ALTERATIONS IN MICROBOME ASSOCIATED WITH ACUTE PANCREATITIS

Abstract

Purpose of review:

This review evaluates the current knowledge of gut microbiome (GM) alterations in acute pancreatitis (AP), including those that can increase AP risk or worsen disease severity, and the mechanisms of GM driven injury in AP.

Recent findings:

Recent observational studies in humans showed the association of GM changes (decreased GM diversity, alterations in relative abundances of certain species, and association of unique species with functional pathways) with AP risk and severity. Furthermore, in vivo studies highlighted the role of GM in the development and severity of AP using FMT models. The gut barrier integrity, immune cell homeostasis and microbial metabolites appear to play key roles in AP risk and severity.

Summary:

Large human cohort studies that assess GM profile, its metabolites and impact on AP risk and severity will be crucial for development of innovative prediction, prevention and treatment strategies.

INTRODUCTION

There is an increasing interest in the gut microbiota as a causative or modifying factor in acute pancreatitis (AP) in recent years (1^••, 2^••, 3^•, 4^•, 5^••, 6^••, 7^•, 8^••, 9^••, 10, 11, 12, 13, 14, 15 ). Gut microbiota, the complex community of microorganisms living in the digestive tracts, have important functions in digestion, immune function and overall health. This review will describe the current knowledge of gut microbiome (GM) alterations in AP, including those which may predispose to the onset of AP or impact disease severity, and those that occur as a result of the inflammatory cascade in AP.

Gut microbiota includes a diverse and complex array of numerous organisms, some of which are considered beneficial colonizing bacteria and others of which can become pathogenic due to shifts in their abundance, where overall the GM can impact the host metabolome and/or host immunity. GM is dynamic and may be influenced by other disease states including infections, dietary changes, medications, and alterations in motility. It can influence host metabolic processes from nutrient absorption to the immune response, affecting permeability across the intestinal wall and invasion of pathogenic species. In fact, fecal microbiota transplantation (FMT) is an effective treatment for recurrent Clostridioides difficile infection, and microbiota-based therapies are being evaluated in other inflammatory diseases such as graft-versus-host disease, metabolic syndrome, and inflammatory bowel diseases (16).

AP prevalence is increasing (17) with mortality rate up to 20 % in those with severe acute pancreatitis (SAP). It leads to $2.6 billion in healthcare costs in the US annually(18), and affects minorities disproportionally with higher risk of AP and organ failure in African Americans (19, 20). The incidence of AP rises with age, though the prevalence is similar between men and women(17). Gallstones are the most common cause of AP accounting for around 40% of cases, followed by alcohol-related AP (nearly 30%), hypertriglyceridemia ( 2–7%), other less common etiologies (21).

While there is evidence linking alterations in GM to the onset of AP, there is an incomplete understanding of the mechanism by which these alterations contribute to the onset of inflammation. Likewise, the severity of illness may also be impacted by GM but these mechanisms – and importantly, the opportunities for preventive or therapeutic interventions - are also in need of further study.

KEY CONSIDERATIONS IN GUT MICROBIOME STUDIES

Changes in an individual’s microbial composition and function over time may have significant health implications, and specific microbial diversity and abundance patterns can be associated with various health outcomes.

Key metrics used in gut microbiome studies

Several metrics are used to quantify the microbiome composition. The relative abundance of each microbial taxon is the proportion relative to the total microbial population in a GM sample. Shannon’s alpha diversity quantifies microbial composition and abundance within a GM sample, while beta diversity compares the differences between two GM samples.

Potential confounders and important considerations while studying gut microbiome

GM can be influenced by diet, age, sex, medication, lifestyle factors (alcohol consumption, smoking, and exercise), genetics, and health status. It is crucial to account for these confounders when analyzing associations to host outcomes at the composition or individual taxon levels (22). For example, while heavy alcohol consumption can lead to alcohol-induced AP, even moderate alcohol intake can change GM composition. Such GM changes may alter production of microbial metabolites, and may increase AP risk. Therefore detailed phenotyping of AP patients and adequate adjustment of key confounders are crucial during analysis in GM studies.

Sequencing methods, bioinformatics, and emerging machine learning models

The most cost-effective method for sequencing is 16S ribosomal RNA (rRNA) gene sequencing which quantifies a sample’s microbial composition and abundance, from which the metabolic capabilities can be predicted computationally. The more comprehensive approach is shotgun metagenomic sequencing technology which sequences all DNAs in a sample, allowing identification of the species present and their functions based on the identified microbial genes.

Bioinformatics tools are essential in microbiome data analysis. QIIME 2 and DADA2 (23) support sample sequence quality filtering, taxonomic analysis, and diversity analyses (24). Linear discriminant analysis Effect Size (LEfSe) quantifies differences in microbial feature abundance among biological groups (25). Multivariate Association with Linear Models (MaAsLin2) is powerful in detecting associations between clinical variables, microbial taxa, and gene features (26). Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) predicts the functional composition of 16S rRNA data using marker gene data and a database of reference genomes (27).

Machine learning (ML) is increasingly used to predict host outcomes based on microbiome (28). Support Vector Machines, Random Forests, and Elastic Net are typical ML models available in the scikit-learn package to build a prediction model from GM samples of a modest size. Recently, neural networks have emerged as commanding ML models for integrating microbiomes, metabolomes, and clinical data to classify host phenotypes and predict continuous clinical variables. The superiority of neural network models lies in their capability to detect patterns and associations. We have used several tools mentioned above to quantify microbial composition differences between GMs of AP and control patients previously (8^••) and Figure 1 provides the framework we utilized.

Figure 1. An overview of our framework of gut microbiome study in acute pancreatitis.

Stool DNA from samples was extracted and sequenced for gut microbiome. The sequencing data were analyzed by the sequence of bioinformatic and statistical tools (DADA2) to quantify microbiome composition and abundance, compare Shannon’s alpha diversity, and detect taxa with the largest effect size in abundance between disease and control groups (LefSe). Finally, machine learning models (random forest) were developed for disease prediction, and important taxa were ranked by the random forest model. (Original submission).

ASSOCIATION OF GUT MICROBIOME CHANGES WITH AP ETIOLOGIES AND SEVERITY

The most common AP etiologies have been associated with alterations in GM (1^••, 7^•, 29). Wang et al showed that AP etiology explained 3.5% variation in species’ relative abundance and was a key driver of GM composition(7^•). In this study, Alistipes shahii and Escherichia coli were enriched in those with biliary AP while pathways related to lipid synthesis was enhanced in those with hypertriglyceridemic pancreatitis (HTGP) (7^•) highlighting the associations of GM differences with functional pathways in AP. In a larger study (n=69), Li et al showed increased E. faecium and E. coli and decreased Bacteroides uniformis abundance in HTGP patients(6^••). Furthermore, LEfSe showed higher abundances of pathogenic Klebsiella, Enterococcus and Escherichia Shigella, and lower abundance of beneficial bacteria, including Faecalibacterium prausnitzii and Bacteroides uniformis in HTGP patients (6^••). The association of AP etiologies with GM composition was further explored by Liu et al (n=197) (1^••). Investigators identified differences in relative abundances of 21 species among various AP etiologies, and a microbiome panel that could distinguish the underlying AP etiology (1^••). This study further revealed how differences in GM can impact host metabolic function by showing contribution of enrichment of Bilophila wadsworthia and depletion of Bifidobacterium spp. to increased triglyceride levels in AP patients with biliary or hyperlipidemic etiologies (1^••).

Multiple human subject and mouse studies investigated the association of GM with AP and AP severity (Table 1 and Table 2). This is important since early prediction of disease severity and development of microbiota-based therapies have the potential to improve outcomes in SAP. Ammer-Herrmenau et al recently showed that GM changes were associated with disease severity and other clinical hallmarks such as mortality in patients with AP (2^••). Abundant species in those with SAP mainly belonged to taxonomic families that produce short chain fatty acids (SCFAs). In addition, they have developed a classifier that included 16 differentially abundant species and systemic inflammatory response syndrome that outperformed current AP severity scores (2^••). In alignment with these findings, Yazici et al also showed that AP patients have significantly higher serum levels of acetate, a SCFA, compared to controls, and developed a microbiome-based ML model that had 96% predictive ability to distinguish AP patients from controls (8^••). Li et al investigated GM driven AP severity using a mouse model (6^••). In this study, FMT from HTGP patients increased recruitment of neutrophils and NETs formation and resulted in worsened pancreatic injury and systemic inflammation compared to FMT from healthy controls or hypertriglyceridemia patients (6^••). Subsequently, Liu et al compared the structural features of gut microbial communities in AP patients and showed a significant effect of disease severity on the overall GM in AP patients (1^••). These results also showed significantly decreased species richness in SAP and identification of 18 GM species that associated with AP severity (1^••). Liu et al further investigated the modulation of AP severity by gut microbiota in a caerulein-induced AP model that included FMT from AP-associated species vs. AP-depleted strains. Mice that received FMT from AP-associated species had higher pancreatic edema, massive infiltration of inflammatory cells, and a pancreatic tissue score (1^••) showing the causation between GM and SAP, and its potential therapeutic implications in future.

Table 1.

Recent human subject studies looking at the role of gut microbiome in acute pancreatitis

Study	Methods	Microbial changes	AP relevance	Clinical implications
2024, Liu et al, (1) *	AP, n=81 & Controls, n=115 Fecal samples Whole-metagenome shotgun sequencing Metabolomics & lipidomic	Differential abundances of 77 species influenced by etiology and severity ↓ species richness in SAP ↑Bilophila wadsworthia and ↓ Bifidobacterium spp. are associated with ↑ TG in biliary or hyperlipidemic AP	GM alterations was the highest in biliary AP patients Identified association of 18 GM species with AP severity Mice that received FMT from AP-associated species had worsened pancreatic injury	Differences in GM can impact host metabolic function Elucidated the role of GM in disease severity and progression
2024, Ammer-Herrmenau et al, (2)	AP, n=424 Buccal and rectal swabs 16S rRNA sequencing	Altered GM in AP patients is associated with disease severity, length of hospital stay and mortality Species abundant in SAP belong to taxonomic families known as producers of SCFAs	SAP had functional pathways suggesting ↑ SCFA A classifier including microbiome species & SIRS outperforms established AP severity scores	GM can predict clinical features of SAP SCFAs as potential diagnostic & therapeutic targets in future
2024, Dike et al, (3)	AP patients, n=30 Controls, n=34 Shotgun metagenomics	AP patients had ↑ R.gnavus, V.parvula, E.faecalis, C.innocuum SAP group had ↑ E.faecalis and C.citroniae	Pediatric AP patients ↓ had alpha-diversity & ↑ amino acid metabolism & fatty acid beta-oxidation	Children with AP have gut microbiome alterations & clinical relevance of these needs to be investigated
2023, Hu et al, (4)	AP-ARDS, n=26; AP-nonARDS, n= 39 & Controls, n=20 Rectal swabs 16S rRNA sequencing	AP-ARDS group had: ↑ Proteobacteria, Enterobacteriaceae, Escherichia-Shigella, Klebsiella pneumoniae (K. pneumoniae) & ↓ Bifidobacterium	↓ Simpson index in both AP-nonARDS and AP-ARDS groups compared to controls Escherichia-Shigella genus could predict ARDS occurrence in AP	Potential predictive role of gut microbiota in ARDS in AP patients
2023, Kostenko et al, (5) *	AP & Controls Blood samples 16S rRNA sequencing	Patients with infected AP had ↑ Bacterial 16S DNA, altered β-diversity & ↑ Pseudomonadales	↑ Circulating unsaturated NEFAs; unbound NEFAs, including linoleic and oleic acid ↓ Bacterial clearance & infected sterile inflammation	Targeting excessive lipolysis to maintain immune cell function to prevent infection of sterile necrosis
2023, Li et al, (6) *	HTGP patients, n=25 TG-matched controls, n=8 Healthy volunteers, n=16 AP patients, n=20 Fecal samples 16S rRNA sequencing	HTGP patients had ↑ Enterococcus faecium (E. faecium) & E. coli & ↓ Bacteroides uniformis (B. uniformis) LeFSe analysis showed ↑ Klebsiella, Enterococcus & Escherichia Shigella & ↓ Faecalibacterium prausnitzii & B. uniformis	↓ Gut microbiome alpha diversity SIRS & ICU admission negatively correlated with Faecalibacterium, B. uniformis & Collinsella aerofaciens abundances *FMT-HTGP worsened pancreatitis & inflammation * ↓ B. uniformis in FMT-HTGP group	HTGP-related microbiome changes can exacerbate pancreatic injury ↑ Immune response through neutrophil infiltration in HTGP-derived microbiota Ameliorating HTGP through gut microbiota modulation
2023, Wang et al, (7)	Mild AP, n=13 & SAP, n= 7 Stool samples Metagenomic sequencing	SAP patients had ↓ Bacteroides xylanisolvens, Clostridium lavalense, and Roseburia inulinivorans	SAP patients had ↑ Oscillibacter sp. 57_20, Parabacteroides johnsonii, Bacteroides stercoris, Methanobrevibacter smithii, Ruminococcus lactaris, Coprococcus comes, and Dorea formicigenerans	Etiology is the main driver of GM composition GM as a potential biomarkers for early SAP prediction
2023, Yazici et al, (8)	AP, n=54 & Controls, n=46 Stool samples 16S rRNA sequencing Machine learning models	AP patients had ↑ abundance of sulfidogenic bacteria including Veillonella & Haemophilus ↓ SCFA producing genera ↓ GM alpha diversity	AP patients had ↑ Serum levels of hydrogen sulfide (H2S), microbial end-product of sulfidogenic bacteria ↑ Serum acetate concentrations ML distinguished AP from control, 96% predictability	Gut microbiota and their end-metabolites may modulate AP risk GM as a therapeutic target in AP GM as a prediction tool in AP
2022, Yan et al, (9) *	SAP, n= 72 & Controls, n = 32 Fecal samples 16S rRNA sequencing	Patients with SAP had ↓ Clostridium butyricum (C. butyricum)	*C. butyricum or butyrate treatment upregulated tight junction protein expression *C. butyricum protected against SAP	*C. butyricum or butyrate as potential therapeutic targets in SAP

Abx: Antibiotic; ARDS: Adult respiratory distress syndrome; AP: Acute pancreatitis; FMT: Fecal microbiota transplantation; GM: Gut microbiome; HTGP: Hypertriglyceridemic pancreatitis; ICU: Intensive care unit; LEfSe: Linear discriminant analysis Effect Size; ML: Machine learning; NEFAs: non esterified fatty acids; rRNA: ribosomal RNA; SAP: Severe acute pancreatitis; SCFA: Short-chain fatty acid; SIRS: Systemic inflammatory response syndrome; TG: Triglyceride

^(*):Study includes in vivo experiments. (Original submission).

Table 2.

Recent animal studies looking at the role of gut microbiome in acute pancreatitis

Study	Methods	Microbial changes	AP relevance	Clinical implications
2024, Wang et al, (10)	Caerulein-induced AP Pretreatment with Akkermansia muciniphila membrane protein (Amuc_1100) Cecal contents (feces) of mice 16S rRNA sequencing	AP mice pretreated with Amuc_1100 had ↓ Bacteroidetes, Proteobacteria, Desulfobacterota & Campilobacterota ↑ Firmicutes & Actinobacteriota	Amuc_1100 decreased AP severity	Gut-microbiota based interventions can decrease inflammatory injury and AP severity
2024, Xu et al, (11)	SAP, L- ornithine intraperitoneal injection Impact of electroacupuncture on GM Cecal contents of rat 16S rRNA sequencing SCFAs-targeted metabolomics	Lactobacillus was the predominant bacteria in the cecum of control group ↑ Lactobacillus in rats in the SAP-EA group compared with SAP group only	EA impacts intestinal barrier injury & GM composition in SAP ↓ SCFAs in cecal contents of SAP rats	Gut-barrier can be targeted especially in SAP to alleviate diseases progression
2023, Zhou et al, (12)	AP (Caerulein & LPS) & Control Fecal samples of mice Metagenomics and untargeted metabolomics	AP mice had ↓ Actinobacteria & Bacteroidetes, and ↑ Proteobacteria and Firmicutes (phylum level) AP mice had ↑ Desnuesiella, Citrobacter, Rhizobium Caldicoprobacter, & Caldanaerovirga (genus level)	Metabolites belonging to lipid and lipid-like molecules & organic acids and their derivatives were found to be different between AP and control mice	Highlights the role of GM and microbial metabolites during AP
2022, Glaubitz et al, (13)	AP, partial duct ligation Intestinal samples Murine & human pancreatic necrosis 16S rRNA sequencing	Escherichia/Shigella was diminished in Treg depleted animals	Treg- activation disturbs the duodenal barrier permitting bacterial translocation into pancreatic necrosis	Targeting Tregs can help to ameliorate disease severity in AP
2022, Liu et al, (14)	SAP, Mild AP & sham operation Rats, sodium taurocholate-induced AP Fecal samples 16S rRNA sequencing	↑ Clostridiaceae 1 & Clostridium sensu stricto 1 in AP ↑ Collinsella, Morganella, and Blautia in SAP Lactobacillus nearly disappeared in SAP at 72 hours Firmicutes/Bacteroidetes (F/B) ratio changed from 24 to 72 hours and was highest in the SAP	F/B ratio & relative abundance of Lactobacillus as potential markers for AP and AP severity	GM can serve as a marker of disease severity in AP
2022, Li et al, (15)	Caerulein-induced AP Antibiotics by oral gavage & subsequent FMT Lactate treatment by oral gavage Fecal samples of mice 16S rRNA sequencing LC-MS/MS for metabolites	↓ Bifidobacterium in the MAP & SAP B. animalis colonization i) alleviated pancreatic pathological changes and ii) improved survival in SAP (caerulein & LPS; and sodium taurocholate & pancreatic-bile duct) models	-B. animalis & its metabolite lactate protected Abx and GF mice from AP -B. animalis relieved macrophage-associated local and systemic inflammation of AP	GM and microbial metabolites as therapeutic targets in AP

Abx: Antibiotic; AP: Acute pancreatitis; Electroacupuncture: EA; FMT: Fecal microbiota transplantation; GF: Germ free; GM: Gut microbiome; LPS: Lipopolysaccharide; MAP: Mild acute pancreatitis; rRNA: ribosomal RNA; SAP: Severe acute pancreatitis; SCFA: Short-chain fatty acid. (Original submission).

MECHANISMS OF MICROBIOME DRIVEN INJURY IN AP

Key observations that deviations from the gut normobiosis exacerbate while microbiota targeted therapies may resolve inflammation and injury in AP collectively indicate that the GM may be a critical determinant in the pathophysiology of AP(15, 30, 31). Emerging experimental evidence has identified intestinal barrier breach, microbial translocation and colonization of the pancreas, disruption of immune homeostasis, and the dysregulated production and activity of microbial metabolites as prominent mechanisms by which the GM impacts AP risk and severity (Figure 2).

Figure 2. Gut microbiome-pancreas crosstalk under homeostasis and in acute pancreatitis.

Alongside a healthy intestine, normal microbiome and pancreas, a dysbiotic gut and inflamed pancreas as in acute pancreatitis are depicted. Under normal physiologic conditions, gut microbial activity (via its metabolites like short chain fatty acids) support a healthy gut barrier and regulate pancreatic function. Antimicrobial peptides released by the pancreatic acinar cells shape the gut microbiome. Several factors like an unbalanced diet, alcohol consumption and other triggers participate in the pathophysiology of acute pancreatitis. These factors promote gut microbiome dysbiosis leading to a loss of beneficial gut metabolites and enhanced production of proinflammatory metabolites like H₂S that damages the barrier. Systemic and pancreatic bacterial translocation aggravates pancreas damage. Dysbiosis also promotes activation and recruitment of immune cells amplifying the inflammatory response. Acinar cell damage in turn disturbs pH homeostasis, which along with the lack of antimicrobial peptides promotes bacterial overgrowth in the duodenum. Finally, this inflammatory and infectious damage inflicted on the pancreas can lead to exocrine and/or endocrine insufficiency. (Original submission).

H₂S-hydrogen sulfide, Trp- tryptophan, LPS- lipopolysaccharide, NMN- nicotinamide mononucleotide, SCFAs- short chain fatty acids.

Intestinal barrier breach

AP results in increased gut permeability due to gut epithelial cell apoptosis, decreased expression of tight junction proteins, claudin 4, zona occludens-1, occludin, loss of mucus producing goblet cells, and dysfunction of Paneth cells (32, 33). This gut barrier damage allows systemic and pancreatic bacterial translocation along with the systemic dissemination of gut bacteria derivatives like lipopolysaccharides (LPS). Altogether these processes promote sustained immune responses, inflammation, and pancreatic necrosis and infection (32, 33). Peripheral organ damage, like acute lung and kidney injury associated with SAP may also be due to a defective gut barrier(34, 35). The role of a dysbiotic and dysfunctional GM leading to gut barrier damage in AP is supported by a distorted Firmicutes to Bacteriodetes ratio, bloom of Proteobacteria and pathogenic strains like Enterobacterium and Enterococcus and altered levels of beneficial microbial metabolites in both human subjects and rodent models(32). Interestingly, how this loss of gut barrier integrity converts a sterile necrosis in AP, to an infected one by amphipathic liponecrosis has recently been revealed(5^••).

Disruption of immune homeostasis

AP patients have excessive visceral fat lipolysis, increased serum levels of non-esterified fatty acids (FAs) and unbound unsaturated FAs (5^••). In face of concurrent hypoalbuminemia in AP, immune cell uptake of these FAs is increased, and this may disturb immune cell mitochondrial energetics, reduce its phagocytic capacity, and bacterial clearance. In fact, amphipathic liponecrosis, death of immune cells by the circulating unbound fatty acids (FA) was shown to worsen AP outcomes (5^••). This mechanism was further substantiated by the fact that infected AP patients had increased serum non-esterified FA and 16S bacterial DNA levels, and significantly lower number of circulating CD3 T cells compared to controls(5^••).

Immune homeostasis can also be affected by the gut microbiota in a more direct manner via their metabolites. SCFAs, like butyrate, confer benefit against AP through reinforcing the gut barrier (36), exerting anti-inflammatory effects through the inhibition of NLRP3 inflammasome, HDAC (32) and MMP9 activities (9^••) and modulating gut microbial tryptophan metabolism(9^••). However, recent discordant findings observed elevated SCFAs (acetate) (9^••) or over-presentation of SCFA producing gut bacteria in patients with SAP (2^••) warrant retrograde translation studies in pre-clinical models to decipher directionality, biological mechanisms and specific effect of these changes. These findings also reiterate redundancy of SCFA production pathways, for example, increased acetate levels in AP subjects from the study of Yazici et al (8^••) signify enhanced activity of acetate-producing genera may be in response to alcohol consumption. Using mouse models, a cause-effect relationship between decreased fecal abundance of Bifidobacterium and increased systemic inflammatory responses in SAP has been recently demonstrated(15). Colonization of antibiotic treated or germ-free mice with Bifidobacterium animalis or restoration of its metabolite, lactate, mitigated local and systemic inflammation of AP by TLR4/MyD88 and NLRP3/Caspase1 dependent mechanisms(15). Similarly, establishing a retrograde translation model by FMT from HTGP patients or healthy controls to antibiotic treated mice, Li et al (6^••) found that the deficiency of Bacteroides uniformis and its metabolite, taurine, exacerbates HTGP by driving IL-17 production, formation of neutrophil extracellular traps and activation of NF-κB signaling suggesting taurine or Bacteroides uniformis supplementation as a novel therapeutic strategy against HTGP.

Microbial metabolites and pancreatic exocrine insufficiency

SCFAs, metabolites generated by gut microbial fermentation, are considered metabolically beneficial for health. However, recent AP studies in humans showed either elevated levels of acetate, a SCFA, (8^••) or increased abundance of SCFA producing bacteria in patients with SAP (2^••). Given conflicting results and redundancy of SCFA production pathways, it is crucial to evaluate the role of SCFAs in AP in a larger paired metagenomic and metabolomic dataset in humans. Several host factors, like diet and exocrine pancreatic function, can also influence the gut microbial functional diversity. AP associated with HFD exhibit distinct GM profiles compared to healthy controls(1^••). Mostly enriched by pathobionts, certain specific strains like sulfidogenic Bilophila wadsworthia are abundant and lactogenic Bifidobacterium spp. are deficient in hyperlipidemic AP subjects(1^••). Conversely, AP subjects consuming diets deficient in fiber, vitamin D, omega-3 and polyunsaturated fatty acids exhibit lower GM alpha diversity and enrichment of sulfidogenic bacterial species like Veillonella spp. and Haemophilus spp. (8^••). These sulfidogenic bacteria can produce pro-inflammatory hydrogen sulfide (H₂S) that has been associated with AP risk (8^••). Indeed, high serum levels of H₂S compared to healthy controls have been reported in AP patients (8^••).

As acinar cell antimicrobial peptides help shape the intestinal microbial population, pancreatic exocrine insufficiency (PEI) is accompanied by reduced antimicrobial peptide production, duodenal pH change and alteration in the GM (37). It is unclear if a similar deficiency in the exocrine pancreatic function affects GM in AP. However, AP patients often have impairments in exocrine and endocrine pancreatic function and are at increased risk for new onset diabetes and PEI(38, 39, 40). Owing to functional GM alterations noted in AP and dependence of GM diversity on PEI, it is plausible that GM and its metabolites participate in the metabolic sequelae of AP(39). Moving forward, more definitive evidence still is needed to demonstrate this experimentally. Other microbial metabolites like nicotinamide mononucleotide and tryptophan metabolites have also been shown to inhibit inflammation and injury in AP mouse models(10, 41). These studies indicate that pancreatitis metabolome is profoundly affected by the gut microbial functional diversity. Nonetheless pancreas-gut microbiota crosstalk in regulating AP is increasingly clear. Systematic metagenomic and metabolomic profiling of AP patients in larger clinical cohorts can allow identification of microbiota-based selective therapeutic strategies.

CONCLUSION

The factors driving AP such as alcohol use or gallstones are well known; however, the development and severity of AP can be influenced by other factors such as GM. Understanding the impact of GM in AP is needed. Key observational studies have associated GM changes with onset and severity of AP. Going further, mouse studies, using FMT models, have further highlighted the role of GM in the development and severity of AP. Mechanisms behind the role of GM in AP are likely multifactorial, including regulating the gut barrier, immune cell homeostasis and the production of metabolites that may impact (positively or negatively) AP development. Important next steps will be leveraging large human cohort studies and assessing the GM profile, its metabolites and impact on AP type, development and severity. These findings can help drive new innovations to be tested in the prevention and treatment of AP.

Key points:

• Recent human subject studies demonstrate the association of GM changes with AP risk, etiology and severity.

• FMT models have been increasingly used in recent mouse studies and highlight the contribution of GM in the development and severity of AP.

• The gut barrier integrity, immune cell homeostasis and microbial metabolites appear to be crucial mechanisms for GM-driven injury in AP.

• Machine learning models integrating GM, metabolome, and clinical data have been successful in predicting host phenotypes and have the potential for prediction of disease severity or recurrence as future clinical applications.

• Large prospective human studies that combine GM, metabolome and clinical data may be helpful for development of targeted prevention and treatment strategies.