image

Differential Microbiome analysis during the recovery phase of AP

Differential species composition analysis showed that unclassified g Enterococcus was the most increased gut microbiota species during the acute phase of AP (Fig. 5A). During the recovery phase, the abundance of beneficial bacteria such as Faecalibacterium prausnitzii and Bifidobacterium longum was low, indicating that the overall abundance of beneficial bacteria had not recovered during this period (Fig. 5B). At the species level, a total of 90 common species were found between AP vs. control and recovery vs. acute phases of AP (Fig. 5C). Among them, 29 species showed an inverse trend in abundance between the acute phase vs. control and recovery vs. acute phases of AP (Table 1). The short-chain fatty acid-producing bacteria, including Anaerococcus, Blautia, Candidatus Blautia pullistercoris, and Eubacterium sp. CAG:86, were low in abundance during the recovery phase. The abundance of the propionic and butyric acid-producing bacterium Roseburia inulinivorans CAG:15 was low during the acute phase of AP, whereas that of Candidatus Eisenbergiella pullicola was high. The levels of anti-inflammatory microbiota began to increase during the recovery phase, while those of other beneficial bacteria did not return to normal.

Fig. 5
figure 5

Difference analysis of gut microbiota species in the recovery phase of AP. A Difference analysis of gut microbiota species in AP vs. Control; B difference analysis of gut microbiota species in the recovery phase of AP vs. the acute phase of AP; C Venn plot of gut microbiota species in the total group. A acute phase, R recovery phase. *0.01 < p ≤ 0.05, ns not significant.

Full size image
Table 1 The 29 kinds of bacteria showed a reverse trend in abundance levels between the acute and recovery phases of AP patients.
Full size table

Comparison of the species between patients with MAP and healthy controls revealed 498 significantly different species in the gut microbiota. The levels of Enterococcus and Bacteroidales significantly increased and decreased, respectively (Fig. 6A). Comparison of the acute and recovery phases of MAP revealed 13 significantly different species of the gut microbiota. The abundances of Oscillibacter, Ruminococcus, and Firmicutes significantly decreased during the recovery phase (Fig. 6B). Comparison of patients with MSAP and healthy controls revealed 1223 significantly different species of the gut microbiota, with Enterococcus being the most significantly increased genus (Fig. 6C). In the MSAP group, 241 significantly different species were identified between the recovery and acute phases. Adlercreutzia equolifaciens showed the most significant decrease in abundance; however, its abundance remained low during the recovery phase (Fig. 6D). Comparison of patients with SAP and healthy controls revealed 397 significantly different species of the gut microbiota. The abundances of Enterococcus and Porphyromonas asaccharolytica significantly increased and decreased, respectively (Figure S3A). No significant differences in gut microbiota species were found between the acute and recovery phases of SAP (Figure S3B). Our results suggest that patients with SAP may have recovered from their clinical symptoms; however, their microbiota profiles have not yet recovered. Enterococcus was the most abundant genus in the acute phase of AP at varying severities.

Fig. 6
figure 6

Difference analysis of gut microbiota species in the recovery phase of MAP and MSAP. A Difference analysis of gut microbiota species in MAP vs. Control; B difference analysis of gut microbiota species in the recovery phase of MAP vs. the acute phase of MAP; C difference analysis of gut microbiota species in MSAP vs. Control; D difference analysis of gut microbiota species in the recovery phase of MSAP vs. the acute phase of MSAP; E Venn plot of gut microbiota species in MAP; F Venn plot of gut microbiota species in MSAP. A_MAP acute phase of MAP, R_MAP recovery phase of MAP, A_MSAP acute phase of MSAP, R_MSAP recovery phase of MSAP. *0.01 < p ≤ 0.05, ns not significant.

Full size image

Two common species were identified between the 498 differential species from the acute phase of MAP and 13 differential species from the recovery phase of MAP (Fig. 6E). Firmicutes bacterium CAG:129_59_24, belonging to the phylum Firmicutes, showed low abundance during the recovery phase of MAP, indicating that its function had not yet recovered. Intestinimonas timonensis, a pro-inflammatory microorganism, showed decreased abundance during the recovery phase of MAP, indicating that the inflammatory function of the species gradually recovered. Between the 1223 differential species identified during the acute phase of MSAP and 241 differential species from the recovery phase, 47 common species were identified (Fig. 6F). Of these, 19 were highly abundant in patients with MSAP during the acute phase but showed low abundance during the recovery phase (Table 2). Among them, Candidatus Eisenbergiella pullicola and Collinsella provencensis, which are related to intestinal wall permeability, Olsenella sp. Marseille-P4518, which is a fructose-consuming strain, Clostridium sp. D53t1 180,928 C8, Blautia sp. AM28-10, and Olsenella sp. Marseille-P4518 were highly abundant during the acute phase of MSAP but decreased in abundance during the recovery phase.

Table 2 The 19 kinds of bacteria showed a reverse trend in abundance levels between the acute and recovery phases of MASP patients.
Full size table

Analysis of differences in functional compositions during the recovery phase of AP

The gut microbiota of patients with acute-phase AP was involved in 87 significantly different functional compositions of the KEGG pathway. Adenosine triphosphate-binding cassette (ABC) transporters, fructose and mannose metabolism, and activation of the phosphotransferase system (PTS) were significantly upregulated, whereas glycine, serine, and threonine metabolism was significantly downregulated during the acute phase of AP (Fig. 7A). Meanwhile, 13 significantly different functional pathways were identified during the recovery phase, among which microbial metabolism was significantly activated in diverse environments. However, RNA degradation, protein processing in the endoplasmic reticulum, ribosome biogenesis in eukaryotes, and other RNA- and protein-related pathways were inhibited or downregulated (Fig. 7B). Only three pathways were shared between the acute and recovery phases: RNA degradation, flavonoid biosynthesis, and stilbenoid, diarylheptanoid, and gingerol biosynthesis (Fig. 7C). Changes in pathway functions for patients with varying AP severities are shown in Fig. 8 and summarised in Table 3. We identified 37, 98, and 23 significantly different KEGG functional compositions in patients with MAP, MSAP, and SAP, respectively.

Fig. 7
figure 7

KEGG functional composition difference analysis of the microbiota in the recovery phase of AP. A Difference analysis of functional compositions of KEGG in AP vs. Control; B difference analysis of functional compositions of KEGG in the recovery phase of AP vs. the acute phase of AP; C Venn plot of KEGG functional compositions in AP. A acute phase, R recovery phase. *0.01 < p ≤ 0.05.

Full size image
Fig. 8
figure 8

KEGG functional composition difference analysis of the microbiota in the recovery phase of AP with different severities. A difference analysis of functional compositions of KEGG in MAP vs. Control; B difference analysis of functional compositions of KEGG in the recovery phase of MAP vs. the acute phase of MAP; C difference analysis of functional compositions of KEGG in MSAP vs. Control; D difference analysis of functional compositions of KEGG in the recovery phase of MSAP vs. the acute phase of MSAP; E difference analysis of functional compositions of KEGG in SAP vs. Control; F difference analysis of functional compositions of KEGG in the recovery phase of SAP vs. the acute phase of SAP. A_MAP acute phase of MAP, R_MAP recovery phase of MAP, A_MSAP acute phase of MSAP, R_MSAP recovery phase of MSAP, A_SAP acute phase of SAP, R_SAP recovery phase of SAP. *0.01 < p ≤ 0.05.

Full size image
Table 3 KEGG functional composition difference analysis of the microbiota in the recovery phase of AP with different severity.
Full size table

Discussion

The onset, development, and recovery of AP are closely related to intestinal microbiota dysbiosis and bacterial translocation27. The intestinal microbiota can participate in the onset and recovery of AP by influencing host metabolism and intestinal mucosal permeability28. We performed shotgun metagenomic sequencing on 12 patients with AP during the acute and recovery phases. Significant microbiome dysbiosis was observed during both phases, suggesting an association between dysbiosis and recovery from AP.

Enterococcus, a harmful bacterium, was the most abundant genus present during the acute and recovery phases of AP. Specifically, the relative abundances of Enterococcus faecium, Enterococcus faecalis, Enterococcus hirae, Enterococcus gallinarum, and Enterococcus casseliflavus were high during the acute phase of AP. Enterococcus can attach to and penetrate host cells, traverse host epithelial barriers, and gain entry into the systemic circulation and other organs, promoting its spread within the host14. The relative abundance of Enterococcus spp. is associated with infection and systemic inflammation29. Biofilm formation by Enterococcus involves a complex interplay between genes and virulence factors, including gelatinases, cytolysins, secretory antigen A, fimbriae, microbial surface components that recognise adhesion matrix molecules, and DNA30. Zhang et al. reported that the relative abundance of the potentially pathogenic genus Enterococcus increased in the SAP group but decreased after Qingyi decoction treatment31. Enterococcus faecium performed well in distinguishing patients with acute necrotizing pancreatitis from those with non-necrotizing pancreatitis32. In patients with Crohn’s disease, ulcerative colitis, or ischemic colitis, the abundance of potentially pathogenic bacteria, such as Enterococcus faecium and Enterococcus faecalis, was increased, indicating their potential diagnostic value33.

During the recovery phase of AP, the abundance of the beneficial bacteria Faecalibacterium prausnitzii, Bifidobacterium longum, and Firmicutes bacterium was significantly lower than that during the acute phase. These results indicate that while the clinical manifestations began to improve, the overall abundance of beneficial bacteria did not recover concomitantly during the AP recovery period. Recovery of the gut microbiota and intestinal function was slower than the overall clinical improvement. Besides, the use of antibiotics could be a significant contributing factor to the slow recovery of microbial diversity observed during the recovery phase. Faecalibacterium prausnitzii is involved in inflammatory processes, blocking nuclear factor kappa B activation and interleukin (IL)−8 production34. The abundance of Faecalibacterium prausnitzii was lower in patients with chronic pancreatitis (CP) and diabetes than in those without diabetes and in controls35. Similarly, McEachron et al. observed a significantly decreased abundance of Faecalibacterium in patients with CP36. Bifidobacterium has been reported to reduce the risk of multiple organ failure in SAP37. Bifidobacteria, with a reduced abundance in feces, is inversely correlated with the severity of the systemic inflammatory response in patients with AP38. Firmicutes, which are pathogenic Gram-positive bacteria, are mainly butyrate producers and have been identified in patients with AP39. Anaerococcus, Blautia, Candidatus Blautia pullistercoris, Eubacterium sp. CAG:86, and other short-chain fatty acid-producing bacteria were significantly lower in abundance during the recovery phase of AP. In our previous study, we observed a significant increase in the abundance of Anaerococcus in MSAP and a significant decrease in the abundance of Blautia in MAP12. The abundance of the propionic and butyric acid-producing bacterium Roseburia inulinivorans CAG:15 was significantly lower during the recovery phase, whereas that of Candidatus Eisenbergiella pullicola was significantly higher during the acute phase of AP. Faecalibacterium and Roseburia are well-known butyrate producers that attenuate intestinal inflammation and are beneficial to the host40,41. The relative abundance of Roseburia decreased in mice with CP42. Eisenbergiella has been suggested to protect mice with ulcerative colitis by reducing the intestinal inflammatory response43. Bacterial abundance decreased significantly during the recovery phase, indicating an improvement in inflammation-related microbiota function. However, the abundance of beneficial bacteria did not fully recover during the AP recovery phase.

In MAP, the abundance of Oscillibacter sp. significantly increased during the acute phase but decreased during the recovery phase. Increased abundance of Oscillibacter is associated with intestinal permeability and host inflammation in patients with large-artery atherosclerotic stroke or transient ischemic attacks44. Oscillibacter sp. 57_20 has been reported to identify SAP at an early stage28. In MSAP, the abundances of Clostridium sp. D53t1_180928_C8 and Blautia sp. AM28-10 significantly increased during the acute phase but decreased during the recovery phase. Clostridium perfringens was isolated from cultures of pancreatic tissue obtained from a patient with AP during hospitalisation45. Significant differences in Collinsella abundance were observed between the MAP and SAP groups46. The abundance of Collinsella is associated with the production of the proinflammatory cytokine IL-17 A47. Decreased abundance of Collinsella has been reported in children with CP48. In addition, Candidatus Eisenbergiella pullicola and Collinsella provencensis were associated with intestinal wall permeability. In SAP, the abundance of Adlercreutzia equolifaciens was high during the acute phase but decreased during recovery. However, no significant difference in its abundance was found between the two phases, suggesting a possible intrinsic role for SAP.

Additionally, we observed that disturbances in gut microbiota affect host metabolism and molecular transport mechanisms. PTS, one of the most efficient sugar transport systems, plays a crucial role in the regulation of glucose metabolism in bacteria49. During the acute phase of AP, PTS was significantly activated, whereas glycine, serine, and threonine metabolism was significantly inhibited. Compared with their levels six weeks after hospital discharge, the levels of glycine, serine, and threonine in patients with AP were significantly lower upon admission50. Additionally, glycine administration improves AP pathomorphology and reduces the severity of MAP and SAP51. During the AP recovery phase, microbial metabolism is significantly activated and upregulated in diverse environmental pathways. However, RNA degradation, protein processing in the endoplasmic reticulum, ribosome biogenesis in eukaryotes, and other functional pathways related to RNA and proteins were inhibited and downregulated. These findings indicate that while intestinal function and the gut microbiota structure start to recover, the restoration of RNA- and protein-related biological processes lags behind.

During the recovery phase of MAP, an increase in the functional composition of glycine, serine, and threonine metabolism, along with a decrease in the functional composition of the PTS compared to the acute phase, were observed. In addition, cardiac muscle contraction was the most significantly reduced functional component during the MAP recovery phase. The PTS regulates carbohydrate uptake and is considered a potential biomarker of inflammatory bowel disease52. Glycine, serine, and threonine metabolism is involved in glycolytic biosynthesis in pancreatic cancer53. Regardless of its severity, AP is accompanied by cardiac injury, with the degree of damage associated with the clinical type of AP54. Thus, activation of the glycine, serine, and threonine metabolic pathways, deactivation of the phosphotransferase system, and reduction of cardiac muscle contraction are all associated with the recovery phase in patients with MAP. Moreover, during the recovery phase of MSAP, an increase in the functional composition of carbon metabolism and a decrease in the functional composition of ABC transporters, compared with the acute phase, was observed. ABC transporters translocate various endogenous metabolites and cytotoxic compounds across the lipid bilayer55. In cancer, overexpression of ABC transporters leads to drug resistance56. In mice with AP, antioxidant treatment with carbon monoxide-bound hemoglobin vesicles effectively reduced the inflammatory response and pancreatic tissue damage57. Therefore, carbon metabolism and ABC transporters may be involved in MSAP recovery. These signalling pathways exhibit an opposite trend during the recovery phase of MAP and MSAP, in comparison with the acute phase, suggesting that they may benefit recovery.

Nonetheless, this study had several limitations. The sample size used in the shotgun metagenomics was relatively small, and further studies with more participants are warranted. Moreover, the average age of control group was relatively young, which may introduce potential limitations. In addition, the causal relationships between the microbiota, functional composition, and AP recovery require further investigation. The way gut microbiome shifts reflect the status of the pancreas also requires further investigation.

In conclusion, we identified several bacteria, along with their functional composition, during the recovery phase of AP through a shotgun metagenomic survey, extending the current understanding of gut microbiota during this phase. This study provides information that may guide the development of treatment strategies and the prognosis of AP.